There has been increasing emphasis on the development and successful execution of disease management strategies to improve the delivery of evidence‐based care for hospitalized patients with heart failure.14 Current care is woefully suboptimal for heart failure patients. Fonorow et al. describe the significant gap in performance on the Joint Commission for the Accreditation of Healthcare Organizations (JCAHO) heart failure core measures, with the median rate of conformity with all measures at only 24% nationally.5 For a variety of clinical and external factors, such as publicly reported quality measures and pay‐for‐performance incentives, institutions are increasingly motivated to identify patients who will make up the denominator of the heart failure metrics. At first glance, system‐level identification of heart failure patients may not seem critical to the delivery of evidence‐based care, but given that the management of patients with heart failure is multidisciplinary, it is critical that patients who have heart failure be clearly identified for all members of the care team. The capability of prospectively identifying inpatients with the principal diagnosis of heart failure is an essential first step in the implementation of performance improvement programs.

Successful interventions have included a multidisciplinary intervention with postdischarge follow‐up.6 However, the interventions described do not fully indicate how patients with heart failure are identified while in the hospital, so those interventions may be difficult to replicate in other settings. It has not been easy to identify these patients in a timely fashion given that a chief complaint of shortness of breath can indicate other clinical conditions in addition to heart failure. Previous studies have used an admission diagnosis of heart failure or suggestive chest X‐ray findings to trigger a clinical evaluation.7, 8 However, the sensitivity and specificity of these case‐finding methods have not been reported. Furthermore, patients presenting with shortness of breath may not have a diagnosis established until a series of diagnostic and therapeutic maneuvers have been performed. The challenge of promoting physician provision and documentation of evidence‐based care is compounded by these patients usually not being housed in a single geographical unit, possibly being attended by any number of medical specialties, and often having short lengths of stay. Given the multiple factors contributing to the complexity of identifying patients hospitalized with heart failure, it is important to delineate case‐finding strategies that efficiently and effectively identify heart failure patients so that clinical care and self‐management interventions are optimized.

With this goal in mind, we hypothesized that the receipt of intravenous loop diuretics may be an important trigger for identifying patients with heart failure. Receipt of intravenous loop diuretics is ubiquitous in the management of decompensated systolic and diastolic heart failure. We compare 2 electronic pharmacy‐based strategies in a tertiary‐care community teaching hospital to identify hospitalized patients early in their stay who were likely to be discharged with a principal diagnosis of heart failure (HF).

METHODS

RESULTS

Salient features of the study population in the first and second study years are shown in Table 1. Mean age was approximately 59 years, and women made up 60% of the patients. The percentage of patients with a principal diagnosis of heart failure was 3% each year. Serum creatinine levels were available for 78% of patients in year 1 and 80% of patients in year 2. Serum BNP levels were available for 7.6% of patients in year 1 and 9% of patients in year 2.

The parameter estimates and 95% confidence intervals of the main effects of the final prediction model are shown in Table 2, with interaction terms noted in the footnote. Examination of the z scores (available from the authors) indicated that by far the most influential risk factor in the multivariable model was receipt of intravenous diuretics, with receipt of spironolactone a very distant second. The probability that a given patient had heart failure increased with the number of risk factors present and the magnitude of their parameter estimates. For example, an older patient who had been hospitalized with heart failure previously and who was currently receiving intravenous diuretics and spironolactone would be more likely to have a principal diagnosis of heart failure than would an older patient receiving intravenous diuretics who had no other risk factors. However, the interaction terms with negative values (see the footnote in Table 2) indicate that certain combinations of risk factors convey a level of risk somewhat less than the sum of their parts.

The identification strategies performed well from one year to the next, as summarized in Table 3. Receipt of intravenous loop diuretics had a strong association with diagnosis of heart failure (OR 51.6, 95% CI 41.7, 63.7, P < .0001), with a c statistic of 0.88, a sensitivity of 0.89, and a specificity of 0.87.

The linear predictor of the multivariable prediction model as described in the Methods section also performed well in year 2 with excellent discrimination (c statistic of 0.96). Calibration was also excellent, as demonstrated by an intercept of 0.03 (standard error 0.05) and a beta coefficient of 1.02 (SE 0.03). If the threshold for identifying potential heart failure cases was defined as a predicted probability of at least 0.02, then the sensitivity of the multivariable predictor was 0.94 and the specificity was 0.86. If the positivity threshold was raised to 0.04, then the predictor's sensitivity dropped slightly, to 0.91, but specificity increased to 0.89.

The principal diagnoses of the 3045 patients in year 2 who were incorrectly predicted as having a principal diagnosis of heart failure (ie, false positives) were cardiac related (1026 of 3045; 34%), pulmonary related (685 of 3045; 22%), and renal‐ or fluid electrolyte related (117 of 3045; 4%), as determined using the multivariable approach with a 0.04 positivity threshold.

DISCUSSION

Identification of patients with heart failure early in their hospitalization is critical for successfully implementing disease management programs targeted at optimizing evidenced‐based care. Furthermore, public reporting of performance measures has increased the scrutiny of care delivered to those having this principal diagnosis. We developed a strategy that used the receipt of intravenous diuretics as a trigger of further clinical evaluation. We subsequently tested the value of other electronically available indicators to improve the sensitivity and specificity of the case‐finding strategy.

The receipt of an intravenous loop diuretic alone had a sensitivity of .89 and a specificity of .87. Incorporation of the additional information available to us electronically improved the sensitivity to .91 and the specificity to .89 (using a positivity threshold of 0.04), although these might be slightly different if BNP levels had been available for more patients. As with all diagnostic testing, there is a trade‐off between improved sensitivity and decreased specificity. At first glance, the resulting number of false positives generated by either prediction strategy may appear problematic. Although a formal cost‐effectiveness analysis of our case‐finding strategies is beyond the scope of this article, the cost of a false positive in this scenario is likely to be small.

For example in our hospital, clinical pharmacists place a reminder on the charts of patients who appear to have heart failure in order to prompt the clinical team to provide the recommended care processes. A list of inpatients treated with an IV diuretic is generated daily. A clinical pharmacist then reviews identified patient medical records to determine whether the diuretic was ordered for heart failure management or for some other purpose. This review consists of reading the completed history and physical and/or progress notes. On average, each medical record review takes 60 seconds to complete, with a range of 30‐90 seconds. At this speed, roughly 3000 minutes per year (or approximately 1 hour per week) would be spent reviewing the medical records of patients who would not have a principal diagnosis of heart failure. Nevertheless, we found that at least one‐third of the nominal false positives (multivariable rule, threshold of 0.04) still had a cardiac‐related diagnosis. Many of these had heart failure as a secondary diagnosis, but other diagnoses such as acute myocardial infarction took precedence in coding algorithms that assigned the principal diagnosis at discharge. Such patients might still benefit from the interventions and so are not truly false positives.

Patients with heart failure missed by this strategy included patients admitted for placement of pacemakers and/or defibrillators. Patients in this specialized population always had a single team managing their care, so the clinical and educational interventions were integrated into that team's daily work flow. Patients on dialysis with volume overload were not identified using this algorithm and constituted a very small number of patients in our heart failure population. Patients with stable heart failure on oral diuretics were not the focus of this case‐finding strategy and became a target for further intervention only if their heart failure worsened and required intravenous diuretics while they were hospitalized.

The identification of predictors for heart failure has allowed us not only to more effectively identify and risk‐stratify patients with heart failure but also to integrate the case‐finding strategy into clinical care delivery. This approach may also be relevant in hospitals that do not have computerized provider order entry (CPOE) systems but may be able to implement this case‐finding strategy by simply requesting a daily report of patients prescribed intravenous diuretics. As more institutions move to adopt CPOE platforms and clinical information such as ejection fractions become available, the predictors studied here may be augmented to form more sophisticated clinical rules and alerts.

Our study had several limitations. Although we validated our predictors in a separate cohort of patients, this is a single‐site study and may not be representative of the diverse institutions that care for patients with heart failure. There may also be interinstitutional differences in how a principal diagnosis of heart failure is assigned. We have demonstrated the stability of locally derived predictors from one year to the next but cannot make claims about how well our parameter estimates would perform in other settings. Finally, the complexity of the multivariable predictor requires an integrated database and computer application of a formula that may not be commonly available elsewhere at this time.

If disease management strategies are to be successful, early identification of at‐risk patients is crucial for both clinical care delivery and patient education regarding self‐management. The strategies tested here may be useful for other community‐based institutions whose care of heart failure patients is decentralized and involves multiple clinicians.

There has been increasing emphasis on the development and successful execution of disease management strategies to improve the delivery of evidence‐based care for hospitalized patients with heart failure.14 Current care is woefully suboptimal for heart failure patients. Fonorow et al. describe the significant gap in performance on the Joint Commission for the Accreditation of Healthcare Organizations (JCAHO) heart failure core measures, with the median rate of conformity with all measures at only 24% nationally.5 For a variety of clinical and external factors, such as publicly reported quality measures and pay‐for‐performance incentives, institutions are increasingly motivated to identify patients who will make up the denominator of the heart failure metrics. At first glance, system‐level identification of heart failure patients may not seem critical to the delivery of evidence‐based care, but given that the management of patients with heart failure is multidisciplinary, it is critical that patients who have heart failure be clearly identified for all members of the care team. The capability of prospectively identifying inpatients with the principal diagnosis of heart failure is an essential first step in the implementation of performance improvement programs.

Successful interventions have included a multidisciplinary intervention with postdischarge follow‐up.6 However, the interventions described do not fully indicate how patients with heart failure are identified while in the hospital, so those interventions may be difficult to replicate in other settings. It has not been easy to identify these patients in a timely fashion given that a chief complaint of shortness of breath can indicate other clinical conditions in addition to heart failure. Previous studies have used an admission diagnosis of heart failure or suggestive chest X‐ray findings to trigger a clinical evaluation.7, 8 However, the sensitivity and specificity of these case‐finding methods have not been reported. Furthermore, patients presenting with shortness of breath may not have a diagnosis established until a series of diagnostic and therapeutic maneuvers have been performed. The challenge of promoting physician provision and documentation of evidence‐based care is compounded by these patients usually not being housed in a single geographical unit, possibly being attended by any number of medical specialties, and often having short lengths of stay. Given the multiple factors contributing to the complexity of identifying patients hospitalized with heart failure, it is important to delineate case‐finding strategies that efficiently and effectively identify heart failure patients so that clinical care and self‐management interventions are optimized.

With this goal in mind, we hypothesized that the receipt of intravenous loop diuretics may be an important trigger for identifying patients with heart failure. Receipt of intravenous loop diuretics is ubiquitous in the management of decompensated systolic and diastolic heart failure. We compare 2 electronic pharmacy‐based strategies in a tertiary‐care community teaching hospital to identify hospitalized patients early in their stay who were likely to be discharged with a principal diagnosis of heart failure (HF).

