Ollier disease, or multiple enchondromatosis, is a rare nonfamilial condition characterized by multiple cartilaginous tumors often beginning in early childhood. There is significant variation in disease distribution, location, size, number of lesions, and behavior, but the tumors are often located unilaterally.¹ Enchondromas are most commonly found in the metacarpals, metatarsals, and phalanges, and develop from metaphyseal bone in close proximity to the physis. They frequently present as painless masses or are incidentally noted during the evaluation of another musculoskeletal condition. Radiographically, enchondromas of the hands and feet appear as oval radiolucencies with thinned, sclerotic rims. The lesions have varying degrees of mineralization and endosteal scalloping, and may expand the bone.² Enchondromas usually enlarge until skeletal maturity and have been observed to ossify spontaneously.^1,3 The clinical course of Ollier disease is variable, and a number of cases of significant hand deformity and malignant transformation have been reported.^4-6

In this case report, we present a mild form of Ollier disease isolated to the patient’s left hand, which we followed for 8 years, demonstrating part of the natural history of these lesions. We discuss the patient’s clinical features, radiologic findings, diagnosis, treatment, prognosis, and follow-up, as well as review the literature. The patient and the patient’s family provided written informed consent for print and electronic publication of this case report.

Case Report

A 10-year-old, right-handed girl was referred to our department for the evaluation of left-hand masses. At age 3 years, the patient underwent a chondroma excision from the middle phalanx of her middle finger on her left hand. No operative or pathology report was available from this surgery, and the patient tolerated the procedure well without any complications. At the time of presentation, the masses did not cause any pain, motor or sensory dysfunction, or any systemic symptoms. No history of recent or distant trauma was elicited. The patient’s medical and family history was unremarkable.

On physical examination, there was a firm, immobile, nontender palpable mass over the dorsal aspect of the distal second metacarpal bone of the left hand. The mass extended medially between the second and third metacarpals. A second small, firm, nontender left-hand mass was palpated over the volar aspect of her proximal phalanx on her index finger. She was neurovascularly intact with full active range of motion of the metacarpophalangeal and proximal and distal interphalangeal joints. There was no angular deformity of the digits. Plain radiographs taken at the time of initial presentation showed a 2.3×1.7-cm radiolucent lesion located in the metaphysis and diaphysis of the second metacarpal of the left hand (Figures 1A-1C). The lesion had varying degrees of mineralization with cortical thinning and expansion in the volar, dorsal, radial, and ulnar directions, consistent with a chondroid lesion. The second and third lesions were oval radiolucencies with sclerotic rims located at the metaphyseal-diaphyseal junction of the proximal phalanx of the index finger and middle phalanx of the middle finger, respectively. No fractures were identified in the radiographs, and the physes were open at this time. The patient was diagnosed with multiple enchondromatosis, or Ollier disease.

Our case showed 1 episode of pain and tenderness to palpation at the second proximal phalanx approximately 6 months after initial presentation. We attributed the pain and tenderness to a small pathologic fracture but did not see radiographic evidence of this. We elected to provide a trial of supportive measures, such as splinting and buddy taping, and to monitor the pain with a tentative plan of open biopsy with curettage and bone grafting if the pain persisted or evidence of fracture was seen on radiographs. The pain and tenderness to palpation resolved at a follow-up visit, and the surgery was deferred.

The patient was treated nonoperatively at initial presentation given the lack of significant cosmetic deformity or functional compromise and was advised close follow-up at 3 and 6 months. Given the absence of disease progression, annual checks (ie, clinical examination and radiographs) in a skeletally immature patient were decided on after consultation with the patient and parent. The family was educated about the possibility of pathologic fracture from minimal trauma to the hand versus the small risk of iatrogenic physeal injury with surgical curettage and bone grafting. No protective splinting was offered. A favorable prognosis and reassurance was provided to the patient and family, given the absence of symptoms, low suspicion and risk of malignant transformation, and stability of the lesion. Serial radiographs showed gradual increases in the lesions’ sizes but were consistent with the stable growth of the metacarpal and phalanges. With the patient nearing skeletal maturity, no pathologic fractures were identified on radiography during follow-up, and the risks of surgery lessened with growth; however, the continued absence of symptoms led to the mutual decision to continue observation.

Nearly 8 years after initial presentation, plain radiographs showed closed physes and partially ossified bone masses (Figures 2A-2C). The metacarpal lesion measured 3.2×1.5 cm, and the cortex appeared thickened and regular. The proximal phalanx lesion had a thickened cortex without periosteal reaction, and the middle phalanx lesion appeared to be completely healed. The patient has been asymptomatic for many years, and she has retained complete function of her left hand without any growth retardation, angular deformity, or pathologic fracture. A small but potential risk of malignant transformation was discussed with the patient and her family, as was the need for lifetime follow-up. We intend to follow the enchondromas clinically and radiographically every 2 years and obtain new radiographs if the mass presents with new clinical findings, such as enlargement or pain, for surveillance of tumor transformation. If the patient desired or symptoms developed, curettage and bone grafting would be offered, and the surgical tissue would be sent for pathologic analysis. A bone scan that was obtained at the request of the patient, when she was 21 years old, showed no other sites of disease besides the fingers.

Discussion

Multiple enchondromatosis was first described by Ollier at the turn of the 19th century and has been estimated to affect one in every 100,000 persons.¹ The low prevalence and variable manifestations of Ollier disease lead clinicians to handle the disease and its complications, namely skeletal deformity and malignant transformation, on a case-by-case approach. Additionally, the prognosis of Ollier disease with malignant transformation is quite variable, with studies reporting the estimated incidence as 5% to 50%.⁷ Muramatsu and colleagues⁶ reported that the occurrence of malignant transformation of multiple enchondromas limited to the bones of the hand was extremely rare, with only 12 cases of malignant transformation. Enchondromas of the pelvis, scapula, and long bones of the extremities have increased risks and rates of secondary transformation to chondrosarcoma.⁸

A recent large European multicenter retrospective study investigating the clinical characteristics and behavior of enchondromas in 144 patients with Ollier disease has provided new information regarding this rare disease.⁷ Verdegaal and colleagues⁷ divided patients into 3 categories depending on their distribution of enchondromas. The development of chondrosarcoma was notably different between individuals with enchondromas limited to the small bones of the hands and feet (15%, group I) versus individuals with enchondromas limited to the long bones and flat bones (43%, group II) or individuals with enchondromas of the short, long, and flat bones (46%, group III).⁷ The only location found to be statistically significant for the development of chondrosarcoma was the pelvis.

The clinical findings associated with risk of malignant transformation of enchondromas are increasing size of the lesion and onset of pain and tenderness. Dahlin and Salvador⁹ reported that only 60% of patients with chondrosarcoma of the hand experience pain. The absence of pain may lead to a delay in patient presentation to the clinician.^5,6 Radiographic findings of malignant transformation include the classic features of temporal increases in the lesion’s size after skeletal maturity and cortical destruction associated with soft-tissue invasion. However, both findings are nonspecific for differentiating enchondromas from grade 1 chondrosarcomas as described by Geirnaerdt and colleagues.¹⁰

Sassoon and colleagues¹¹ reported on a series of hand enchondromas treated operatively. Subgroup analysis between pathologic fractures treated primarily or in delayed fashion showed similar outcomes for achieving full motion and similar number of complications; however, they noted that the delayed group required 7 more weeks of immobilization. Additionally, review of the whole series showed 1 episode of metacarpal shortening and 1 occurrence of angular malalignment. In our patient, we were concerned about introducing an iatrogenic cosmetic deformity, and we believed a pathologic fracture could be managed expectantly. Overall, patients without pathologic fracture treated surgically experienced a complication rate of 12%, whereas patients with a fracture had a complication rate of 20%.¹¹ The majority of patients with multiple enchondromatosis treated with surgical curettage and grafting had successful outcomes, with 86% of patients regaining full motion, but the recurrence rate was 21%.¹¹ Patients with expansile lesions regained less motion than patients with nonexpansile lesions. There was a single lesion believed preoperatively to be an enchondroma, but it underwent malignant transformation, as confirmed on intraoperative pathology. This patient had Maffucci syndrome and was treated with an amputation through the metacarpophalangeal joint.

There are 3 options for treating hand enchondromas: observation, curettage alone, or curettage with bone grafting. There is no consensus about conservative management, timing of intervention, or risk of pathologic fracture. Each patient is treated individually with attention to reason for presentation, number of lesions, associated pain, deformity, or pathologic fracture. Operative criteria include high risk of pathologic fracture based on location of enchondroma, cortical thinning, and previous pathologic fracture with resulting angular deformity. Nonoperative management may increase the risk of pathologic fracture, particularly in patients involved in aggressive contact sports, but the physician may offer protective splinting or counsel the patient on activity modification. Our case provides a study of the natural history of multiple enchondromatosis and shows mild increases in the lesions’ sizes during the 8-year follow-up. This was an expected finding given the patient’s immature skeleton. The lesions’ cortices continued to ossify after the physes closed and now provides an excellent comparison for the identification of future malignant changes. 

Histologic analysis of biopsied or surgically treated lesions contributes to the differentiation between benign hand enchondromas and chondrosarcoma. Pathologic findings must be correlated with clinical and radiographic findings because hand enchondromas contain cytologic features of chondrosarcoma.¹² In a series of 55 patients with chondrosarcoma, Liu and colleagues⁸ reported no cases from the hand. Verdegaal and colleagues⁷ reported a total of 13 chondrosarcomas in the metacarpals and hand phalanges in 97 group I and III patients. Five of these lesions were grade 1, 2 were grade 2, 1 was grade 3, and 5 lesions were unknown.

For patients with multiple enchondromatosis limited to the hands, prognosis is relatively good with respect to risk of secondary chondrosarcoma transformation, metastasis of secondary chondrosarcoma, and death. Verdegaal and colleagues⁷ reported the rate of secondary transformation in the hand to be 15%. Patil and colleagues¹³ reported no distant metastases in 23 patients with hand chondrosarcoma at mean follow-up of 8.5 years (range, 2-19 years), although none of their patients had Ollier disease. Verdegaal and colleagues⁷ reported 7 of the 8 deaths in their study were related to development of pulmonary metastases; however, none originated from chondrosarcomas in the hand. Additionally, there were no disease-related deaths in 29 group I patients. Herget and colleagues,¹⁴ in summarizing the literature, postulated that the overall survival rate of patients with secondary chondrosarcoma at 5 years is approximately 90%.

In our case, the patient, who had 3 enchondromas isolated to the left hand, can be categorized in group I. Thus, this case highlights the natural history of a patient with hand enchondromas and demonstrates that enchondromatosis of the short tubular bones of the hands can mature and ossify.

Ollier disease, or multiple enchondromatosis, is a rare nonfamilial condition characterized by multiple cartilaginous tumors often beginning in early childhood. There is significant variation in disease distribution, location, size, number of lesions, and behavior, but the tumors are often located unilaterally.¹ Enchondromas are most commonly found in the metacarpals, metatarsals, and phalanges, and develop from metaphyseal bone in close proximity to the physis. They frequently present as painless masses or are incidentally noted during the evaluation of another musculoskeletal condition. Radiographically, enchondromas of the hands and feet appear as oval radiolucencies with thinned, sclerotic rims. The lesions have varying degrees of mineralization and endosteal scalloping, and may expand the bone.² Enchondromas usually enlarge until skeletal maturity and have been observed to ossify spontaneously.^1,3 The clinical course of Ollier disease is variable, and a number of cases of significant hand deformity and malignant transformation have been reported.^4-6

In this case report, we present a mild form of Ollier disease isolated to the patient’s left hand, which we followed for 8 years, demonstrating part of the natural history of these lesions. We discuss the patient’s clinical features, radiologic findings, diagnosis, treatment, prognosis, and follow-up, as well as review the literature. The patient and the patient’s family provided written informed consent for print and electronic publication of this case report.

Case Report

A 10-year-old, right-handed girl was referred to our department for the evaluation of left-hand masses. At age 3 years, the patient underwent a chondroma excision from the middle phalanx of her middle finger on her left hand. No operative or pathology report was available from this surgery, and the patient tolerated the procedure well without any complications. At the time of presentation, the masses did not cause any pain, motor or sensory dysfunction, or any systemic symptoms. No history of recent or distant trauma was elicited. The patient’s medical and family history was unremarkable.

On physical examination, there was a firm, immobile, nontender palpable mass over the dorsal aspect of the distal second metacarpal bone of the left hand. The mass extended medially between the second and third metacarpals. A second small, firm, nontender left-hand mass was palpated over the volar aspect of her proximal phalanx on her index finger. She was neurovascularly intact with full active range of motion of the metacarpophalangeal and proximal and distal interphalangeal joints. There was no angular deformity of the digits. Plain radiographs taken at the time of initial presentation showed a 2.3×1.7-cm radiolucent lesion located in the metaphysis and diaphysis of the second metacarpal of the left hand (Figures 1A-1C). The lesion had varying degrees of mineralization with cortical thinning and expansion in the volar, dorsal, radial, and ulnar directions, consistent with a chondroid lesion. The second and third lesions were oval radiolucencies with sclerotic rims located at the metaphyseal-diaphyseal junction of the proximal phalanx of the index finger and middle phalanx of the middle finger, respectively. No fractures were identified in the radiographs, and the physes were open at this time. The patient was diagnosed with multiple enchondromatosis, or Ollier disease.

Our case showed 1 episode of pain and tenderness to palpation at the second proximal phalanx approximately 6 months after initial presentation. We attributed the pain and tenderness to a small pathologic fracture but did not see radiographic evidence of this. We elected to provide a trial of supportive measures, such as splinting and buddy taping, and to monitor the pain with a tentative plan of open biopsy with curettage and bone grafting if the pain persisted or evidence of fracture was seen on radiographs. The pain and tenderness to palpation resolved at a follow-up visit, and the surgery was deferred.

The patient was treated nonoperatively at initial presentation given the lack of significant cosmetic deformity or functional compromise and was advised close follow-up at 3 and 6 months. Given the absence of disease progression, annual checks (ie, clinical examination and radiographs) in a skeletally immature patient were decided on after consultation with the patient and parent. The family was educated about the possibility of pathologic fracture from minimal trauma to the hand versus the small risk of iatrogenic physeal injury with surgical curettage and bone grafting. No protective splinting was offered. A favorable prognosis and reassurance was provided to the patient and family, given the absence of symptoms, low suspicion and risk of malignant transformation, and stability of the lesion. Serial radiographs showed gradual increases in the lesions’ sizes but were consistent with the stable growth of the metacarpal and phalanges. With the patient nearing skeletal maturity, no pathologic fractures were identified on radiography during follow-up, and the risks of surgery lessened with growth; however, the continued absence of symptoms led to the mutual decision to continue observation.

Nearly 8 years after initial presentation, plain radiographs showed closed physes and partially ossified bone masses (Figures 2A-2C). The metacarpal lesion measured 3.2×1.5 cm, and the cortex appeared thickened and regular. The proximal phalanx lesion had a thickened cortex without periosteal reaction, and the middle phalanx lesion appeared to be completely healed. The patient has been asymptomatic for many years, and she has retained complete function of her left hand without any growth retardation, angular deformity, or pathologic fracture. A small but potential risk of malignant transformation was discussed with the patient and her family, as was the need for lifetime follow-up. We intend to follow the enchondromas clinically and radiographically every 2 years and obtain new radiographs if the mass presents with new clinical findings, such as enlargement or pain, for surveillance of tumor transformation. If the patient desired or symptoms developed, curettage and bone grafting would be offered, and the surgical tissue would be sent for pathologic analysis. A bone scan that was obtained at the request of the patient, when she was 21 years old, showed no other sites of disease besides the fingers.

Discussion

Multiple enchondromatosis was first described by Ollier at the turn of the 19th century and has been estimated to affect one in every 100,000 persons.¹ The low prevalence and variable manifestations of Ollier disease lead clinicians to handle the disease and its complications, namely skeletal deformity and malignant transformation, on a case-by-case approach. Additionally, the prognosis of Ollier disease with malignant transformation is quite variable, with studies reporting the estimated incidence as 5% to 50%.⁷ Muramatsu and colleagues⁶ reported that the occurrence of malignant transformation of multiple enchondromas limited to the bones of the hand was extremely rare, with only 12 cases of malignant transformation. Enchondromas of the pelvis, scapula, and long bones of the extremities have increased risks and rates of secondary transformation to chondrosarcoma.⁸

A recent large European multicenter retrospective study investigating the clinical characteristics and behavior of enchondromas in 144 patients with Ollier disease has provided new information regarding this rare disease.⁷ Verdegaal and colleagues⁷ divided patients into 3 categories depending on their distribution of enchondromas. The development of chondrosarcoma was notably different between individuals with enchondromas limited to the small bones of the hands and feet (15%, group I) versus individuals with enchondromas limited to the long bones and flat bones (43%, group II) or individuals with enchondromas of the short, long, and flat bones (46%, group III).⁷ The only location found to be statistically significant for the development of chondrosarcoma was the pelvis.

The clinical findings associated with risk of malignant transformation of enchondromas are increasing size of the lesion and onset of pain and tenderness. Dahlin and Salvador⁹ reported that only 60% of patients with chondrosarcoma of the hand experience pain. The absence of pain may lead to a delay in patient presentation to the clinician.^5,6 Radiographic findings of malignant transformation include the classic features of temporal increases in the lesion’s size after skeletal maturity and cortical destruction associated with soft-tissue invasion. However, both findings are nonspecific for differentiating enchondromas from grade 1 chondrosarcomas as described by Geirnaerdt and colleagues.¹⁰

Sassoon and colleagues¹¹ reported on a series of hand enchondromas treated operatively. Subgroup analysis between pathologic fractures treated primarily or in delayed fashion showed similar outcomes for achieving full motion and similar number of complications; however, they noted that the delayed group required 7 more weeks of immobilization. Additionally, review of the whole series showed 1 episode of metacarpal shortening and 1 occurrence of angular malalignment. In our patient, we were concerned about introducing an iatrogenic cosmetic deformity, and we believed a pathologic fracture could be managed expectantly. Overall, patients without pathologic fracture treated surgically experienced a complication rate of 12%, whereas patients with a fracture had a complication rate of 20%.¹¹ The majority of patients with multiple enchondromatosis treated with surgical curettage and grafting had successful outcomes, with 86% of patients regaining full motion, but the recurrence rate was 21%.¹¹ Patients with expansile lesions regained less motion than patients with nonexpansile lesions. There was a single lesion believed preoperatively to be an enchondroma, but it underwent malignant transformation, as confirmed on intraoperative pathology. This patient had Maffucci syndrome and was treated with an amputation through the metacarpophalangeal joint.

There are 3 options for treating hand enchondromas: observation, curettage alone, or curettage with bone grafting. There is no consensus about conservative management, timing of intervention, or risk of pathologic fracture. Each patient is treated individually with attention to reason for presentation, number of lesions, associated pain, deformity, or pathologic fracture. Operative criteria include high risk of pathologic fracture based on location of enchondroma, cortical thinning, and previous pathologic fracture with resulting angular deformity. Nonoperative management may increase the risk of pathologic fracture, particularly in patients involved in aggressive contact sports, but the physician may offer protective splinting or counsel the patient on activity modification. Our case provides a study of the natural history of multiple enchondromatosis and shows mild increases in the lesions’ sizes during the 8-year follow-up. This was an expected finding given the patient’s immature skeleton. The lesions’ cortices continued to ossify after the physes closed and now provides an excellent comparison for the identification of future malignant changes. 

Histologic analysis of biopsied or surgically treated lesions contributes to the differentiation between benign hand enchondromas and chondrosarcoma. Pathologic findings must be correlated with clinical and radiographic findings because hand enchondromas contain cytologic features of chondrosarcoma.¹² In a series of 55 patients with chondrosarcoma, Liu and colleagues⁸ reported no cases from the hand. Verdegaal and colleagues⁷ reported a total of 13 chondrosarcomas in the metacarpals and hand phalanges in 97 group I and III patients. Five of these lesions were grade 1, 2 were grade 2, 1 was grade 3, and 5 lesions were unknown.

For patients with multiple enchondromatosis limited to the hands, prognosis is relatively good with respect to risk of secondary chondrosarcoma transformation, metastasis of secondary chondrosarcoma, and death. Verdegaal and colleagues⁷ reported the rate of secondary transformation in the hand to be 15%. Patil and colleagues¹³ reported no distant metastases in 23 patients with hand chondrosarcoma at mean follow-up of 8.5 years (range, 2-19 years), although none of their patients had Ollier disease. Verdegaal and colleagues⁷ reported 7 of the 8 deaths in their study were related to development of pulmonary metastases; however, none originated from chondrosarcomas in the hand. Additionally, there were no disease-related deaths in 29 group I patients. Herget and colleagues,¹⁴ in summarizing the literature, postulated that the overall survival rate of patients with secondary chondrosarcoma at 5 years is approximately 90%.

In our case, the patient, who had 3 enchondromas isolated to the left hand, can be categorized in group I. Thus, this case highlights the natural history of a patient with hand enchondromas and demonstrates that enchondromatosis of the short tubular bones of the hands can mature and ossify.

Ollier disease, or multiple enchondromatosis, is a rare nonfamilial condition characterized by multiple cartilaginous tumors often beginning in early childhood. There is significant variation in disease distribution, location, size, number of lesions, and behavior, but the tumors are often located unilaterally.¹ Enchondromas are most commonly found in the metacarpals, metatarsals, and phalanges, and develop from metaphyseal bone in close proximity to the physis. They frequently present as painless masses or are incidentally noted during the evaluation of another musculoskeletal condition. Radiographically, enchondromas of the hands and feet appear as oval radiolucencies with thinned, sclerotic rims. The lesions have varying degrees of mineralization and endosteal scalloping, and may expand the bone.² Enchondromas usually enlarge until skeletal maturity and have been observed to ossify spontaneously.^1,3 The clinical course of Ollier disease is variable, and a number of cases of significant hand deformity and malignant transformation have been reported.^4-6

In this case report, we present a mild form of Ollier disease isolated to the patient’s left hand, which we followed for 8 years, demonstrating part of the natural history of these lesions. We discuss the patient’s clinical features, radiologic findings, diagnosis, treatment, prognosis, and follow-up, as well as review the literature. The patient and the patient’s family provided written informed consent for print and electronic publication of this case report.

Case Report

A 10-year-old, right-handed girl was referred to our department for the evaluation of left-hand masses. At age 3 years, the patient underwent a chondroma excision from the middle phalanx of her middle finger on her left hand. No operative or pathology report was available from this surgery, and the patient tolerated the procedure well without any complications. At the time of presentation, the masses did not cause any pain, motor or sensory dysfunction, or any systemic symptoms. No history of recent or distant trauma was elicited. The patient’s medical and family history was unremarkable.

On physical examination, there was a firm, immobile, nontender palpable mass over the dorsal aspect of the distal second metacarpal bone of the left hand. The mass extended medially between the second and third metacarpals. A second small, firm, nontender left-hand mass was palpated over the volar aspect of her proximal phalanx on her index finger. She was neurovascularly intact with full active range of motion of the metacarpophalangeal and proximal and distal interphalangeal joints. There was no angular deformity of the digits. Plain radiographs taken at the time of initial presentation showed a 2.3×1.7-cm radiolucent lesion located in the metaphysis and diaphysis of the second metacarpal of the left hand (Figures 1A-1C). The lesion had varying degrees of mineralization with cortical thinning and expansion in the volar, dorsal, radial, and ulnar directions, consistent with a chondroid lesion. The second and third lesions were oval radiolucencies with sclerotic rims located at the metaphyseal-diaphyseal junction of the proximal phalanx of the index finger and middle phalanx of the middle finger, respectively. No fractures were identified in the radiographs, and the physes were open at this time. The patient was diagnosed with multiple enchondromatosis, or Ollier disease.

Our case showed 1 episode of pain and tenderness to palpation at the second proximal phalanx approximately 6 months after initial presentation. We attributed the pain and tenderness to a small pathologic fracture but did not see radiographic evidence of this. We elected to provide a trial of supportive measures, such as splinting and buddy taping, and to monitor the pain with a tentative plan of open biopsy with curettage and bone grafting if the pain persisted or evidence of fracture was seen on radiographs. The pain and tenderness to palpation resolved at a follow-up visit, and the surgery was deferred.

The patient was treated nonoperatively at initial presentation given the lack of significant cosmetic deformity or functional compromise and was advised close follow-up at 3 and 6 months. Given the absence of disease progression, annual checks (ie, clinical examination and radiographs) in a skeletally immature patient were decided on after consultation with the patient and parent. The family was educated about the possibility of pathologic fracture from minimal trauma to the hand versus the small risk of iatrogenic physeal injury with surgical curettage and bone grafting. No protective splinting was offered. A favorable prognosis and reassurance was provided to the patient and family, given the absence of symptoms, low suspicion and risk of malignant transformation, and stability of the lesion. Serial radiographs showed gradual increases in the lesions’ sizes but were consistent with the stable growth of the metacarpal and phalanges. With the patient nearing skeletal maturity, no pathologic fractures were identified on radiography during follow-up, and the risks of surgery lessened with growth; however, the continued absence of symptoms led to the mutual decision to continue observation.

Nearly 8 years after initial presentation, plain radiographs showed closed physes and partially ossified bone masses (Figures 2A-2C). The metacarpal lesion measured 3.2×1.5 cm, and the cortex appeared thickened and regular. The proximal phalanx lesion had a thickened cortex without periosteal reaction, and the middle phalanx lesion appeared to be completely healed. The patient has been asymptomatic for many years, and she has retained complete function of her left hand without any growth retardation, angular deformity, or pathologic fracture. A small but potential risk of malignant transformation was discussed with the patient and her family, as was the need for lifetime follow-up. We intend to follow the enchondromas clinically and radiographically every 2 years and obtain new radiographs if the mass presents with new clinical findings, such as enlargement or pain, for surveillance of tumor transformation. If the patient desired or symptoms developed, curettage and bone grafting would be offered, and the surgical tissue would be sent for pathologic analysis. A bone scan that was obtained at the request of the patient, when she was 21 years old, showed no other sites of disease besides the fingers.

Discussion

Multiple enchondromatosis was first described by Ollier at the turn of the 19th century and has been estimated to affect one in every 100,000 persons.¹ The low prevalence and variable manifestations of Ollier disease lead clinicians to handle the disease and its complications, namely skeletal deformity and malignant transformation, on a case-by-case approach. Additionally, the prognosis of Ollier disease with malignant transformation is quite variable, with studies reporting the estimated incidence as 5% to 50%.⁷ Muramatsu and colleagues⁶ reported that the occurrence of malignant transformation of multiple enchondromas limited to the bones of the hand was extremely rare, with only 12 cases of malignant transformation. Enchondromas of the pelvis, scapula, and long bones of the extremities have increased risks and rates of secondary transformation to chondrosarcoma.⁸

A recent large European multicenter retrospective study investigating the clinical characteristics and behavior of enchondromas in 144 patients with Ollier disease has provided new information regarding this rare disease.⁷ Verdegaal and colleagues⁷ divided patients into 3 categories depending on their distribution of enchondromas. The development of chondrosarcoma was notably different between individuals with enchondromas limited to the small bones of the hands and feet (15%, group I) versus individuals with enchondromas limited to the long bones and flat bones (43%, group II) or individuals with enchondromas of the short, long, and flat bones (46%, group III).⁷ The only location found to be statistically significant for the development of chondrosarcoma was the pelvis.

The clinical findings associated with risk of malignant transformation of enchondromas are increasing size of the lesion and onset of pain and tenderness. Dahlin and Salvador⁹ reported that only 60% of patients with chondrosarcoma of the hand experience pain. The absence of pain may lead to a delay in patient presentation to the clinician.^5,6 Radiographic findings of malignant transformation include the classic features of temporal increases in the lesion’s size after skeletal maturity and cortical destruction associated with soft-tissue invasion. However, both findings are nonspecific for differentiating enchondromas from grade 1 chondrosarcomas as described by Geirnaerdt and colleagues.¹⁰

Sassoon and colleagues¹¹ reported on a series of hand enchondromas treated operatively. Subgroup analysis between pathologic fractures treated primarily or in delayed fashion showed similar outcomes for achieving full motion and similar number of complications; however, they noted that the delayed group required 7 more weeks of immobilization. Additionally, review of the whole series showed 1 episode of metacarpal shortening and 1 occurrence of angular malalignment. In our patient, we were concerned about introducing an iatrogenic cosmetic deformity, and we believed a pathologic fracture could be managed expectantly. Overall, patients without pathologic fracture treated surgically experienced a complication rate of 12%, whereas patients with a fracture had a complication rate of 20%.¹¹ The majority of patients with multiple enchondromatosis treated with surgical curettage and grafting had successful outcomes, with 86% of patients regaining full motion, but the recurrence rate was 21%.¹¹ Patients with expansile lesions regained less motion than patients with nonexpansile lesions. There was a single lesion believed preoperatively to be an enchondroma, but it underwent malignant transformation, as confirmed on intraoperative pathology. This patient had Maffucci syndrome and was treated with an amputation through the metacarpophalangeal joint.