METHODS

RESULTS

Salient features of the study population in the first and second study years are shown in Table 1. Mean age was approximately 59 years, and women made up 60% of the patients. The percentage of patients with a principal diagnosis of heart failure was 3% each year. Serum creatinine levels were available for 78% of patients in year 1 and 80% of patients in year 2. Serum BNP levels were available for 7.6% of patients in year 1 and 9% of patients in year 2.

The parameter estimates and 95% confidence intervals of the main effects of the final prediction model are shown in Table 2, with interaction terms noted in the footnote. Examination of the z scores (available from the authors) indicated that by far the most influential risk factor in the multivariable model was receipt of intravenous diuretics, with receipt of spironolactone a very distant second. The probability that a given patient had heart failure increased with the number of risk factors present and the magnitude of their parameter estimates. For example, an older patient who had been hospitalized with heart failure previously and who was currently receiving intravenous diuretics and spironolactone would be more likely to have a principal diagnosis of heart failure than would an older patient receiving intravenous diuretics who had no other risk factors. However, the interaction terms with negative values (see the footnote in Table 2) indicate that certain combinations of risk factors convey a level of risk somewhat less than the sum of their parts.

The identification strategies performed well from one year to the next, as summarized in Table 3. Receipt of intravenous loop diuretics had a strong association with diagnosis of heart failure (OR 51.6, 95% CI 41.7, 63.7, P < .0001), with a c statistic of 0.88, a sensitivity of 0.89, and a specificity of 0.87.

The linear predictor of the multivariable prediction model as described in the Methods section also performed well in year 2 with excellent discrimination (c statistic of 0.96). Calibration was also excellent, as demonstrated by an intercept of 0.03 (standard error 0.05) and a beta coefficient of 1.02 (SE 0.03). If the threshold for identifying potential heart failure cases was defined as a predicted probability of at least 0.02, then the sensitivity of the multivariable predictor was 0.94 and the specificity was 0.86. If the positivity threshold was raised to 0.04, then the predictor's sensitivity dropped slightly, to 0.91, but specificity increased to 0.89.

The principal diagnoses of the 3045 patients in year 2 who were incorrectly predicted as having a principal diagnosis of heart failure (ie, false positives) were cardiac related (1026 of 3045; 34%), pulmonary related (685 of 3045; 22%), and renal‐ or fluid electrolyte related (117 of 3045; 4%), as determined using the multivariable approach with a 0.04 positivity threshold.

DISCUSSION

Identification of patients with heart failure early in their hospitalization is critical for successfully implementing disease management programs targeted at optimizing evidenced‐based care. Furthermore, public reporting of performance measures has increased the scrutiny of care delivered to those having this principal diagnosis. We developed a strategy that used the receipt of intravenous diuretics as a trigger of further clinical evaluation. We subsequently tested the value of other electronically available indicators to improve the sensitivity and specificity of the case‐finding strategy.

The receipt of an intravenous loop diuretic alone had a sensitivity of .89 and a specificity of .87. Incorporation of the additional information available to us electronically improved the sensitivity to .91 and the specificity to .89 (using a positivity threshold of 0.04), although these might be slightly different if BNP levels had been available for more patients. As with all diagnostic testing, there is a trade‐off between improved sensitivity and decreased specificity. At first glance, the resulting number of false positives generated by either prediction strategy may appear problematic. Although a formal cost‐effectiveness analysis of our case‐finding strategies is beyond the scope of this article, the cost of a false positive in this scenario is likely to be small.

For example in our hospital, clinical pharmacists place a reminder on the charts of patients who appear to have heart failure in order to prompt the clinical team to provide the recommended care processes. A list of inpatients treated with an IV diuretic is generated daily. A clinical pharmacist then reviews identified patient medical records to determine whether the diuretic was ordered for heart failure management or for some other purpose. This review consists of reading the completed history and physical and/or progress notes. On average, each medical record review takes 60 seconds to complete, with a range of 30‐90 seconds. At this speed, roughly 3000 minutes per year (or approximately 1 hour per week) would be spent reviewing the medical records of patients who would not have a principal diagnosis of heart failure. Nevertheless, we found that at least one‐third of the nominal false positives (multivariable rule, threshold of 0.04) still had a cardiac‐related diagnosis. Many of these had heart failure as a secondary diagnosis, but other diagnoses such as acute myocardial infarction took precedence in coding algorithms that assigned the principal diagnosis at discharge. Such patients might still benefit from the interventions and so are not truly false positives.

Patients with heart failure missed by this strategy included patients admitted for placement of pacemakers and/or defibrillators. Patients in this specialized population always had a single team managing their care, so the clinical and educational interventions were integrated into that team's daily work flow. Patients on dialysis with volume overload were not identified using this algorithm and constituted a very small number of patients in our heart failure population. Patients with stable heart failure on oral diuretics were not the focus of this case‐finding strategy and became a target for further intervention only if their heart failure worsened and required intravenous diuretics while they were hospitalized.

The identification of predictors for heart failure has allowed us not only to more effectively identify and risk‐stratify patients with heart failure but also to integrate the case‐finding strategy into clinical care delivery. This approach may also be relevant in hospitals that do not have computerized provider order entry (CPOE) systems but may be able to implement this case‐finding strategy by simply requesting a daily report of patients prescribed intravenous diuretics. As more institutions move to adopt CPOE platforms and clinical information such as ejection fractions become available, the predictors studied here may be augmented to form more sophisticated clinical rules and alerts.

Our study had several limitations. Although we validated our predictors in a separate cohort of patients, this is a single‐site study and may not be representative of the diverse institutions that care for patients with heart failure. There may also be interinstitutional differences in how a principal diagnosis of heart failure is assigned. We have demonstrated the stability of locally derived predictors from one year to the next but cannot make claims about how well our parameter estimates would perform in other settings. Finally, the complexity of the multivariable predictor requires an integrated database and computer application of a formula that may not be commonly available elsewhere at this time.

If disease management strategies are to be successful, early identification of at‐risk patients is crucial for both clinical care delivery and patient education regarding self‐management. The strategies tested here may be useful for other community‐based institutions whose care of heart failure patients is decentralized and involves multiple clinicians.

There has been increasing emphasis on the development and successful execution of disease management strategies to improve the delivery of evidence‐based care for hospitalized patients with heart failure.14 Current care is woefully suboptimal for heart failure patients. Fonorow et al. describe the significant gap in performance on the Joint Commission for the Accreditation of Healthcare Organizations (JCAHO) heart failure core measures, with the median rate of conformity with all measures at only 24% nationally.5 For a variety of clinical and external factors, such as publicly reported quality measures and pay‐for‐performance incentives, institutions are increasingly motivated to identify patients who will make up the denominator of the heart failure metrics. At first glance, system‐level identification of heart failure patients may not seem critical to the delivery of evidence‐based care, but given that the management of patients with heart failure is multidisciplinary, it is critical that patients who have heart failure be clearly identified for all members of the care team. The capability of prospectively identifying inpatients with the principal diagnosis of heart failure is an essential first step in the implementation of performance improvement programs.

Successful interventions have included a multidisciplinary intervention with postdischarge follow‐up.6 However, the interventions described do not fully indicate how patients with heart failure are identified while in the hospital, so those interventions may be difficult to replicate in other settings. It has not been easy to identify these patients in a timely fashion given that a chief complaint of shortness of breath can indicate other clinical conditions in addition to heart failure. Previous studies have used an admission diagnosis of heart failure or suggestive chest X‐ray findings to trigger a clinical evaluation.7, 8 However, the sensitivity and specificity of these case‐finding methods have not been reported. Furthermore, patients presenting with shortness of breath may not have a diagnosis established until a series of diagnostic and therapeutic maneuvers have been performed. The challenge of promoting physician provision and documentation of evidence‐based care is compounded by these patients usually not being housed in a single geographical unit, possibly being attended by any number of medical specialties, and often having short lengths of stay. Given the multiple factors contributing to the complexity of identifying patients hospitalized with heart failure, it is important to delineate case‐finding strategies that efficiently and effectively identify heart failure patients so that clinical care and self‐management interventions are optimized.

With this goal in mind, we hypothesized that the receipt of intravenous loop diuretics may be an important trigger for identifying patients with heart failure. Receipt of intravenous loop diuretics is ubiquitous in the management of decompensated systolic and diastolic heart failure. We compare 2 electronic pharmacy‐based strategies in a tertiary‐care community teaching hospital to identify hospitalized patients early in their stay who were likely to be discharged with a principal diagnosis of heart failure (HF).

METHODS

RESULTS

Salient features of the study population in the first and second study years are shown in Table 1. Mean age was approximately 59 years, and women made up 60% of the patients. The percentage of patients with a principal diagnosis of heart failure was 3% each year. Serum creatinine levels were available for 78% of patients in year 1 and 80% of patients in year 2. Serum BNP levels were available for 7.6% of patients in year 1 and 9% of patients in year 2.

The parameter estimates and 95% confidence intervals of the main effects of the final prediction model are shown in Table 2, with interaction terms noted in the footnote. Examination of the z scores (available from the authors) indicated that by far the most influential risk factor in the multivariable model was receipt of intravenous diuretics, with receipt of spironolactone a very distant second. The probability that a given patient had heart failure increased with the number of risk factors present and the magnitude of their parameter estimates. For example, an older patient who had been hospitalized with heart failure previously and who was currently receiving intravenous diuretics and spironolactone would be more likely to have a principal diagnosis of heart failure than would an older patient receiving intravenous diuretics who had no other risk factors. However, the interaction terms with negative values (see the footnote in Table 2) indicate that certain combinations of risk factors convey a level of risk somewhat less than the sum of their parts.

The identification strategies performed well from one year to the next, as summarized in Table 3. Receipt of intravenous loop diuretics had a strong association with diagnosis of heart failure (OR 51.6, 95% CI 41.7, 63.7, P < .0001), with a c statistic of 0.88, a sensitivity of 0.89, and a specificity of 0.87.

The linear predictor of the multivariable prediction model as described in the Methods section also performed well in year 2 with excellent discrimination (c statistic of 0.96). Calibration was also excellent, as demonstrated by an intercept of 0.03 (standard error 0.05) and a beta coefficient of 1.02 (SE 0.03). If the threshold for identifying potential heart failure cases was defined as a predicted probability of at least 0.02, then the sensitivity of the multivariable predictor was 0.94 and the specificity was 0.86. If the positivity threshold was raised to 0.04, then the predictor's sensitivity dropped slightly, to 0.91, but specificity increased to 0.89.

The principal diagnoses of the 3045 patients in year 2 who were incorrectly predicted as having a principal diagnosis of heart failure (ie, false positives) were cardiac related (1026 of 3045; 34%), pulmonary related (685 of 3045; 22%), and renal‐ or fluid electrolyte related (117 of 3045; 4%), as determined using the multivariable approach with a 0.04 positivity threshold.

DISCUSSION

Identification of patients with heart failure early in their hospitalization is critical for successfully implementing disease management programs targeted at optimizing evidenced‐based care. Furthermore, public reporting of performance measures has increased the scrutiny of care delivered to those having this principal diagnosis. We developed a strategy that used the receipt of intravenous diuretics as a trigger of further clinical evaluation. We subsequently tested the value of other electronically available indicators to improve the sensitivity and specificity of the case‐finding strategy.