There are 3 options for treating hand enchondromas: observation, curettage alone, or curettage with bone grafting. There is no consensus about conservative management, timing of intervention, or risk of pathologic fracture. Each patient is treated individually with attention to reason for presentation, number of lesions, associated pain, deformity, or pathologic fracture. Operative criteria include high risk of pathologic fracture based on location of enchondroma, cortical thinning, and previous pathologic fracture with resulting angular deformity. Nonoperative management may increase the risk of pathologic fracture, particularly in patients involved in aggressive contact sports, but the physician may offer protective splinting or counsel the patient on activity modification. Our case provides a study of the natural history of multiple enchondromatosis and shows mild increases in the lesions’ sizes during the 8-year follow-up. This was an expected finding given the patient’s immature skeleton. The lesions’ cortices continued to ossify after the physes closed and now provides an excellent comparison for the identification of future malignant changes. 

Histologic analysis of biopsied or surgically treated lesions contributes to the differentiation between benign hand enchondromas and chondrosarcoma. Pathologic findings must be correlated with clinical and radiographic findings because hand enchondromas contain cytologic features of chondrosarcoma.¹² In a series of 55 patients with chondrosarcoma, Liu and colleagues⁸ reported no cases from the hand. Verdegaal and colleagues⁷ reported a total of 13 chondrosarcomas in the metacarpals and hand phalanges in 97 group I and III patients. Five of these lesions were grade 1, 2 were grade 2, 1 was grade 3, and 5 lesions were unknown.

For patients with multiple enchondromatosis limited to the hands, prognosis is relatively good with respect to risk of secondary chondrosarcoma transformation, metastasis of secondary chondrosarcoma, and death. Verdegaal and colleagues⁷ reported the rate of secondary transformation in the hand to be 15%. Patil and colleagues¹³ reported no distant metastases in 23 patients with hand chondrosarcoma at mean follow-up of 8.5 years (range, 2-19 years), although none of their patients had Ollier disease. Verdegaal and colleagues⁷ reported 7 of the 8 deaths in their study were related to development of pulmonary metastases; however, none originated from chondrosarcomas in the hand. Additionally, there were no disease-related deaths in 29 group I patients. Herget and colleagues,¹⁴ in summarizing the literature, postulated that the overall survival rate of patients with secondary chondrosarcoma at 5 years is approximately 90%.

In our case, the patient, who had 3 enchondromas isolated to the left hand, can be categorized in group I. Thus, this case highlights the natural history of a patient with hand enchondromas and demonstrates that enchondromatosis of the short tubular bones of the hands can mature and ossify.

EVIDENCE SUMMARY

A 2012 Cochrane review of 52 studies (44 RCTs and 8 cluster-randomized trials; N=56,451) assessed the overall effectiveness of multiple supportive measures on decreasing cessation of “any” (partial and exclusive) and “exclusive” breastfeeding compared with usual care.¹ Participants were healthy breastfeeding mothers of healthy term babies. Support interventions were defined broadly but included individual and group interactions, as well as contact in person or over the phone by professionals or lay volunteers. Patients were approached proactively or reactively upon request, and the interventions occurred one or more times.

The interventions reduced discontinuation rates among both “exclusive” and “any” breastfeeding mothers (TABLE¹). The review found lay and professional support to be equally effective at promoting continuation of breastfeeding. Limitations include a moderate to high amount of heterogeneity, as well as the inherent difficulty of blinding subjects in the studies.

Lay support can make a significant difference in the short term

A 2008 systematic review of 38 RCTs (N=29,020) compared any counseling or behavioral intervention initiated from a clinician’s practice (office or hospital) with usual care.² The review excluded community and peer-initiated interventions. The reviewers defined breastfeeding duration as follows: initiation (up to 2 weeks), short-term (one to 3 months), intermediate-term (4 to 5 months), long-term (6 to 8 months), and prolonged (9 or more months). Investigators also analyzed breastfeeding rates by “exclusive” and “nonexclusive” (formula supplementation) regimens.

For nonexclusive breastfeeding, the review found interventions to promote breastfeeding improved rates only at initiation (18 RCTs, N=7688; relative risk [RR] for cessation of breastfeeding=1.04; 95% confidence interval [CI], 1.0-1.08; number needed to treat [NNT]=38) and in the short term (18 RCTs, N= 19,358; RR=1.10; 95% CI, 1.02-1.19; NNT=7). For exclusive breastfeeding, interventions improved rates only in the short term (17 RCTs, N=20,552; RR=1.72; 95% CI, 1.0-2.97; NNT=3).

The review found that lay support (defined as counseling or social support from peers) but not professional support was significantly associated with improving rates of both “nonexclusive” and “exclusive’ breastfeeding, but only over the short term (5 RCTs, N not provided; RR=1.22; 95% CI, 1.08-1.37; and 4 RCTs, N not provided; RR=1.65; 95% CI, 1.03-2.63; respectively). As with the Cochrane review, the results for all study groups demonstrated moderate to significant heterogeneity.

RECOMMENDATIONS

The Surgeon General, the American Academy of Family Physicians, the American Academy of Pediatrics, and the American College of Obstetricians and Gynecologists all recommend that women be educated about the benefits of breastfeeding and receive supportive interventions before and after delivery.^3-6

EVIDENCE SUMMARY

A 2012 Cochrane review of 52 studies (44 RCTs and 8 cluster-randomized trials; N=56,451) assessed the overall effectiveness of multiple supportive measures on decreasing cessation of “any” (partial and exclusive) and “exclusive” breastfeeding compared with usual care.¹ Participants were healthy breastfeeding mothers of healthy term babies. Support interventions were defined broadly but included individual and group interactions, as well as contact in person or over the phone by professionals or lay volunteers. Patients were approached proactively or reactively upon request, and the interventions occurred one or more times.

The interventions reduced discontinuation rates among both “exclusive” and “any” breastfeeding mothers (TABLE¹). The review found lay and professional support to be equally effective at promoting continuation of breastfeeding. Limitations include a moderate to high amount of heterogeneity, as well as the inherent difficulty of blinding subjects in the studies.

Lay support can make a significant difference in the short term

A 2008 systematic review of 38 RCTs (N=29,020) compared any counseling or behavioral intervention initiated from a clinician’s practice (office or hospital) with usual care.² The review excluded community and peer-initiated interventions. The reviewers defined breastfeeding duration as follows: initiation (up to 2 weeks), short-term (one to 3 months), intermediate-term (4 to 5 months), long-term (6 to 8 months), and prolonged (9 or more months). Investigators also analyzed breastfeeding rates by “exclusive” and “nonexclusive” (formula supplementation) regimens.

For nonexclusive breastfeeding, the review found interventions to promote breastfeeding improved rates only at initiation (18 RCTs, N=7688; relative risk [RR] for cessation of breastfeeding=1.04; 95% confidence interval [CI], 1.0-1.08; number needed to treat [NNT]=38) and in the short term (18 RCTs, N= 19,358; RR=1.10; 95% CI, 1.02-1.19; NNT=7). For exclusive breastfeeding, interventions improved rates only in the short term (17 RCTs, N=20,552; RR=1.72; 95% CI, 1.0-2.97; NNT=3).

The review found that lay support (defined as counseling or social support from peers) but not professional support was significantly associated with improving rates of both “nonexclusive” and “exclusive’ breastfeeding, but only over the short term (5 RCTs, N not provided; RR=1.22; 95% CI, 1.08-1.37; and 4 RCTs, N not provided; RR=1.65; 95% CI, 1.03-2.63; respectively). As with the Cochrane review, the results for all study groups demonstrated moderate to significant heterogeneity.

RECOMMENDATIONS

The Surgeon General, the American Academy of Family Physicians, the American Academy of Pediatrics, and the American College of Obstetricians and Gynecologists all recommend that women be educated about the benefits of breastfeeding and receive supportive interventions before and after delivery.^3-6

EVIDENCE SUMMARY

A 2012 Cochrane review of 52 studies (44 RCTs and 8 cluster-randomized trials; N=56,451) assessed the overall effectiveness of multiple supportive measures on decreasing cessation of “any” (partial and exclusive) and “exclusive” breastfeeding compared with usual care.¹ Participants were healthy breastfeeding mothers of healthy term babies. Support interventions were defined broadly but included individual and group interactions, as well as contact in person or over the phone by professionals or lay volunteers. Patients were approached proactively or reactively upon request, and the interventions occurred one or more times.

The interventions reduced discontinuation rates among both “exclusive” and “any” breastfeeding mothers (TABLE¹). The review found lay and professional support to be equally effective at promoting continuation of breastfeeding. Limitations include a moderate to high amount of heterogeneity, as well as the inherent difficulty of blinding subjects in the studies.

Lay support can make a significant difference in the short term

A 2008 systematic review of 38 RCTs (N=29,020) compared any counseling or behavioral intervention initiated from a clinician’s practice (office or hospital) with usual care.² The review excluded community and peer-initiated interventions. The reviewers defined breastfeeding duration as follows: initiation (up to 2 weeks), short-term (one to 3 months), intermediate-term (4 to 5 months), long-term (6 to 8 months), and prolonged (9 or more months). Investigators also analyzed breastfeeding rates by “exclusive” and “nonexclusive” (formula supplementation) regimens.

For nonexclusive breastfeeding, the review found interventions to promote breastfeeding improved rates only at initiation (18 RCTs, N=7688; relative risk [RR] for cessation of breastfeeding=1.04; 95% confidence interval [CI], 1.0-1.08; number needed to treat [NNT]=38) and in the short term (18 RCTs, N= 19,358; RR=1.10; 95% CI, 1.02-1.19; NNT=7). For exclusive breastfeeding, interventions improved rates only in the short term (17 RCTs, N=20,552; RR=1.72; 95% CI, 1.0-2.97; NNT=3).

The review found that lay support (defined as counseling or social support from peers) but not professional support was significantly associated with improving rates of both “nonexclusive” and “exclusive’ breastfeeding, but only over the short term (5 RCTs, N not provided; RR=1.22; 95% CI, 1.08-1.37; and 4 RCTs, N not provided; RR=1.65; 95% CI, 1.03-2.63; respectively). As with the Cochrane review, the results for all study groups demonstrated moderate to significant heterogeneity.

RECOMMENDATIONS

The Surgeon General, the American Academy of Family Physicians, the American Academy of Pediatrics, and the American College of Obstetricians and Gynecologists all recommend that women be educated about the benefits of breastfeeding and receive supportive interventions before and after delivery.^3-6

Should these complaints have prompted a colonoscopy?

A 45-YEAR-OLD WOMAN went to her primary care physician due to cramping abdominal pain after eating. She hadn’t seen her physician in 5 years and noted that her bowel movements were somewhat smaller than usual. Her physician suspected an ulcer and treated her with acid-reducing medication.

A month later, the patient returned with similar symptoms and said that her bowel movements were somewhat loose. The physician increased the dosage of the acid-reducing medication. The patient returned again a month later and reported constipation. The stomach issues continued and she was referred to a gynecologist. Ultimately, she went to a gastroenterologist and underwent a colonoscopy 8 months after her first visit. She was diagnosed with stage IV colon cancer with metastasis to the ovaries. The patient passed away 8 years later.

PLAINTIFF’S CLAIM The physician was negligent in failing to suspect colon cancer and perform a colonoscopy, digital rectal exam, or fecal occult blood test.

THE DEFENSE The decedent’s symptoms were inconsistent with cancer and did not indicate the need for a colonoscopy. The cancer was already advanced and the outcome would not have changed.

VERDICT $2.16 million Massachusetts verdict.

COMMENT Wow, this is a tough one! I am not at all sure I would have diagnosed this correctly. Is there a lesson here? Perhaps the history was not sufficiently thorough? Perhaps these were totally new symptoms that should have demanded a more thorough investigation? Or perhaps it would have taken 4 to 6 months for any of us to make this diagnosis in a 45-year-old woman.

Complication of pregnancy goes undetected after delivery 

A 31-YEAR-OLD WOMAN went to the emergency department (ED) complaining of tightness in her chest, difficulty breathing, and swelling in her lower legs 4 days after she delivered a child. The ED physician ruled out a pulmonary embolism and discharged her. Three days later, she returned with the same symptoms, but her legs were more swollen and her systolic blood pressure was above 160 mm Hg. She was sent home again. The woman had a seizure 4 days later. In the ambulance on the way to the hospital and following her arrival, she suffered more seizures. A few days later, she was transferred to a different facility and died soon after.

PLAINTIFF’S CLAIM The hospital and 2 ED physicians were negligent in failing to diagnose and treat postpartum preeclampsia during the ED visits. This led to the seizures, brain damage, and death. Antihypertensive and anti-seizure medications would have prevented her death.

THE DEFENSE The actions taken were reasonable, especially because the decedent had no symptoms of preeclampsia during pregnancy or delivery.

VERDICT $6.9 million Illinois settlement.

COMMENT This case speaks for itself. The physicians involved appear to have had a knowledge gap since they apparently did not consider preeclampsia in the differential. Primary care physicians and emergency physicians must be trained to recognize complications of pregnancy.

Should these complaints have prompted a colonoscopy?

A 45-YEAR-OLD WOMAN went to her primary care physician due to cramping abdominal pain after eating. She hadn’t seen her physician in 5 years and noted that her bowel movements were somewhat smaller than usual. Her physician suspected an ulcer and treated her with acid-reducing medication.

A month later, the patient returned with similar symptoms and said that her bowel movements were somewhat loose. The physician increased the dosage of the acid-reducing medication. The patient returned again a month later and reported constipation. The stomach issues continued and she was referred to a gynecologist. Ultimately, she went to a gastroenterologist and underwent a colonoscopy 8 months after her first visit. She was diagnosed with stage IV colon cancer with metastasis to the ovaries. The patient passed away 8 years later.

PLAINTIFF’S CLAIM The physician was negligent in failing to suspect colon cancer and perform a colonoscopy, digital rectal exam, or fecal occult blood test.

THE DEFENSE The decedent’s symptoms were inconsistent with cancer and did not indicate the need for a colonoscopy. The cancer was already advanced and the outcome would not have changed.

VERDICT $2.16 million Massachusetts verdict.

COMMENT Wow, this is a tough one! I am not at all sure I would have diagnosed this correctly. Is there a lesson here? Perhaps the history was not sufficiently thorough? Perhaps these were totally new symptoms that should have demanded a more thorough investigation? Or perhaps it would have taken 4 to 6 months for any of us to make this diagnosis in a 45-year-old woman.

Complication of pregnancy goes undetected after delivery 

A 31-YEAR-OLD WOMAN went to the emergency department (ED) complaining of tightness in her chest, difficulty breathing, and swelling in her lower legs 4 days after she delivered a child. The ED physician ruled out a pulmonary embolism and discharged her. Three days later, she returned with the same symptoms, but her legs were more swollen and her systolic blood pressure was above 160 mm Hg. She was sent home again. The woman had a seizure 4 days later. In the ambulance on the way to the hospital and following her arrival, she suffered more seizures. A few days later, she was transferred to a different facility and died soon after.

PLAINTIFF’S CLAIM The hospital and 2 ED physicians were negligent in failing to diagnose and treat postpartum preeclampsia during the ED visits. This led to the seizures, brain damage, and death. Antihypertensive and anti-seizure medications would have prevented her death.

THE DEFENSE The actions taken were reasonable, especially because the decedent had no symptoms of preeclampsia during pregnancy or delivery.

VERDICT $6.9 million Illinois settlement.

COMMENT This case speaks for itself. The physicians involved appear to have had a knowledge gap since they apparently did not consider preeclampsia in the differential. Primary care physicians and emergency physicians must be trained to recognize complications of pregnancy.

Should these complaints have prompted a colonoscopy?

A 45-YEAR-OLD WOMAN went to her primary care physician due to cramping abdominal pain after eating. She hadn’t seen her physician in 5 years and noted that her bowel movements were somewhat smaller than usual. Her physician suspected an ulcer and treated her with acid-reducing medication.

A month later, the patient returned with similar symptoms and said that her bowel movements were somewhat loose. The physician increased the dosage of the acid-reducing medication. The patient returned again a month later and reported constipation. The stomach issues continued and she was referred to a gynecologist. Ultimately, she went to a gastroenterologist and underwent a colonoscopy 8 months after her first visit. She was diagnosed with stage IV colon cancer with metastasis to the ovaries. The patient passed away 8 years later.

PLAINTIFF’S CLAIM The physician was negligent in failing to suspect colon cancer and perform a colonoscopy, digital rectal exam, or fecal occult blood test.

THE DEFENSE The decedent’s symptoms were inconsistent with cancer and did not indicate the need for a colonoscopy. The cancer was already advanced and the outcome would not have changed.

VERDICT $2.16 million Massachusetts verdict.

COMMENT Wow, this is a tough one! I am not at all sure I would have diagnosed this correctly. Is there a lesson here? Perhaps the history was not sufficiently thorough? Perhaps these were totally new symptoms that should have demanded a more thorough investigation? Or perhaps it would have taken 4 to 6 months for any of us to make this diagnosis in a 45-year-old woman.

Complication of pregnancy goes undetected after delivery 

A 31-YEAR-OLD WOMAN went to the emergency department (ED) complaining of tightness in her chest, difficulty breathing, and swelling in her lower legs 4 days after she delivered a child. The ED physician ruled out a pulmonary embolism and discharged her. Three days later, she returned with the same symptoms, but her legs were more swollen and her systolic blood pressure was above 160 mm Hg. She was sent home again. The woman had a seizure 4 days later. In the ambulance on the way to the hospital and following her arrival, she suffered more seizures. A few days later, she was transferred to a different facility and died soon after.

PLAINTIFF’S CLAIM The hospital and 2 ED physicians were negligent in failing to diagnose and treat postpartum preeclampsia during the ED visits. This led to the seizures, brain damage, and death. Antihypertensive and anti-seizure medications would have prevented her death.

THE DEFENSE The actions taken were reasonable, especially because the decedent had no symptoms of preeclampsia during pregnancy or delivery.

VERDICT $6.9 million Illinois settlement.

COMMENT This case speaks for itself. The physicians involved appear to have had a knowledge gap since they apparently did not consider preeclampsia in the differential. Primary care physicians and emergency physicians must be trained to recognize complications of pregnancy.

Dr. Hickner’s editorial “Let’s talk about the evidence” (J Fam Pract. 2015;64:337) struck a chord with me. I am very supportive of evidence-based medicine (EBM), but am often dismayed by the lack of humility expressed by EBM leaders, including the US Preventive Services Task Force. We have so little evidence about much of what we do in family medicine, and most evidence comes from studies that are narrow by nature (reductionist research).

Increasingly, I see patients become annoyed and critical of physicians who do not examine them.

For example, doing a physical exam is part of “laying on of hands” that is part of the art of medicine. Abraham Verghese, MD, MACP, has written and spoken about the importance of examining the patient and not just depending on data.¹ Yet elements of the physical exam, such as the pelvic exam example Dr. Hickner mentioned in his editorial, do not stand up well in EBM due to a lack of diagnostic accuracy. I’ll ask this: Who has studied the harm that may be caused by not examining our patients?

My physical exam “ritual” takes less than 10 minutes, and the value in the relationship I have with patients is more than a diagnostic exercise. Increasingly, I see patients become annoyed and critical of physicians who do not examine them.

Joseph E. Scherger, MD, MPH
Rancho Mirage, Calif

1. TED Talks. Abraham Verghese: A Doctor’s Touch. TED Web site. Available at: http://www.ted.com/talks/abraham_verghese_a_doctor_s_touch. Accessed July 20, 2015.

Author’s response:
Dr. Scherger makes an excellent point about the importance of physical touch for the doctor-patient relationship. The question is: What touching is appropriate? In my own experience, I have noticed that most—but not all—of the women I see are quite relieved that they don’t need yearly pelvic exams, and women I see for pap smears do not seem put off if I do not do a bimanual exam. The data are actually quite strong that routine pelvic exams in asymptomatic women lead to more harm than good. They uncover way too many false positives and almost no true positive findings, leading to unnecessary testing and treatment.^1,2

John Hickner, MD, MSc
Chicago, Ill

Dr. Hickner is the editor-in-chief of The Journal of Family Practice

1. Ebell MH, Culp M, Lastinger K, et al. A systematic review of the bimanual examination as a test for ovarian cancer. Am J Prev Med. 2015;48:350–356.

2. Well-woman visit. Committee Opinion No. 534. American College of Obstetricians and Gynecologists. Obstet Gynecol. 2012;120:421-424.

Dr. Hickner’s editorial “Let’s talk about the evidence” (J Fam Pract. 2015;64:337) struck a chord with me. I am very supportive of evidence-based medicine (EBM), but am often dismayed by the lack of humility expressed by EBM leaders, including the US Preventive Services Task Force. We have so little evidence about much of what we do in family medicine, and most evidence comes from studies that are narrow by nature (reductionist research).

Increasingly, I see patients become annoyed and critical of physicians who do not examine them.

For example, doing a physical exam is part of “laying on of hands” that is part of the art of medicine. Abraham Verghese, MD, MACP, has written and spoken about the importance of examining the patient and not just depending on data.¹ Yet elements of the physical exam, such as the pelvic exam example Dr. Hickner mentioned in his editorial, do not stand up well in EBM due to a lack of diagnostic accuracy. I’ll ask this: Who has studied the harm that may be caused by not examining our patients?

My physical exam “ritual” takes less than 10 minutes, and the value in the relationship I have with patients is more than a diagnostic exercise. Increasingly, I see patients become annoyed and critical of physicians who do not examine them.

Joseph E. Scherger, MD, MPH
Rancho Mirage, Calif

1. TED Talks. Abraham Verghese: A Doctor’s Touch. TED Web site. Available at: http://www.ted.com/talks/abraham_verghese_a_doctor_s_touch. Accessed July 20, 2015.

Author’s response:
Dr. Scherger makes an excellent point about the importance of physical touch for the doctor-patient relationship. The question is: What touching is appropriate? In my own experience, I have noticed that most—but not all—of the women I see are quite relieved that they don’t need yearly pelvic exams, and women I see for pap smears do not seem put off if I do not do a bimanual exam. The data are actually quite strong that routine pelvic exams in asymptomatic women lead to more harm than good. They uncover way too many false positives and almost no true positive findings, leading to unnecessary testing and treatment.^1,2

John Hickner, MD, MSc
Chicago, Ill

Dr. Hickner is the editor-in-chief of The Journal of Family Practice

1. Ebell MH, Culp M, Lastinger K, et al. A systematic review of the bimanual examination as a test for ovarian cancer. Am J Prev Med. 2015;48:350–356.

2. Well-woman visit. Committee Opinion No. 534. American College of Obstetricians and Gynecologists. Obstet Gynecol. 2012;120:421-424.

Dr. Hickner’s editorial “Let’s talk about the evidence” (J Fam Pract. 2015;64:337) struck a chord with me. I am very supportive of evidence-based medicine (EBM), but am often dismayed by the lack of humility expressed by EBM leaders, including the US Preventive Services Task Force. We have so little evidence about much of what we do in family medicine, and most evidence comes from studies that are narrow by nature (reductionist research).

Increasingly, I see patients become annoyed and critical of physicians who do not examine them.

For example, doing a physical exam is part of “laying on of hands” that is part of the art of medicine. Abraham Verghese, MD, MACP, has written and spoken about the importance of examining the patient and not just depending on data.¹ Yet elements of the physical exam, such as the pelvic exam example Dr. Hickner mentioned in his editorial, do not stand up well in EBM due to a lack of diagnostic accuracy. I’ll ask this: Who has studied the harm that may be caused by not examining our patients?

My physical exam “ritual” takes less than 10 minutes, and the value in the relationship I have with patients is more than a diagnostic exercise. Increasingly, I see patients become annoyed and critical of physicians who do not examine them.

Joseph E. Scherger, MD, MPH
Rancho Mirage, Calif

1. TED Talks. Abraham Verghese: A Doctor’s Touch. TED Web site. Available at: http://www.ted.com/talks/abraham_verghese_a_doctor_s_touch. Accessed July 20, 2015.

Author’s response:
Dr. Scherger makes an excellent point about the importance of physical touch for the doctor-patient relationship. The question is: What touching is appropriate? In my own experience, I have noticed that most—but not all—of the women I see are quite relieved that they don’t need yearly pelvic exams, and women I see for pap smears do not seem put off if I do not do a bimanual exam. The data are actually quite strong that routine pelvic exams in asymptomatic women lead to more harm than good. They uncover way too many false positives and almost no true positive findings, leading to unnecessary testing and treatment.^1,2

John Hickner, MD, MSc
Chicago, Ill

Dr. Hickner is the editor-in-chief of The Journal of Family Practice

1. Ebell MH, Culp M, Lastinger K, et al. A systematic review of the bimanual examination as a test for ovarian cancer. Am J Prev Med. 2015;48:350–356.

2. Well-woman visit. Committee Opinion No. 534. American College of Obstetricians and Gynecologists. Obstet Gynecol. 2012;120:421-424.

EVIDENCE SUMMARY

A systematic review and meta-analysis of 15 RCTs with a total enrollment of 5540 assessed the efficacy of opioids in adults with chronic low back pain of at least 12 weeks’ duration.¹ Five low-quality studies (1378 patients) that compared tramadol with placebo found tramadol to be moderately superior to placebo for relieving pain (standard mean difference [SMD]= -0.55; 95% confidence interval [CI], -0.66 to -0.44) but only modestly better for improving function (SMD= −0.18; 95% CI, -0.29 to -0.07).

Six trials with 1887 patients compared strong opioids (morphine, hydromorphone, oxycodone, oxymorphone, and tapentadol) with placebo. The opioids were better than placebo for improving pain (SMD= -0.43; 95% CI, -0.52 to -0.33) and function (SMD= -0.26; 95% CI, -0.37 to -0.15). The general interpretation of SMD effect size is 0.2=small, 0.5=medium, 0.8=large. In this case, larger negative numbers correlate with greater improvement.

How opioids stack up against NSAIDs

Two separate double-blind, double-dummy studies randomized adults with low back pain of at least 12 weeks’ duration to receive celecoxib 200 mg twice daily (404 and 398 patients, respectively) or tramadol 50 mg 4 times daily (392 and 404 patients, respectively) for 6 weeks.² The primary outcome measure was at least a 30% improvement in pain using a 0 (no pain) to 10 (worst possible pain) scale. In both studies, more patients taking celecoxib had positive responses than patients taking tramadol (63% vs 50%, P<.001, and 64% vs 55%, P<.008, respectively).

A small RCT (36 patients who had suffered low back pain for more than 6 months) randomized patients to one of 3 treatment groups for 16 weeks: oxycodone as much as 20 mg/d (13 patients); naproxen as much as 1 g/d (12 patients); or oxycodone and sustained-release morphine (titrated up to 200 mg morphine equivalent/d (11 patients).³ After 16 weeks, patients receiving oxycodone or naproxen were treated with oxycodone and sustained-release morphine for another 16 weeks, as were patients already receiving this therapy. Pain was assessed on a 0 (none) to 100 (worst possible pain) scale.

Both opioid groups had significantly less pain on average (59.8 for oxycodone, 54.9 for titrated morphine) than the naproxen group (65.5; F=16.07; P<.001) but no significant difference in activity level. However, an independent analysis of the naproxen group and titrated morphine group found no significant difference in either pain relief (SMD= -0.58; 95% CI, -1.42 to 0.26) or disability (SMD= -0.06; 95% CI, -0.88 to 0.76) between the 2 groups.⁴

How does long-term opioid use affect pain and function?

Two prospective cohort studies have evaluated long-term opioid use. The first (715 patients) used a Roland-Morris Disability Questionnaire (RMDQ) to assess disability at 6 months in patients taking opioids compared with patients not taking opioids.⁵ Patients using opioids showed an increase in RMDQ score of 1.18 units (95% CI, 0.17-2.19) on a 0 to 24 scale, with 24 representing greatest disability.

Short-term treatment with opioids provides modest relief of chronic low back pain, but only minimal improvement in function compared with placebo.

The second study evaluated pain and function in 1843 adults with acute back injuries taking opioids for a year.⁶ Pain, rated on a 0 to 10 scale, decreased from 7.7 at baseline to 6.8 at one year (no P value). At the end of the first quarter, the RMDQ score decreased from 18.8 at baseline (the end of the first quarter) to 17.5 at one year (no P value). Clinically meaningful improvement in pain and function (30% or more) occurred in 26% (95% CI, 18%-36%) and 16% (95% CI, 10%-25%) of patients, respectively.

RECOMMENDATIONS

The 2007 clinical practice guideline on low back pain from The American College of Physicians and American Pain Society recommends opioids, including tramadol, for patients with severe back pain who don’t get adequate relief from acetaminophen or NSAIDs.⁷

The 2009 National Institute for Health and Care Excellence (NICE) guidelines for early management of persistent, nonspecific low back pain recommend considering strong opioids (buprenorphine, fentanyl, and oxycodone) for short-term use in severe pain and referral to a specialist for patients requiring prolonged use of strong opioids.⁸

The 2013 British Pain Society guidelines for low back and radicular pain recommend tight restrictions on the use of strong opioids. They also recommend giving the lowest possible dose of opioids for the shortest time possible.⁹

EVIDENCE SUMMARY

A systematic review and meta-analysis of 15 RCTs with a total enrollment of 5540 assessed the efficacy of opioids in adults with chronic low back pain of at least 12 weeks’ duration.¹ Five low-quality studies (1378 patients) that compared tramadol with placebo found tramadol to be moderately superior to placebo for relieving pain (standard mean difference [SMD]= -0.55; 95% confidence interval [CI], -0.66 to -0.44) but only modestly better for improving function (SMD= −0.18; 95% CI, -0.29 to -0.07).

Six trials with 1887 patients compared strong opioids (morphine, hydromorphone, oxycodone, oxymorphone, and tapentadol) with placebo. The opioids were better than placebo for improving pain (SMD= -0.43; 95% CI, -0.52 to -0.33) and function (SMD= -0.26; 95% CI, -0.37 to -0.15). The general interpretation of SMD effect size is 0.2=small, 0.5=medium, 0.8=large. In this case, larger negative numbers correlate with greater improvement.

How opioids stack up against NSAIDs

Two separate double-blind, double-dummy studies randomized adults with low back pain of at least 12 weeks’ duration to receive celecoxib 200 mg twice daily (404 and 398 patients, respectively) or tramadol 50 mg 4 times daily (392 and 404 patients, respectively) for 6 weeks.² The primary outcome measure was at least a 30% improvement in pain using a 0 (no pain) to 10 (worst possible pain) scale. In both studies, more patients taking celecoxib had positive responses than patients taking tramadol (63% vs 50%, P<.001, and 64% vs 55%, P<.008, respectively).

A small RCT (36 patients who had suffered low back pain for more than 6 months) randomized patients to one of 3 treatment groups for 16 weeks: oxycodone as much as 20 mg/d (13 patients); naproxen as much as 1 g/d (12 patients); or oxycodone and sustained-release morphine (titrated up to 200 mg morphine equivalent/d (11 patients).³ After 16 weeks, patients receiving oxycodone or naproxen were treated with oxycodone and sustained-release morphine for another 16 weeks, as were patients already receiving this therapy. Pain was assessed on a 0 (none) to 100 (worst possible pain) scale.

Both opioid groups had significantly less pain on average (59.8 for oxycodone, 54.9 for titrated morphine) than the naproxen group (65.5; F=16.07; P<.001) but no significant difference in activity level. However, an independent analysis of the naproxen group and titrated morphine group found no significant difference in either pain relief (SMD= -0.58; 95% CI, -1.42 to 0.26) or disability (SMD= -0.06; 95% CI, -0.88 to 0.76) between the 2 groups.⁴

How does long-term opioid use affect pain and function?

Two prospective cohort studies have evaluated long-term opioid use. The first (715 patients) used a Roland-Morris Disability Questionnaire (RMDQ) to assess disability at 6 months in patients taking opioids compared with patients not taking opioids.⁵ Patients using opioids showed an increase in RMDQ score of 1.18 units (95% CI, 0.17-2.19) on a 0 to 24 scale, with 24 representing greatest disability.

Short-term treatment with opioids provides modest relief of chronic low back pain, but only minimal improvement in function compared with placebo.

The second study evaluated pain and function in 1843 adults with acute back injuries taking opioids for a year.⁶ Pain, rated on a 0 to 10 scale, decreased from 7.7 at baseline to 6.8 at one year (no P value). At the end of the first quarter, the RMDQ score decreased from 18.8 at baseline (the end of the first quarter) to 17.5 at one year (no P value). Clinically meaningful improvement in pain and function (30% or more) occurred in 26% (95% CI, 18%-36%) and 16% (95% CI, 10%-25%) of patients, respectively.

RECOMMENDATIONS

The 2007 clinical practice guideline on low back pain from The American College of Physicians and American Pain Society recommends opioids, including tramadol, for patients with severe back pain who don’t get adequate relief from acetaminophen or NSAIDs.⁷

The 2009 National Institute for Health and Care Excellence (NICE) guidelines for early management of persistent, nonspecific low back pain recommend considering strong opioids (buprenorphine, fentanyl, and oxycodone) for short-term use in severe pain and referral to a specialist for patients requiring prolonged use of strong opioids.⁸

The 2013 British Pain Society guidelines for low back and radicular pain recommend tight restrictions on the use of strong opioids. They also recommend giving the lowest possible dose of opioids for the shortest time possible.⁹

EVIDENCE SUMMARY

A systematic review and meta-analysis of 15 RCTs with a total enrollment of 5540 assessed the efficacy of opioids in adults with chronic low back pain of at least 12 weeks’ duration.¹ Five low-quality studies (1378 patients) that compared tramadol with placebo found tramadol to be moderately superior to placebo for relieving pain (standard mean difference [SMD]= -0.55; 95% confidence interval [CI], -0.66 to -0.44) but only modestly better for improving function (SMD= −0.18; 95% CI, -0.29 to -0.07).

Six trials with 1887 patients compared strong opioids (morphine, hydromorphone, oxycodone, oxymorphone, and tapentadol) with placebo. The opioids were better than placebo for improving pain (SMD= -0.43; 95% CI, -0.52 to -0.33) and function (SMD= -0.26; 95% CI, -0.37 to -0.15). The general interpretation of SMD effect size is 0.2=small, 0.5=medium, 0.8=large. In this case, larger negative numbers correlate with greater improvement.

How opioids stack up against NSAIDs

Two separate double-blind, double-dummy studies randomized adults with low back pain of at least 12 weeks’ duration to receive celecoxib 200 mg twice daily (404 and 398 patients, respectively) or tramadol 50 mg 4 times daily (392 and 404 patients, respectively) for 6 weeks.² The primary outcome measure was at least a 30% improvement in pain using a 0 (no pain) to 10 (worst possible pain) scale. In both studies, more patients taking celecoxib had positive responses than patients taking tramadol (63% vs 50%, P<.001, and 64% vs 55%, P<.008, respectively).

A small RCT (36 patients who had suffered low back pain for more than 6 months) randomized patients to one of 3 treatment groups for 16 weeks: oxycodone as much as 20 mg/d (13 patients); naproxen as much as 1 g/d (12 patients); or oxycodone and sustained-release morphine (titrated up to 200 mg morphine equivalent/d (11 patients).³ After 16 weeks, patients receiving oxycodone or naproxen were treated with oxycodone and sustained-release morphine for another 16 weeks, as were patients already receiving this therapy. Pain was assessed on a 0 (none) to 100 (worst possible pain) scale.

Both opioid groups had significantly less pain on average (59.8 for oxycodone, 54.9 for titrated morphine) than the naproxen group (65.5; F=16.07; P<.001) but no significant difference in activity level. However, an independent analysis of the naproxen group and titrated morphine group found no significant difference in either pain relief (SMD= -0.58; 95% CI, -1.42 to 0.26) or disability (SMD= -0.06; 95% CI, -0.88 to 0.76) between the 2 groups.⁴

How does long-term opioid use affect pain and function?

Two prospective cohort studies have evaluated long-term opioid use. The first (715 patients) used a Roland-Morris Disability Questionnaire (RMDQ) to assess disability at 6 months in patients taking opioids compared with patients not taking opioids.⁵ Patients using opioids showed an increase in RMDQ score of 1.18 units (95% CI, 0.17-2.19) on a 0 to 24 scale, with 24 representing greatest disability.

Short-term treatment with opioids provides modest relief of chronic low back pain, but only minimal improvement in function compared with placebo.

The second study evaluated pain and function in 1843 adults with acute back injuries taking opioids for a year.⁶ Pain, rated on a 0 to 10 scale, decreased from 7.7 at baseline to 6.8 at one year (no P value). At the end of the first quarter, the RMDQ score decreased from 18.8 at baseline (the end of the first quarter) to 17.5 at one year (no P value). Clinically meaningful improvement in pain and function (30% or more) occurred in 26% (95% CI, 18%-36%) and 16% (95% CI, 10%-25%) of patients, respectively.

RECOMMENDATIONS

The 2007 clinical practice guideline on low back pain from The American College of Physicians and American Pain Society recommends opioids, including tramadol, for patients with severe back pain who don’t get adequate relief from acetaminophen or NSAIDs.⁷

The 2009 National Institute for Health and Care Excellence (NICE) guidelines for early management of persistent, nonspecific low back pain recommend considering strong opioids (buprenorphine, fentanyl, and oxycodone) for short-term use in severe pain and referral to a specialist for patients requiring prolonged use of strong opioids.⁸

The 2013 British Pain Society guidelines for low back and radicular pain recommend tight restrictions on the use of strong opioids. They also recommend giving the lowest possible dose of opioids for the shortest time possible.⁹

Finger injuries are common, representing an estimated 3 million emergency department visits per year in the United States, with 44% of these diagnosed as lacerations.¹ Amputations of the finger (partial and complete) in non-work-related accidents alone are estimated at 30,000 per year.¹ The fingertip is a highly specialized structure that contributes to precision function of the hand through tactile feedback and fine motor control as well as hand aesthetics. An injury can compromise a variety of fingertip structures, including the distal phalanx, which provides length and structural support; the fingernail, germinal matrix, and sterile matrix, which protect the fingertip and function as tools; and the volar skin pad, which is important for sensation and fine motor activity.

There is considerable debate regarding optimal management of fingertip amputations, and to date there have been no prospective, randomly controlled trials to guide treatment.² Injury characteristics, amputation levels, and patient priorities all contribute to management decisions. Treatment goals are to maintain length when possible; to provide stable, supple, and sensate skin coverage; to ensure the nail plate regrows without complication; and to maintain normal overall finger shape and cosmesis. In addition, a simple, cost-effective treatment with short recovery time and no donor-site morbidity is desired.

Treatment recommendations are wide-ranging, and evidence-based literature is sparse. About 30 years ago, 2 retrospective comparative studies found no difference in outcomes between simpler treatments (primary closure, secondary wound healing) and various operative strategies.^3,4 Since then, most of the scientific studies have been retrospective noncomparative case series, all reporting good to excellent results.^5-17 Investigators generally implied superior results of a studied procedure over those of more conservative treatments. Recommended treatments include secondary wound healing, simple flaps, staged flaps, pedicle flaps, allograft and autograft coverage, composite grafting, and replantation, for all levels of fingertip injury.

Given our surgical advances, improved techniques, and accumulating experience, we may have expected better outcomes with newer and more complex reconstructive efforts. Unfortunately, in a recent review of 53 fingertip injuries treated with a reconstructive procedure, bone shortening with closure, or secondary healing, Wang and colleagues¹⁸ found no discernible differences in outcomes at 4.5-year follow-up. They questioned whether complex reconstructive procedures are worth the time, expense, and risk. In the absence of prospective, comparative studies, surgeons must rely on anecdotal evidence (including predominantly level IV evidence), training bias, previous experience, and the prevailing common wisdom.

Toward that end, we became interested in identifying treatment preferences for fingertip amputations. We conducted a study to better understand how surgeon and patient factors influence the treatment preferences for distal fingertip amputations among a cross section of US and international hand surgeons. We hypothesized that hand surgeons’ treatment preferences would be varied and influenced by surgeon and patient demographics.

Materials and Methods

An online multiple-choice survey was created and powered by Constant Contact. The survey consisted of 6 surgeon demographic questions; 5 treatment preference questions regarding patient age, sex, occupation, and germinal matrix management; and 5 clinical scenarios based on Allen levels 2, 3 (with and without exposed distal phalanx), and 4 and volar oblique middle-finger amputations. The Allen classification designates level 2 injuries as those involving only the distal pulp and nail.¹⁹  Level 3 injuries also involve the terminal distal phalanx, and level 4 injuries extend to the lunula. The survey questions are listed in the Appendix. For the clinical scenario questions, treatment choices included wound care, skeletal shortening and closure, composite graft, autograft, allograft, V-Y/Kutler flap, advancement flap, thenar flap, cross-finger flap, pedicle and homodigital flap, replantation, and other.

An email invitation was sent to members of the American Association for Hand Surgery (AAHS). The survey was also submitted to personal contacts of international hand societies named on the AAHS website to expand the international response. A reminder email was sent 1 week after the original invitation. The survey was closed 5 weeks later, and the responses were analyzed with all non-US hand surgeons grouped collectively as an international group, compared with the US group. Institutional review board approval was not needed for this survey study.

Statistics

A generalized linear regression model was used to implement logistic regression with random effects for question and respondent. This approach accounts for multiple observations from the same respondent, assuming that both respondent and question are random samples from a larger population. The model estimated the probability that a given surgical approach (eg, skeletal shortening, wound care) would be selected, based on the predictors of the US versus international respondent, time in practice, practice type, and whether the fingertip was available. The model returned adjusted odds ratios (ORs) for each predictor, controlling for all the others. By convention, P < .05 was considered significant. No attempt was made to prune the model of nonsignificant factors. Analyses were performed using the lme4 package on the R statistical platform (R Foundation for Statistical Computing).

Results

One hundred ninety-eight responses were recorded. Of the 1054 AAHS members invited to take the survey, 174 (US, international) responded (17% response rate). One hundred twenty-three responses and 62% of the total were generated from US hand surgeons. Fifty-eight percent of US responses were from the Mid-South, Midwest, or Mid-Atlantic region. Fifty-seven percent of international responses were from Brazil and Europe. Respondents’ demographic data are listed in Tables 1 and 2.

Responses to the 5 clinical scenarios showed a wide variation in treatment preferences. The top 6 preferred treatment selections for an acute, clean long-finger amputation in a healthy 40-year-old office worker are shown in Figures 1 to 5. When surgeons who preferred replant were asked what they would do if the amputated part was not available, they indicated flap coverage more often than less complex treatments, such as skeletal shortening/primary closure or wound care.

There were statistically significant differences in treatment preferences between US and international hand surgeons when controlling for all other demographic variables. Adjusted ORs and their confidence intervals (CIs) for the aggregate clinical scenarios are presented in a forest plot in Figure 6. Figure 4 shows that US surgeons were more likely to choose wound care (OR, 3.6; P < .0004) and less likely to attempt a replant (OR, 0.01; P < .0001). US surgeons were also less likely to use a pedicle or homodigital island flap when the amputated fingertip was both available (OR, 0.04; P = .039) and unavailable (OR, 0.47; Ps = .029).

Among all respondents and across all clinical scenarios, skeletal shortening with closure was favored among hand surgeons in practice less than 5 years compared with those in practice longer (OR, 2.11; 95% CI, 1.36-3.25; P = .0008). Similarly, surgeons with more than 30 years of experience were the least likely to favor wound care (OR, 0.2; 95% CI, 0.09-0.93; P = .037). Compared with orthopedic surgeons, plastic surgeons opted for wound care less often (OR, 0.44; 95% CI, 0.23-0.98; P = .018) and appeared to prefer replantation, but the difference was not statistically significant (OR, 8.86; 95% CI, 0.99-79.61; P = .054).

Replantation was less often chosen by private practice versus full-time academic surgeons (OR, 0.09; 95% CI, 0.01-0.91; P = .041.) Part-time academics were no more or less likely to perform replantation than full-time academics were (OR, 0.52; 95% CI, 0.05-5.41; P = .58). Of the 59 respondents who performed more than 10 microvascular cases a year, 18 (31%) chose replant for Allen level 4 amputations. In comparison, 9 (20%) of the 45 respondents who performed fewer than 3 microvascular cases a year chose replant for amputations at this level. Amount of time working with fellows did not affect treatment preferences.

Patient demographics (age, sex, occupation) also played a role in treatment decisions (Table 3). The most significant factors appeared to be age and occupation. Regarding age, 41% of respondents chose more complex procedures for patients younger than 15, and 62% chose less complex procedures for patients older than 70 years. Regarding occupation, 61% chose more complex procedures for professional musicians, and 60% chose less complex procedures for manual laborers. Sex did not influence clinical decisions for 78% of respondents. There was also substantial variation in both the indications for germinal matrix ablation and the frequency of sterile matrix transplant (Table 3).

Discussion

Although there is a variety of treatment options and published treatment guidelines for distal fingertip amputations, few comparative studies support use of one treatment over another. In our experience, treatment decisions are based mainly on injury parameters, but surgeon preference and patient factors (age, sex, occupation) can also influence care. Our goal in this study was to better understand how surgeon and patient factors influence treatment preferences for distal fingertip amputations among a cross section of US and international hand surgeons. Our survey results showed lack of consensus among hand surgeons and highlighted several trends.

As expected, we found a wide range of treatment preferences for each clinical scenario queried, ranging from more simple treatments (eg, wound care) to more complex ones (eg, replantation). With patient parameters (age, profession, finger, acuity, injury type, tissue preservation, smoking status) standardized in the clinical scenarios, the treatment differences noted should reflect surgeon preference. However, other patient factors (eg, cultural differences, religious beliefs, surgeon setting, practice pattern, resource availability) that were not included in the clinical scenarios could also affect treatment preferences.

One particularly interesting finding was that international hand surgeons were 6.8 times more likely to replant a distal fingertip amputation. One possible explanation for this variation is the influence of cultural differences. For example, in East Asian countries, there can be a cultural stigma associated with loss of a fingertip, and therefore more of a desire on the part of the patient to restore the original finger.^20,21 In addition, the international respondents were biased toward academic practices—which could skew the treatment preference toward replantation, as we found that academic surgeons were more inclined to replantation.

Our finding that replantation was more commonly preferred by academic versus private practice surgeons may suggest a training bias, an affinity for more complex or interesting procedures, or access to hospital equipment and staff, including residents and fellows, not usually found at smaller community hospitals, where private practice surgeons are more commonly based. Jazayeri and colleagues²² found that institutions specializing in microsurgery often produced better outcomes than nonspecializing institutions. Therefore, it is not surprising that private practice hand surgeons may less often opt to replant a distal fingertip amputation. It is also not surprising that plastic surgeons are more inclined to perform a replantation or flap coverage, as their training is more microsurgery-intensive and their practice more focused on aesthetics compared with the other specialists.

Distal fingertip replantation is accepted by most as technically demanding, but it seems that the additional effort and resources would be justified if the procedure provided a superior outcome. However, other factors, such as cost of treatment and length of recovery, should also be considered. Average replantation cost has been estimated to range from $7500 to $14,000, compared with $2800 for non-replantation-related care, and median stay is about 4 days longer for replantation-related care.^23,24 These estimates do not include indirect costs, such as for postoperative rehabilitation, which is likely longer and more expensive, even in distal fingertip replantation. These disparities may not justify the outcome (of having a complete fingertip) if more conservative treatments yield similar results.^17,18 In addition, there is the expected failure rate of limb replantation surgery. In analysis of the overall societal costs and benefits of larger upper extremity limb replantation, the loss of invested resources sustained with failed limb replantation may be outweighed by the benefit of another patient having a successful outcome. In the case of fingertip replantation, however, does the undefined benefit of the successful patient outcome outweigh the investment of resources lost in cases of replantation failure? Understandably, there is a need for more robust clinical outcome and cost-comparative evidence to better inform decisions regarding distal fingertip amputation.

We found that wound care and skeletal shortening with primary closure (particularly with Allen level 3 injuries) were preferred more by surgeons within the first 5 years of practice. This finding seems to imply a lack of experience or confidence on the part of younger surgeons performing more complex procedures, such as flap coverage. Conversely, this finding may indicate a shift in treatment principle based on recent literature suggesting equivalent outcomes with simpler procedures.^17,18 Although our survey study did not provide an option for treatment combinations or staged procedures, several respondents wrote in that skeletal shortening supplemented with various types of autografts and allografts would be their preferred treatment.

Patient factors also play a significant role in clinical decisions. Age and profession seem to be important determinants, with more than 50% of respondents, on average, changing their treatment recommendation based on these 2 factors. A majority of respondents would perform a less involved procedure for a manual laborer, suggesting a quicker return to work is prioritized over a perceived improved clinical outcome. Interestingly, for patients younger than 15 years, the preference was divided, with 41% of surgeons opting for a more complex procedure. This suggests the importance of restoring anatomy in a younger patient, or the perceived decreased risk or failure rate with more involved treatment. Twenty percent preferred a less complex procedure in a younger patient, perhaps relying on the patient’s developmental potential for a good outcome or suggesting a concern for patient intolerance or compliance with complex surgery.

Nail plate regrowth can be a problem with fingertip amputations. Nail deformity is highly correlated with injury level, with amputations proximal to the lunula more likely to cause nail plate deformity.^25,26 Jebson and colleagues²⁷ recommended germinal matrix ablation for amputations proximal to the lunula. We found respondents often performed ablations for other indications, including injured or minimal remaining sterile matrix and lack of bony support for the sterile matrix. Forty-six percent of respondents had never performed sterile matrix transplant, which could indicate that they were unfamiliar with the technique or had donor-site concerns, or that postinjury nail deformities are uncommon, well tolerated, or treated along with other procedures, such as germinal matrix ablation.

Several weaknesses of this study must be highlighted. First, our response rate was smaller than desired. Although this work incorporated a large number of surgeon responses, nearly 200, the response rate was only 17%. In addition, although number of responses was likely adequate to show the diversity of opinion, the preferences and trends reported might not be representative of all hand surgeons. We could not perform a nonresponder analysis because of a lack of specific demographic data for the AAHS and international hand society members. However, AAHS has an approximate 50/50 mix of plastic and orthopedic surgeons, similar to our responder demographic, suggesting our smaller subset of responses might be representative of the whole. According to AAHS, a majority of its members are “academic” hand surgeons, so our results might not adequately reflect the preferences of community hand surgeons and ultimately might overstate the frequency of more complex treatments. Last, our international response was limited to a few countries. A larger, more broadly distributed response would provide a better understanding of regional preferences, which could shed light on the importance of cultural differences.

Variations in patient insurance status were not queried in this survey but might also affect treatment decisions. More involved, costly, and highly reimbursing procedures might be deemed reasonable options for a small perceived clinical benefit for insured patients.

When multiple digits or the thumb is injured, or there are other concomitant injuries, surgeons may alter their choice of intervention. In mangled extremities, preservation of salvageable functional units takes precedence over aesthetics and likely affects choice of treatment for the amputated fingertips. Similarly, multiple fingertip amputations, even if all at the same level, may be differently regarded than a solitary injury.

Conclusion

For distal fingertip amputations, there is little evidence supporting one approach over another. Without level I comparative data guiding treatment, anecdotal evidence and surgeon personal preferences likely contribute to the large variation noted in this survey. Our study results showed the disparity of fingertip treatment preferences among a cross section of US and international hand surgeons. More important, results underscored the need for a well-designed comparative study to determine the most effective treatments for distal fingertip amputations.

Finger injuries are common, representing an estimated 3 million emergency department visits per year in the United States, with 44% of these diagnosed as lacerations.¹ Amputations of the finger (partial and complete) in non-work-related accidents alone are estimated at 30,000 per year.¹ The fingertip is a highly specialized structure that contributes to precision function of the hand through tactile feedback and fine motor control as well as hand aesthetics. An injury can compromise a variety of fingertip structures, including the distal phalanx, which provides length and structural support; the fingernail, germinal matrix, and sterile matrix, which protect the fingertip and function as tools; and the volar skin pad, which is important for sensation and fine motor activity.

There is considerable debate regarding optimal management of fingertip amputations, and to date there have been no prospective, randomly controlled trials to guide treatment.² Injury characteristics, amputation levels, and patient priorities all contribute to management decisions. Treatment goals are to maintain length when possible; to provide stable, supple, and sensate skin coverage; to ensure the nail plate regrows without complication; and to maintain normal overall finger shape and cosmesis. In addition, a simple, cost-effective treatment with short recovery time and no donor-site morbidity is desired.

Treatment recommendations are wide-ranging, and evidence-based literature is sparse. About 30 years ago, 2 retrospective comparative studies found no difference in outcomes between simpler treatments (primary closure, secondary wound healing) and various operative strategies.^3,4 Since then, most of the scientific studies have been retrospective noncomparative case series, all reporting good to excellent results.^5-17 Investigators generally implied superior results of a studied procedure over those of more conservative treatments. Recommended treatments include secondary wound healing, simple flaps, staged flaps, pedicle flaps, allograft and autograft coverage, composite grafting, and replantation, for all levels of fingertip injury.

Given our surgical advances, improved techniques, and accumulating experience, we may have expected better outcomes with newer and more complex reconstructive efforts. Unfortunately, in a recent review of 53 fingertip injuries treated with a reconstructive procedure, bone shortening with closure, or secondary healing, Wang and colleagues¹⁸ found no discernible differences in outcomes at 4.5-year follow-up. They questioned whether complex reconstructive procedures are worth the time, expense, and risk. In the absence of prospective, comparative studies, surgeons must rely on anecdotal evidence (including predominantly level IV evidence), training bias, previous experience, and the prevailing common wisdom.

Toward that end, we became interested in identifying treatment preferences for fingertip amputations. We conducted a study to better understand how surgeon and patient factors influence the treatment preferences for distal fingertip amputations among a cross section of US and international hand surgeons. We hypothesized that hand surgeons’ treatment preferences would be varied and influenced by surgeon and patient demographics.

Materials and Methods

An online multiple-choice survey was created and powered by Constant Contact. The survey consisted of 6 surgeon demographic questions; 5 treatment preference questions regarding patient age, sex, occupation, and germinal matrix management; and 5 clinical scenarios based on Allen levels 2, 3 (with and without exposed distal phalanx), and 4 and volar oblique middle-finger amputations. The Allen classification designates level 2 injuries as those involving only the distal pulp and nail.¹⁹  Level 3 injuries also involve the terminal distal phalanx, and level 4 injuries extend to the lunula. The survey questions are listed in the Appendix. For the clinical scenario questions, treatment choices included wound care, skeletal shortening and closure, composite graft, autograft, allograft, V-Y/Kutler flap, advancement flap, thenar flap, cross-finger flap, pedicle and homodigital flap, replantation, and other.

An email invitation was sent to members of the American Association for Hand Surgery (AAHS). The survey was also submitted to personal contacts of international hand societies named on the AAHS website to expand the international response. A reminder email was sent 1 week after the original invitation. The survey was closed 5 weeks later, and the responses were analyzed with all non-US hand surgeons grouped collectively as an international group, compared with the US group. Institutional review board approval was not needed for this survey study.

Statistics

A generalized linear regression model was used to implement logistic regression with random effects for question and respondent. This approach accounts for multiple observations from the same respondent, assuming that both respondent and question are random samples from a larger population. The model estimated the probability that a given surgical approach (eg, skeletal shortening, wound care) would be selected, based on the predictors of the US versus international respondent, time in practice, practice type, and whether the fingertip was available. The model returned adjusted odds ratios (ORs) for each predictor, controlling for all the others. By convention, P < .05 was considered significant. No attempt was made to prune the model of nonsignificant factors. Analyses were performed using the lme4 package on the R statistical platform (R Foundation for Statistical Computing).

Results

One hundred ninety-eight responses were recorded. Of the 1054 AAHS members invited to take the survey, 174 (US, international) responded (17% response rate). One hundred twenty-three responses and 62% of the total were generated from US hand surgeons. Fifty-eight percent of US responses were from the Mid-South, Midwest, or Mid-Atlantic region. Fifty-seven percent of international responses were from Brazil and Europe. Respondents’ demographic data are listed in Tables 1 and 2.

Responses to the 5 clinical scenarios showed a wide variation in treatment preferences. The top 6 preferred treatment selections for an acute, clean long-finger amputation in a healthy 40-year-old office worker are shown in Figures 1 to 5. When surgeons who preferred replant were asked what they would do if the amputated part was not available, they indicated flap coverage more often than less complex treatments, such as skeletal shortening/primary closure or wound care.