The receipt of an intravenous loop diuretic alone had a sensitivity of .89 and a specificity of .87. Incorporation of the additional information available to us electronically improved the sensitivity to .91 and the specificity to .89 (using a positivity threshold of 0.04), although these might be slightly different if BNP levels had been available for more patients. As with all diagnostic testing, there is a trade‐off between improved sensitivity and decreased specificity. At first glance, the resulting number of false positives generated by either prediction strategy may appear problematic. Although a formal cost‐effectiveness analysis of our case‐finding strategies is beyond the scope of this article, the cost of a false positive in this scenario is likely to be small.

For example in our hospital, clinical pharmacists place a reminder on the charts of patients who appear to have heart failure in order to prompt the clinical team to provide the recommended care processes. A list of inpatients treated with an IV diuretic is generated daily. A clinical pharmacist then reviews identified patient medical records to determine whether the diuretic was ordered for heart failure management or for some other purpose. This review consists of reading the completed history and physical and/or progress notes. On average, each medical record review takes 60 seconds to complete, with a range of 30‐90 seconds. At this speed, roughly 3000 minutes per year (or approximately 1 hour per week) would be spent reviewing the medical records of patients who would not have a principal diagnosis of heart failure. Nevertheless, we found that at least one‐third of the nominal false positives (multivariable rule, threshold of 0.04) still had a cardiac‐related diagnosis. Many of these had heart failure as a secondary diagnosis, but other diagnoses such as acute myocardial infarction took precedence in coding algorithms that assigned the principal diagnosis at discharge. Such patients might still benefit from the interventions and so are not truly false positives.

Patients with heart failure missed by this strategy included patients admitted for placement of pacemakers and/or defibrillators. Patients in this specialized population always had a single team managing their care, so the clinical and educational interventions were integrated into that team's daily work flow. Patients on dialysis with volume overload were not identified using this algorithm and constituted a very small number of patients in our heart failure population. Patients with stable heart failure on oral diuretics were not the focus of this case‐finding strategy and became a target for further intervention only if their heart failure worsened and required intravenous diuretics while they were hospitalized.

The identification of predictors for heart failure has allowed us not only to more effectively identify and risk‐stratify patients with heart failure but also to integrate the case‐finding strategy into clinical care delivery. This approach may also be relevant in hospitals that do not have computerized provider order entry (CPOE) systems but may be able to implement this case‐finding strategy by simply requesting a daily report of patients prescribed intravenous diuretics. As more institutions move to adopt CPOE platforms and clinical information such as ejection fractions become available, the predictors studied here may be augmented to form more sophisticated clinical rules and alerts.

Our study had several limitations. Although we validated our predictors in a separate cohort of patients, this is a single‐site study and may not be representative of the diverse institutions that care for patients with heart failure. There may also be interinstitutional differences in how a principal diagnosis of heart failure is assigned. We have demonstrated the stability of locally derived predictors from one year to the next but cannot make claims about how well our parameter estimates would perform in other settings. Finally, the complexity of the multivariable predictor requires an integrated database and computer application of a formula that may not be commonly available elsewhere at this time.

If disease management strategies are to be successful, early identification of at‐risk patients is crucial for both clinical care delivery and patient education regarding self‐management. The strategies tested here may be useful for other community‐based institutions whose care of heart failure patients is decentralized and involves multiple clinicians.

A 37‐year‐old man presented to an ophthalmologist in July 2004 with a history of slowly decreasing vision in both eyes for several weeks. His vision on presentation was 20/400 in the right eye and 20/200 in the left eye. Slit‐lamp examination showed a bilateral anterior uveitis with 360 degrees of posterior synechiae (adhesions) and a dense vitritis (posterior uveitis) that obscured the view of the retina in both eyes. He was diagnosed with panuveitis and started on topical steroid and cycloplegic drops. He was referred to a uveitis specialist for investigation but missed his appointments.

One year later he presented to the emergency room with fever and severe pain in his left eye. On initial assessment he had no complaints of mouth or genital ulcers, recent or remote rashes, joint symptoms, or penile discharge. He denied any prior eye trauma or surgery. He reported that his last sexual encounter had been 8 months prior with a male and that his most recent HIV screen was negative 6 months ago. His family history was negative for autoimmune disorders.

On inspection, he appeared cachectic, lethargic, and very ill. He was febrile and tachycardic; the remainder of his vital signs were normal. There was no lymphadenopathy. His neck was supple with no meningismal signs. There were no heart murmurs, oral ulcers, swollen joints, mucosal eschar, or skin lesions. Respiratory and abdominal examinations were unremarkable.

His visual acuity was light perception in right eye and no light perception in the left eye. There was significant eyelid edema, erythema and purulent discharge with mild proptosis of the left eye (Fig. 1). Pupils were 3 mm and fixed, with 360 of posterior synechiae. Intraocular pressure was elevated in the left eye (42 mm Hg, where normal is 20 mm Hg). There was moderate uveitis in both eyes, with a 1‐mm hypopyon in the left eye and forward bowing of the iris (iris bomb). A dense vitritis was present in both eyes, preventing visualization of the retina. B‐scan ultrasound examination showed bilateral retinal detachments, worse in the left eye.

Figure 1

Because of the high intraocular pressure in the left eye, the patient was given topical Cosopt (dorzolamide hydrochloride‐timolol maleate), bromonidine 0.15%, and oral acetazolamide to lower intraocular pressures. He was started on a preliminary treatment of hourly topical prednisolone acetate 1%, atropine 1% 4 times daily, and topical moxifloxacin 0.5%. He was admitted to hospital to investigate the source of his panophthalmitis (suppurative infection of the eye and sclera, extending to involve the orbit).

Blood and urine cultures, HIV, rapid plasma reagin test (RPR), HLA B27, toxoplasmosis serology, and ANA rheumatoid factor were sent. Overnight, he developed classic Janeway lesions on his palms and soles, and both blood and urine cultures grew gram‐positive cocci in clusters. Repeat blood cultures were taken. He was started on IV vancomycin empirically. Ultimately, all 3 blood cultures grew Staphylococcus aureus.

A transesophageal echocardiogram diagnosed endocarditis with a pedunculated mobile mass identified on the posterior mitral valve leaflet. Mild mitral regurgitation was noted. The aortic valve was normal, as were ventricular size and function. Antibiotics were modified to cloxacillin and gentamicin IV 2 days later, once sensitivities were reported.

A CT scan of the orbits revealed diffuse orbital inflammation with no evidence of an orbital abscess (Fig. 2). The inflammation and proptosis of the left eye continued to worsen, and a vitreous paracentesis of the left eye was performed for 1.5 mL of dark brown fluid. The aspirated sample was sent for C&S, PCR (for HSV, CMV), acid‐fast stain, and fungal, viral, and mycobacterial cultures. Intravitreal injections of vancomycin and ceftazidime were given. Bacterial cultures showed a heavy intraocular growth of S. aureus, giving the diagnosis of endophthalmitis (bacterial or fungal infection of the vitreous or aqueous humor); all remaining stains and cultures were negative.

Figure 2

Over the next several days, the initial blood work returned with the following abnormal results: CD4 count was 70/L, and HIV serology was positive. The rapid plasma reagin test (RPR) was positive (titer 1:64). The enzyme immunoassay (EIA) and Treponema pallidum particle agglutination (TPPA) were also positive.

A lumbar puncture was performed, and CSF analysis indicated CSF fluid was clear, 2 erythrocytes and 2 leukocytes in the fourth tube, CSF glucose of 2.7 mmol/L (serum glucose 8.2 mmol/L), and CSF total protein of 1100 mg/L. There were no bacteria seen on the gram stain, and a rapid agglutination test for cryptococcal antigen was negative. The CSF RPR titer was 1:2, and the Treponema pallidum particle agglutination assay (TP‐PA) was reactive. The MRI of the brain indicated diffuse white matter disease but no meningeal enhancement. In combination, these results were indicative of neurosyphilis, and penicillin G IV therapy was initiated. He received a total of 14 days of IV therapy, followed by 3 weekly IM doses of benzathine penicillin. He also received a total of 28 days of IV cloxacillin therapy with 5 days of concomitant IV gentamicin for endocarditis treatment.

Over 8 weeks, the patient's panophthalmitis slowly improved. However, he maintained only light perception in the right eye and did not regain any vision in the left eye. He was discharged home to follow‐up with the infectious diseases and ophthalmology departments. The issue of initiating antiretroviral therapy, deferred during hospital admission because of his poor compliance history and the threat of immune reconstitution symptoms, was to be readdressed at this time. He missed both appointments and returned to the emergency room several months later with widespread Kaposi's sarcoma.

DISCUSSION

One of the key learning points from this case underlines that panuveitis carries a broad differential including inflammatory and infectious conditions, as well as lymphoma. Systemic infections include tuberculosis, syphilis, and in cases of severe immunosuppression, toxoplasmosis. Cytomegalovirus and candidiasis are less likely as they are not associated with intraocular inflammation. HIV is also on the differential, although it rarely causes severe panuveitis on its own. Inflammatory disorders such as Behcet syndrome, sarcoidosis, and, rarely, lens‐associated uveitis (if presented with a history of lens trauma or surgery) are also included on the differential. A systematic approach to the history and physical examination must be undertaken to narrow the search. A syphilis screen should always be included in the differential when investigating uveitis,1 especially given the resurgence of syphilis since 2000.2

Our patient presents an interesting study as he was coinfected with both syphilis and HIV. The progression of syphilis is far more aggressive in this scenario,3 as there is a higher frequency of initial presentation as secondary syphilis4 and with multiple persisting chancres.5 Secondary‐stage skin lesions are also more aggressive in coinfected patients (nodular or ulcerative lesions with necrotic centers), although the same dermatological presentations can be seen in HIV‐negative patients.6 It has not been definitively established whether HIV‐positive patients develop neurological complications of syphilis more frequently or earlier in disease, but most patients present with early neurosyphilis at the time of diagnosis.7 In keeping with these findings, our patient's initial presentation included both ocular and neurosyphilis as diagnostic features.

An atypical link highlighted by our case is that of endogenous, bacterial endophthalmitis secondary to endocarditis. Although traumatic or surgical complications are the most common causes of endophthalmitis, seeding from an endogenous infective source, although rare, is possible.810 Staphylococcus aureus endocarditis is one of the most common causes of endogenous spread.9 In our patient, his chronic uveitis and decompensated blood‐ocular barrier may have contributed to S. aureus seeding of his eye. As is the case with many patients diagnosed with S. aureus endocarditis, the source of infection was unknown, although several risk factors for S. aureus bacteremia have been documented. These risk factors include hospitalization, dialysis, transplantation, HIV‐positive status, heart disease, cancer, diabetes, and intravenous drug use. In a population‐based surveillance study from 1999 to 2000, 550 invasive isolates of S. aureus were obtained; the relative risk in HIV‐positive patients was 23.7.11 In a similar study, the source of the S. aureus bacteremia/endocarditis was not identified in 26% of patients with underlying medical conditions such as HIV infection.12

This case has demonstrated several intertwined disease presentations in a patient coinfected with multiple organisms. In an immunocompromised patient, Occam's razor does not necessarily hold true, and the possibility of multiple diagnoses must be entertained. Thus, clinicians must maintain a high index of suspicion for atypical presentations of typical diseases if their patients are to survive in the eye of the storm.