There were statistically significant differences in treatment preferences between US and international hand surgeons when controlling for all other demographic variables. Adjusted ORs and their confidence intervals (CIs) for the aggregate clinical scenarios are presented in a forest plot in Figure 6. Figure 4 shows that US surgeons were more likely to choose wound care (OR, 3.6; P < .0004) and less likely to attempt a replant (OR, 0.01; P < .0001). US surgeons were also less likely to use a pedicle or homodigital island flap when the amputated fingertip was both available (OR, 0.04; P = .039) and unavailable (OR, 0.47; Ps = .029).

Among all respondents and across all clinical scenarios, skeletal shortening with closure was favored among hand surgeons in practice less than 5 years compared with those in practice longer (OR, 2.11; 95% CI, 1.36-3.25; P = .0008). Similarly, surgeons with more than 30 years of experience were the least likely to favor wound care (OR, 0.2; 95% CI, 0.09-0.93; P = .037). Compared with orthopedic surgeons, plastic surgeons opted for wound care less often (OR, 0.44; 95% CI, 0.23-0.98; P = .018) and appeared to prefer replantation, but the difference was not statistically significant (OR, 8.86; 95% CI, 0.99-79.61; P = .054).

Replantation was less often chosen by private practice versus full-time academic surgeons (OR, 0.09; 95% CI, 0.01-0.91; P = .041.) Part-time academics were no more or less likely to perform replantation than full-time academics were (OR, 0.52; 95% CI, 0.05-5.41; P = .58). Of the 59 respondents who performed more than 10 microvascular cases a year, 18 (31%) chose replant for Allen level 4 amputations. In comparison, 9 (20%) of the 45 respondents who performed fewer than 3 microvascular cases a year chose replant for amputations at this level. Amount of time working with fellows did not affect treatment preferences.

Patient demographics (age, sex, occupation) also played a role in treatment decisions (Table 3). The most significant factors appeared to be age and occupation. Regarding age, 41% of respondents chose more complex procedures for patients younger than 15, and 62% chose less complex procedures for patients older than 70 years. Regarding occupation, 61% chose more complex procedures for professional musicians, and 60% chose less complex procedures for manual laborers. Sex did not influence clinical decisions for 78% of respondents. There was also substantial variation in both the indications for germinal matrix ablation and the frequency of sterile matrix transplant (Table 3).

Discussion

Although there is a variety of treatment options and published treatment guidelines for distal fingertip amputations, few comparative studies support use of one treatment over another. In our experience, treatment decisions are based mainly on injury parameters, but surgeon preference and patient factors (age, sex, occupation) can also influence care. Our goal in this study was to better understand how surgeon and patient factors influence treatment preferences for distal fingertip amputations among a cross section of US and international hand surgeons. Our survey results showed lack of consensus among hand surgeons and highlighted several trends.

As expected, we found a wide range of treatment preferences for each clinical scenario queried, ranging from more simple treatments (eg, wound care) to more complex ones (eg, replantation). With patient parameters (age, profession, finger, acuity, injury type, tissue preservation, smoking status) standardized in the clinical scenarios, the treatment differences noted should reflect surgeon preference. However, other patient factors (eg, cultural differences, religious beliefs, surgeon setting, practice pattern, resource availability) that were not included in the clinical scenarios could also affect treatment preferences.

One particularly interesting finding was that international hand surgeons were 6.8 times more likely to replant a distal fingertip amputation. One possible explanation for this variation is the influence of cultural differences. For example, in East Asian countries, there can be a cultural stigma associated with loss of a fingertip, and therefore more of a desire on the part of the patient to restore the original finger.^20,21 In addition, the international respondents were biased toward academic practices—which could skew the treatment preference toward replantation, as we found that academic surgeons were more inclined to replantation.

Our finding that replantation was more commonly preferred by academic versus private practice surgeons may suggest a training bias, an affinity for more complex or interesting procedures, or access to hospital equipment and staff, including residents and fellows, not usually found at smaller community hospitals, where private practice surgeons are more commonly based. Jazayeri and colleagues²² found that institutions specializing in microsurgery often produced better outcomes than nonspecializing institutions. Therefore, it is not surprising that private practice hand surgeons may less often opt to replant a distal fingertip amputation. It is also not surprising that plastic surgeons are more inclined to perform a replantation or flap coverage, as their training is more microsurgery-intensive and their practice more focused on aesthetics compared with the other specialists.

Distal fingertip replantation is accepted by most as technically demanding, but it seems that the additional effort and resources would be justified if the procedure provided a superior outcome. However, other factors, such as cost of treatment and length of recovery, should also be considered. Average replantation cost has been estimated to range from $7500 to $14,000, compared with $2800 for non-replantation-related care, and median stay is about 4 days longer for replantation-related care.^23,24 These estimates do not include indirect costs, such as for postoperative rehabilitation, which is likely longer and more expensive, even in distal fingertip replantation. These disparities may not justify the outcome (of having a complete fingertip) if more conservative treatments yield similar results.^17,18 In addition, there is the expected failure rate of limb replantation surgery. In analysis of the overall societal costs and benefits of larger upper extremity limb replantation, the loss of invested resources sustained with failed limb replantation may be outweighed by the benefit of another patient having a successful outcome. In the case of fingertip replantation, however, does the undefined benefit of the successful patient outcome outweigh the investment of resources lost in cases of replantation failure? Understandably, there is a need for more robust clinical outcome and cost-comparative evidence to better inform decisions regarding distal fingertip amputation.

We found that wound care and skeletal shortening with primary closure (particularly with Allen level 3 injuries) were preferred more by surgeons within the first 5 years of practice. This finding seems to imply a lack of experience or confidence on the part of younger surgeons performing more complex procedures, such as flap coverage. Conversely, this finding may indicate a shift in treatment principle based on recent literature suggesting equivalent outcomes with simpler procedures.^17,18 Although our survey study did not provide an option for treatment combinations or staged procedures, several respondents wrote in that skeletal shortening supplemented with various types of autografts and allografts would be their preferred treatment.

Patient factors also play a significant role in clinical decisions. Age and profession seem to be important determinants, with more than 50% of respondents, on average, changing their treatment recommendation based on these 2 factors. A majority of respondents would perform a less involved procedure for a manual laborer, suggesting a quicker return to work is prioritized over a perceived improved clinical outcome. Interestingly, for patients younger than 15 years, the preference was divided, with 41% of surgeons opting for a more complex procedure. This suggests the importance of restoring anatomy in a younger patient, or the perceived decreased risk or failure rate with more involved treatment. Twenty percent preferred a less complex procedure in a younger patient, perhaps relying on the patient’s developmental potential for a good outcome or suggesting a concern for patient intolerance or compliance with complex surgery.

Nail plate regrowth can be a problem with fingertip amputations. Nail deformity is highly correlated with injury level, with amputations proximal to the lunula more likely to cause nail plate deformity.^25,26 Jebson and colleagues²⁷ recommended germinal matrix ablation for amputations proximal to the lunula. We found respondents often performed ablations for other indications, including injured or minimal remaining sterile matrix and lack of bony support for the sterile matrix. Forty-six percent of respondents had never performed sterile matrix transplant, which could indicate that they were unfamiliar with the technique or had donor-site concerns, or that postinjury nail deformities are uncommon, well tolerated, or treated along with other procedures, such as germinal matrix ablation.

Several weaknesses of this study must be highlighted. First, our response rate was smaller than desired. Although this work incorporated a large number of surgeon responses, nearly 200, the response rate was only 17%. In addition, although number of responses was likely adequate to show the diversity of opinion, the preferences and trends reported might not be representative of all hand surgeons. We could not perform a nonresponder analysis because of a lack of specific demographic data for the AAHS and international hand society members. However, AAHS has an approximate 50/50 mix of plastic and orthopedic surgeons, similar to our responder demographic, suggesting our smaller subset of responses might be representative of the whole. According to AAHS, a majority of its members are “academic” hand surgeons, so our results might not adequately reflect the preferences of community hand surgeons and ultimately might overstate the frequency of more complex treatments. Last, our international response was limited to a few countries. A larger, more broadly distributed response would provide a better understanding of regional preferences, which could shed light on the importance of cultural differences.

Variations in patient insurance status were not queried in this survey but might also affect treatment decisions. More involved, costly, and highly reimbursing procedures might be deemed reasonable options for a small perceived clinical benefit for insured patients.

When multiple digits or the thumb is injured, or there are other concomitant injuries, surgeons may alter their choice of intervention. In mangled extremities, preservation of salvageable functional units takes precedence over aesthetics and likely affects choice of treatment for the amputated fingertips. Similarly, multiple fingertip amputations, even if all at the same level, may be differently regarded than a solitary injury.

Conclusion

For distal fingertip amputations, there is little evidence supporting one approach over another. Without level I comparative data guiding treatment, anecdotal evidence and surgeon personal preferences likely contribute to the large variation noted in this survey. Our study results showed the disparity of fingertip treatment preferences among a cross section of US and international hand surgeons. More important, results underscored the need for a well-designed comparative study to determine the most effective treatments for distal fingertip amputations.

Finger injuries are common, representing an estimated 3 million emergency department visits per year in the United States, with 44% of these diagnosed as lacerations.¹ Amputations of the finger (partial and complete) in non-work-related accidents alone are estimated at 30,000 per year.¹ The fingertip is a highly specialized structure that contributes to precision function of the hand through tactile feedback and fine motor control as well as hand aesthetics. An injury can compromise a variety of fingertip structures, including the distal phalanx, which provides length and structural support; the fingernail, germinal matrix, and sterile matrix, which protect the fingertip and function as tools; and the volar skin pad, which is important for sensation and fine motor activity.

There is considerable debate regarding optimal management of fingertip amputations, and to date there have been no prospective, randomly controlled trials to guide treatment.² Injury characteristics, amputation levels, and patient priorities all contribute to management decisions. Treatment goals are to maintain length when possible; to provide stable, supple, and sensate skin coverage; to ensure the nail plate regrows without complication; and to maintain normal overall finger shape and cosmesis. In addition, a simple, cost-effective treatment with short recovery time and no donor-site morbidity is desired.

Treatment recommendations are wide-ranging, and evidence-based literature is sparse. About 30 years ago, 2 retrospective comparative studies found no difference in outcomes between simpler treatments (primary closure, secondary wound healing) and various operative strategies.^3,4 Since then, most of the scientific studies have been retrospective noncomparative case series, all reporting good to excellent results.^5-17 Investigators generally implied superior results of a studied procedure over those of more conservative treatments. Recommended treatments include secondary wound healing, simple flaps, staged flaps, pedicle flaps, allograft and autograft coverage, composite grafting, and replantation, for all levels of fingertip injury.

Given our surgical advances, improved techniques, and accumulating experience, we may have expected better outcomes with newer and more complex reconstructive efforts. Unfortunately, in a recent review of 53 fingertip injuries treated with a reconstructive procedure, bone shortening with closure, or secondary healing, Wang and colleagues¹⁸ found no discernible differences in outcomes at 4.5-year follow-up. They questioned whether complex reconstructive procedures are worth the time, expense, and risk. In the absence of prospective, comparative studies, surgeons must rely on anecdotal evidence (including predominantly level IV evidence), training bias, previous experience, and the prevailing common wisdom.

Toward that end, we became interested in identifying treatment preferences for fingertip amputations. We conducted a study to better understand how surgeon and patient factors influence the treatment preferences for distal fingertip amputations among a cross section of US and international hand surgeons. We hypothesized that hand surgeons’ treatment preferences would be varied and influenced by surgeon and patient demographics.

Materials and Methods

An online multiple-choice survey was created and powered by Constant Contact. The survey consisted of 6 surgeon demographic questions; 5 treatment preference questions regarding patient age, sex, occupation, and germinal matrix management; and 5 clinical scenarios based on Allen levels 2, 3 (with and without exposed distal phalanx), and 4 and volar oblique middle-finger amputations. The Allen classification designates level 2 injuries as those involving only the distal pulp and nail.¹⁹  Level 3 injuries also involve the terminal distal phalanx, and level 4 injuries extend to the lunula. The survey questions are listed in the Appendix. For the clinical scenario questions, treatment choices included wound care, skeletal shortening and closure, composite graft, autograft, allograft, V-Y/Kutler flap, advancement flap, thenar flap, cross-finger flap, pedicle and homodigital flap, replantation, and other.

An email invitation was sent to members of the American Association for Hand Surgery (AAHS). The survey was also submitted to personal contacts of international hand societies named on the AAHS website to expand the international response. A reminder email was sent 1 week after the original invitation. The survey was closed 5 weeks later, and the responses were analyzed with all non-US hand surgeons grouped collectively as an international group, compared with the US group. Institutional review board approval was not needed for this survey study.

Statistics

A generalized linear regression model was used to implement logistic regression with random effects for question and respondent. This approach accounts for multiple observations from the same respondent, assuming that both respondent and question are random samples from a larger population. The model estimated the probability that a given surgical approach (eg, skeletal shortening, wound care) would be selected, based on the predictors of the US versus international respondent, time in practice, practice type, and whether the fingertip was available. The model returned adjusted odds ratios (ORs) for each predictor, controlling for all the others. By convention, P < .05 was considered significant. No attempt was made to prune the model of nonsignificant factors. Analyses were performed using the lme4 package on the R statistical platform (R Foundation for Statistical Computing).

Results

One hundred ninety-eight responses were recorded. Of the 1054 AAHS members invited to take the survey, 174 (US, international) responded (17% response rate). One hundred twenty-three responses and 62% of the total were generated from US hand surgeons. Fifty-eight percent of US responses were from the Mid-South, Midwest, or Mid-Atlantic region. Fifty-seven percent of international responses were from Brazil and Europe. Respondents’ demographic data are listed in Tables 1 and 2.

Responses to the 5 clinical scenarios showed a wide variation in treatment preferences. The top 6 preferred treatment selections for an acute, clean long-finger amputation in a healthy 40-year-old office worker are shown in Figures 1 to 5. When surgeons who preferred replant were asked what they would do if the amputated part was not available, they indicated flap coverage more often than less complex treatments, such as skeletal shortening/primary closure or wound care.

There were statistically significant differences in treatment preferences between US and international hand surgeons when controlling for all other demographic variables. Adjusted ORs and their confidence intervals (CIs) for the aggregate clinical scenarios are presented in a forest plot in Figure 6. Figure 4 shows that US surgeons were more likely to choose wound care (OR, 3.6; P < .0004) and less likely to attempt a replant (OR, 0.01; P < .0001). US surgeons were also less likely to use a pedicle or homodigital island flap when the amputated fingertip was both available (OR, 0.04; P = .039) and unavailable (OR, 0.47; Ps = .029).

Among all respondents and across all clinical scenarios, skeletal shortening with closure was favored among hand surgeons in practice less than 5 years compared with those in practice longer (OR, 2.11; 95% CI, 1.36-3.25; P = .0008). Similarly, surgeons with more than 30 years of experience were the least likely to favor wound care (OR, 0.2; 95% CI, 0.09-0.93; P = .037). Compared with orthopedic surgeons, plastic surgeons opted for wound care less often (OR, 0.44; 95% CI, 0.23-0.98; P = .018) and appeared to prefer replantation, but the difference was not statistically significant (OR, 8.86; 95% CI, 0.99-79.61; P = .054).

Replantation was less often chosen by private practice versus full-time academic surgeons (OR, 0.09; 95% CI, 0.01-0.91; P = .041.) Part-time academics were no more or less likely to perform replantation than full-time academics were (OR, 0.52; 95% CI, 0.05-5.41; P = .58). Of the 59 respondents who performed more than 10 microvascular cases a year, 18 (31%) chose replant for Allen level 4 amputations. In comparison, 9 (20%) of the 45 respondents who performed fewer than 3 microvascular cases a year chose replant for amputations at this level. Amount of time working with fellows did not affect treatment preferences.

Patient demographics (age, sex, occupation) also played a role in treatment decisions (Table 3). The most significant factors appeared to be age and occupation. Regarding age, 41% of respondents chose more complex procedures for patients younger than 15, and 62% chose less complex procedures for patients older than 70 years. Regarding occupation, 61% chose more complex procedures for professional musicians, and 60% chose less complex procedures for manual laborers. Sex did not influence clinical decisions for 78% of respondents. There was also substantial variation in both the indications for germinal matrix ablation and the frequency of sterile matrix transplant (Table 3).

Discussion

Although there is a variety of treatment options and published treatment guidelines for distal fingertip amputations, few comparative studies support use of one treatment over another. In our experience, treatment decisions are based mainly on injury parameters, but surgeon preference and patient factors (age, sex, occupation) can also influence care. Our goal in this study was to better understand how surgeon and patient factors influence treatment preferences for distal fingertip amputations among a cross section of US and international hand surgeons. Our survey results showed lack of consensus among hand surgeons and highlighted several trends.

As expected, we found a wide range of treatment preferences for each clinical scenario queried, ranging from more simple treatments (eg, wound care) to more complex ones (eg, replantation). With patient parameters (age, profession, finger, acuity, injury type, tissue preservation, smoking status) standardized in the clinical scenarios, the treatment differences noted should reflect surgeon preference. However, other patient factors (eg, cultural differences, religious beliefs, surgeon setting, practice pattern, resource availability) that were not included in the clinical scenarios could also affect treatment preferences.

One particularly interesting finding was that international hand surgeons were 6.8 times more likely to replant a distal fingertip amputation. One possible explanation for this variation is the influence of cultural differences. For example, in East Asian countries, there can be a cultural stigma associated with loss of a fingertip, and therefore more of a desire on the part of the patient to restore the original finger.^20,21 In addition, the international respondents were biased toward academic practices—which could skew the treatment preference toward replantation, as we found that academic surgeons were more inclined to replantation.

Our finding that replantation was more commonly preferred by academic versus private practice surgeons may suggest a training bias, an affinity for more complex or interesting procedures, or access to hospital equipment and staff, including residents and fellows, not usually found at smaller community hospitals, where private practice surgeons are more commonly based. Jazayeri and colleagues²² found that institutions specializing in microsurgery often produced better outcomes than nonspecializing institutions. Therefore, it is not surprising that private practice hand surgeons may less often opt to replant a distal fingertip amputation. It is also not surprising that plastic surgeons are more inclined to perform a replantation or flap coverage, as their training is more microsurgery-intensive and their practice more focused on aesthetics compared with the other specialists.

Distal fingertip replantation is accepted by most as technically demanding, but it seems that the additional effort and resources would be justified if the procedure provided a superior outcome. However, other factors, such as cost of treatment and length of recovery, should also be considered. Average replantation cost has been estimated to range from $7500 to $14,000, compared with $2800 for non-replantation-related care, and median stay is about 4 days longer for replantation-related care.^23,24 These estimates do not include indirect costs, such as for postoperative rehabilitation, which is likely longer and more expensive, even in distal fingertip replantation. These disparities may not justify the outcome (of having a complete fingertip) if more conservative treatments yield similar results.^17,18 In addition, there is the expected failure rate of limb replantation surgery. In analysis of the overall societal costs and benefits of larger upper extremity limb replantation, the loss of invested resources sustained with failed limb replantation may be outweighed by the benefit of another patient having a successful outcome. In the case of fingertip replantation, however, does the undefined benefit of the successful patient outcome outweigh the investment of resources lost in cases of replantation failure? Understandably, there is a need for more robust clinical outcome and cost-comparative evidence to better inform decisions regarding distal fingertip amputation.

We found that wound care and skeletal shortening with primary closure (particularly with Allen level 3 injuries) were preferred more by surgeons within the first 5 years of practice. This finding seems to imply a lack of experience or confidence on the part of younger surgeons performing more complex procedures, such as flap coverage. Conversely, this finding may indicate a shift in treatment principle based on recent literature suggesting equivalent outcomes with simpler procedures.^17,18 Although our survey study did not provide an option for treatment combinations or staged procedures, several respondents wrote in that skeletal shortening supplemented with various types of autografts and allografts would be their preferred treatment.

Patient factors also play a significant role in clinical decisions. Age and profession seem to be important determinants, with more than 50% of respondents, on average, changing their treatment recommendation based on these 2 factors. A majority of respondents would perform a less involved procedure for a manual laborer, suggesting a quicker return to work is prioritized over a perceived improved clinical outcome. Interestingly, for patients younger than 15 years, the preference was divided, with 41% of surgeons opting for a more complex procedure. This suggests the importance of restoring anatomy in a younger patient, or the perceived decreased risk or failure rate with more involved treatment. Twenty percent preferred a less complex procedure in a younger patient, perhaps relying on the patient’s developmental potential for a good outcome or suggesting a concern for patient intolerance or compliance with complex surgery.

Nail plate regrowth can be a problem with fingertip amputations. Nail deformity is highly correlated with injury level, with amputations proximal to the lunula more likely to cause nail plate deformity.^25,26 Jebson and colleagues²⁷ recommended germinal matrix ablation for amputations proximal to the lunula. We found respondents often performed ablations for other indications, including injured or minimal remaining sterile matrix and lack of bony support for the sterile matrix. Forty-six percent of respondents had never performed sterile matrix transplant, which could indicate that they were unfamiliar with the technique or had donor-site concerns, or that postinjury nail deformities are uncommon, well tolerated, or treated along with other procedures, such as germinal matrix ablation.

Several weaknesses of this study must be highlighted. First, our response rate was smaller than desired. Although this work incorporated a large number of surgeon responses, nearly 200, the response rate was only 17%. In addition, although number of responses was likely adequate to show the diversity of opinion, the preferences and trends reported might not be representative of all hand surgeons. We could not perform a nonresponder analysis because of a lack of specific demographic data for the AAHS and international hand society members. However, AAHS has an approximate 50/50 mix of plastic and orthopedic surgeons, similar to our responder demographic, suggesting our smaller subset of responses might be representative of the whole. According to AAHS, a majority of its members are “academic” hand surgeons, so our results might not adequately reflect the preferences of community hand surgeons and ultimately might overstate the frequency of more complex treatments. Last, our international response was limited to a few countries. A larger, more broadly distributed response would provide a better understanding of regional preferences, which could shed light on the importance of cultural differences.

Variations in patient insurance status were not queried in this survey but might also affect treatment decisions. More involved, costly, and highly reimbursing procedures might be deemed reasonable options for a small perceived clinical benefit for insured patients.

When multiple digits or the thumb is injured, or there are other concomitant injuries, surgeons may alter their choice of intervention. In mangled extremities, preservation of salvageable functional units takes precedence over aesthetics and likely affects choice of treatment for the amputated fingertips. Similarly, multiple fingertip amputations, even if all at the same level, may be differently regarded than a solitary injury.

Conclusion

For distal fingertip amputations, there is little evidence supporting one approach over another. Without level I comparative data guiding treatment, anecdotal evidence and surgeon personal preferences likely contribute to the large variation noted in this survey. Our study results showed the disparity of fingertip treatment preferences among a cross section of US and international hand surgeons. More important, results underscored the need for a well-designed comparative study to determine the most effective treatments for distal fingertip amputations.

EVIDENCE SUMMARY

Two RCTs that evaluated qHPV in young men for preventing outcomes associated with the 4 HPV subtypes in the vaccine (6, 11, 16, and 18) found that it reduced them by 50% to 66% using an intention-to-treat protocol (TABLE^1-4).

Vaccination reduces AIN and persistent infection in MSM

The first RCT evaluated a subset of 602 MSM from the second, larger RCT for preventing AIN and persistent HPV infection.¹ The intention-to-treat population included men with 5 or fewer lifetime sexual partners who had engaged in insertive or receptive anal intercourse or oral sex within the last year, were not necessarily HPV-negative at enrollment, and received at least one dose of vaccine (or placebo).

The vaccine reduced AIN associated with the 4 HPV types (6.3 vs 12.6 events per 100 person-years; relative risk reduction [RRR]=50.3%; 95% confidence interval [CI], 25.7-67.2; number needed to treat [NNT]=16 to prevent one AIN case per year) and with HPV of any type (13 vs 17.5 events per 100 person-years; RRR=25.7%; 95% CI, -1.1 to 45.6). It also reduced the rate of persistent HPV infection with the 4 HPV vaccine subtypes (8.8 vs 21.6 events per 100 person-years; RRR=59.4%; 95% CI, 43%-71%; NNT=8 to prevent one persistent HPV infection per year).

Investigators in the study also evaluated vaccine efficacy in a smaller subset (194 men) using per-protocol analysis and found higher prevention rates (78% for AIN due to HPV types 6, 11, 16, and 18). Investigators followed these subjects every 6 months for 36 months with polymerase chain reaction testing for HPV DNA, high-resolution anoscopy with anal cytology, and anal biopsy and histology if there were atypia.

The vaccine decreases persistent HPV infection and external genital lesions

The second RCT, including both MSM and heterosexual men, found that qHPV vaccine reduced rates of persistent HPV infection by 48%, and external genital lesions (condylomata or intraepithelial neoplasia involving the penis, perineum, or perianal area) by 66% associated with HPV types 6, 11, 16, and 18 using the intention-to-treat protocol.²

Investigators used the same protocols used in the first RCT, and the per-protocol population again had higher prevention rates (84% for any HPV type, 90% against the 4 vaccine types). The only adverse effect of the vaccine was injection site pain (57% vs 51% with placebo; P<.001).

The vaccine also helps older MSM

A nonconcurrent cohort study that evaluated qHPV vaccination among older MSM with previously treated high-grade AIN found a 50% decrease in recurrence rates in the 2 years after vaccination.³ Investigators recruited HIV-negative men, some of whom chose vaccination (not randomized), and followed them for 2 years. Study limitations included using medical records for data collection and the predominance of white, nonsmoking men with private insurance.

A post-hoc analysis of older men without previous anal condylomata (210 men) or with treated condylomata and no recurrence in the year before vaccination (103 men) found that qHPV vaccination was associated with 55% lower rates of anal condylomata.⁴

RECOMMENDATIONS

The Centers for Disease Control and Prevention’s Advisory Committee on Immunization Practices recommends routine use of qHPV vaccine in males ages 11 through 21 years, and optional use in unvaccinated men as old as 26 years.⁵

EVIDENCE SUMMARY

Two RCTs that evaluated qHPV in young men for preventing outcomes associated with the 4 HPV subtypes in the vaccine (6, 11, 16, and 18) found that it reduced them by 50% to 66% using an intention-to-treat protocol (TABLE^1-4).

Vaccination reduces AIN and persistent infection in MSM

The first RCT evaluated a subset of 602 MSM from the second, larger RCT for preventing AIN and persistent HPV infection.¹ The intention-to-treat population included men with 5 or fewer lifetime sexual partners who had engaged in insertive or receptive anal intercourse or oral sex within the last year, were not necessarily HPV-negative at enrollment, and received at least one dose of vaccine (or placebo).

The vaccine reduced AIN associated with the 4 HPV types (6.3 vs 12.6 events per 100 person-years; relative risk reduction [RRR]=50.3%; 95% confidence interval [CI], 25.7-67.2; number needed to treat [NNT]=16 to prevent one AIN case per year) and with HPV of any type (13 vs 17.5 events per 100 person-years; RRR=25.7%; 95% CI, -1.1 to 45.6). It also reduced the rate of persistent HPV infection with the 4 HPV vaccine subtypes (8.8 vs 21.6 events per 100 person-years; RRR=59.4%; 95% CI, 43%-71%; NNT=8 to prevent one persistent HPV infection per year).

Investigators in the study also evaluated vaccine efficacy in a smaller subset (194 men) using per-protocol analysis and found higher prevention rates (78% for AIN due to HPV types 6, 11, 16, and 18). Investigators followed these subjects every 6 months for 36 months with polymerase chain reaction testing for HPV DNA, high-resolution anoscopy with anal cytology, and anal biopsy and histology if there were atypia.

The vaccine decreases persistent HPV infection and external genital lesions

The second RCT, including both MSM and heterosexual men, found that qHPV vaccine reduced rates of persistent HPV infection by 48%, and external genital lesions (condylomata or intraepithelial neoplasia involving the penis, perineum, or perianal area) by 66% associated with HPV types 6, 11, 16, and 18 using the intention-to-treat protocol.²

Investigators used the same protocols used in the first RCT, and the per-protocol population again had higher prevention rates (84% for any HPV type, 90% against the 4 vaccine types). The only adverse effect of the vaccine was injection site pain (57% vs 51% with placebo; P<.001).

The vaccine also helps older MSM

A nonconcurrent cohort study that evaluated qHPV vaccination among older MSM with previously treated high-grade AIN found a 50% decrease in recurrence rates in the 2 years after vaccination.³ Investigators recruited HIV-negative men, some of whom chose vaccination (not randomized), and followed them for 2 years. Study limitations included using medical records for data collection and the predominance of white, nonsmoking men with private insurance.