A 37‐year‐old man presented to an ophthalmologist in July 2004 with a history of slowly decreasing vision in both eyes for several weeks. His vision on presentation was 20/400 in the right eye and 20/200 in the left eye. Slit‐lamp examination showed a bilateral anterior uveitis with 360 degrees of posterior synechiae (adhesions) and a dense vitritis (posterior uveitis) that obscured the view of the retina in both eyes. He was diagnosed with panuveitis and started on topical steroid and cycloplegic drops. He was referred to a uveitis specialist for investigation but missed his appointments.

One year later he presented to the emergency room with fever and severe pain in his left eye. On initial assessment he had no complaints of mouth or genital ulcers, recent or remote rashes, joint symptoms, or penile discharge. He denied any prior eye trauma or surgery. He reported that his last sexual encounter had been 8 months prior with a male and that his most recent HIV screen was negative 6 months ago. His family history was negative for autoimmune disorders.

On inspection, he appeared cachectic, lethargic, and very ill. He was febrile and tachycardic; the remainder of his vital signs were normal. There was no lymphadenopathy. His neck was supple with no meningismal signs. There were no heart murmurs, oral ulcers, swollen joints, mucosal eschar, or skin lesions. Respiratory and abdominal examinations were unremarkable.

His visual acuity was light perception in right eye and no light perception in the left eye. There was significant eyelid edema, erythema and purulent discharge with mild proptosis of the left eye (Fig. 1). Pupils were 3 mm and fixed, with 360 of posterior synechiae. Intraocular pressure was elevated in the left eye (42 mm Hg, where normal is 20 mm Hg). There was moderate uveitis in both eyes, with a 1‐mm hypopyon in the left eye and forward bowing of the iris (iris bomb). A dense vitritis was present in both eyes, preventing visualization of the retina. B‐scan ultrasound examination showed bilateral retinal detachments, worse in the left eye.

Figure 1

Because of the high intraocular pressure in the left eye, the patient was given topical Cosopt (dorzolamide hydrochloride‐timolol maleate), bromonidine 0.15%, and oral acetazolamide to lower intraocular pressures. He was started on a preliminary treatment of hourly topical prednisolone acetate 1%, atropine 1% 4 times daily, and topical moxifloxacin 0.5%. He was admitted to hospital to investigate the source of his panophthalmitis (suppurative infection of the eye and sclera, extending to involve the orbit).

Blood and urine cultures, HIV, rapid plasma reagin test (RPR), HLA B27, toxoplasmosis serology, and ANA rheumatoid factor were sent. Overnight, he developed classic Janeway lesions on his palms and soles, and both blood and urine cultures grew gram‐positive cocci in clusters. Repeat blood cultures were taken. He was started on IV vancomycin empirically. Ultimately, all 3 blood cultures grew Staphylococcus aureus.

A transesophageal echocardiogram diagnosed endocarditis with a pedunculated mobile mass identified on the posterior mitral valve leaflet. Mild mitral regurgitation was noted. The aortic valve was normal, as were ventricular size and function. Antibiotics were modified to cloxacillin and gentamicin IV 2 days later, once sensitivities were reported.

A CT scan of the orbits revealed diffuse orbital inflammation with no evidence of an orbital abscess (Fig. 2). The inflammation and proptosis of the left eye continued to worsen, and a vitreous paracentesis of the left eye was performed for 1.5 mL of dark brown fluid. The aspirated sample was sent for C&S, PCR (for HSV, CMV), acid‐fast stain, and fungal, viral, and mycobacterial cultures. Intravitreal injections of vancomycin and ceftazidime were given. Bacterial cultures showed a heavy intraocular growth of S. aureus, giving the diagnosis of endophthalmitis (bacterial or fungal infection of the vitreous or aqueous humor); all remaining stains and cultures were negative.

Figure 2

Over the next several days, the initial blood work returned with the following abnormal results: CD4 count was 70/L, and HIV serology was positive. The rapid plasma reagin test (RPR) was positive (titer 1:64). The enzyme immunoassay (EIA) and Treponema pallidum particle agglutination (TPPA) were also positive.

A lumbar puncture was performed, and CSF analysis indicated CSF fluid was clear, 2 erythrocytes and 2 leukocytes in the fourth tube, CSF glucose of 2.7 mmol/L (serum glucose 8.2 mmol/L), and CSF total protein of 1100 mg/L. There were no bacteria seen on the gram stain, and a rapid agglutination test for cryptococcal antigen was negative. The CSF RPR titer was 1:2, and the Treponema pallidum particle agglutination assay (TP‐PA) was reactive. The MRI of the brain indicated diffuse white matter disease but no meningeal enhancement. In combination, these results were indicative of neurosyphilis, and penicillin G IV therapy was initiated. He received a total of 14 days of IV therapy, followed by 3 weekly IM doses of benzathine penicillin. He also received a total of 28 days of IV cloxacillin therapy with 5 days of concomitant IV gentamicin for endocarditis treatment.

Over 8 weeks, the patient's panophthalmitis slowly improved. However, he maintained only light perception in the right eye and did not regain any vision in the left eye. He was discharged home to follow‐up with the infectious diseases and ophthalmology departments. The issue of initiating antiretroviral therapy, deferred during hospital admission because of his poor compliance history and the threat of immune reconstitution symptoms, was to be readdressed at this time. He missed both appointments and returned to the emergency room several months later with widespread Kaposi's sarcoma.

DISCUSSION

One of the key learning points from this case underlines that panuveitis carries a broad differential including inflammatory and infectious conditions, as well as lymphoma. Systemic infections include tuberculosis, syphilis, and in cases of severe immunosuppression, toxoplasmosis. Cytomegalovirus and candidiasis are less likely as they are not associated with intraocular inflammation. HIV is also on the differential, although it rarely causes severe panuveitis on its own. Inflammatory disorders such as Behcet syndrome, sarcoidosis, and, rarely, lens‐associated uveitis (if presented with a history of lens trauma or surgery) are also included on the differential. A systematic approach to the history and physical examination must be undertaken to narrow the search. A syphilis screen should always be included in the differential when investigating uveitis,1 especially given the resurgence of syphilis since 2000.2

Our patient presents an interesting study as he was coinfected with both syphilis and HIV. The progression of syphilis is far more aggressive in this scenario,3 as there is a higher frequency of initial presentation as secondary syphilis4 and with multiple persisting chancres.5 Secondary‐stage skin lesions are also more aggressive in coinfected patients (nodular or ulcerative lesions with necrotic centers), although the same dermatological presentations can be seen in HIV‐negative patients.6 It has not been definitively established whether HIV‐positive patients develop neurological complications of syphilis more frequently or earlier in disease, but most patients present with early neurosyphilis at the time of diagnosis.7 In keeping with these findings, our patient's initial presentation included both ocular and neurosyphilis as diagnostic features.

An atypical link highlighted by our case is that of endogenous, bacterial endophthalmitis secondary to endocarditis. Although traumatic or surgical complications are the most common causes of endophthalmitis, seeding from an endogenous infective source, although rare, is possible.810 Staphylococcus aureus endocarditis is one of the most common causes of endogenous spread.9 In our patient, his chronic uveitis and decompensated blood‐ocular barrier may have contributed to S. aureus seeding of his eye. As is the case with many patients diagnosed with S. aureus endocarditis, the source of infection was unknown, although several risk factors for S. aureus bacteremia have been documented. These risk factors include hospitalization, dialysis, transplantation, HIV‐positive status, heart disease, cancer, diabetes, and intravenous drug use. In a population‐based surveillance study from 1999 to 2000, 550 invasive isolates of S. aureus were obtained; the relative risk in HIV‐positive patients was 23.7.11 In a similar study, the source of the S. aureus bacteremia/endocarditis was not identified in 26% of patients with underlying medical conditions such as HIV infection.12

This case has demonstrated several intertwined disease presentations in a patient coinfected with multiple organisms. In an immunocompromised patient, Occam's razor does not necessarily hold true, and the possibility of multiple diagnoses must be entertained. Thus, clinicians must maintain a high index of suspicion for atypical presentations of typical diseases if their patients are to survive in the eye of the storm.

A 37‐year‐old man presented to an ophthalmologist in July 2004 with a history of slowly decreasing vision in both eyes for several weeks. His vision on presentation was 20/400 in the right eye and 20/200 in the left eye. Slit‐lamp examination showed a bilateral anterior uveitis with 360 degrees of posterior synechiae (adhesions) and a dense vitritis (posterior uveitis) that obscured the view of the retina in both eyes. He was diagnosed with panuveitis and started on topical steroid and cycloplegic drops. He was referred to a uveitis specialist for investigation but missed his appointments.

One year later he presented to the emergency room with fever and severe pain in his left eye. On initial assessment he had no complaints of mouth or genital ulcers, recent or remote rashes, joint symptoms, or penile discharge. He denied any prior eye trauma or surgery. He reported that his last sexual encounter had been 8 months prior with a male and that his most recent HIV screen was negative 6 months ago. His family history was negative for autoimmune disorders.

On inspection, he appeared cachectic, lethargic, and very ill. He was febrile and tachycardic; the remainder of his vital signs were normal. There was no lymphadenopathy. His neck was supple with no meningismal signs. There were no heart murmurs, oral ulcers, swollen joints, mucosal eschar, or skin lesions. Respiratory and abdominal examinations were unremarkable.

His visual acuity was light perception in right eye and no light perception in the left eye. There was significant eyelid edema, erythema and purulent discharge with mild proptosis of the left eye (Fig. 1). Pupils were 3 mm and fixed, with 360 of posterior synechiae. Intraocular pressure was elevated in the left eye (42 mm Hg, where normal is 20 mm Hg). There was moderate uveitis in both eyes, with a 1‐mm hypopyon in the left eye and forward bowing of the iris (iris bomb). A dense vitritis was present in both eyes, preventing visualization of the retina. B‐scan ultrasound examination showed bilateral retinal detachments, worse in the left eye.

Figure 1

Because of the high intraocular pressure in the left eye, the patient was given topical Cosopt (dorzolamide hydrochloride‐timolol maleate), bromonidine 0.15%, and oral acetazolamide to lower intraocular pressures. He was started on a preliminary treatment of hourly topical prednisolone acetate 1%, atropine 1% 4 times daily, and topical moxifloxacin 0.5%. He was admitted to hospital to investigate the source of his panophthalmitis (suppurative infection of the eye and sclera, extending to involve the orbit).

Blood and urine cultures, HIV, rapid plasma reagin test (RPR), HLA B27, toxoplasmosis serology, and ANA rheumatoid factor were sent. Overnight, he developed classic Janeway lesions on his palms and soles, and both blood and urine cultures grew gram‐positive cocci in clusters. Repeat blood cultures were taken. He was started on IV vancomycin empirically. Ultimately, all 3 blood cultures grew Staphylococcus aureus.

A transesophageal echocardiogram diagnosed endocarditis with a pedunculated mobile mass identified on the posterior mitral valve leaflet. Mild mitral regurgitation was noted. The aortic valve was normal, as were ventricular size and function. Antibiotics were modified to cloxacillin and gentamicin IV 2 days later, once sensitivities were reported.

A CT scan of the orbits revealed diffuse orbital inflammation with no evidence of an orbital abscess (Fig. 2). The inflammation and proptosis of the left eye continued to worsen, and a vitreous paracentesis of the left eye was performed for 1.5 mL of dark brown fluid. The aspirated sample was sent for C&S, PCR (for HSV, CMV), acid‐fast stain, and fungal, viral, and mycobacterial cultures. Intravitreal injections of vancomycin and ceftazidime were given. Bacterial cultures showed a heavy intraocular growth of S. aureus, giving the diagnosis of endophthalmitis (bacterial or fungal infection of the vitreous or aqueous humor); all remaining stains and cultures were negative.

Figure 2

Over the next several days, the initial blood work returned with the following abnormal results: CD4 count was 70/L, and HIV serology was positive. The rapid plasma reagin test (RPR) was positive (titer 1:64). The enzyme immunoassay (EIA) and Treponema pallidum particle agglutination (TPPA) were also positive.

A lumbar puncture was performed, and CSF analysis indicated CSF fluid was clear, 2 erythrocytes and 2 leukocytes in the fourth tube, CSF glucose of 2.7 mmol/L (serum glucose 8.2 mmol/L), and CSF total protein of 1100 mg/L. There were no bacteria seen on the gram stain, and a rapid agglutination test for cryptococcal antigen was negative. The CSF RPR titer was 1:2, and the Treponema pallidum particle agglutination assay (TP‐PA) was reactive. The MRI of the brain indicated diffuse white matter disease but no meningeal enhancement. In combination, these results were indicative of neurosyphilis, and penicillin G IV therapy was initiated. He received a total of 14 days of IV therapy, followed by 3 weekly IM doses of benzathine penicillin. He also received a total of 28 days of IV cloxacillin therapy with 5 days of concomitant IV gentamicin for endocarditis treatment.

Over 8 weeks, the patient's panophthalmitis slowly improved. However, he maintained only light perception in the right eye and did not regain any vision in the left eye. He was discharged home to follow‐up with the infectious diseases and ophthalmology departments. The issue of initiating antiretroviral therapy, deferred during hospital admission because of his poor compliance history and the threat of immune reconstitution symptoms, was to be readdressed at this time. He missed both appointments and returned to the emergency room several months later with widespread Kaposi's sarcoma.

DISCUSSION

One of the key learning points from this case underlines that panuveitis carries a broad differential including inflammatory and infectious conditions, as well as lymphoma. Systemic infections include tuberculosis, syphilis, and in cases of severe immunosuppression, toxoplasmosis. Cytomegalovirus and candidiasis are less likely as they are not associated with intraocular inflammation. HIV is also on the differential, although it rarely causes severe panuveitis on its own. Inflammatory disorders such as Behcet syndrome, sarcoidosis, and, rarely, lens‐associated uveitis (if presented with a history of lens trauma or surgery) are also included on the differential. A systematic approach to the history and physical examination must be undertaken to narrow the search. A syphilis screen should always be included in the differential when investigating uveitis,1 especially given the resurgence of syphilis since 2000.2

Our patient presents an interesting study as he was coinfected with both syphilis and HIV. The progression of syphilis is far more aggressive in this scenario,3 as there is a higher frequency of initial presentation as secondary syphilis4 and with multiple persisting chancres.5 Secondary‐stage skin lesions are also more aggressive in coinfected patients (nodular or ulcerative lesions with necrotic centers), although the same dermatological presentations can be seen in HIV‐negative patients.6 It has not been definitively established whether HIV‐positive patients develop neurological complications of syphilis more frequently or earlier in disease, but most patients present with early neurosyphilis at the time of diagnosis.7 In keeping with these findings, our patient's initial presentation included both ocular and neurosyphilis as diagnostic features.

An atypical link highlighted by our case is that of endogenous, bacterial endophthalmitis secondary to endocarditis. Although traumatic or surgical complications are the most common causes of endophthalmitis, seeding from an endogenous infective source, although rare, is possible.810 Staphylococcus aureus endocarditis is one of the most common causes of endogenous spread.9 In our patient, his chronic uveitis and decompensated blood‐ocular barrier may have contributed to S. aureus seeding of his eye. As is the case with many patients diagnosed with S. aureus endocarditis, the source of infection was unknown, although several risk factors for S. aureus bacteremia have been documented. These risk factors include hospitalization, dialysis, transplantation, HIV‐positive status, heart disease, cancer, diabetes, and intravenous drug use. In a population‐based surveillance study from 1999 to 2000, 550 invasive isolates of S. aureus were obtained; the relative risk in HIV‐positive patients was 23.7.11 In a similar study, the source of the S. aureus bacteremia/endocarditis was not identified in 26% of patients with underlying medical conditions such as HIV infection.12

This case has demonstrated several intertwined disease presentations in a patient coinfected with multiple organisms. In an immunocompromised patient, Occam's razor does not necessarily hold true, and the possibility of multiple diagnoses must be entertained. Thus, clinicians must maintain a high index of suspicion for atypical presentations of typical diseases if their patients are to survive in the eye of the storm.

Klebsiella pneumoniae liver abscess is known to be associated with metastatic endophthalmitis,1 although most cases have been clustered in Taiwan, with few reports in the United States.2 The first reported case of Klebsiella liver abscess with endophthalmitis in the United States was in a 38‐year‐old man with a new diagnosis of diabetes, a known risk factor for hematogenous spread of Klebsiella to metastatic sites.3

CASE REPORT

A previously healthy 43‐year old Haitian man presented after experiencing 5 days of right eye pain with associated fever and swelling. The patient denied preceding trauma, manipulation of the eye, contact lens use, or illicit drug use and had no significant medical history. He had moved to the United States from Haiti more than 15 years ago and had not traveled out of the state of Florida since that time.

Physical exam showed tachycardia (rate = 110/min), tachypnea (rate = 20/min), and a temperature of 101.5F. The right eye had injected conjunctiva, a swollen lid, and decreased palpebral fissure, and visual acuity on the left was 20/60, whereas visual acuity on the right was recorded as the ability to count fingers at 3 feet. The remainder of his physical exam was within normal limits including the abdominal exam.

Laboratory data on admission included a white blood cell count of 37,500/L significant for 12% bands, total bilirubin of 2.8 mg/dL, AST of 141 U/L, ALT of 130 U/L, and alkaline phosphatase of 196 U/L. HIV testing was negative, and urine toxicology did not detect the presence of any illicit drugs. Vitreous cultures grew Klebsiella pneumoniae.

The initial CT scan of the orbit (Fig. 1) showed periorbital swelling and a preseptal collection anterior to the right globe consistent with an abscess. Because of the abnormal results of the liver panel in the presence of the ophthalmologic infection, an abdominal CT was obtained that showed an 11.5 by 8.0 cm lesion involving all segments of the right lobe of the liver with a 0.9‐mm cylindrical extension toward the right hepatic vein.

Figure 1

Percutaneous drainage of the liver abscess was performed, yielding positive cultures for K. pneumoniae. The patient was treated with oral gatifloxacin and intravenous ceftriaxone. Gatifloxacin therapy was chosen for its excellent penetrance into the vitreous.4 Despite antibiotic therapy, a repeat CT scan of the orbit showed further extension of the collection, and the decision was made to drain the abscess and perform right eye enucleation. The patient was discharged home on oral gatifloxacin. A follow‐up abdominal ultrasound 2 weeks after discharge showed complete resolution of the liver abscess.

DISCUSSION

Bacterial endophthalmitis is a rare infection involving the vitreous humor and other deep intraocular structures. It is most commonly exogenous in origin, caused by intraocular surgery, penetrating injury, a corneal ulcer, or periocular infection. Endogenous endophthalmitis occurs when organisms reach the eye hematogenously and accounts for fewer than 6% of all cases of endophthalmitis.5

Klebsiella liver abscesses have been increasing in incidence worldwide and since the mid‐1990s have become a common cause of liver abscess in the United States, along with Escherichia coli. The association with endophthalmitis was first reported in a series of 7 cases from Taiwan in 1986,1 and subsequent East Asian cases have been reported, usually in diabetic patients.6, 7 The association of Klebsiella liver abscesses with endogenous endophthalmitis has been rarely reported in the United States, with review of the literature from 1966 to 2003 revealing only 3 reported cases.2 One of these patients had diabetes, whereas another had beta‐thalassemia with previous splenectomy. Another study looking at only pyogenic liver abscess found biliary disease, hypertension, intraabdominal infection, and diabetes to be the most common underlying or concurrent conditions.8 Our patient did not appear to have any of these risk factors.

Our patient had no known risk factors to promote metastatic spread of the causative organisms. The patient was HIV negative, had no personal or family history of diabetes, and was not found to have elevated glucose levels at any point during admission. Although the ultimate etiology may never be determined, the possibility of undetected malignancy or cardiovascular or inflammatory disease cannot be excluded.

Physicians need to be aware of the global emergence of a hypervirulent strain of K. pneumonia causing liver abscesses and metastatic complications, especially endophthalmitis.9 Mucoviscosity associated gene A (magA) has been found in some liver isolates of K. pneumoniae.10 It has been suggested that as many as one‐third of patients infected with hyperviscous strains of K. pneumoniae will develop an invasive infection.11 Although it is unclear why metastatic endophthalmitis from Klebsiella liver abscess would be more common in East Asia, the magA gene may account for the observed difference. It was not possible to determine if the infectious organism that had infected our patient had the magA gene, although the clinical use of this information may not have changed management because the patient presented with metastatic infection. If this patient's particular organism had tested positive for the magA gene, it might explain why an apparently immunocompetent patient developed metastatic endophthalmitis not simply a liver abscess.

Patients with evidence of endogenous endophthalmitis without clear risk factors should be covered for K. pneumoniae, and extraocular sources should be sought, particularly the liver, even in the absence of diabetes. Early recognition and prompt initiation of antimicrobial therapy is essential if the patient's vision is to be preserved.

Klebsiella pneumoniae liver abscess is known to be associated with metastatic endophthalmitis,1 although most cases have been clustered in Taiwan, with few reports in the United States.2 The first reported case of Klebsiella liver abscess with endophthalmitis in the United States was in a 38‐year‐old man with a new diagnosis of diabetes, a known risk factor for hematogenous spread of Klebsiella to metastatic sites.3

CASE REPORT

A previously healthy 43‐year old Haitian man presented after experiencing 5 days of right eye pain with associated fever and swelling. The patient denied preceding trauma, manipulation of the eye, contact lens use, or illicit drug use and had no significant medical history. He had moved to the United States from Haiti more than 15 years ago and had not traveled out of the state of Florida since that time.