A post-hoc analysis of older men without previous anal condylomata (210 men) or with treated condylomata and no recurrence in the year before vaccination (103 men) found that qHPV vaccination was associated with 55% lower rates of anal condylomata.⁴

RECOMMENDATIONS

The Centers for Disease Control and Prevention’s Advisory Committee on Immunization Practices recommends routine use of qHPV vaccine in males ages 11 through 21 years, and optional use in unvaccinated men as old as 26 years.⁵

EVIDENCE SUMMARY

Two RCTs that evaluated qHPV in young men for preventing outcomes associated with the 4 HPV subtypes in the vaccine (6, 11, 16, and 18) found that it reduced them by 50% to 66% using an intention-to-treat protocol (TABLE^1-4).

Vaccination reduces AIN and persistent infection in MSM

The first RCT evaluated a subset of 602 MSM from the second, larger RCT for preventing AIN and persistent HPV infection.¹ The intention-to-treat population included men with 5 or fewer lifetime sexual partners who had engaged in insertive or receptive anal intercourse or oral sex within the last year, were not necessarily HPV-negative at enrollment, and received at least one dose of vaccine (or placebo).

The vaccine reduced AIN associated with the 4 HPV types (6.3 vs 12.6 events per 100 person-years; relative risk reduction [RRR]=50.3%; 95% confidence interval [CI], 25.7-67.2; number needed to treat [NNT]=16 to prevent one AIN case per year) and with HPV of any type (13 vs 17.5 events per 100 person-years; RRR=25.7%; 95% CI, -1.1 to 45.6). It also reduced the rate of persistent HPV infection with the 4 HPV vaccine subtypes (8.8 vs 21.6 events per 100 person-years; RRR=59.4%; 95% CI, 43%-71%; NNT=8 to prevent one persistent HPV infection per year).

Investigators in the study also evaluated vaccine efficacy in a smaller subset (194 men) using per-protocol analysis and found higher prevention rates (78% for AIN due to HPV types 6, 11, 16, and 18). Investigators followed these subjects every 6 months for 36 months with polymerase chain reaction testing for HPV DNA, high-resolution anoscopy with anal cytology, and anal biopsy and histology if there were atypia.

The vaccine decreases persistent HPV infection and external genital lesions

The second RCT, including both MSM and heterosexual men, found that qHPV vaccine reduced rates of persistent HPV infection by 48%, and external genital lesions (condylomata or intraepithelial neoplasia involving the penis, perineum, or perianal area) by 66% associated with HPV types 6, 11, 16, and 18 using the intention-to-treat protocol.²

Investigators used the same protocols used in the first RCT, and the per-protocol population again had higher prevention rates (84% for any HPV type, 90% against the 4 vaccine types). The only adverse effect of the vaccine was injection site pain (57% vs 51% with placebo; P<.001).

The vaccine also helps older MSM

A nonconcurrent cohort study that evaluated qHPV vaccination among older MSM with previously treated high-grade AIN found a 50% decrease in recurrence rates in the 2 years after vaccination.³ Investigators recruited HIV-negative men, some of whom chose vaccination (not randomized), and followed them for 2 years. Study limitations included using medical records for data collection and the predominance of white, nonsmoking men with private insurance.

A post-hoc analysis of older men without previous anal condylomata (210 men) or with treated condylomata and no recurrence in the year before vaccination (103 men) found that qHPV vaccination was associated with 55% lower rates of anal condylomata.⁴

RECOMMENDATIONS

The Centers for Disease Control and Prevention’s Advisory Committee on Immunization Practices recommends routine use of qHPV vaccine in males ages 11 through 21 years, and optional use in unvaccinated men as old as 26 years.⁵

Rotator cuff tears are a common condition affecting the shoulder joint. Initial open repair techniques were associated with several complications, including severe early postoperative pain, deltoid detachment and/or weakness, risk for infection, and arthrofibrosis.^1-3 In addition, open procedures cannot address other possible diagnoses, such as labral tears and loose bodies. These disadvantages promoted the development of an arthroscopically assisted mini-open technique.⁴ Superior long-term results, with more than 90% of patients achieving good to excellent results,^5-13 established the mini-open rotator cuff repair (RCR) as the gold standard.^{3,6,10,12,14-16}

Recently, as instrumentation for arthroscopy has improved, enthusiasm for all-arthroscopic techniques (hereafter referred to as arthroscopic repair) has grown. The appeal of arthroscopic repair includes potentially less initial pain, ability to treat intra-articular lesions concurrently, smaller skin incisions with better cosmesis, less soft-tissue dissection, and low risk for deltoid detachment.^3,17 The potential advantages of arthroscopic repair can lead to perceptions of quicker healing and shorter recovery, which are not supported by the literature. However, arthroscopic repair is technically more challenging, time-consuming, and expensive than open or mini-open repairs,^18,19 and though some investigators have reported a trend toward fewer complications,³ the long-term outcome of arthroscopic RCRs has not been shown to be better than that of other techniques.

Given that no differences have been shown between the emerging arthroscopic repair technique and mini-open repair with respect to range of motion or clinical scores in the short term,³ it is unclear what perceptions influence choice of technique for one’s own personal RCR.

We conducted a study to determine which RCR technique medical professionals (orthopedic attendings and residents, anesthesiologists, internal medicine attendings, main operating room nurses, and physical therapists) preferred for their own surgery and to analyze perceptions shaping those opinions. Orthopedic surgeons have the best concept of rotator cuff surgery, but anesthesiologists and nurses have a “front row seat” and opinions on types of rotator cuff surgery. Physical therapists, who treat patients with rotator cuff tears, also have a working knowledge of rotator cuff surgery. Finally, internists represent a rotator cuff injury referral service and may have patients who have undergone rotator cuff surgery. We hypothesized that most medical professionals, irrespective of specialty or career length, would prefer arthroscopic RCR because of its perceived superior outcome and fast recovery.

Materials and Methods

This cross-sectional, descriptive, survey-based study was approved by our institutional review board (IRB) and offered via 3 emails between April 2011 and June 2011 to attendings (orthopedists, internists, anesthesiologists), residents, and allied health professionals (AHPs; operating room nurses, physical therapists) involved in orthopedic care at our institution. Each email contained a hyperlink to the online survey (Appendix), which took about 10 minutes to complete and explored respondent demographics, exposure to the different techniques, and opinions regarding different aspects of RCR surgery and recovery.

There were 84 respondents. The sexes were equally represented, and age ranged from 25 to 78 years (Table 1). Of the respondents, 41 (49%) were attendings, 20 (24%) were residents, and 23 (27%) were AHPs. Of the attendings, 13 (32%) were orthopedic surgeons, 26 (63%) were primary care physicians, and 2 (5%) did not specify their specialty. Four orthopedic surgeons had fellowship training in sports medicine or shoulder and elbow surgery. The attendings were overall more experienced in their profession than the other groups were, with 68% reporting more than 5 years of experience.

Descriptive statistics, including means and standard errors, were calculated. Fisher exact test was used to compare preferences of RCR type according to type of training and years of experience. Significance was set at P ≤ .05.

Results

Overall Responses (Table 2)

Of the 84 respondents, almost half (46%) preferred deferring their choice of RCR to their surgeon. Most of the other respondents preferred the arthroscopic technique (26%) or the mini-open repair (23%). There was no association between technique preference and medical professional type. Most respondents (63%) had never assisted in or performed rotator cuff surgery.

Seventy-four percent of all respondents indicated they thought arthroscopic, mini-open, and open RCRs are safe, and about half thought these procedures are fast. About half expressed no opinion about the cost-effectiveness of arthroscopic, mini-open, or open RCRs (54%, 52%, and 48%, respectively), and slightly more than half expressed no opinion about whether arthroscopic, mini-open, or open RCR provide the best outcome (58%, 60%, and 62%, respectively). Significantly (P < .05) more respondents thought arthroscopic and mini-open repairs, rather than open repairs, promote quick healing (64% and 45%, respectively, vs 15%), good cosmetic results (81% and 51%, respectively, vs 10%), and patient satisfaction (50% and 48%, respectively, vs 30%). However, a significant (P < .05) number also thought arthroscopic and mini-open repairs are harder to learn/more challenging to perform than open repairs (52% and 38%, respectively, vs 17%).

Of all factors considered, safety of arthroscopic repair garnered the highest consensus: 82%. Respondents were least opinionated about the outcome of the open repair technique, with more than 62% expressing no opinion about the outcome. The responses to the questions on the learning curves for the 3 techniques varied the most.

Responses by Group (Table 2)

Attendings. Of the 41 attendings, 24 (59%) responded they would defer to their surgeon’s technique preference for RCR. Of the other 17 who expressed a preference, most indicated arthroscopic or mini-open repair (17% each). There was a difference (P < .05) between years of experience and RCR preference: of the 13 attendings with less than 5 years of experience, arthroscopic repair was preferred by 31%; in contrast, of the 28 attendings with more than 5 years of experience, only 11% preferred arthroscopic repair.

Of the 11 attendings who performed rotator cuff surgery, 55% used the open technique, but most (8) preferred to have their own rotator cuff fixed arthroscopically or according to their surgeon’s preference. Only 1 surgeon preferred open repair for his own rotator cuff. Of the 4 surgeons who performed arthroscopic RCRs, 3 had less than 5 years of experience. Conversely, all 7 surgeons who performed mini-open or open repairs had more than 5 years of experience.

Of the 30 attendings who did not perform rotator cuff surgery, most (20) responded they would defer to their surgeon’s technique preference for RCR.

The attendings’ opinions on factors affecting rotator cuff surgery were similar to those of the other respondents with respect to safety, cost-effectiveness, recovery, cosmesis, patient satisfaction, outcome, and technical difficulty. Unlike the others, however, attendings considered all 3 repair techniques fast.

Residents. Of the 20 residents, 7 preferred arthroscopic, 5 preferred mini-open, and 1 preferred open repair; the other 7 responded they would defer to their surgeon’s preference. Residents’ opinions on each factor were more polarized and consistent across categories than those of the other groups. Residents overwhelmingly thought all 3 techniques (arthroscopic, mini-open, open) are safe (19, 19, and 18, respectively) and cost-effective (12, 14, and 14, respectively). Although most residents considered the open and mini-open repair techniques fast (19 and 15, respectively), only 8 considered arthroscopic RCR fast, and 4 considered it slow. Residents’ opinions about the technique that produces the best outcome were mixed. As with the other respondents, residents thought arthroscopic RCRs heal fast and produce great cosmetic results, but are challenging to perform and have a steep learning curve. Unlike the other respondents, most residents (12) considered open RCR easy to learn (P = .006), with a learning curve of fewer than 20 procedures.

AHPs. No AHP expressed a preference for open RCR. This group was evenly divided among 3 choices: deferring to their surgeon’s preference, arthroscopic repair, and mini-open repair. The 23 AHPs thought arthroscopic, mini-open, and open repairs are safe (17, 15, and 12, respectively), but most indicated they were “equivocal” about which techniques are cost-effective, challenging to perform, and produce the best outcomes. A significantly (P = .014) larger number of AHPs (7) considered open rotator cuff surgery slow compared with arthroscopic (0) and mini-open (2) repair techniques. As with the overall cohort, AHPs reported arthroscopic and mini-open repairs promote quick healing and good cosmetic results, but are challenging to perform.

Discussion

As our population ages and continues to remain active, the demand for RCR has accelerated. National data show that 272,148 ambulatory RCRs and 20,433 inpatient RCRs were performed in 2006—an overall 141% increase in RCR since 1996.²⁰ In 1996, 41 per 100,000 population underwent RCR.²⁰ By 2006, this number ballooned to 98 per 100,000 population.²⁰ There are 3 predominant techniques for repairing the rotator cuff: open, mini-open, and arthroscopic. As RCR use increases, we should consider the factors that medical professionals consider important when choosing a method for their own RCR.

Of the 84 medical professionals in our cohort, 39 (46%) indicated they would defer to their surgeon’s technique preference for RCR. Of the other 45, about equal numbers preferred arthroscopic and mini-open RCRs; only 2 preferred open RCRs. This finding suggests that the individual opinions of surgeons who perform RCRs have a substantial influence on a large proportion of medical professionals’ ultimate choice of RCR method. Interestingly, of the attendings who performed open RCR, only 1 expressed a preference for the open technique for his own RCR. This finding might suggest a shift in opinion and an emerging perception among surgeons performing RCR about the value of this technique.

Several factors may account for these evolving beliefs. We hypothesized that a biased favorable view of arthroscopic repair outcome might influence opinions. However, our results did not support the hypothesis. Medical professionals in our cohort were equivocal about the best RCR technique. No consensus was evident among attendings, residents, or AHPs. This lack of clinical agreement about rotator cuff surgery has been observed elsewhere—for example, among members of the American Academy of Orthopaedic Surgeons (AAOS)²¹ and the European Society of Sports Traumatology, Knee Surgery, and Arthroscopy.²² Despite theoretical advantages of arthroscopic repair, there has been no documented significant difference in patient outcomes when compared with other techniques.²³ To our knowledge, there have been only a few clinical studies comparing the different RCR techniques. A meta-analysis of 5 clinical studies comparing arthroscopic and mini-open RCR techniques showed no difference in clinical outcomes or complication rates.⁸ The 2012 AAOS clinical practice guidelines for RCR reflect these observations.²⁴ That consortium of leading shoulder surgeons could not recommend a modality of surgical rotator cuff tear repair given the lack of conclusive evidence.²⁴

At our institution, arthroscopic, mini-open, and open RCRs were performed by 36%, 9%, and 55% of our surgeons, respectively. A survey of AAOS surgeons showed that, of those who perform RCRs, 14.5%, 46.2%, and 36.6% used arthroscopic, mini-open, and open techniques, respectively.²¹ The greater use of open repairs at our institution might reflect the seniority of our faculty. Dunn and colleagues²¹ found that surgeons who preferred open RCR had been in practice longer than those who preferred the arthroscopic or mini-open technique. Of our 4 faculty who performed arthroscopic repairs, 3 were less than 5 years from completing their training. In contrast, all faculty who performed mini-open or open repairs were more than 5 years from completing their training. Furthermore, mean age of the surgeons who performed arthroscopic repair was 39.8 years (range, 32-51 years), and these surgeons were significantly younger than those who performed mini-open or open repair (mean age, 56.3 years; range, 41-78 years). Younger surgeon age has been associated with higher rates of arthroscopic repair.²⁵

Attendings unaccustomed to arthroscopy may find it more challenging than the younger generation of surgeons, who are exposed to it early in training. Dunn and colleagues²¹ noted that the likelihood of performing an arthroscopic repair was influenced by the surgeon’s experience level. Fellowship-trained shoulder and sports medicine surgeons are also more likely to perform arthroscopic repairs than those with training limited to orthopedic residency.²⁵ Arthroscopic RCR demands a high level of technical skill that many acquire in fellowship training.²⁶ Mauro and colleagues²⁶ found that surgeons trained in a sports medicine fellowship performed 82.6% of subacromial decompression and/or RCR procedures arthroscopically, compared with 54.5% to 70.1% for surgeons trained in other fellowships. In our cohort, with the exception of 1 surgeon, all fellowship-trained shoulder and sports medicine surgeons performed arthroscopic RCRs.

Although no conclusive evidence in the literature supports arthroscopic over the other repair types, the demand for arthroscopic RCR has rapidly increased relative to that for the others. Between 1996 and 2006, use of arthroscopic RCR increased 600%, from 8 to 58 per 100,000 population.²⁰ In that same period, use of open RCR increased by only 34%.²⁰ Similarly, Mauro and colleagues²⁶ found that the proportion of subacromial decompression and RCRs performed arthroscopically rose from 58.3% in 2004 to 83.7% in 2009. Using the 2006 New York State Ambulatory Surgery Database, Churchill and Ghorai²⁷ found that 74.5% of RCRs with acromioplasty were performed arthroscopically.

Respondent-indicated factors that may have contributed to the more favorable opinion of arthroscopic and mini-open repair include quick healing, good cosmetic results, and better perceived patient satisfaction. The literature supports these perceptions. Baker and Liu¹⁴ found shorter hospital stays and quicker return to activity with arthroscopic repair compared with open repair. Vitale and colleagues²⁵ also noted that, compared with open or mini-open repair techniques, arthroscopic repair resulted in shorter hospitalization and quicker overall recovery.

If these selected health care professionals with some inside information on rotator cuff surgery have biases that affect their selection of rotator cuff procedures, we should acknowledge that nonmedical personnel, in particular our patients, also have biases. The knowledge base of patients may be further influenced by friends or family members who have had rotator cuff surgery, by lay publications, and by the Internet. Satisfaction with any surgical procedure depends not only on the success of the surgery and the rehabilitation but also on patient and provider expectations. Such expectations are influenced, in part, by biases.

Our medical professionals had similar opinions on safety, recovery, cosmesis, and overall outcome of the RCR techniques, but different opinions on procedure durations and associated training requirements. All residents except one indicated open repair was a quick procedure. In contrast, a significant number of AHPs thought open repair was time-consuming. The attendings considered all the methods fast. The residents’ opinions were the most consistent with the true operating times reported. According to the literature, total operating time for mini-open repair ranges from 10 to 16 minutes faster than that for arthroscopic repair.^18,20,27 Ultimately, procedure duration did not affect the respondents’ technique preference for RCR.

There was substantial disagreement about the number of procedures needed to become proficient in the different repair techniques. Overall, however, there was consensus that arthroscopic and mini-open repairs had longer learning curves than open repair. Given the lack of agreement among orthopedic department chairmen and sports medicine fellowship directors regarding the minimum exposure needed (during residency) to become proficient in diagnostic shoulder arthroscopy,²⁸ this finding is not surprising. Guttmann and colleagues²⁹ attempted to quantify the learning curve for arthroscopic RCR by tracking operating time as a surrogate measure. They found that RCR operative time decreased rapidly during the initial block of 10 cases to the second block of 10 cases, but thereafter improvement continued at a much lower rate.²⁹ None of our respondents thought the learning curve for arthroscopic RCR was under 10 cases, but no group, not even the attendings who performed RCRs, could agree on the minimum number of cases needed for proficiency. The longer learning curve for arthroscopic RCR did not discourage the respondents who preferred arthroscopic or mini-open RCR.

Cost was not an influential factor in opinions about which RCR method is optimal. Medical professionals were ambivalent about the cost-effectiveness of the different procedures, with most expressing no opinion on cost. Multiple investigators have shown that arthroscopic RCR costs as much as $1144 more than mini-open RCR,^18,27 which has many of the advantages of arthroscopic repair but not the costly implants and instruments. As our medical community becomes more cost-conscious, concern about this factor may increase among medical professionals.

Our study had several limitations. Its results must be interpreted carefully, given they represent the viewpoints of a nonrandomized sample of motivated respondents at one institution. A selection bias excluded surgeons who were uncomfortable with RCR and unwilling to report any shortcomings. The conclusions cannot be generalized to other medical professionals or to other institutions. Furthermore, to develop a simple, straightforward survey focused on a specific type of rotator cuff tear, and to avoid confusion, we assumed that the treatment preference for the described tear was generalizable to all encountered tears. However, some surgeons have reported different repair techniques for different types and sizes of rotator cuff tears.²⁵

Conclusion

Most of our surveyed medical professionals were willing to defer to their surgeon’s decision about which technique would be appropriate for their own personal RCR. There is a trend nationally, and at our institution, for increased use of arthroscopic RCR. Although medical professionals readily acknowledge it is unclear which repair method provides the best ultimate outcome, many perceive fast recovery and good cosmetic results with arthroscopic and mini-open repairs. When medical professionals are counseling patients, we need to recognize these personal biases because many patients defer to their surgeon’s counsel. For some medical professionals, cosmesis can be an important factor, but cost, procedure duration, potential technical challenges of arthroscopic repair, and other considerations may make other techniques more desirable for others.

Rotator cuff tears are a common condition affecting the shoulder joint. Initial open repair techniques were associated with several complications, including severe early postoperative pain, deltoid detachment and/or weakness, risk for infection, and arthrofibrosis.^1-3 In addition, open procedures cannot address other possible diagnoses, such as labral tears and loose bodies. These disadvantages promoted the development of an arthroscopically assisted mini-open technique.⁴ Superior long-term results, with more than 90% of patients achieving good to excellent results,^5-13 established the mini-open rotator cuff repair (RCR) as the gold standard.^{3,6,10,12,14-16}

Recently, as instrumentation for arthroscopy has improved, enthusiasm for all-arthroscopic techniques (hereafter referred to as arthroscopic repair) has grown. The appeal of arthroscopic repair includes potentially less initial pain, ability to treat intra-articular lesions concurrently, smaller skin incisions with better cosmesis, less soft-tissue dissection, and low risk for deltoid detachment.^3,17 The potential advantages of arthroscopic repair can lead to perceptions of quicker healing and shorter recovery, which are not supported by the literature. However, arthroscopic repair is technically more challenging, time-consuming, and expensive than open or mini-open repairs,^18,19 and though some investigators have reported a trend toward fewer complications,³ the long-term outcome of arthroscopic RCRs has not been shown to be better than that of other techniques.

Given that no differences have been shown between the emerging arthroscopic repair technique and mini-open repair with respect to range of motion or clinical scores in the short term,³ it is unclear what perceptions influence choice of technique for one’s own personal RCR.

We conducted a study to determine which RCR technique medical professionals (orthopedic attendings and residents, anesthesiologists, internal medicine attendings, main operating room nurses, and physical therapists) preferred for their own surgery and to analyze perceptions shaping those opinions. Orthopedic surgeons have the best concept of rotator cuff surgery, but anesthesiologists and nurses have a “front row seat” and opinions on types of rotator cuff surgery. Physical therapists, who treat patients with rotator cuff tears, also have a working knowledge of rotator cuff surgery. Finally, internists represent a rotator cuff injury referral service and may have patients who have undergone rotator cuff surgery. We hypothesized that most medical professionals, irrespective of specialty or career length, would prefer arthroscopic RCR because of its perceived superior outcome and fast recovery.

Materials and Methods

This cross-sectional, descriptive, survey-based study was approved by our institutional review board (IRB) and offered via 3 emails between April 2011 and June 2011 to attendings (orthopedists, internists, anesthesiologists), residents, and allied health professionals (AHPs; operating room nurses, physical therapists) involved in orthopedic care at our institution. Each email contained a hyperlink to the online survey (Appendix), which took about 10 minutes to complete and explored respondent demographics, exposure to the different techniques, and opinions regarding different aspects of RCR surgery and recovery.

There were 84 respondents. The sexes were equally represented, and age ranged from 25 to 78 years (Table 1). Of the respondents, 41 (49%) were attendings, 20 (24%) were residents, and 23 (27%) were AHPs. Of the attendings, 13 (32%) were orthopedic surgeons, 26 (63%) were primary care physicians, and 2 (5%) did not specify their specialty. Four orthopedic surgeons had fellowship training in sports medicine or shoulder and elbow surgery. The attendings were overall more experienced in their profession than the other groups were, with 68% reporting more than 5 years of experience.

Descriptive statistics, including means and standard errors, were calculated. Fisher exact test was used to compare preferences of RCR type according to type of training and years of experience. Significance was set at P ≤ .05.

Results

Overall Responses (Table 2)

Of the 84 respondents, almost half (46%) preferred deferring their choice of RCR to their surgeon. Most of the other respondents preferred the arthroscopic technique (26%) or the mini-open repair (23%). There was no association between technique preference and medical professional type. Most respondents (63%) had never assisted in or performed rotator cuff surgery.

Seventy-four percent of all respondents indicated they thought arthroscopic, mini-open, and open RCRs are safe, and about half thought these procedures are fast. About half expressed no opinion about the cost-effectiveness of arthroscopic, mini-open, or open RCRs (54%, 52%, and 48%, respectively), and slightly more than half expressed no opinion about whether arthroscopic, mini-open, or open RCR provide the best outcome (58%, 60%, and 62%, respectively). Significantly (P < .05) more respondents thought arthroscopic and mini-open repairs, rather than open repairs, promote quick healing (64% and 45%, respectively, vs 15%), good cosmetic results (81% and 51%, respectively, vs 10%), and patient satisfaction (50% and 48%, respectively, vs 30%). However, a significant (P < .05) number also thought arthroscopic and mini-open repairs are harder to learn/more challenging to perform than open repairs (52% and 38%, respectively, vs 17%).

Of all factors considered, safety of arthroscopic repair garnered the highest consensus: 82%. Respondents were least opinionated about the outcome of the open repair technique, with more than 62% expressing no opinion about the outcome. The responses to the questions on the learning curves for the 3 techniques varied the most.

Responses by Group (Table 2)

Attendings. Of the 41 attendings, 24 (59%) responded they would defer to their surgeon’s technique preference for RCR. Of the other 17 who expressed a preference, most indicated arthroscopic or mini-open repair (17% each). There was a difference (P < .05) between years of experience and RCR preference: of the 13 attendings with less than 5 years of experience, arthroscopic repair was preferred by 31%; in contrast, of the 28 attendings with more than 5 years of experience, only 11% preferred arthroscopic repair.

Of the 11 attendings who performed rotator cuff surgery, 55% used the open technique, but most (8) preferred to have their own rotator cuff fixed arthroscopically or according to their surgeon’s preference. Only 1 surgeon preferred open repair for his own rotator cuff. Of the 4 surgeons who performed arthroscopic RCRs, 3 had less than 5 years of experience. Conversely, all 7 surgeons who performed mini-open or open repairs had more than 5 years of experience.

Of the 30 attendings who did not perform rotator cuff surgery, most (20) responded they would defer to their surgeon’s technique preference for RCR.

The attendings’ opinions on factors affecting rotator cuff surgery were similar to those of the other respondents with respect to safety, cost-effectiveness, recovery, cosmesis, patient satisfaction, outcome, and technical difficulty. Unlike the others, however, attendings considered all 3 repair techniques fast.

Residents. Of the 20 residents, 7 preferred arthroscopic, 5 preferred mini-open, and 1 preferred open repair; the other 7 responded they would defer to their surgeon’s preference. Residents’ opinions on each factor were more polarized and consistent across categories than those of the other groups. Residents overwhelmingly thought all 3 techniques (arthroscopic, mini-open, open) are safe (19, 19, and 18, respectively) and cost-effective (12, 14, and 14, respectively). Although most residents considered the open and mini-open repair techniques fast (19 and 15, respectively), only 8 considered arthroscopic RCR fast, and 4 considered it slow. Residents’ opinions about the technique that produces the best outcome were mixed. As with the other respondents, residents thought arthroscopic RCRs heal fast and produce great cosmetic results, but are challenging to perform and have a steep learning curve. Unlike the other respondents, most residents (12) considered open RCR easy to learn (P = .006), with a learning curve of fewer than 20 procedures.

AHPs. No AHP expressed a preference for open RCR. This group was evenly divided among 3 choices: deferring to their surgeon’s preference, arthroscopic repair, and mini-open repair. The 23 AHPs thought arthroscopic, mini-open, and open repairs are safe (17, 15, and 12, respectively), but most indicated they were “equivocal” about which techniques are cost-effective, challenging to perform, and produce the best outcomes. A significantly (P = .014) larger number of AHPs (7) considered open rotator cuff surgery slow compared with arthroscopic (0) and mini-open (2) repair techniques. As with the overall cohort, AHPs reported arthroscopic and mini-open repairs promote quick healing and good cosmetic results, but are challenging to perform.

Discussion

As our population ages and continues to remain active, the demand for RCR has accelerated. National data show that 272,148 ambulatory RCRs and 20,433 inpatient RCRs were performed in 2006—an overall 141% increase in RCR since 1996.²⁰ In 1996, 41 per 100,000 population underwent RCR.²⁰ By 2006, this number ballooned to 98 per 100,000 population.²⁰ There are 3 predominant techniques for repairing the rotator cuff: open, mini-open, and arthroscopic. As RCR use increases, we should consider the factors that medical professionals consider important when choosing a method for their own RCR.

Of the 84 medical professionals in our cohort, 39 (46%) indicated they would defer to their surgeon’s technique preference for RCR. Of the other 45, about equal numbers preferred arthroscopic and mini-open RCRs; only 2 preferred open RCRs. This finding suggests that the individual opinions of surgeons who perform RCRs have a substantial influence on a large proportion of medical professionals’ ultimate choice of RCR method. Interestingly, of the attendings who performed open RCR, only 1 expressed a preference for the open technique for his own RCR. This finding might suggest a shift in opinion and an emerging perception among surgeons performing RCR about the value of this technique.