Physical exam showed tachycardia (rate = 110/min), tachypnea (rate = 20/min), and a temperature of 101.5F. The right eye had injected conjunctiva, a swollen lid, and decreased palpebral fissure, and visual acuity on the left was 20/60, whereas visual acuity on the right was recorded as the ability to count fingers at 3 feet. The remainder of his physical exam was within normal limits including the abdominal exam.

Laboratory data on admission included a white blood cell count of 37,500/L significant for 12% bands, total bilirubin of 2.8 mg/dL, AST of 141 U/L, ALT of 130 U/L, and alkaline phosphatase of 196 U/L. HIV testing was negative, and urine toxicology did not detect the presence of any illicit drugs. Vitreous cultures grew Klebsiella pneumoniae.

The initial CT scan of the orbit (Fig. 1) showed periorbital swelling and a preseptal collection anterior to the right globe consistent with an abscess. Because of the abnormal results of the liver panel in the presence of the ophthalmologic infection, an abdominal CT was obtained that showed an 11.5 by 8.0 cm lesion involving all segments of the right lobe of the liver with a 0.9‐mm cylindrical extension toward the right hepatic vein.

Figure 1

Percutaneous drainage of the liver abscess was performed, yielding positive cultures for K. pneumoniae. The patient was treated with oral gatifloxacin and intravenous ceftriaxone. Gatifloxacin therapy was chosen for its excellent penetrance into the vitreous.4 Despite antibiotic therapy, a repeat CT scan of the orbit showed further extension of the collection, and the decision was made to drain the abscess and perform right eye enucleation. The patient was discharged home on oral gatifloxacin. A follow‐up abdominal ultrasound 2 weeks after discharge showed complete resolution of the liver abscess.

DISCUSSION

Bacterial endophthalmitis is a rare infection involving the vitreous humor and other deep intraocular structures. It is most commonly exogenous in origin, caused by intraocular surgery, penetrating injury, a corneal ulcer, or periocular infection. Endogenous endophthalmitis occurs when organisms reach the eye hematogenously and accounts for fewer than 6% of all cases of endophthalmitis.5

Klebsiella liver abscesses have been increasing in incidence worldwide and since the mid‐1990s have become a common cause of liver abscess in the United States, along with Escherichia coli. The association with endophthalmitis was first reported in a series of 7 cases from Taiwan in 1986,1 and subsequent East Asian cases have been reported, usually in diabetic patients.6, 7 The association of Klebsiella liver abscesses with endogenous endophthalmitis has been rarely reported in the United States, with review of the literature from 1966 to 2003 revealing only 3 reported cases.2 One of these patients had diabetes, whereas another had beta‐thalassemia with previous splenectomy. Another study looking at only pyogenic liver abscess found biliary disease, hypertension, intraabdominal infection, and diabetes to be the most common underlying or concurrent conditions.8 Our patient did not appear to have any of these risk factors.

Our patient had no known risk factors to promote metastatic spread of the causative organisms. The patient was HIV negative, had no personal or family history of diabetes, and was not found to have elevated glucose levels at any point during admission. Although the ultimate etiology may never be determined, the possibility of undetected malignancy or cardiovascular or inflammatory disease cannot be excluded.

Physicians need to be aware of the global emergence of a hypervirulent strain of K. pneumonia causing liver abscesses and metastatic complications, especially endophthalmitis.9 Mucoviscosity associated gene A (magA) has been found in some liver isolates of K. pneumoniae.10 It has been suggested that as many as one‐third of patients infected with hyperviscous strains of K. pneumoniae will develop an invasive infection.11 Although it is unclear why metastatic endophthalmitis from Klebsiella liver abscess would be more common in East Asia, the magA gene may account for the observed difference. It was not possible to determine if the infectious organism that had infected our patient had the magA gene, although the clinical use of this information may not have changed management because the patient presented with metastatic infection. If this patient's particular organism had tested positive for the magA gene, it might explain why an apparently immunocompetent patient developed metastatic endophthalmitis not simply a liver abscess.

Patients with evidence of endogenous endophthalmitis without clear risk factors should be covered for K. pneumoniae, and extraocular sources should be sought, particularly the liver, even in the absence of diabetes. Early recognition and prompt initiation of antimicrobial therapy is essential if the patient's vision is to be preserved.

Klebsiella pneumoniae liver abscess is known to be associated with metastatic endophthalmitis,1 although most cases have been clustered in Taiwan, with few reports in the United States.2 The first reported case of Klebsiella liver abscess with endophthalmitis in the United States was in a 38‐year‐old man with a new diagnosis of diabetes, a known risk factor for hematogenous spread of Klebsiella to metastatic sites.3

CASE REPORT

A previously healthy 43‐year old Haitian man presented after experiencing 5 days of right eye pain with associated fever and swelling. The patient denied preceding trauma, manipulation of the eye, contact lens use, or illicit drug use and had no significant medical history. He had moved to the United States from Haiti more than 15 years ago and had not traveled out of the state of Florida since that time.

Physical exam showed tachycardia (rate = 110/min), tachypnea (rate = 20/min), and a temperature of 101.5F. The right eye had injected conjunctiva, a swollen lid, and decreased palpebral fissure, and visual acuity on the left was 20/60, whereas visual acuity on the right was recorded as the ability to count fingers at 3 feet. The remainder of his physical exam was within normal limits including the abdominal exam.

Laboratory data on admission included a white blood cell count of 37,500/L significant for 12% bands, total bilirubin of 2.8 mg/dL, AST of 141 U/L, ALT of 130 U/L, and alkaline phosphatase of 196 U/L. HIV testing was negative, and urine toxicology did not detect the presence of any illicit drugs. Vitreous cultures grew Klebsiella pneumoniae.

The initial CT scan of the orbit (Fig. 1) showed periorbital swelling and a preseptal collection anterior to the right globe consistent with an abscess. Because of the abnormal results of the liver panel in the presence of the ophthalmologic infection, an abdominal CT was obtained that showed an 11.5 by 8.0 cm lesion involving all segments of the right lobe of the liver with a 0.9‐mm cylindrical extension toward the right hepatic vein.

Figure 1

Percutaneous drainage of the liver abscess was performed, yielding positive cultures for K. pneumoniae. The patient was treated with oral gatifloxacin and intravenous ceftriaxone. Gatifloxacin therapy was chosen for its excellent penetrance into the vitreous.4 Despite antibiotic therapy, a repeat CT scan of the orbit showed further extension of the collection, and the decision was made to drain the abscess and perform right eye enucleation. The patient was discharged home on oral gatifloxacin. A follow‐up abdominal ultrasound 2 weeks after discharge showed complete resolution of the liver abscess.

DISCUSSION

Bacterial endophthalmitis is a rare infection involving the vitreous humor and other deep intraocular structures. It is most commonly exogenous in origin, caused by intraocular surgery, penetrating injury, a corneal ulcer, or periocular infection. Endogenous endophthalmitis occurs when organisms reach the eye hematogenously and accounts for fewer than 6% of all cases of endophthalmitis.5

Klebsiella liver abscesses have been increasing in incidence worldwide and since the mid‐1990s have become a common cause of liver abscess in the United States, along with Escherichia coli. The association with endophthalmitis was first reported in a series of 7 cases from Taiwan in 1986,1 and subsequent East Asian cases have been reported, usually in diabetic patients.6, 7 The association of Klebsiella liver abscesses with endogenous endophthalmitis has been rarely reported in the United States, with review of the literature from 1966 to 2003 revealing only 3 reported cases.2 One of these patients had diabetes, whereas another had beta‐thalassemia with previous splenectomy. Another study looking at only pyogenic liver abscess found biliary disease, hypertension, intraabdominal infection, and diabetes to be the most common underlying or concurrent conditions.8 Our patient did not appear to have any of these risk factors.

Our patient had no known risk factors to promote metastatic spread of the causative organisms. The patient was HIV negative, had no personal or family history of diabetes, and was not found to have elevated glucose levels at any point during admission. Although the ultimate etiology may never be determined, the possibility of undetected malignancy or cardiovascular or inflammatory disease cannot be excluded.

Physicians need to be aware of the global emergence of a hypervirulent strain of K. pneumonia causing liver abscesses and metastatic complications, especially endophthalmitis.9 Mucoviscosity associated gene A (magA) has been found in some liver isolates of K. pneumoniae.10 It has been suggested that as many as one‐third of patients infected with hyperviscous strains of K. pneumoniae will develop an invasive infection.11 Although it is unclear why metastatic endophthalmitis from Klebsiella liver abscess would be more common in East Asia, the magA gene may account for the observed difference. It was not possible to determine if the infectious organism that had infected our patient had the magA gene, although the clinical use of this information may not have changed management because the patient presented with metastatic infection. If this patient's particular organism had tested positive for the magA gene, it might explain why an apparently immunocompetent patient developed metastatic endophthalmitis not simply a liver abscess.

Patients with evidence of endogenous endophthalmitis without clear risk factors should be covered for K. pneumoniae, and extraocular sources should be sought, particularly the liver, even in the absence of diabetes. Early recognition and prompt initiation of antimicrobial therapy is essential if the patient's vision is to be preserved.

A 38‐year‐old HIV+ Ohio man with a recent CD4+ count of 534 cells/mL presented to his physician with 3 weeks of fever as high as 102F. He noted mild myalgias, pruritus, and an occasional cough but no headache, sore throat, dyspnea, rash, or gastrointestinal or genitourinary complaints. He had been seen elsewhere 2 weeks previously, when he had reported a single episode of receptive oral sex with a male partner several weeks earlier. He had been prescribed ciprofloxacin and azithromycin, but a throat swab came back negative for Chlamydia and Neisseria gonorrhoeae, and he reported no change in his symptoms after the course of antibiotics. He denied smoking or using street drugs. His only medications were citalopram and trazodone for depression.

This is a HIV+ man with a mild degree of immunosuppression with a fever of unknown origin (FUO). It is not yet known if the requisite basic infectious evaluation has been completed to meet this definition, but the duration certainly qualifies, and regardless of semantics, the FUO framework is a helpful starting point. The primary considerations in FUO are infections, neoplasms, and autoimmune illnesses. Autoimmune diseases are relatively less common in HIV patients. Although pruritis is quite common in HIV alone, it may also herald renal failure, cholestasis, or a malignancy (usually hematologic). Drugs must also be considered as a cause of unexplained fever; the pruritis might suggest an allergic reaction, although I do not think of citalopram or trazodone as having this effect. The failure to respond to broad‐spectrum antimicrobials (along with the duration) lowers my suspicion for common infections such as pneumonia, urinary tract infection, or cellulitis. Among sexually transmitted diseases, syphilis can be protean and merits consideration.