Several factors may account for these evolving beliefs. We hypothesized that a biased favorable view of arthroscopic repair outcome might influence opinions. However, our results did not support the hypothesis. Medical professionals in our cohort were equivocal about the best RCR technique. No consensus was evident among attendings, residents, or AHPs. This lack of clinical agreement about rotator cuff surgery has been observed elsewhere—for example, among members of the American Academy of Orthopaedic Surgeons (AAOS)²¹ and the European Society of Sports Traumatology, Knee Surgery, and Arthroscopy.²² Despite theoretical advantages of arthroscopic repair, there has been no documented significant difference in patient outcomes when compared with other techniques.²³ To our knowledge, there have been only a few clinical studies comparing the different RCR techniques. A meta-analysis of 5 clinical studies comparing arthroscopic and mini-open RCR techniques showed no difference in clinical outcomes or complication rates.⁸ The 2012 AAOS clinical practice guidelines for RCR reflect these observations.²⁴ That consortium of leading shoulder surgeons could not recommend a modality of surgical rotator cuff tear repair given the lack of conclusive evidence.²⁴

At our institution, arthroscopic, mini-open, and open RCRs were performed by 36%, 9%, and 55% of our surgeons, respectively. A survey of AAOS surgeons showed that, of those who perform RCRs, 14.5%, 46.2%, and 36.6% used arthroscopic, mini-open, and open techniques, respectively.²¹ The greater use of open repairs at our institution might reflect the seniority of our faculty. Dunn and colleagues²¹ found that surgeons who preferred open RCR had been in practice longer than those who preferred the arthroscopic or mini-open technique. Of our 4 faculty who performed arthroscopic repairs, 3 were less than 5 years from completing their training. In contrast, all faculty who performed mini-open or open repairs were more than 5 years from completing their training. Furthermore, mean age of the surgeons who performed arthroscopic repair was 39.8 years (range, 32-51 years), and these surgeons were significantly younger than those who performed mini-open or open repair (mean age, 56.3 years; range, 41-78 years). Younger surgeon age has been associated with higher rates of arthroscopic repair.²⁵

Attendings unaccustomed to arthroscopy may find it more challenging than the younger generation of surgeons, who are exposed to it early in training. Dunn and colleagues²¹ noted that the likelihood of performing an arthroscopic repair was influenced by the surgeon’s experience level. Fellowship-trained shoulder and sports medicine surgeons are also more likely to perform arthroscopic repairs than those with training limited to orthopedic residency.²⁵ Arthroscopic RCR demands a high level of technical skill that many acquire in fellowship training.²⁶ Mauro and colleagues²⁶ found that surgeons trained in a sports medicine fellowship performed 82.6% of subacromial decompression and/or RCR procedures arthroscopically, compared with 54.5% to 70.1% for surgeons trained in other fellowships. In our cohort, with the exception of 1 surgeon, all fellowship-trained shoulder and sports medicine surgeons performed arthroscopic RCRs.

Although no conclusive evidence in the literature supports arthroscopic over the other repair types, the demand for arthroscopic RCR has rapidly increased relative to that for the others. Between 1996 and 2006, use of arthroscopic RCR increased 600%, from 8 to 58 per 100,000 population.²⁰ In that same period, use of open RCR increased by only 34%.²⁰ Similarly, Mauro and colleagues²⁶ found that the proportion of subacromial decompression and RCRs performed arthroscopically rose from 58.3% in 2004 to 83.7% in 2009. Using the 2006 New York State Ambulatory Surgery Database, Churchill and Ghorai²⁷ found that 74.5% of RCRs with acromioplasty were performed arthroscopically.

Respondent-indicated factors that may have contributed to the more favorable opinion of arthroscopic and mini-open repair include quick healing, good cosmetic results, and better perceived patient satisfaction. The literature supports these perceptions. Baker and Liu¹⁴ found shorter hospital stays and quicker return to activity with arthroscopic repair compared with open repair. Vitale and colleagues²⁵ also noted that, compared with open or mini-open repair techniques, arthroscopic repair resulted in shorter hospitalization and quicker overall recovery.

If these selected health care professionals with some inside information on rotator cuff surgery have biases that affect their selection of rotator cuff procedures, we should acknowledge that nonmedical personnel, in particular our patients, also have biases. The knowledge base of patients may be further influenced by friends or family members who have had rotator cuff surgery, by lay publications, and by the Internet. Satisfaction with any surgical procedure depends not only on the success of the surgery and the rehabilitation but also on patient and provider expectations. Such expectations are influenced, in part, by biases.

Our medical professionals had similar opinions on safety, recovery, cosmesis, and overall outcome of the RCR techniques, but different opinions on procedure durations and associated training requirements. All residents except one indicated open repair was a quick procedure. In contrast, a significant number of AHPs thought open repair was time-consuming. The attendings considered all the methods fast. The residents’ opinions were the most consistent with the true operating times reported. According to the literature, total operating time for mini-open repair ranges from 10 to 16 minutes faster than that for arthroscopic repair.^18,20,27 Ultimately, procedure duration did not affect the respondents’ technique preference for RCR.

There was substantial disagreement about the number of procedures needed to become proficient in the different repair techniques. Overall, however, there was consensus that arthroscopic and mini-open repairs had longer learning curves than open repair. Given the lack of agreement among orthopedic department chairmen and sports medicine fellowship directors regarding the minimum exposure needed (during residency) to become proficient in diagnostic shoulder arthroscopy,²⁸ this finding is not surprising. Guttmann and colleagues²⁹ attempted to quantify the learning curve for arthroscopic RCR by tracking operating time as a surrogate measure. They found that RCR operative time decreased rapidly during the initial block of 10 cases to the second block of 10 cases, but thereafter improvement continued at a much lower rate.²⁹ None of our respondents thought the learning curve for arthroscopic RCR was under 10 cases, but no group, not even the attendings who performed RCRs, could agree on the minimum number of cases needed for proficiency. The longer learning curve for arthroscopic RCR did not discourage the respondents who preferred arthroscopic or mini-open RCR.

Cost was not an influential factor in opinions about which RCR method is optimal. Medical professionals were ambivalent about the cost-effectiveness of the different procedures, with most expressing no opinion on cost. Multiple investigators have shown that arthroscopic RCR costs as much as $1144 more than mini-open RCR,^18,27 which has many of the advantages of arthroscopic repair but not the costly implants and instruments. As our medical community becomes more cost-conscious, concern about this factor may increase among medical professionals.

Our study had several limitations. Its results must be interpreted carefully, given they represent the viewpoints of a nonrandomized sample of motivated respondents at one institution. A selection bias excluded surgeons who were uncomfortable with RCR and unwilling to report any shortcomings. The conclusions cannot be generalized to other medical professionals or to other institutions. Furthermore, to develop a simple, straightforward survey focused on a specific type of rotator cuff tear, and to avoid confusion, we assumed that the treatment preference for the described tear was generalizable to all encountered tears. However, some surgeons have reported different repair techniques for different types and sizes of rotator cuff tears.²⁵

Conclusion

Most of our surveyed medical professionals were willing to defer to their surgeon’s decision about which technique would be appropriate for their own personal RCR. There is a trend nationally, and at our institution, for increased use of arthroscopic RCR. Although medical professionals readily acknowledge it is unclear which repair method provides the best ultimate outcome, many perceive fast recovery and good cosmetic results with arthroscopic and mini-open repairs. When medical professionals are counseling patients, we need to recognize these personal biases because many patients defer to their surgeon’s counsel. For some medical professionals, cosmesis can be an important factor, but cost, procedure duration, potential technical challenges of arthroscopic repair, and other considerations may make other techniques more desirable for others.

Rotator cuff tears are a common condition affecting the shoulder joint. Initial open repair techniques were associated with several complications, including severe early postoperative pain, deltoid detachment and/or weakness, risk for infection, and arthrofibrosis.^1-3 In addition, open procedures cannot address other possible diagnoses, such as labral tears and loose bodies. These disadvantages promoted the development of an arthroscopically assisted mini-open technique.⁴ Superior long-term results, with more than 90% of patients achieving good to excellent results,^5-13 established the mini-open rotator cuff repair (RCR) as the gold standard.^{3,6,10,12,14-16}

Recently, as instrumentation for arthroscopy has improved, enthusiasm for all-arthroscopic techniques (hereafter referred to as arthroscopic repair) has grown. The appeal of arthroscopic repair includes potentially less initial pain, ability to treat intra-articular lesions concurrently, smaller skin incisions with better cosmesis, less soft-tissue dissection, and low risk for deltoid detachment.^3,17 The potential advantages of arthroscopic repair can lead to perceptions of quicker healing and shorter recovery, which are not supported by the literature. However, arthroscopic repair is technically more challenging, time-consuming, and expensive than open or mini-open repairs,^18,19 and though some investigators have reported a trend toward fewer complications,³ the long-term outcome of arthroscopic RCRs has not been shown to be better than that of other techniques.

Given that no differences have been shown between the emerging arthroscopic repair technique and mini-open repair with respect to range of motion or clinical scores in the short term,³ it is unclear what perceptions influence choice of technique for one’s own personal RCR.

We conducted a study to determine which RCR technique medical professionals (orthopedic attendings and residents, anesthesiologists, internal medicine attendings, main operating room nurses, and physical therapists) preferred for their own surgery and to analyze perceptions shaping those opinions. Orthopedic surgeons have the best concept of rotator cuff surgery, but anesthesiologists and nurses have a “front row seat” and opinions on types of rotator cuff surgery. Physical therapists, who treat patients with rotator cuff tears, also have a working knowledge of rotator cuff surgery. Finally, internists represent a rotator cuff injury referral service and may have patients who have undergone rotator cuff surgery. We hypothesized that most medical professionals, irrespective of specialty or career length, would prefer arthroscopic RCR because of its perceived superior outcome and fast recovery.

Materials and Methods

This cross-sectional, descriptive, survey-based study was approved by our institutional review board (IRB) and offered via 3 emails between April 2011 and June 2011 to attendings (orthopedists, internists, anesthesiologists), residents, and allied health professionals (AHPs; operating room nurses, physical therapists) involved in orthopedic care at our institution. Each email contained a hyperlink to the online survey (Appendix), which took about 10 minutes to complete and explored respondent demographics, exposure to the different techniques, and opinions regarding different aspects of RCR surgery and recovery.

There were 84 respondents. The sexes were equally represented, and age ranged from 25 to 78 years (Table 1). Of the respondents, 41 (49%) were attendings, 20 (24%) were residents, and 23 (27%) were AHPs. Of the attendings, 13 (32%) were orthopedic surgeons, 26 (63%) were primary care physicians, and 2 (5%) did not specify their specialty. Four orthopedic surgeons had fellowship training in sports medicine or shoulder and elbow surgery. The attendings were overall more experienced in their profession than the other groups were, with 68% reporting more than 5 years of experience.

Descriptive statistics, including means and standard errors, were calculated. Fisher exact test was used to compare preferences of RCR type according to type of training and years of experience. Significance was set at P ≤ .05.

Results

Overall Responses (Table 2)

Of the 84 respondents, almost half (46%) preferred deferring their choice of RCR to their surgeon. Most of the other respondents preferred the arthroscopic technique (26%) or the mini-open repair (23%). There was no association between technique preference and medical professional type. Most respondents (63%) had never assisted in or performed rotator cuff surgery.

Seventy-four percent of all respondents indicated they thought arthroscopic, mini-open, and open RCRs are safe, and about half thought these procedures are fast. About half expressed no opinion about the cost-effectiveness of arthroscopic, mini-open, or open RCRs (54%, 52%, and 48%, respectively), and slightly more than half expressed no opinion about whether arthroscopic, mini-open, or open RCR provide the best outcome (58%, 60%, and 62%, respectively). Significantly (P < .05) more respondents thought arthroscopic and mini-open repairs, rather than open repairs, promote quick healing (64% and 45%, respectively, vs 15%), good cosmetic results (81% and 51%, respectively, vs 10%), and patient satisfaction (50% and 48%, respectively, vs 30%). However, a significant (P < .05) number also thought arthroscopic and mini-open repairs are harder to learn/more challenging to perform than open repairs (52% and 38%, respectively, vs 17%).

Of all factors considered, safety of arthroscopic repair garnered the highest consensus: 82%. Respondents were least opinionated about the outcome of the open repair technique, with more than 62% expressing no opinion about the outcome. The responses to the questions on the learning curves for the 3 techniques varied the most.

Responses by Group (Table 2)

Attendings. Of the 41 attendings, 24 (59%) responded they would defer to their surgeon’s technique preference for RCR. Of the other 17 who expressed a preference, most indicated arthroscopic or mini-open repair (17% each). There was a difference (P < .05) between years of experience and RCR preference: of the 13 attendings with less than 5 years of experience, arthroscopic repair was preferred by 31%; in contrast, of the 28 attendings with more than 5 years of experience, only 11% preferred arthroscopic repair.

Of the 11 attendings who performed rotator cuff surgery, 55% used the open technique, but most (8) preferred to have their own rotator cuff fixed arthroscopically or according to their surgeon’s preference. Only 1 surgeon preferred open repair for his own rotator cuff. Of the 4 surgeons who performed arthroscopic RCRs, 3 had less than 5 years of experience. Conversely, all 7 surgeons who performed mini-open or open repairs had more than 5 years of experience.

Of the 30 attendings who did not perform rotator cuff surgery, most (20) responded they would defer to their surgeon’s technique preference for RCR.

The attendings’ opinions on factors affecting rotator cuff surgery were similar to those of the other respondents with respect to safety, cost-effectiveness, recovery, cosmesis, patient satisfaction, outcome, and technical difficulty. Unlike the others, however, attendings considered all 3 repair techniques fast.

Residents. Of the 20 residents, 7 preferred arthroscopic, 5 preferred mini-open, and 1 preferred open repair; the other 7 responded they would defer to their surgeon’s preference. Residents’ opinions on each factor were more polarized and consistent across categories than those of the other groups. Residents overwhelmingly thought all 3 techniques (arthroscopic, mini-open, open) are safe (19, 19, and 18, respectively) and cost-effective (12, 14, and 14, respectively). Although most residents considered the open and mini-open repair techniques fast (19 and 15, respectively), only 8 considered arthroscopic RCR fast, and 4 considered it slow. Residents’ opinions about the technique that produces the best outcome were mixed. As with the other respondents, residents thought arthroscopic RCRs heal fast and produce great cosmetic results, but are challenging to perform and have a steep learning curve. Unlike the other respondents, most residents (12) considered open RCR easy to learn (P = .006), with a learning curve of fewer than 20 procedures.

AHPs. No AHP expressed a preference for open RCR. This group was evenly divided among 3 choices: deferring to their surgeon’s preference, arthroscopic repair, and mini-open repair. The 23 AHPs thought arthroscopic, mini-open, and open repairs are safe (17, 15, and 12, respectively), but most indicated they were “equivocal” about which techniques are cost-effective, challenging to perform, and produce the best outcomes. A significantly (P = .014) larger number of AHPs (7) considered open rotator cuff surgery slow compared with arthroscopic (0) and mini-open (2) repair techniques. As with the overall cohort, AHPs reported arthroscopic and mini-open repairs promote quick healing and good cosmetic results, but are challenging to perform.

Discussion

As our population ages and continues to remain active, the demand for RCR has accelerated. National data show that 272,148 ambulatory RCRs and 20,433 inpatient RCRs were performed in 2006—an overall 141% increase in RCR since 1996.²⁰ In 1996, 41 per 100,000 population underwent RCR.²⁰ By 2006, this number ballooned to 98 per 100,000 population.²⁰ There are 3 predominant techniques for repairing the rotator cuff: open, mini-open, and arthroscopic. As RCR use increases, we should consider the factors that medical professionals consider important when choosing a method for their own RCR.

Of the 84 medical professionals in our cohort, 39 (46%) indicated they would defer to their surgeon’s technique preference for RCR. Of the other 45, about equal numbers preferred arthroscopic and mini-open RCRs; only 2 preferred open RCRs. This finding suggests that the individual opinions of surgeons who perform RCRs have a substantial influence on a large proportion of medical professionals’ ultimate choice of RCR method. Interestingly, of the attendings who performed open RCR, only 1 expressed a preference for the open technique for his own RCR. This finding might suggest a shift in opinion and an emerging perception among surgeons performing RCR about the value of this technique.

Several factors may account for these evolving beliefs. We hypothesized that a biased favorable view of arthroscopic repair outcome might influence opinions. However, our results did not support the hypothesis. Medical professionals in our cohort were equivocal about the best RCR technique. No consensus was evident among attendings, residents, or AHPs. This lack of clinical agreement about rotator cuff surgery has been observed elsewhere—for example, among members of the American Academy of Orthopaedic Surgeons (AAOS)²¹ and the European Society of Sports Traumatology, Knee Surgery, and Arthroscopy.²² Despite theoretical advantages of arthroscopic repair, there has been no documented significant difference in patient outcomes when compared with other techniques.²³ To our knowledge, there have been only a few clinical studies comparing the different RCR techniques. A meta-analysis of 5 clinical studies comparing arthroscopic and mini-open RCR techniques showed no difference in clinical outcomes or complication rates.⁸ The 2012 AAOS clinical practice guidelines for RCR reflect these observations.²⁴ That consortium of leading shoulder surgeons could not recommend a modality of surgical rotator cuff tear repair given the lack of conclusive evidence.²⁴

At our institution, arthroscopic, mini-open, and open RCRs were performed by 36%, 9%, and 55% of our surgeons, respectively. A survey of AAOS surgeons showed that, of those who perform RCRs, 14.5%, 46.2%, and 36.6% used arthroscopic, mini-open, and open techniques, respectively.²¹ The greater use of open repairs at our institution might reflect the seniority of our faculty. Dunn and colleagues²¹ found that surgeons who preferred open RCR had been in practice longer than those who preferred the arthroscopic or mini-open technique. Of our 4 faculty who performed arthroscopic repairs, 3 were less than 5 years from completing their training. In contrast, all faculty who performed mini-open or open repairs were more than 5 years from completing their training. Furthermore, mean age of the surgeons who performed arthroscopic repair was 39.8 years (range, 32-51 years), and these surgeons were significantly younger than those who performed mini-open or open repair (mean age, 56.3 years; range, 41-78 years). Younger surgeon age has been associated with higher rates of arthroscopic repair.²⁵

Attendings unaccustomed to arthroscopy may find it more challenging than the younger generation of surgeons, who are exposed to it early in training. Dunn and colleagues²¹ noted that the likelihood of performing an arthroscopic repair was influenced by the surgeon’s experience level. Fellowship-trained shoulder and sports medicine surgeons are also more likely to perform arthroscopic repairs than those with training limited to orthopedic residency.²⁵ Arthroscopic RCR demands a high level of technical skill that many acquire in fellowship training.²⁶ Mauro and colleagues²⁶ found that surgeons trained in a sports medicine fellowship performed 82.6% of subacromial decompression and/or RCR procedures arthroscopically, compared with 54.5% to 70.1% for surgeons trained in other fellowships. In our cohort, with the exception of 1 surgeon, all fellowship-trained shoulder and sports medicine surgeons performed arthroscopic RCRs.

Although no conclusive evidence in the literature supports arthroscopic over the other repair types, the demand for arthroscopic RCR has rapidly increased relative to that for the others. Between 1996 and 2006, use of arthroscopic RCR increased 600%, from 8 to 58 per 100,000 population.²⁰ In that same period, use of open RCR increased by only 34%.²⁰ Similarly, Mauro and colleagues²⁶ found that the proportion of subacromial decompression and RCRs performed arthroscopically rose from 58.3% in 2004 to 83.7% in 2009. Using the 2006 New York State Ambulatory Surgery Database, Churchill and Ghorai²⁷ found that 74.5% of RCRs with acromioplasty were performed arthroscopically.

Respondent-indicated factors that may have contributed to the more favorable opinion of arthroscopic and mini-open repair include quick healing, good cosmetic results, and better perceived patient satisfaction. The literature supports these perceptions. Baker and Liu¹⁴ found shorter hospital stays and quicker return to activity with arthroscopic repair compared with open repair. Vitale and colleagues²⁵ also noted that, compared with open or mini-open repair techniques, arthroscopic repair resulted in shorter hospitalization and quicker overall recovery.

If these selected health care professionals with some inside information on rotator cuff surgery have biases that affect their selection of rotator cuff procedures, we should acknowledge that nonmedical personnel, in particular our patients, also have biases. The knowledge base of patients may be further influenced by friends or family members who have had rotator cuff surgery, by lay publications, and by the Internet. Satisfaction with any surgical procedure depends not only on the success of the surgery and the rehabilitation but also on patient and provider expectations. Such expectations are influenced, in part, by biases.

Our medical professionals had similar opinions on safety, recovery, cosmesis, and overall outcome of the RCR techniques, but different opinions on procedure durations and associated training requirements. All residents except one indicated open repair was a quick procedure. In contrast, a significant number of AHPs thought open repair was time-consuming. The attendings considered all the methods fast. The residents’ opinions were the most consistent with the true operating times reported. According to the literature, total operating time for mini-open repair ranges from 10 to 16 minutes faster than that for arthroscopic repair.^18,20,27 Ultimately, procedure duration did not affect the respondents’ technique preference for RCR.

There was substantial disagreement about the number of procedures needed to become proficient in the different repair techniques. Overall, however, there was consensus that arthroscopic and mini-open repairs had longer learning curves than open repair. Given the lack of agreement among orthopedic department chairmen and sports medicine fellowship directors regarding the minimum exposure needed (during residency) to become proficient in diagnostic shoulder arthroscopy,²⁸ this finding is not surprising. Guttmann and colleagues²⁹ attempted to quantify the learning curve for arthroscopic RCR by tracking operating time as a surrogate measure. They found that RCR operative time decreased rapidly during the initial block of 10 cases to the second block of 10 cases, but thereafter improvement continued at a much lower rate.²⁹ None of our respondents thought the learning curve for arthroscopic RCR was under 10 cases, but no group, not even the attendings who performed RCRs, could agree on the minimum number of cases needed for proficiency. The longer learning curve for arthroscopic RCR did not discourage the respondents who preferred arthroscopic or mini-open RCR.

Cost was not an influential factor in opinions about which RCR method is optimal. Medical professionals were ambivalent about the cost-effectiveness of the different procedures, with most expressing no opinion on cost. Multiple investigators have shown that arthroscopic RCR costs as much as $1144 more than mini-open RCR,^18,27 which has many of the advantages of arthroscopic repair but not the costly implants and instruments. As our medical community becomes more cost-conscious, concern about this factor may increase among medical professionals.

Our study had several limitations. Its results must be interpreted carefully, given they represent the viewpoints of a nonrandomized sample of motivated respondents at one institution. A selection bias excluded surgeons who were uncomfortable with RCR and unwilling to report any shortcomings. The conclusions cannot be generalized to other medical professionals or to other institutions. Furthermore, to develop a simple, straightforward survey focused on a specific type of rotator cuff tear, and to avoid confusion, we assumed that the treatment preference for the described tear was generalizable to all encountered tears. However, some surgeons have reported different repair techniques for different types and sizes of rotator cuff tears.²⁵

Conclusion

Most of our surveyed medical professionals were willing to defer to their surgeon’s decision about which technique would be appropriate for their own personal RCR. There is a trend nationally, and at our institution, for increased use of arthroscopic RCR. Although medical professionals readily acknowledge it is unclear which repair method provides the best ultimate outcome, many perceive fast recovery and good cosmetic results with arthroscopic and mini-open repairs. When medical professionals are counseling patients, we need to recognize these personal biases because many patients defer to their surgeon’s counsel. For some medical professionals, cosmesis can be an important factor, but cost, procedure duration, potential technical challenges of arthroscopic repair, and other considerations may make other techniques more desirable for others.

THE CASE

A 72-year-old woman came to our internal medicine department clinic for a follow-up appointment for her fibromyalgia. Thirteen months earlier, she had sought care at our facility not only for fibromyalgia, but for insomnia, anxiety, depression, and urinary incontinence. At the time, we prescribed amitriptyline 10 mg/d—for her pain and depression—as well as clonazepam 10 mg/d and paracetamol 650 mg, as needed.

When she came in for the follow-up, she indicated that for the past 8 months, she’d been experiencing urinary retention that required her to self-catheterize 2 to 3 times a day. She said she hadn’t used other medicines or herbal products during this time.

The patient had visited her family physician several times over the previous few months, and had been referred to a urologist. During an episode of acute urinary retention, she went to the emergency department (ED), where the ED physician performed urinary catheterization and referred her to the hospital’s Urology Department. After 48 hours, she was evaluated by a urologist, who diagnosed chronic urinary retention related to a hypercontractile bladder, without any particular cause. She was advised to continue to catheterize herself when needed. She was also prescribed pyridostigmine bromide, but she stopped taking it because of abdominal pain and bloating.

Two months prior to her visit with us, the patient suffered a second acute urinary retention episode and returned to the ED. Urinary catheterization was performed for 72 hours. At her next visit to her urologist, she was told to continue self-catheterization and was prescribed silodosin 8 mg/d.

THE DIAGNOSIS

Based on the patient’s history, we suspected the urinary retention was secondary to the anticholinergic effects of amitriptyline. We were able to determine that the patient’s urinary retention was likely the result of an adverse drug reaction (ADR) by using the causality algorithm of the Spanish Pharmacovigilance System, which suggests the following criteria:¹ a) a positive time sequence (ie, onset of symptoms closely followed administration of the medication), b) the existence of an ADR that is well known and consistent with the mechanism of action of the drug,² c) symptoms that resolve after suspending the drug; d) no repeat exposure (to the adverse effects of amitriptyline) due to ethical reasons; and e) the absence of an alternative explanation for the symptoms.³

DISCUSSION

Although indicated for depression, amitriptyline is also used for other conditions, including nocturnal enuresis and chronic neuropathic pain.⁴ Amitriptyline exhibits anticholinergic effects that can cause symptoms related to the nervous system (agitation, disorientation, sleepiness, delirium, cognitive impairment), ocular system (blurred vision, dry eye, accommodation disturbances, increased intraocular pressure), cardiovascular system (tachycardia), gastrointestinal tract (dry mouth, paralytic ileus, constipation), urinary system (urinary retention); and skin and mucosal membranes (dryness).^5,6 Anticholinergic effects can also induce hyperthermia or increase the risk of falls.^5,6

Four other physicians had seen our patient, and none had considered the possibility that this was an adverse drug effect.

Anticholinergic medications can cause ADRs in high-risk older patients and thus are usually considered inappropriate for this patient population.⁶ The Anticholinergic Risk Scale (ARS) can be used to categorize medications based on their potential for anticholinergic adverse effects (TABLE).⁷ Amitriptyline is included in the group with the highest risk of ADRs. Amitriptyline is also included in the list of drugs that should be avoided in older adults, according to the 2012 American Geriatrics Society Beers Criteria.⁸

Our patient. We instructed her to stop taking amitriptyline, and her urinary retention disappeared within 48 hours. Two months later, she remained asymptomatic.

THE TAKEAWAY

Although many medications are known to cause adverse events, they can be missed when clinicians fail to pinpoint exactly when a new sign, symptom, or health problem appeared. This often leads to a chain reaction of unnecessary explorations, harmful treatment, patient suffering, and unjustified costs.^9-11 Our patient had seen 4 different health care providers (a family physician, urologist, and 2 ED physicians) before we saw her and ultimately made the diagnosis. Family physicians can prevent anticholinergic ADRs by using a scale, such as the ARS, before prescribing a medication.

THE CASE

A 72-year-old woman came to our internal medicine department clinic for a follow-up appointment for her fibromyalgia. Thirteen months earlier, she had sought care at our facility not only for fibromyalgia, but for insomnia, anxiety, depression, and urinary incontinence. At the time, we prescribed amitriptyline 10 mg/d—for her pain and depression—as well as clonazepam 10 mg/d and paracetamol 650 mg, as needed.

When she came in for the follow-up, she indicated that for the past 8 months, she’d been experiencing urinary retention that required her to self-catheterize 2 to 3 times a day. She said she hadn’t used other medicines or herbal products during this time.

The patient had visited her family physician several times over the previous few months, and had been referred to a urologist. During an episode of acute urinary retention, she went to the emergency department (ED), where the ED physician performed urinary catheterization and referred her to the hospital’s Urology Department. After 48 hours, she was evaluated by a urologist, who diagnosed chronic urinary retention related to a hypercontractile bladder, without any particular cause. She was advised to continue to catheterize herself when needed. She was also prescribed pyridostigmine bromide, but she stopped taking it because of abdominal pain and bloating.

Two months prior to her visit with us, the patient suffered a second acute urinary retention episode and returned to the ED. Urinary catheterization was performed for 72 hours. At her next visit to her urologist, she was told to continue self-catheterization and was prescribed silodosin 8 mg/d.