On examination he appeared well. His temperature was 102.4F, pulse 111 beats/min, blood pressure 138/78 mm Hg. The head, neck, cardiovascular system, and lungs appeared normal on examination. The abdomen was soft and nontender without organomegaly; skin, extremities, and neurological system were unremarkable. Rectal examination showed small anal condylomata. Hemoglobin was 14.3 g/dL, white blood cell count 6200/cm³, and platelet count 230,000/cm³. Serum electrolytes and lactate dehydrogenase were normal. The results of his liver function tests (LFTs) demonstrated a serum aspartate transaminase of 60 U/L (normal, 7‐40 U/L), alanine transaminase of 125 U/L (normal, 5‐50 U/L), alkaline phosphatase 218 U/L (normal, 40‐150 U/L), and total bilirubin 2.1 mg/dL (normal, 0.0‐1.5 mg/dL). Urinalysis demonstrated 2+ bilirubin and was otherwise normal. His erythrocyte sedimentation rate was 32 mm/hr (normal, 0‐15 mm/hr).

After 3 weeks of illness, his CBC demonstrates no signs of chronic illness (such as anemia of a chronic disease or a reactive leukocytosis or thrombocytosis). The results of his liver function tests showed moderate elevation, slightly more cholestatic than hepatocellular. This finding may reflect a disease process involving the liver, but such abnormal findings are often nonspecific in acute and chronic illnesses. With an unremitting fever, infectious complications in the liver merit early consideration. The time course rules out common biliary disorders such as cholangitis or cholecystitis. Pyogenic or amoebic liver abscesses are possible (homosexual men are at increased risk for the latter), but the absence of pain or abdominal tenderness is atypical. This biochemical profile can also be seen in chronic (but not acute) viral infections of the liver. Chronic hepatitis B and C predispose to hepatocellular carcinoma (HCC), which can be associated with fever. Cancers that infiltrate the liver, such as lymphoma or carcinoma, could also account for this picture. Indolent infections such as tuberculosis (TB) and syphilis are also possible, so associated signs of these systemic diseases should be sought. I do not believe either of his antibiotics is commonly associated with LFT abnormalities, and his CD4 count is too high for HIV cholangiopathy. In sum, a host of liver diseases are possible, but an extrahepatic systemic disease deserves equal attention.

His CD4+ count was 537 cells/mL, and his HIV RNA viral load was 44,300 copies/mL. Radiographs of the chest were normal. Two sets of blood cultures were negative. The rapid plasma reagin (RPR) was nonreactive. The results of serologies for acute hepatitis A, B, C, and E, chronic hepatitis B and C, and toxoplasmosis were negative. Testing for both Epstein‐Barr virus and cytomegalovirus showed evidence of remote infection. Results of serologies for bartonella species, human herpesviruses 6 and 7, and parvovirus B19 were negative.

The negative RPR makes disseminated (secondary) syphilis improbable, provided the prozone phenomenon has been excluded. An extensive serological workup is common in the evaluation of fever of unknown origin, although the threat of false‐positive results always looms when many studies are sent simultaneously. This must be considered in advance here, as his relatively preserved CD4 count affords him significant protection against many opportunistic infections. His HIV infection, however, regardless of CD4 count, increases his risk for TB and lymphoma, which remain high on my list. Both may be residing primarily in the liver. In FUO, the abdominal CT is frequently a high‐yield test (primarily by demonstrating unsuspected tumors and abscesses), even in the absence of symptoms, and would certainly be of interest here given the liver function test results. Imaging could diagnose febrile tumors such as lymphoma, HCC, or renal cell carcinoma. In the event that imaging is unrevealing, causes of granulomatous hepatitis should be entertained. The constellation of cough, LFT abnormalities, and fever is compatible with Q fever. As with any FUO case, I would also carefully revisit this patient's history to discern where he was born, where he has been, and what activities or exposures he is engaged in.

He was seen 2 days later with fever of 104F and new papules over his sternal area. Over the next week, he had intermittent fevers and severe fatigue. The rash progressed, predominantly involving his chest and back, but also his legs, arms, and face (see Fig. 1). The lesions spared his palms and soles. The exanthem was intensely pruritic and maculopapular, consisting of lesions with a diameter of 0.5 cm or less, with some scaling. There were no vesicles or pustular lesions. There were no other new findings on examination. His transaminase and bilirubin had normalized, and his CBC and electrolytes were unchanged. Repeat blood cultures held for extended incubation were negative. Computerized tomography of the chest, abdomen, and pelvis demonstrated mild lymphadenopathy at the porta hepatis with increased portocaval and periaortic lymphadenopathy.

Figure 1

The only LFT abnormality that persists is the elevated alkaline phosphatase, which suggests (1) that liver involvement was not specific and that there is a disease process involving the bone, (2) that there is a persistent infiltrative disorder of the liver such as infection or malignancy or, less likely, amyloidosis or sarcoidosis, or (3) that the porta hepatis lymphadenopathy is causing biliary obstruction. The underlying diagnosis must explain the rash, intraabdominal lymphadenopathy, and fever. The time course does somewhat limit the extensive differential of fever and rash. After 3 weeks of illness, some of the most life‐threatening entities such as meningococcal disease, Rocky Mountain spotted fever, and toxic shock syndrome are unlikely. Concern remains for infections that are more indolent, such as mycobacteria, fungi, or spirochetes. The most striking elements of the rash are the extensive distribution, rapid progression, large number, and discreteness of the lesions, which collectively point more toward disseminated fungal (eg, histoplasmosis, as he lives in Ohio), spirochetal, rickettsial, or viral etiologies, rather than bacterial or mycobacterial entities. The absence of vesicles detracts from the diagnosis of a disseminated herpes virus such as herpes simplex or varicella. I believe that this rash is too disseminated to be caused by a common mycobacterial illness. This extent of cutaneous metastases would usually accompany a far more ill patient with an obvious primary cancer (none is seen on imaging, including the liver), and it appears too extensive to be caused by a paraneoplastic phenomenon such as Sweet's syndrome. A systemic vasculitis or another autoimmune disease remains possible, but there is minimal evidence of visceral organ involvement. All the aforementioned diseases could explain the intraabdominal lymphadenopathy, but my suspicion is highest for infection. I would biopsy and culture the skin lesions, repeat the RPR and/or send a treponemal‐specific test, place a PPD skin test, and send fungal studies (serum serologies and urine antigens) for evaluation. If the results of these noninvasive studies are unrevealing, I would consider a liver biopsy.

The patient's medications were discontinued, and a skin biopsy of the rash from his chest showed atypical lymphohistiocytic infiltrates without acute inflammatory cells and with negative Gomori methenamine silver (GMS), acid‐fast bacilli (AFB), and Fite (for Nocardia) stains. The infiltrates were predominantly T cells with a 1:1 CD4:CD8 ratio. This was read as suspicious for cytotoxic (CD8) mycosis fungoides.

I do not have reason to doubt the pathologist's impression of mycosis fungoides on histopathologic grounds, but from a clinical standpoint, I do not think mycosis fungoides is a disease that has a prolonged febrile prodrome or an explosive cutaneous onset. Rather, it is frequently preceded by nonspecific skin findings over a long period. Thinking broadly and pathophysiologically and noting that T cells are the predominant lymphocytes in skin, I wonder if they could represent a nonmalignant, immunological reaction in the skin. The stains, although not perfectly sensitive, make mycobacterial and fungal diseases less likely, although incubation of cultures is necessary.

Over the next 10 days (bringing the total duration of the patient's illness to 6 weeks), the skin lesions increased in number. In the physician's office at his next follow‐up, the patient had a temperature of 104.1F, was uncomfortable, shivering, and ill‐appearing. His blood pressure was 108/66 mm Hg, and his pulse 114 beats/min. He complained of severe shooting pains, predominantly in his pretibial regions and arms. Examination showed no other new findings, including no focal neurological findings. The results of the T‐cell rearrangement study from the skin biopsy showed evidence of a monoclonal T‐cell population. He was admitted to the hospital for further evaluation and treatment.

The extremity dysesthesias could represent a lesion of the spinal cord (including the CSF/meninges), a polyradiculopathy, or a polyneuropathy. Unfortunately, this does not add a tremendous amount of diagnostic resolution, as infection, malignancy, and autoimmune syndromes, such as vasculitis, may all involve the nervous system in these ways. In general, I associate monoclonal lymphocyte responses with hematological malignancies and polyclonal responses with the less specific inflammation that could accompany infection, autoimmunity, or solid malignancies. His age, fever, and rapid progression seem atypical for mycosis fungoides, but given the monoclonal T cells, this must now be considered. Adult T‐cell leukemia/lymphoma, with its prominent skin manifestations and its association with HLTV‐1, is an alternative T‐cell malignancy that could explain the fever, neurological symptoms, and possible visceral involvement (elevated alkaline phosphatase, which could reflect liver or bone). In cases that are diagnostic challenges, one of the highest‐yield maneuvers is to repeat the preceding evaluation, starting with the history, exam, and basic labs, and if necessary, to review or repeat the imaging or skin biopsy. Given the elevated alkaline phosphatase, disseminated rash, new neurological symptoms, and his HIV status, I remain particularly concerned about syphilis and would do further testing (accounting for the prozone phenomenon) before proceeding with the malignancy evaluation.

A lumbar puncture demonstrated clear cerebrospinal fluid, with 2 leukocytes and 195 erythrocytes/cm³, protein of 26 mg/dL, and glucose of 52 mg/dL. Bacterial and fungal cultures of the fluid were negative. The results of colonoscopy were normal. A bone marrow biopsy demonstrated ring granulomas. GMS, AFB, Fite, and Steiner (for spirochetes) stains were negative, cultures of the aspirate were negative for bacteria, and smears were negative for fungi and mycobacteria. Antibody tests for human T‐cell lymphotropic virus types I and II, Coxiella burnetii, and Bartonella henselae were negative. The dermatology consultant believed the absence of lymphadenopathy and the pruritic nature of the lesions was atypical for cytotoxic T‐cell lymphoma (CTCL). Before initiating therapy for CTCL, she suggested repeating the skin biopsy and RPR.

The repeat RPR was positive at 1:64 dilutions, and a confirmatory fluorescent treponemal antibody absorption test showed a positive result. He was prescribed intramuscular benzathine pencillin 2.4 million units weekly for 3 weeks, with almost immediate defervescence and slower resolution of his rash and shooting pains in his limbs. The repeat skin biopsy done during the hospitalization demonstrated lichenoid‐type dermatitis with interstitial and perivascular lymphohistiocytic infiltrates and granulomas. Steiner stains for spirochetes were positive. Immunohistochemical stains ruled out a lymphoproliferative process. One year later his RPR was nonreactive.

COMMENTARY

Fever of unknown origin (FUO) was first defined by Petersdorf and Beeson in 1961 as a temperature higher than 38.3C on several occasions lasting longer than 3 weeks and defying diagnosis despite 1 week of inpatient investigation.1 Dramatic changes in medical practice have rendered this definition outdated, with more recent proposals allowing thoughtful outpatient investigation to serve as a surrogate for hospitalization. Some have proposed that HIV‐associated FUO be considered a distinct entity, with the most complete North American series finding the etiology of the HIV‐associated FUO in 56 of 70 patients.2 The mean CD4+ count in this series was 58/cm³. Disseminated M. avium was the most frequently diagnosed cause, followed by P. jirovecii pneumonia, cytomegalovirus infection, disseminated histoplasmosis, and lymphoma. Of 14 patients with fever of no definable etiology, 12 eventually proved to have self‐limiting illness.