THE DIAGNOSIS

Based on the patient’s history, we suspected the urinary retention was secondary to the anticholinergic effects of amitriptyline. We were able to determine that the patient’s urinary retention was likely the result of an adverse drug reaction (ADR) by using the causality algorithm of the Spanish Pharmacovigilance System, which suggests the following criteria:¹ a) a positive time sequence (ie, onset of symptoms closely followed administration of the medication), b) the existence of an ADR that is well known and consistent with the mechanism of action of the drug,² c) symptoms that resolve after suspending the drug; d) no repeat exposure (to the adverse effects of amitriptyline) due to ethical reasons; and e) the absence of an alternative explanation for the symptoms.³

DISCUSSION

Although indicated for depression, amitriptyline is also used for other conditions, including nocturnal enuresis and chronic neuropathic pain.⁴ Amitriptyline exhibits anticholinergic effects that can cause symptoms related to the nervous system (agitation, disorientation, sleepiness, delirium, cognitive impairment), ocular system (blurred vision, dry eye, accommodation disturbances, increased intraocular pressure), cardiovascular system (tachycardia), gastrointestinal tract (dry mouth, paralytic ileus, constipation), urinary system (urinary retention); and skin and mucosal membranes (dryness).^5,6 Anticholinergic effects can also induce hyperthermia or increase the risk of falls.^5,6

Four other physicians had seen our patient, and none had considered the possibility that this was an adverse drug effect.

Anticholinergic medications can cause ADRs in high-risk older patients and thus are usually considered inappropriate for this patient population.⁶ The Anticholinergic Risk Scale (ARS) can be used to categorize medications based on their potential for anticholinergic adverse effects (TABLE).⁷ Amitriptyline is included in the group with the highest risk of ADRs. Amitriptyline is also included in the list of drugs that should be avoided in older adults, according to the 2012 American Geriatrics Society Beers Criteria.⁸

Our patient. We instructed her to stop taking amitriptyline, and her urinary retention disappeared within 48 hours. Two months later, she remained asymptomatic.

THE TAKEAWAY

Although many medications are known to cause adverse events, they can be missed when clinicians fail to pinpoint exactly when a new sign, symptom, or health problem appeared. This often leads to a chain reaction of unnecessary explorations, harmful treatment, patient suffering, and unjustified costs.^9-11 Our patient had seen 4 different health care providers (a family physician, urologist, and 2 ED physicians) before we saw her and ultimately made the diagnosis. Family physicians can prevent anticholinergic ADRs by using a scale, such as the ARS, before prescribing a medication.

THE CASE

A 72-year-old woman came to our internal medicine department clinic for a follow-up appointment for her fibromyalgia. Thirteen months earlier, she had sought care at our facility not only for fibromyalgia, but for insomnia, anxiety, depression, and urinary incontinence. At the time, we prescribed amitriptyline 10 mg/d—for her pain and depression—as well as clonazepam 10 mg/d and paracetamol 650 mg, as needed.

When she came in for the follow-up, she indicated that for the past 8 months, she’d been experiencing urinary retention that required her to self-catheterize 2 to 3 times a day. She said she hadn’t used other medicines or herbal products during this time.

The patient had visited her family physician several times over the previous few months, and had been referred to a urologist. During an episode of acute urinary retention, she went to the emergency department (ED), where the ED physician performed urinary catheterization and referred her to the hospital’s Urology Department. After 48 hours, she was evaluated by a urologist, who diagnosed chronic urinary retention related to a hypercontractile bladder, without any particular cause. She was advised to continue to catheterize herself when needed. She was also prescribed pyridostigmine bromide, but she stopped taking it because of abdominal pain and bloating.

Two months prior to her visit with us, the patient suffered a second acute urinary retention episode and returned to the ED. Urinary catheterization was performed for 72 hours. At her next visit to her urologist, she was told to continue self-catheterization and was prescribed silodosin 8 mg/d.

THE DIAGNOSIS

Based on the patient’s history, we suspected the urinary retention was secondary to the anticholinergic effects of amitriptyline. We were able to determine that the patient’s urinary retention was likely the result of an adverse drug reaction (ADR) by using the causality algorithm of the Spanish Pharmacovigilance System, which suggests the following criteria:¹ a) a positive time sequence (ie, onset of symptoms closely followed administration of the medication), b) the existence of an ADR that is well known and consistent with the mechanism of action of the drug,² c) symptoms that resolve after suspending the drug; d) no repeat exposure (to the adverse effects of amitriptyline) due to ethical reasons; and e) the absence of an alternative explanation for the symptoms.³

DISCUSSION

Although indicated for depression, amitriptyline is also used for other conditions, including nocturnal enuresis and chronic neuropathic pain.⁴ Amitriptyline exhibits anticholinergic effects that can cause symptoms related to the nervous system (agitation, disorientation, sleepiness, delirium, cognitive impairment), ocular system (blurred vision, dry eye, accommodation disturbances, increased intraocular pressure), cardiovascular system (tachycardia), gastrointestinal tract (dry mouth, paralytic ileus, constipation), urinary system (urinary retention); and skin and mucosal membranes (dryness).^5,6 Anticholinergic effects can also induce hyperthermia or increase the risk of falls.^5,6

Four other physicians had seen our patient, and none had considered the possibility that this was an adverse drug effect.

Anticholinergic medications can cause ADRs in high-risk older patients and thus are usually considered inappropriate for this patient population.⁶ The Anticholinergic Risk Scale (ARS) can be used to categorize medications based on their potential for anticholinergic adverse effects (TABLE).⁷ Amitriptyline is included in the group with the highest risk of ADRs. Amitriptyline is also included in the list of drugs that should be avoided in older adults, according to the 2012 American Geriatrics Society Beers Criteria.⁸

Our patient. We instructed her to stop taking amitriptyline, and her urinary retention disappeared within 48 hours. Two months later, she remained asymptomatic.

THE TAKEAWAY

Although many medications are known to cause adverse events, they can be missed when clinicians fail to pinpoint exactly when a new sign, symptom, or health problem appeared. This often leads to a chain reaction of unnecessary explorations, harmful treatment, patient suffering, and unjustified costs.^9-11 Our patient had seen 4 different health care providers (a family physician, urologist, and 2 ED physicians) before we saw her and ultimately made the diagnosis. Family physicians can prevent anticholinergic ADRs by using a scale, such as the ARS, before prescribing a medication.

Tourniquets are often used in total knee arthroplasty (TKA) to improve visualization of structures, shorten operative time, reduce intraoperative bleeding, and improve cementing technique. Despite these advantages, controversy remains regarding the safety of tourniquet use. Tourniquets have been associated with nerve palsies, vascular injury, and muscle damage.^1-5 Some have hypothesized they may cause venous stasis or direct endothelial damage that may develop into deep vein thrombosis (DVT). Abdel-Salam and Eyres⁶ found an increased incidence of postoperative wound complications and DVTs associated with tourniquet use.

Moreover, investigators have analyzed the role of tourniquets in populations at high risk for wound complications. DeLaurentis and colleagues⁷ performed a prospective and retrospective analysis of 1182 TKA patients, 24 (2%) of whom had preexisting peripheral vascular disease (PVD), defined as a history of arterial insufficiency, absent dorsalis pedis and/or absent posterior tibial pulsations, and arterial calcifications. A tourniquet was used in each case. Arterial complications occurred in 6 of the 24 patients with PVD. As expected, the authors found that a history of intermittent claudication, pain at rest, and arterial ulcers resulted in a high risk for vascular complications. Further studies have supported this finding and expanded the list of predisposing factors to include previous vascular surgery and absent and asymmetric pedal pulsations.^7-11 Of particular concern to total joint arthroplasty surgeons was the finding by DeLaurentis and colleagues⁷ that patients with radiographic evidence of calcification of the distal superficial femoral artery and/or popliteal artery were at risk for arterial complications. This finding is also supported by other studies.^8,11 In TKA, damage to arterial structures proximal to the surgical field could manifest as impaired postoperative wound healing or an ischemic limb. Wound healing depends on adequate blood flow to the healing tissue, and any damage to arterial or venous structures can theoretically compromise this process.

Added to vascular/wound complications as concerning complications in orthopedic surgery is venous thromboembolism (VTE). The role of tourniquets in the formation of VTEs is controversial. A tourniquet has the potential to increase the risk for DVT because of the stasis of venous blood in the lower limb or possible damage to calcified blood vessels. Callam and colleagues¹² studied the connection between artery disease and chronic leg ulcers and found that half the patients diagnosed with peripheral artery disease also had stigmata of chronic venous insufficiency. Therefore, the entities can occur in tandem, and surgeons should keep this in mind.

Here we report on a study we conducted to determine whether tourniquet use in TKA in patients with preexisting radiographic evidence of vascular disease increases the risk for wound complications or VTE.

Patients and Methods

We retrospectively reviewed 461 consecutive primary TKAs (373 patients) performed between January 2007 and June 2012 by 2 attending orthopedic surgeons specializing in adult reconstruction. Medical records and operative reports of 583 patients were examined after receiving institutional review board approval. Of these patients, 373 (64%) had a minimum of 12-month follow-up data available. Twelve months was deemed long enough to discover wound complications or DVTs secondary to the index procedure. Most of these outcomes manifest within the first 3 months after surgery and certainly by 12 months. Follow-up longer than 12 months may become a confounder, as wound complications outside the acute to subacute postoperative window could be related to patients’ underlying PVD and not directly to tourniquet use during surgery. Patient demographics and comorbidities were recorded. Comorbidities were obtained from preoperative medical evaluations and surgeons’ preoperative evaluations. All patients had preoperative palpable dorsalis pedis and posterior tibialis arterial pulses. No patient required preoperative vascular studies based on preoperative examination or comorbidities. No patient had prior vascular bypass surgery or stenting.

TKA was performed in a nonlaminar flow, positive-pressure, high-efficiency particulate air-filtered room with sterile toga/surgical helmet systems. For all patients, a pneumatic thigh tourniquet was applied, and the patient was prepared and draped. After limb exsanguination using a rubber bandage, the limb was elevated and the tourniquet inflated to a pressure of 250 to 300 mm Hg. The tourniquet was released either just before closure or immediately after closure in all cases; it was always let down before placement of final bandages.

Prophylactic chemical anticoagulation consisting of warfarin, aspirin, or enoxaparin was used in all patients and continued for 4 to 6 weeks after surgery. All patients received mechanical DVT prophylaxis with sequential compression devices, and all were mobilized out of bed beginning either the day of surgery or the next day. All patients received perioperative intravenous antibiotics, with the preoperative dose given before tourniquet inflation and the last postoperative dose stopped within 24 hours of surgery.

All patients who had primary TKA underwent preoperative medical evaluation and optimization. The patient’s hospital course was monitored closely, and complications noted by the orthopedic team were documented. Follow-up documentation was retrospectively reviewed for evidence of wound complications or VTE. Wound complications were defined as cellulitis, delayed wound healing, wound dehiscence, and/or periprosthetic joint infection. In the case of VTE, physical examination findings were not sufficient for inclusion. Venous duplex ultrasonography demonstrating the clot was reviewed before inclusion.

Preoperative radiographs were examined for arterial calcification (Figure). We refer to calcification seen above the knee joint as proximal calcification and to calcification observed below the joint as distal calcification. Patients exhibited calcification proximally only, distally only, or both proximally and distally. The 373 patients were placed into 2 groups based on whether they had preoperative arterial calcification on plain radiography of the knee. One group (285 patients with no radiographic evidence of preoperative knee arterial calcification) underwent 365 TKAs, and the other group (88 patients with radiographic evidence of preoperative knee arterial calcification) underwent 96 TKAs.

A sample size calculation was performed to determine how many patients were needed in each group with 80% power and an α of 0.05. With an estimated difference in VTE/wound complication rate between the calcification and no-calcification groups of 12%, we needed to review 316 TKAs total. This 12% difference was based on study findings of a 25% complication rate in PVD patients who underwent tourniquet-assisted TKA, and the rate of VTE/wound complication after TKA in patients overall, which can be up to 12%.^7,13,14 We exceeded minimal enrollment and had 461 TKAs. Descriptive statistics were reported, with means and ranges provided where appropriate. Independent t test was used to evaluate the differences in continuous data (age) between the groups. Univariate analysis (using Pearson χ² and Fisher exact tests) and multivariate logistic regression analysis were used to evaluate the effects of categorical variables (sex, comorbidity, calcification [presence, absence], and location of calcification [proximal only, distal only, both]) on wound complication and VTE rates. All tests were 2-tailed and performed with a type I error rate of 0.05. Data analysis was performed with SPSS Version 19.0 (SPSS).

Results

Patient characteristics are summarized in Table 1. Of the 373 patients, 285 lacked calcification, and 88 had calcification. Mean age was 67.73 years (range, 24-92 years) for all patients, 65.99 years (range, 24-89 years) for the no-calcification group, and 74.32 years (range, 54-92 years) for the calcification group; the calcification group demonstrated a trend toward older age, but the difference was not significantly different (P = .07). Of the 373 patients, 156 (41.82%) were male: 110 in the no-calcification group (38.60%) and 46 in the calcification group (52.27%); sex was significantly (P = .002) different between groups, with more males in the calcification group.

Data on total preoperative comorbidities are summarized in Table 2. Hypertension, hyperlipidemia, diabetes, and coronary artery disease (CAD) were the most common comorbidities, and they were all significantly (P ≤ .05) increased in the calcification group.

No patients had reported arterial complications, such as arterial bleeding, aneurysm, intimal tears, or loss of distal pulses. Wound complication after TKA was detected in 3.04% of all cases (Table 3). Rate of DVT after TKA was 2.60% of all cases, and rate of pulmonary embolism after TKA was 2.17% of all cases. Of the 96 TKAs with preoperative radiographic evidence of calcification, 47 (48.96%) had proximal calcification only, 11 (11.46%) had distal calcification only, and 38 (39.58%) had both proximal and distal calcification (Table 4). There was no significant difference between the rate of wound complication or VTE based on location of vascular calcification.

Univariate analysis demonstrated that presence of arterial knee calcification did not increase the risk for postoperative wound complication (odds ratio [OR], 1.04; 95% confidence interval [CI], 0.28-3.80; P > .05) (Table 5). Location of arterial knee calcification also did not increase the risk for postoperative wound complication. In addition, univariate analysis demonstrated that presence of arterial knee calcification did not increase the risk for postoperative VTE (OR, 1.20; 95% CI, 0.43-3.36; P > .05 (Table 6).

Of the 14 wound complications, the most common infections were cellulitis (5/14 cases; 35.71%) and infected hardware that required component revision (5/14 cases; 35.71%). Mean time from TKA to infection was 137.93 days (range, 5-783 days). The most common organism grown in culture from the wound was Staphylococcus (5/14 cases; 35.71%).

Additional univariate statistical analysis revealed that presence of diabetes, hypertension, prior VTE, CAD, and male sex was linked to higher incidence of wound complication (P < .05) (Table 5). When multivariate analysis was performed, hypertension, prior VTE, and male sex remained significant (P < .05) (Table 5).

Discussion

TKA is a safe and effective procedure used to treat osteoarthritis of the knee and improve patients’ quality of life.¹⁵ About 700,000 TKAs are performed annually in the United States.¹⁶ Because of improvements in preventive medicine and medical technology, life expectancy is increasing, and TKAs are now being performed in higher numbers and in an older patient population. Over the next few decades, these developments will lead to more postoperative complications. It is projected that, by 2030, the need for TKAs in the United States will increase by 673% to 3.48 million.¹⁷ Postoperative complications are rare but unfortunately often lead to poor outcomes or even mortality.¹⁸ To help minimize the number of postoperative complications, we must understand the safety of tourniquet use in TKA. Other investigators have concluded that tourniquet use is unsafe in patients with preoperative vascular calcifications on plain radiographs.^7,8,11 The present study, designed to elucidate whether preoperative evidence of knee arterial calcification may predispose TKA patients to postoperative wound complication or VTE, had some important findings.

In our study, wound complication and VTE occurred in a considerable number of patients after TKA. Despite exceeding the number of patients calculated by the power analysis, our population may have been inadequate to fully detect statistical significance. Thus, our conclusion of failing to reject the null hypothesis may have been because of sample size, a type II error. We found that, after primary TKA, 3.04% of patients developed wound complications and 4.77% VTE. According to the literature, the incidence of infection after primary TKA is between 0.5% and 12%, and that of VTE reported within 3 months after TKA is 1.3% to 10%.^13,14 Although we had 100% VTE prophylaxis, meeting the standard of care, VTE after TKA remains a postoperative complication.¹⁹ This study also found that a considerable percentage of primary TKA patients (23.59%) had preoperative calcification of the knee arteries. To our knowledge, this study was the first to quantify the incidence of knee arterial calcification in patients who underwent TKA.

Preoperative calcification of the knee arteries in patients who underwent TKA did not increase the risk for wound complication, VTE, or arterial damage. These calcifications, however, do pose an increased systemic vascular risk.²⁰ Calcification of the vascular wall predicts increased cardiovascular risk, independent of classical cardiovascular risk factors.^3,18,21-24 Clinically, patients who have both diabetes and calcifications are at significant excess risk for total mortality, stroke mortality, and cardiovascular mortality, compared with patients with diabetes but without such calcifications. They also had a significantly higher incidence of coronary heart disease events, stroke events, and lower extremity amputations.^25,26

All our patients underwent tourniquet-assisted TKA. Although previous studies have indicated that tourniquet use may increase arterial complications and wound complications or even limb loss in patients with calcified arteries, we did not find this link.^7,27 Our population had no reported arterial complications related to tourniquet use. Other, smaller studies have had similar findings. Vandenbussche and colleagues²⁸ prospectively studied 80 TKA cases randomized to tourniquet use or no tourniquet use and found no postoperative nerve palsies, wound infections, wound healing problems, or hematomas. Our study is also in accord with studies that have reported tourniquet use did not increase risk for DVT.²⁹ Therefore, unlike earlier data, our data demonstrated that tourniquet use in patients with knee arterial calcification was safe.^7,27,30,31

Patients with calcification were more likely to have the medical comorbidities of hypertension, diabetes, hyperlipidemia, and CAD. All these comorbidities are linked to the development of arterial calcification, or atherosclerotic occlusive disease.^32,33 As life expectancy and the need for TKA increase, it is likely that a larger percentage of TKA patients will have preoperative radiographic evidence of knee arterial calcification. Although current dogma is that tourniquet-assisted TKA is contraindicated for patients with preoperative radiographic evidence of femoral-popliteal calcification, our study results showed that this calcification should not affect preoperative TKA planning for these patients.

We divided our patients into 3 categories: those with proximal calcification (above the joint line), those with distal calcification (below the joint line), and those with both proximal and distal calcification. Location of arterial calcification did not have an effect on their rates of postoperative wound complication or VTE. We hypothesized that patients with proximal calcification would be at increased risk for direct arterial injury and subsequent wound complication because the tourniquet is placed proximally. Previous research has indicated that arterial occlusion and subsequent wound complication can occur because of low blood flow stemming from tourniquet use.⁷ Further, intraoperative manipulation (flexing) of a knee with calcified vessels causes arterial complications after TKA because these vessels are less elastic than nonatheromatous vessels.³¹ However, we found no such effect. At the same time, having arterial calcification might also be an indication of venous disease in this location,¹² which may be especially important for proximal calcifications. Proximal DVT more likely is a precursor to pulmonary embolic events than distal DVT is.^31,34 However, we found no difference in VTE rates among the 3 arterial location groups, which is supported by studies that have found that tourniquet use does not increase DVT incidence.^29,35-40

Risk for wound complications was higher in male patients and in patients with diabetes, prior VTE, hypertension, or CAD. This finding is important because, with the increasing age of patients who undergo TKA, those with serious medical comorbidities will continue to need and have this surgery.¹⁷ Diabetes may increase the rate of wound complication because patients with diabetes have poor microcirculation, poor collagen synthesis, and reduced wound strength.⁴¹ Malinzak and colleagues⁴² demonstrated that, compared with patients without diabetes, those with diabetes had a significantly higher risk for infection after TKA. Prior VTE, specifically DVT, may increase the rate of wound complication because after DVT the deep veins may be damaged and exhibit valvular dysfunction. Labropoulos and colleagues⁴³ showed that DVT history was strongly associated with ulcer nonhealing. Perhaps hypertension has been overlooked as a risk factor for wound complication in TKA. No previous studies have assessed the link between hypertension and wound complications after TKA. However, a study of wound healing after total hip arthroplasty found that, compared with normotensive patients, hypertensive patients had delayed wound healing, putting them at higher risk for infection.⁴⁴ In addition, we found that patients with CAD were at increased risk for wound complications—an unexpected finding, as CAD traditionally is not a risk factor for infection or poor wound healing. Recently, however, CAD was identified as an independent risk factor for surgical site infections in posterior lumbar–instrumented arthrodesis.⁴⁵ The etiology of this association is unknown. Also, male patients were at increased risk for wound complication. Male sex has been implicated as an independent risk factor for development of surgical site infections and has been established as an important predisposing factor for periprosthetic joint infections.⁴⁶

It is possible that patients who present with diabetes, VTE, hypertension, or CAD before TKA should have a consultation with a vascular surgeon or should have TKA performed without a tourniquet, but this conclusion cannot be considered definitive without a large prospective randomized trial or possibly registry data. Our data indicate that patients with these comorbidities have higher rates of wound complications irrespective of preoperative radiographic calcifications. On the basis of our study results, however, we certainly recommend that patients with these risk factors have preoperative medical optimization. Orthopedic surgeons should take a thorough history and perform a meticulous physical examination on these patients to look for evidence of PVD. We recommend that, if vascular claudication is elicited in the history, or if there is evidence of peripheral arterial disease—such as hair loss, skin discoloration, dystrophic nail changes, or absent or unequal peripheral pulses—the ankle-brachial index test should be performed. If the index value is less than 0.9, then a preoperative vascular surgery consultation should be obtained.

This study had some weaknesses. First, it was retrospective, so it is possible that some wound or VTE complications were not reported and thus not found in the paper charts or electronic medical records. Some patients may have had VTE diagnostic scans at other hospitals, and their results may not have been recorded across databases. Moreover, some patients may have seen wound specialists for wound infections or wound healing problems, and these may not have been reported to the orthopedic surgeons. Second, though our patient population was not small, it may not have been of adequate size to fully detect statistical significance. We met our enrollment numbers based on our sample size calculations from an a priori power analysis; however, we still draw conclusions with the possibility of committing a type II error in mind by failing to reject the null hypothesis when in reality a statistically significant difference does exist. Third, none of our consecutive patients carried the preoperative diagnosis of PVD, and none had preoperative vascular surgery. Therefore, though calcifications were noted on radiographs, clinically our patients were asymptomatic with respect to vascular health. Last, the 2 groups were not randomized. All patients underwent tourniquet-assisted TKA.           

Conclusion

To our knowledge, this is the largest study to examine the effect of preoperative knee arterial calcification on wound complication and VTE after tourniquet-assisted TKA. Contrary to previously published recommendations, we conclude that TKA can be safely performed with a tourniquet in the presence of preoperative radiographic evidence of such calcification. However, we recommend that patients with diabetes, hypertension, CAD, or prior VTE undergo an appropriate physical examination to elicit any signs or symptoms of vascular disease. If before surgery there is any question of vascular competence, a vascular surgeon should be consulted.

Tourniquets are often used in total knee arthroplasty (TKA) to improve visualization of structures, shorten operative time, reduce intraoperative bleeding, and improve cementing technique. Despite these advantages, controversy remains regarding the safety of tourniquet use. Tourniquets have been associated with nerve palsies, vascular injury, and muscle damage.^1-5 Some have hypothesized they may cause venous stasis or direct endothelial damage that may develop into deep vein thrombosis (DVT). Abdel-Salam and Eyres⁶ found an increased incidence of postoperative wound complications and DVTs associated with tourniquet use.

Moreover, investigators have analyzed the role of tourniquets in populations at high risk for wound complications. DeLaurentis and colleagues⁷ performed a prospective and retrospective analysis of 1182 TKA patients, 24 (2%) of whom had preexisting peripheral vascular disease (PVD), defined as a history of arterial insufficiency, absent dorsalis pedis and/or absent posterior tibial pulsations, and arterial calcifications. A tourniquet was used in each case. Arterial complications occurred in 6 of the 24 patients with PVD. As expected, the authors found that a history of intermittent claudication, pain at rest, and arterial ulcers resulted in a high risk for vascular complications. Further studies have supported this finding and expanded the list of predisposing factors to include previous vascular surgery and absent and asymmetric pedal pulsations.^7-11 Of particular concern to total joint arthroplasty surgeons was the finding by DeLaurentis and colleagues⁷ that patients with radiographic evidence of calcification of the distal superficial femoral artery and/or popliteal artery were at risk for arterial complications. This finding is also supported by other studies.^8,11 In TKA, damage to arterial structures proximal to the surgical field could manifest as impaired postoperative wound healing or an ischemic limb. Wound healing depends on adequate blood flow to the healing tissue, and any damage to arterial or venous structures can theoretically compromise this process.

Added to vascular/wound complications as concerning complications in orthopedic surgery is venous thromboembolism (VTE). The role of tourniquets in the formation of VTEs is controversial. A tourniquet has the potential to increase the risk for DVT because of the stasis of venous blood in the lower limb or possible damage to calcified blood vessels. Callam and colleagues¹² studied the connection between artery disease and chronic leg ulcers and found that half the patients diagnosed with peripheral artery disease also had stigmata of chronic venous insufficiency. Therefore, the entities can occur in tandem, and surgeons should keep this in mind.

Here we report on a study we conducted to determine whether tourniquet use in TKA in patients with preexisting radiographic evidence of vascular disease increases the risk for wound complications or VTE.

Patients and Methods

We retrospectively reviewed 461 consecutive primary TKAs (373 patients) performed between January 2007 and June 2012 by 2 attending orthopedic surgeons specializing in adult reconstruction. Medical records and operative reports of 583 patients were examined after receiving institutional review board approval. Of these patients, 373 (64%) had a minimum of 12-month follow-up data available. Twelve months was deemed long enough to discover wound complications or DVTs secondary to the index procedure. Most of these outcomes manifest within the first 3 months after surgery and certainly by 12 months. Follow-up longer than 12 months may become a confounder, as wound complications outside the acute to subacute postoperative window could be related to patients’ underlying PVD and not directly to tourniquet use during surgery. Patient demographics and comorbidities were recorded. Comorbidities were obtained from preoperative medical evaluations and surgeons’ preoperative evaluations. All patients had preoperative palpable dorsalis pedis and posterior tibialis arterial pulses. No patient required preoperative vascular studies based on preoperative examination or comorbidities. No patient had prior vascular bypass surgery or stenting.

TKA was performed in a nonlaminar flow, positive-pressure, high-efficiency particulate air-filtered room with sterile toga/surgical helmet systems. For all patients, a pneumatic thigh tourniquet was applied, and the patient was prepared and draped. After limb exsanguination using a rubber bandage, the limb was elevated and the tourniquet inflated to a pressure of 250 to 300 mm Hg. The tourniquet was released either just before closure or immediately after closure in all cases; it was always let down before placement of final bandages.

Prophylactic chemical anticoagulation consisting of warfarin, aspirin, or enoxaparin was used in all patients and continued for 4 to 6 weeks after surgery. All patients received mechanical DVT prophylaxis with sequential compression devices, and all were mobilized out of bed beginning either the day of surgery or the next day. All patients received perioperative intravenous antibiotics, with the preoperative dose given before tourniquet inflation and the last postoperative dose stopped within 24 hours of surgery.

All patients who had primary TKA underwent preoperative medical evaluation and optimization. The patient’s hospital course was monitored closely, and complications noted by the orthopedic team were documented. Follow-up documentation was retrospectively reviewed for evidence of wound complications or VTE. Wound complications were defined as cellulitis, delayed wound healing, wound dehiscence, and/or periprosthetic joint infection. In the case of VTE, physical examination findings were not sufficient for inclusion. Venous duplex ultrasonography demonstrating the clot was reviewed before inclusion.

Preoperative radiographs were examined for arterial calcification (Figure). We refer to calcification seen above the knee joint as proximal calcification and to calcification observed below the joint as distal calcification. Patients exhibited calcification proximally only, distally only, or both proximally and distally. The 373 patients were placed into 2 groups based on whether they had preoperative arterial calcification on plain radiography of the knee. One group (285 patients with no radiographic evidence of preoperative knee arterial calcification) underwent 365 TKAs, and the other group (88 patients with radiographic evidence of preoperative knee arterial calcification) underwent 96 TKAs.

A sample size calculation was performed to determine how many patients were needed in each group with 80% power and an α of 0.05. With an estimated difference in VTE/wound complication rate between the calcification and no-calcification groups of 12%, we needed to review 316 TKAs total. This 12% difference was based on study findings of a 25% complication rate in PVD patients who underwent tourniquet-assisted TKA, and the rate of VTE/wound complication after TKA in patients overall, which can be up to 12%.^7,13,14 We exceeded minimal enrollment and had 461 TKAs. Descriptive statistics were reported, with means and ranges provided where appropriate. Independent t test was used to evaluate the differences in continuous data (age) between the groups. Univariate analysis (using Pearson χ² and Fisher exact tests) and multivariate logistic regression analysis were used to evaluate the effects of categorical variables (sex, comorbidity, calcification [presence, absence], and location of calcification [proximal only, distal only, both]) on wound complication and VTE rates. All tests were 2-tailed and performed with a type I error rate of 0.05. Data analysis was performed with SPSS Version 19.0 (SPSS).

Results

Patient characteristics are summarized in Table 1. Of the 373 patients, 285 lacked calcification, and 88 had calcification. Mean age was 67.73 years (range, 24-92 years) for all patients, 65.99 years (range, 24-89 years) for the no-calcification group, and 74.32 years (range, 54-92 years) for the calcification group; the calcification group demonstrated a trend toward older age, but the difference was not significantly different (P = .07). Of the 373 patients, 156 (41.82%) were male: 110 in the no-calcification group (38.60%) and 46 in the calcification group (52.27%); sex was significantly (P = .002) different between groups, with more males in the calcification group.

Data on total preoperative comorbidities are summarized in Table 2. Hypertension, hyperlipidemia, diabetes, and coronary artery disease (CAD) were the most common comorbidities, and they were all significantly (P ≤ .05) increased in the calcification group.

No patients had reported arterial complications, such as arterial bleeding, aneurysm, intimal tears, or loss of distal pulses. Wound complication after TKA was detected in 3.04% of all cases (Table 3). Rate of DVT after TKA was 2.60% of all cases, and rate of pulmonary embolism after TKA was 2.17% of all cases. Of the 96 TKAs with preoperative radiographic evidence of calcification, 47 (48.96%) had proximal calcification only, 11 (11.46%) had distal calcification only, and 38 (39.58%) had both proximal and distal calcification (Table 4). There was no significant difference between the rate of wound complication or VTE based on location of vascular calcification.

Univariate analysis demonstrated that presence of arterial knee calcification did not increase the risk for postoperative wound complication (odds ratio [OR], 1.04; 95% confidence interval [CI], 0.28-3.80; P > .05) (Table 5). Location of arterial knee calcification also did not increase the risk for postoperative wound complication. In addition, univariate analysis demonstrated that presence of arterial knee calcification did not increase the risk for postoperative VTE (OR, 1.20; 95% CI, 0.43-3.36; P > .05 (Table 6).

Of the 14 wound complications, the most common infections were cellulitis (5/14 cases; 35.71%) and infected hardware that required component revision (5/14 cases; 35.71%). Mean time from TKA to infection was 137.93 days (range, 5-783 days). The most common organism grown in culture from the wound was Staphylococcus (5/14 cases; 35.71%).

Additional univariate statistical analysis revealed that presence of diabetes, hypertension, prior VTE, CAD, and male sex was linked to higher incidence of wound complication (P < .05) (Table 5). When multivariate analysis was performed, hypertension, prior VTE, and male sex remained significant (P < .05) (Table 5).

Discussion

TKA is a safe and effective procedure used to treat osteoarthritis of the knee and improve patients’ quality of life.¹⁵ About 700,000 TKAs are performed annually in the United States.¹⁶ Because of improvements in preventive medicine and medical technology, life expectancy is increasing, and TKAs are now being performed in higher numbers and in an older patient population. Over the next few decades, these developments will lead to more postoperative complications. It is projected that, by 2030, the need for TKAs in the United States will increase by 673% to 3.48 million.¹⁷ Postoperative complications are rare but unfortunately often lead to poor outcomes or even mortality.¹⁸ To help minimize the number of postoperative complications, we must understand the safety of tourniquet use in TKA. Other investigators have concluded that tourniquet use is unsafe in patients with preoperative vascular calcifications on plain radiographs.^7,8,11 The present study, designed to elucidate whether preoperative evidence of knee arterial calcification may predispose TKA patients to postoperative wound complication or VTE, had some important findings.

In our study, wound complication and VTE occurred in a considerable number of patients after TKA. Despite exceeding the number of patients calculated by the power analysis, our population may have been inadequate to fully detect statistical significance. Thus, our conclusion of failing to reject the null hypothesis may have been because of sample size, a type II error. We found that, after primary TKA, 3.04% of patients developed wound complications and 4.77% VTE. According to the literature, the incidence of infection after primary TKA is between 0.5% and 12%, and that of VTE reported within 3 months after TKA is 1.3% to 10%.^13,14 Although we had 100% VTE prophylaxis, meeting the standard of care, VTE after TKA remains a postoperative complication.¹⁹ This study also found that a considerable percentage of primary TKA patients (23.59%) had preoperative calcification of the knee arteries. To our knowledge, this study was the first to quantify the incidence of knee arterial calcification in patients who underwent TKA.

Preoperative calcification of the knee arteries in patients who underwent TKA did not increase the risk for wound complication, VTE, or arterial damage. These calcifications, however, do pose an increased systemic vascular risk.²⁰ Calcification of the vascular wall predicts increased cardiovascular risk, independent of classical cardiovascular risk factors.^3,18,21-24 Clinically, patients who have both diabetes and calcifications are at significant excess risk for total mortality, stroke mortality, and cardiovascular mortality, compared with patients with diabetes but without such calcifications. They also had a significantly higher incidence of coronary heart disease events, stroke events, and lower extremity amputations.^25,26

All our patients underwent tourniquet-assisted TKA. Although previous studies have indicated that tourniquet use may increase arterial complications and wound complications or even limb loss in patients with calcified arteries, we did not find this link.^7,27 Our population had no reported arterial complications related to tourniquet use. Other, smaller studies have had similar findings. Vandenbussche and colleagues²⁸ prospectively studied 80 TKA cases randomized to tourniquet use or no tourniquet use and found no postoperative nerve palsies, wound infections, wound healing problems, or hematomas. Our study is also in accord with studies that have reported tourniquet use did not increase risk for DVT.²⁹ Therefore, unlike earlier data, our data demonstrated that tourniquet use in patients with knee arterial calcification was safe.^7,27,30,31

Patients with calcification were more likely to have the medical comorbidities of hypertension, diabetes, hyperlipidemia, and CAD. All these comorbidities are linked to the development of arterial calcification, or atherosclerotic occlusive disease.^32,33 As life expectancy and the need for TKA increase, it is likely that a larger percentage of TKA patients will have preoperative radiographic evidence of knee arterial calcification. Although current dogma is that tourniquet-assisted TKA is contraindicated for patients with preoperative radiographic evidence of femoral-popliteal calcification, our study results showed that this calcification should not affect preoperative TKA planning for these patients.

We divided our patients into 3 categories: those with proximal calcification (above the joint line), those with distal calcification (below the joint line), and those with both proximal and distal calcification. Location of arterial calcification did not have an effect on their rates of postoperative wound complication or VTE. We hypothesized that patients with proximal calcification would be at increased risk for direct arterial injury and subsequent wound complication because the tourniquet is placed proximally. Previous research has indicated that arterial occlusion and subsequent wound complication can occur because of low blood flow stemming from tourniquet use.⁷ Further, intraoperative manipulation (flexing) of a knee with calcified vessels causes arterial complications after TKA because these vessels are less elastic than nonatheromatous vessels.³¹ However, we found no such effect. At the same time, having arterial calcification might also be an indication of venous disease in this location,¹² which may be especially important for proximal calcifications. Proximal DVT more likely is a precursor to pulmonary embolic events than distal DVT is.^31,34 However, we found no difference in VTE rates among the 3 arterial location groups, which is supported by studies that have found that tourniquet use does not increase DVT incidence.^29,35-40

Risk for wound complications was higher in male patients and in patients with diabetes, prior VTE, hypertension, or CAD. This finding is important because, with the increasing age of patients who undergo TKA, those with serious medical comorbidities will continue to need and have this surgery.¹⁷ Diabetes may increase the rate of wound complication because patients with diabetes have poor microcirculation, poor collagen synthesis, and reduced wound strength.⁴¹ Malinzak and colleagues⁴² demonstrated that, compared with patients without diabetes, those with diabetes had a significantly higher risk for infection after TKA. Prior VTE, specifically DVT, may increase the rate of wound complication because after DVT the deep veins may be damaged and exhibit valvular dysfunction. Labropoulos and colleagues⁴³ showed that DVT history was strongly associated with ulcer nonhealing. Perhaps hypertension has been overlooked as a risk factor for wound complication in TKA. No previous studies have assessed the link between hypertension and wound complications after TKA. However, a study of wound healing after total hip arthroplasty found that, compared with normotensive patients, hypertensive patients had delayed wound healing, putting them at higher risk for infection.⁴⁴ In addition, we found that patients with CAD were at increased risk for wound complications—an unexpected finding, as CAD traditionally is not a risk factor for infection or poor wound healing. Recently, however, CAD was identified as an independent risk factor for surgical site infections in posterior lumbar–instrumented arthrodesis.⁴⁵ The etiology of this association is unknown. Also, male patients were at increased risk for wound complication. Male sex has been implicated as an independent risk factor for development of surgical site infections and has been established as an important predisposing factor for periprosthetic joint infections.⁴⁶

It is possible that patients who present with diabetes, VTE, hypertension, or CAD before TKA should have a consultation with a vascular surgeon or should have TKA performed without a tourniquet, but this conclusion cannot be considered definitive without a large prospective randomized trial or possibly registry data. Our data indicate that patients with these comorbidities have higher rates of wound complications irrespective of preoperative radiographic calcifications. On the basis of our study results, however, we certainly recommend that patients with these risk factors have preoperative medical optimization. Orthopedic surgeons should take a thorough history and perform a meticulous physical examination on these patients to look for evidence of PVD. We recommend that, if vascular claudication is elicited in the history, or if there is evidence of peripheral arterial disease—such as hair loss, skin discoloration, dystrophic nail changes, or absent or unequal peripheral pulses—the ankle-brachial index test should be performed. If the index value is less than 0.9, then a preoperative vascular surgery consultation should be obtained.

This study had some weaknesses. First, it was retrospective, so it is possible that some wound or VTE complications were not reported and thus not found in the paper charts or electronic medical records. Some patients may have had VTE diagnostic scans at other hospitals, and their results may not have been recorded across databases. Moreover, some patients may have seen wound specialists for wound infections or wound healing problems, and these may not have been reported to the orthopedic surgeons. Second, though our patient population was not small, it may not have been of adequate size to fully detect statistical significance. We met our enrollment numbers based on our sample size calculations from an a priori power analysis; however, we still draw conclusions with the possibility of committing a type II error in mind by failing to reject the null hypothesis when in reality a statistically significant difference does exist. Third, none of our consecutive patients carried the preoperative diagnosis of PVD, and none had preoperative vascular surgery. Therefore, though calcifications were noted on radiographs, clinically our patients were asymptomatic with respect to vascular health. Last, the 2 groups were not randomized. All patients underwent tourniquet-assisted TKA.           

Conclusion

To our knowledge, this is the largest study to examine the effect of preoperative knee arterial calcification on wound complication and VTE after tourniquet-assisted TKA. Contrary to previously published recommendations, we conclude that TKA can be safely performed with a tourniquet in the presence of preoperative radiographic evidence of such calcification. However, we recommend that patients with diabetes, hypertension, CAD, or prior VTE undergo an appropriate physical examination to elicit any signs or symptoms of vascular disease. If before surgery there is any question of vascular competence, a vascular surgeon should be consulted.

Tourniquets are often used in total knee arthroplasty (TKA) to improve visualization of structures, shorten operative time, reduce intraoperative bleeding, and improve cementing technique. Despite these advantages, controversy remains regarding the safety of tourniquet use. Tourniquets have been associated with nerve palsies, vascular injury, and muscle damage.^1-5 Some have hypothesized they may cause venous stasis or direct endothelial damage that may develop into deep vein thrombosis (DVT). Abdel-Salam and Eyres⁶ found an increased incidence of postoperative wound complications and DVTs associated with tourniquet use.

Moreover, investigators have analyzed the role of tourniquets in populations at high risk for wound complications. DeLaurentis and colleagues⁷ performed a prospective and retrospective analysis of 1182 TKA patients, 24 (2%) of whom had preexisting peripheral vascular disease (PVD), defined as a history of arterial insufficiency, absent dorsalis pedis and/or absent posterior tibial pulsations, and arterial calcifications. A tourniquet was used in each case. Arterial complications occurred in 6 of the 24 patients with PVD. As expected, the authors found that a history of intermittent claudication, pain at rest, and arterial ulcers resulted in a high risk for vascular complications. Further studies have supported this finding and expanded the list of predisposing factors to include previous vascular surgery and absent and asymmetric pedal pulsations.^7-11 Of particular concern to total joint arthroplasty surgeons was the finding by DeLaurentis and colleagues⁷ that patients with radiographic evidence of calcification of the distal superficial femoral artery and/or popliteal artery were at risk for arterial complications. This finding is also supported by other studies.^8,11 In TKA, damage to arterial structures proximal to the surgical field could manifest as impaired postoperative wound healing or an ischemic limb. Wound healing depends on adequate blood flow to the healing tissue, and any damage to arterial or venous structures can theoretically compromise this process.

Added to vascular/wound complications as concerning complications in orthopedic surgery is venous thromboembolism (VTE). The role of tourniquets in the formation of VTEs is controversial. A tourniquet has the potential to increase the risk for DVT because of the stasis of venous blood in the lower limb or possible damage to calcified blood vessels. Callam and colleagues¹² studied the connection between artery disease and chronic leg ulcers and found that half the patients diagnosed with peripheral artery disease also had stigmata of chronic venous insufficiency. Therefore, the entities can occur in tandem, and surgeons should keep this in mind.

Here we report on a study we conducted to determine whether tourniquet use in TKA in patients with preexisting radiographic evidence of vascular disease increases the risk for wound complications or VTE.

Patients and Methods

We retrospectively reviewed 461 consecutive primary TKAs (373 patients) performed between January 2007 and June 2012 by 2 attending orthopedic surgeons specializing in adult reconstruction. Medical records and operative reports of 583 patients were examined after receiving institutional review board approval. Of these patients, 373 (64%) had a minimum of 12-month follow-up data available. Twelve months was deemed long enough to discover wound complications or DVTs secondary to the index procedure. Most of these outcomes manifest within the first 3 months after surgery and certainly by 12 months. Follow-up longer than 12 months may become a confounder, as wound complications outside the acute to subacute postoperative window could be related to patients’ underlying PVD and not directly to tourniquet use during surgery. Patient demographics and comorbidities were recorded. Comorbidities were obtained from preoperative medical evaluations and surgeons’ preoperative evaluations. All patients had preoperative palpable dorsalis pedis and posterior tibialis arterial pulses. No patient required preoperative vascular studies based on preoperative examination or comorbidities. No patient had prior vascular bypass surgery or stenting.

TKA was performed in a nonlaminar flow, positive-pressure, high-efficiency particulate air-filtered room with sterile toga/surgical helmet systems. For all patients, a pneumatic thigh tourniquet was applied, and the patient was prepared and draped. After limb exsanguination using a rubber bandage, the limb was elevated and the tourniquet inflated to a pressure of 250 to 300 mm Hg. The tourniquet was released either just before closure or immediately after closure in all cases; it was always let down before placement of final bandages.

Prophylactic chemical anticoagulation consisting of warfarin, aspirin, or enoxaparin was used in all patients and continued for 4 to 6 weeks after surgery. All patients received mechanical DVT prophylaxis with sequential compression devices, and all were mobilized out of bed beginning either the day of surgery or the next day. All patients received perioperative intravenous antibiotics, with the preoperative dose given before tourniquet inflation and the last postoperative dose stopped within 24 hours of surgery.

All patients who had primary TKA underwent preoperative medical evaluation and optimization. The patient’s hospital course was monitored closely, and complications noted by the orthopedic team were documented. Follow-up documentation was retrospectively reviewed for evidence of wound complications or VTE. Wound complications were defined as cellulitis, delayed wound healing, wound dehiscence, and/or periprosthetic joint infection. In the case of VTE, physical examination findings were not sufficient for inclusion. Venous duplex ultrasonography demonstrating the clot was reviewed before inclusion.

Preoperative radiographs were examined for arterial calcification (Figure). We refer to calcification seen above the knee joint as proximal calcification and to calcification observed below the joint as distal calcification. Patients exhibited calcification proximally only, distally only, or both proximally and distally. The 373 patients were placed into 2 groups based on whether they had preoperative arterial calcification on plain radiography of the knee. One group (285 patients with no radiographic evidence of preoperative knee arterial calcification) underwent 365 TKAs, and the other group (88 patients with radiographic evidence of preoperative knee arterial calcification) underwent 96 TKAs.

A sample size calculation was performed to determine how many patients were needed in each group with 80% power and an α of 0.05. With an estimated difference in VTE/wound complication rate between the calcification and no-calcification groups of 12%, we needed to review 316 TKAs total. This 12% difference was based on study findings of a 25% complication rate in PVD patients who underwent tourniquet-assisted TKA, and the rate of VTE/wound complication after TKA in patients overall, which can be up to 12%.^7,13,14 We exceeded minimal enrollment and had 461 TKAs. Descriptive statistics were reported, with means and ranges provided where appropriate. Independent t test was used to evaluate the differences in continuous data (age) between the groups. Univariate analysis (using Pearson χ² and Fisher exact tests) and multivariate logistic regression analysis were used to evaluate the effects of categorical variables (sex, comorbidity, calcification [presence, absence], and location of calcification [proximal only, distal only, both]) on wound complication and VTE rates. All tests were 2-tailed and performed with a type I error rate of 0.05. Data analysis was performed with SPSS Version 19.0 (SPSS).

Results

Patient characteristics are summarized in Table 1. Of the 373 patients, 285 lacked calcification, and 88 had calcification. Mean age was 67.73 years (range, 24-92 years) for all patients, 65.99 years (range, 24-89 years) for the no-calcification group, and 74.32 years (range, 54-92 years) for the calcification group; the calcification group demonstrated a trend toward older age, but the difference was not significantly different (P = .07). Of the 373 patients, 156 (41.82%) were male: 110 in the no-calcification group (38.60%) and 46 in the calcification group (52.27%); sex was significantly (P = .002) different between groups, with more males in the calcification group.

Data on total preoperative comorbidities are summarized in Table 2. Hypertension, hyperlipidemia, diabetes, and coronary artery disease (CAD) were the most common comorbidities, and they were all significantly (P ≤ .05) increased in the calcification group.

No patients had reported arterial complications, such as arterial bleeding, aneurysm, intimal tears, or loss of distal pulses. Wound complication after TKA was detected in 3.04% of all cases (Table 3). Rate of DVT after TKA was 2.60% of all cases, and rate of pulmonary embolism after TKA was 2.17% of all cases. Of the 96 TKAs with preoperative radiographic evidence of calcification, 47 (48.96%) had proximal calcification only, 11 (11.46%) had distal calcification only, and 38 (39.58%) had both proximal and distal calcification (Table 4). There was no significant difference between the rate of wound complication or VTE based on location of vascular calcification.

Univariate analysis demonstrated that presence of arterial knee calcification did not increase the risk for postoperative wound complication (odds ratio [OR], 1.04; 95% confidence interval [CI], 0.28-3.80; P > .05) (Table 5). Location of arterial knee calcification also did not increase the risk for postoperative wound complication. In addition, univariate analysis demonstrated that presence of arterial knee calcification did not increase the risk for postoperative VTE (OR, 1.20; 95% CI, 0.43-3.36; P > .05 (Table 6).

Of the 14 wound complications, the most common infections were cellulitis (5/14 cases; 35.71%) and infected hardware that required component revision (5/14 cases; 35.71%). Mean time from TKA to infection was 137.93 days (range, 5-783 days). The most common organism grown in culture from the wound was Staphylococcus (5/14 cases; 35.71%).

Additional univariate statistical analysis revealed that presence of diabetes, hypertension, prior VTE, CAD, and male sex was linked to higher incidence of wound complication (P < .05) (Table 5). When multivariate analysis was performed, hypertension, prior VTE, and male sex remained significant (P < .05) (Table 5).

Discussion

TKA is a safe and effective procedure used to treat osteoarthritis of the knee and improve patients’ quality of life.¹⁵ About 700,000 TKAs are performed annually in the United States.¹⁶ Because of improvements in preventive medicine and medical technology, life expectancy is increasing, and TKAs are now being performed in higher numbers and in an older patient population. Over the next few decades, these developments will lead to more postoperative complications. It is projected that, by 2030, the need for TKAs in the United States will increase by 673% to 3.48 million.¹⁷ Postoperative complications are rare but unfortunately often lead to poor outcomes or even mortality.¹⁸ To help minimize the number of postoperative complications, we must understand the safety of tourniquet use in TKA. Other investigators have concluded that tourniquet use is unsafe in patients with preoperative vascular calcifications on plain radiographs.^7,8,11 The present study, designed to elucidate whether preoperative evidence of knee arterial calcification may predispose TKA patients to postoperative wound complication or VTE, had some important findings.

In our study, wound complication and VTE occurred in a considerable number of patients after TKA. Despite exceeding the number of patients calculated by the power analysis, our population may have been inadequate to fully detect statistical significance. Thus, our conclusion of failing to reject the null hypothesis may have been because of sample size, a type II error. We found that, after primary TKA, 3.04% of patients developed wound complications and 4.77% VTE. According to the literature, the incidence of infection after primary TKA is between 0.5% and 12%, and that of VTE reported within 3 months after TKA is 1.3% to 10%.^13,14 Although we had 100% VTE prophylaxis, meeting the standard of care, VTE after TKA remains a postoperative complication.¹⁹ This study also found that a considerable percentage of primary TKA patients (23.59%) had preoperative calcification of the knee arteries. To our knowledge, this study was the first to quantify the incidence of knee arterial calcification in patients who underwent TKA.

Preoperative calcification of the knee arteries in patients who underwent TKA did not increase the risk for wound complication, VTE, or arterial damage. These calcifications, however, do pose an increased systemic vascular risk.²⁰ Calcification of the vascular wall predicts increased cardiovascular risk, independent of classical cardiovascular risk factors.^3,18,21-24 Clinically, patients who have both diabetes and calcifications are at significant excess risk for total mortality, stroke mortality, and cardiovascular mortality, compared with patients with diabetes but without such calcifications. They also had a significantly higher incidence of coronary heart disease events, stroke events, and lower extremity amputations.^25,26

All our patients underwent tourniquet-assisted TKA. Although previous studies have indicated that tourniquet use may increase arterial complications and wound complications or even limb loss in patients with calcified arteries, we did not find this link.^7,27 Our population had no reported arterial complications related to tourniquet use. Other, smaller studies have had similar findings. Vandenbussche and colleagues²⁸ prospectively studied 80 TKA cases randomized to tourniquet use or no tourniquet use and found no postoperative nerve palsies, wound infections, wound healing problems, or hematomas. Our study is also in accord with studies that have reported tourniquet use did not increase risk for DVT.²⁹ Therefore, unlike earlier data, our data demonstrated that tourniquet use in patients with knee arterial calcification was safe.^7,27,30,31

Patients with calcification were more likely to have the medical comorbidities of hypertension, diabetes, hyperlipidemia, and CAD. All these comorbidities are linked to the development of arterial calcification, or atherosclerotic occlusive disease.^32,33 As life expectancy and the need for TKA increase, it is likely that a larger percentage of TKA patients will have preoperative radiographic evidence of knee arterial calcification. Although current dogma is that tourniquet-assisted TKA is contraindicated for patients with preoperative radiographic evidence of femoral-popliteal calcification, our study results showed that this calcification should not affect preoperative TKA planning for these patients.

We divided our patients into 3 categories: those with proximal calcification (above the joint line), those with distal calcification (below the joint line), and those with both proximal and distal calcification. Location of arterial calcification did not have an effect on their rates of postoperative wound complication or VTE. We hypothesized that patients with proximal calcification would be at increased risk for direct arterial injury and subsequent wound complication because the tourniquet is placed proximally. Previous research has indicated that arterial occlusion and subsequent wound complication can occur because of low blood flow stemming from tourniquet use.⁷ Further, intraoperative manipulation (flexing) of a knee with calcified vessels causes arterial complications after TKA because these vessels are less elastic than nonatheromatous vessels.³¹ However, we found no such effect. At the same time, having arterial calcification might also be an indication of venous disease in this location,¹² which may be especially important for proximal calcifications. Proximal DVT more likely is a precursor to pulmonary embolic events than distal DVT is.^31,34 However, we found no difference in VTE rates among the 3 arterial location groups, which is supported by studies that have found that tourniquet use does not increase DVT incidence.^29,35-40

Risk for wound complications was higher in male patients and in patients with diabetes, prior VTE, hypertension, or CAD. This finding is important because, with the increasing age of patients who undergo TKA, those with serious medical comorbidities will continue to need and have this surgery.¹⁷ Diabetes may increase the rate of wound complication because patients with diabetes have poor microcirculation, poor collagen synthesis, and reduced wound strength.⁴¹ Malinzak and colleagues⁴² demonstrated that, compared with patients without diabetes, those with diabetes had a significantly higher risk for infection after TKA. Prior VTE, specifically DVT, may increase the rate of wound complication because after DVT the deep veins may be damaged and exhibit valvular dysfunction. Labropoulos and colleagues⁴³ showed that DVT history was strongly associated with ulcer nonhealing. Perhaps hypertension has been overlooked as a risk factor for wound complication in TKA. No previous studies have assessed the link between hypertension and wound complications after TKA. However, a study of wound healing after total hip arthroplasty found that, compared with normotensive patients, hypertensive patients had delayed wound healing, putting them at higher risk for infection.⁴⁴ In addition, we found that patients with CAD were at increased risk for wound complications—an unexpected finding, as CAD traditionally is not a risk factor for infection or poor wound healing. Recently, however, CAD was identified as an independent risk factor for surgical site infections in posterior lumbar–instrumented arthrodesis.⁴⁵ The etiology of this association is unknown. Also, male patients were at increased risk for wound complication. Male sex has been implicated as an independent risk factor for development of surgical site infections and has been established as an important predisposing factor for periprosthetic joint infections.⁴⁶

It is possible that patients who present with diabetes, VTE, hypertension, or CAD before TKA should have a consultation with a vascular surgeon or should have TKA performed without a tourniquet, but this conclusion cannot be considered definitive without a large prospective randomized trial or possibly registry data. Our data indicate that patients with these comorbidities have higher rates of wound complications irrespective of preoperative radiographic calcifications. On the basis of our study results, however, we certainly recommend that patients with these risk factors have preoperative medical optimization. Orthopedic surgeons should take a thorough history and perform a meticulous physical examination on these patients to look for evidence of PVD. We recommend that, if vascular claudication is elicited in the history, or if there is evidence of peripheral arterial disease—such as hair loss, skin discoloration, dystrophic nail changes, or absent or unequal peripheral pulses—the ankle-brachial index test should be performed. If the index value is less than 0.9, then a preoperative vascular surgery consultation should be obtained.

This study had some weaknesses. First, it was retrospective, so it is possible that some wound or VTE complications were not reported and thus not found in the paper charts or electronic medical records. Some patients may have had VTE diagnostic scans at other hospitals, and their results may not have been recorded across databases. Moreover, some patients may have seen wound specialists for wound infections or wound healing problems, and these may not have been reported to the orthopedic surgeons. Second, though our patient population was not small, it may not have been of adequate size to fully detect statistical significance. We met our enrollment numbers based on our sample size calculations from an a priori power analysis; however, we still draw conclusions with the possibility of committing a type II error in mind by failing to reject the null hypothesis when in reality a statistically significant difference does exist. Third, none of our consecutive patients carried the preoperative diagnosis of PVD, and none had preoperative vascular surgery. Therefore, though calcifications were noted on radiographs, clinically our patients were asymptomatic with respect to vascular health. Last, the 2 groups were not randomized. All patients underwent tourniquet-assisted TKA.           

Conclusion

To our knowledge, this is the largest study to examine the effect of preoperative knee arterial calcification on wound complication and VTE after tourniquet-assisted TKA. Contrary to previously published recommendations, we conclude that TKA can be safely performed with a tourniquet in the presence of preoperative radiographic evidence of such calcification. However, we recommend that patients with diabetes, hypertension, CAD, or prior VTE undergo an appropriate physical examination to elicit any signs or symptoms of vascular disease. If before surgery there is any question of vascular competence, a vascular surgeon should be consulted.