Despite numerous attempts to reduce the investigation of the patient with FUO to an algorithm, the approach must be individualized. A thorough history and careful, serial physical examinations are frequently and appropriately stressed as the foundation, followed by thoughtful selection of laboratory and imaging studies. Although FUO has a lengthy differential diagnosis, it often proves to be, as Mackowiak and Durack stress, an unusual manifestation of a common disease, rather than a typical presentation of a rare disease.3 A relatively uncommon disease in conjunction with an initially negative diagnostic test result, as was the case with this patient, may lead to a protracted diagnostic puzzle.

Syphilis is a rare cause of FUO. In 6 large studies of a total of 947 patients published over a 40‐year period, only 2 cases of syphilis (1 secondary and 1 neurosyphilis) were reported.1, 48 Syphilis as a cause of prolonged cryptic fever appears to have been seen with greater frequency in the preantibiotic era.9 In the first half of the 20th century, syphilis was known as the great imitator, with its unusual manifestations recognized and indeed expected. As a result of the dramatically lower incidence of syphilis in recent decades, these lessons have largely been forgotten, however, which may lead to diagnostic confusion when syphilis presents atypically. The manifestations of secondary syphilis are protean, including a variety of rashes, aphthous ulcers, arthralgias, pharyngitis, weight loss, fever, meningitis, ocular symptoms, cranial nerve palsies, glomerulonephritis, hepatitis, and periostitis (which afflicted this patient, who complained of severe shooting pains in his arms and shins).

After declining in the last decade of the 20th century, the rates of primary and secondary syphilis are rising in the United States.10 Oral sex is a clear risk factor for syphilis transmission, particularly for men who have sex with men.11 Because of the patient's exposure history and clinical picture, his outpatient physician considered the diagnosis of secondary syphilis early in the course of his illness. The diagnosis was not entertained further when an RPR test, highly sensitive at this stage of the disease, returned nonreactive. Likewise, when a rash subsequently appeared, the lack of palm and sole involvement dissuaded multiple clinicians from reconsidering the diagnosis of syphilis. A skin biopsy that appeared to lead in a distinctly different direction understandably confused the picture still further. Even at the time of the lumbar puncture, VDRL of the CSF was not ordered.

In retrospect, the chief confounder in the case was the false‐negative RPR test, as the discussant suspected early on. Although nontreponemal tests are generally accurate in individuals with HIV, delayed seropositivity and false‐negatives have been reported in this population.12 The false‐negative could have also been a result of the prozone phenomenon, an unusual event, occurring in fewer than 2% of cases of secondary syphilis and attributed to a mismatch between antibody and very high antigen level. The prozone reaction can be corrected for by requesting dilution of the serum prior to repeating the test. Simple lab error must be considered as well, but without access to this patient's serum from his original testing, the cause of his initial false‐negative test cannot be known with certainty.

An unusual presentation in conjunction with failure to recognize the causes of rare false‐negative testing for secondary syphilis led to a delayed diagnosis in this patient. Although syphilis and mycosis fungoides have previously been reported to mimic one another both clinically and histopathologically, the potential for secondary syphilis to be misdiagnosed in this fashion is not generally appreciated.1315 Recognition of the possibility of secondary syphilis occurred just in time to spare this patient the rash decision of treating him with cytotoxic therapy directed against CTCL.

A 38‐year‐old HIV+ Ohio man with a recent CD4+ count of 534 cells/mL presented to his physician with 3 weeks of fever as high as 102F. He noted mild myalgias, pruritus, and an occasional cough but no headache, sore throat, dyspnea, rash, or gastrointestinal or genitourinary complaints. He had been seen elsewhere 2 weeks previously, when he had reported a single episode of receptive oral sex with a male partner several weeks earlier. He had been prescribed ciprofloxacin and azithromycin, but a throat swab came back negative for Chlamydia and Neisseria gonorrhoeae, and he reported no change in his symptoms after the course of antibiotics. He denied smoking or using street drugs. His only medications were citalopram and trazodone for depression.

This is a HIV+ man with a mild degree of immunosuppression with a fever of unknown origin (FUO). It is not yet known if the requisite basic infectious evaluation has been completed to meet this definition, but the duration certainly qualifies, and regardless of semantics, the FUO framework is a helpful starting point. The primary considerations in FUO are infections, neoplasms, and autoimmune illnesses. Autoimmune diseases are relatively less common in HIV patients. Although pruritis is quite common in HIV alone, it may also herald renal failure, cholestasis, or a malignancy (usually hematologic). Drugs must also be considered as a cause of unexplained fever; the pruritis might suggest an allergic reaction, although I do not think of citalopram or trazodone as having this effect. The failure to respond to broad‐spectrum antimicrobials (along with the duration) lowers my suspicion for common infections such as pneumonia, urinary tract infection, or cellulitis. Among sexually transmitted diseases, syphilis can be protean and merits consideration.

On examination he appeared well. His temperature was 102.4F, pulse 111 beats/min, blood pressure 138/78 mm Hg. The head, neck, cardiovascular system, and lungs appeared normal on examination. The abdomen was soft and nontender without organomegaly; skin, extremities, and neurological system were unremarkable. Rectal examination showed small anal condylomata. Hemoglobin was 14.3 g/dL, white blood cell count 6200/cm³, and platelet count 230,000/cm³. Serum electrolytes and lactate dehydrogenase were normal. The results of his liver function tests (LFTs) demonstrated a serum aspartate transaminase of 60 U/L (normal, 7‐40 U/L), alanine transaminase of 125 U/L (normal, 5‐50 U/L), alkaline phosphatase 218 U/L (normal, 40‐150 U/L), and total bilirubin 2.1 mg/dL (normal, 0.0‐1.5 mg/dL). Urinalysis demonstrated 2+ bilirubin and was otherwise normal. His erythrocyte sedimentation rate was 32 mm/hr (normal, 0‐15 mm/hr).

After 3 weeks of illness, his CBC demonstrates no signs of chronic illness (such as anemia of a chronic disease or a reactive leukocytosis or thrombocytosis). The results of his liver function tests showed moderate elevation, slightly more cholestatic than hepatocellular. This finding may reflect a disease process involving the liver, but such abnormal findings are often nonspecific in acute and chronic illnesses. With an unremitting fever, infectious complications in the liver merit early consideration. The time course rules out common biliary disorders such as cholangitis or cholecystitis. Pyogenic or amoebic liver abscesses are possible (homosexual men are at increased risk for the latter), but the absence of pain or abdominal tenderness is atypical. This biochemical profile can also be seen in chronic (but not acute) viral infections of the liver. Chronic hepatitis B and C predispose to hepatocellular carcinoma (HCC), which can be associated with fever. Cancers that infiltrate the liver, such as lymphoma or carcinoma, could also account for this picture. Indolent infections such as tuberculosis (TB) and syphilis are also possible, so associated signs of these systemic diseases should be sought. I do not believe either of his antibiotics is commonly associated with LFT abnormalities, and his CD4 count is too high for HIV cholangiopathy. In sum, a host of liver diseases are possible, but an extrahepatic systemic disease deserves equal attention.

His CD4+ count was 537 cells/mL, and his HIV RNA viral load was 44,300 copies/mL. Radiographs of the chest were normal. Two sets of blood cultures were negative. The rapid plasma reagin (RPR) was nonreactive. The results of serologies for acute hepatitis A, B, C, and E, chronic hepatitis B and C, and toxoplasmosis were negative. Testing for both Epstein‐Barr virus and cytomegalovirus showed evidence of remote infection. Results of serologies for bartonella species, human herpesviruses 6 and 7, and parvovirus B19 were negative.

The negative RPR makes disseminated (secondary) syphilis improbable, provided the prozone phenomenon has been excluded. An extensive serological workup is common in the evaluation of fever of unknown origin, although the threat of false‐positive results always looms when many studies are sent simultaneously. This must be considered in advance here, as his relatively preserved CD4 count affords him significant protection against many opportunistic infections. His HIV infection, however, regardless of CD4 count, increases his risk for TB and lymphoma, which remain high on my list. Both may be residing primarily in the liver. In FUO, the abdominal CT is frequently a high‐yield test (primarily by demonstrating unsuspected tumors and abscesses), even in the absence of symptoms, and would certainly be of interest here given the liver function test results. Imaging could diagnose febrile tumors such as lymphoma, HCC, or renal cell carcinoma. In the event that imaging is unrevealing, causes of granulomatous hepatitis should be entertained. The constellation of cough, LFT abnormalities, and fever is compatible with Q fever. As with any FUO case, I would also carefully revisit this patient's history to discern where he was born, where he has been, and what activities or exposures he is engaged in.

He was seen 2 days later with fever of 104F and new papules over his sternal area. Over the next week, he had intermittent fevers and severe fatigue. The rash progressed, predominantly involving his chest and back, but also his legs, arms, and face (see Fig. 1). The lesions spared his palms and soles. The exanthem was intensely pruritic and maculopapular, consisting of lesions with a diameter of 0.5 cm or less, with some scaling. There were no vesicles or pustular lesions. There were no other new findings on examination. His transaminase and bilirubin had normalized, and his CBC and electrolytes were unchanged. Repeat blood cultures held for extended incubation were negative. Computerized tomography of the chest, abdomen, and pelvis demonstrated mild lymphadenopathy at the porta hepatis with increased portocaval and periaortic lymphadenopathy.

Figure 1

The only LFT abnormality that persists is the elevated alkaline phosphatase, which suggests (1) that liver involvement was not specific and that there is a disease process involving the bone, (2) that there is a persistent infiltrative disorder of the liver such as infection or malignancy or, less likely, amyloidosis or sarcoidosis, or (3) that the porta hepatis lymphadenopathy is causing biliary obstruction. The underlying diagnosis must explain the rash, intraabdominal lymphadenopathy, and fever. The time course does somewhat limit the extensive differential of fever and rash. After 3 weeks of illness, some of the most life‐threatening entities such as meningococcal disease, Rocky Mountain spotted fever, and toxic shock syndrome are unlikely. Concern remains for infections that are more indolent, such as mycobacteria, fungi, or spirochetes. The most striking elements of the rash are the extensive distribution, rapid progression, large number, and discreteness of the lesions, which collectively point more toward disseminated fungal (eg, histoplasmosis, as he lives in Ohio), spirochetal, rickettsial, or viral etiologies, rather than bacterial or mycobacterial entities. The absence of vesicles detracts from the diagnosis of a disseminated herpes virus such as herpes simplex or varicella. I believe that this rash is too disseminated to be caused by a common mycobacterial illness. This extent of cutaneous metastases would usually accompany a far more ill patient with an obvious primary cancer (none is seen on imaging, including the liver), and it appears too extensive to be caused by a paraneoplastic phenomenon such as Sweet's syndrome. A systemic vasculitis or another autoimmune disease remains possible, but there is minimal evidence of visceral organ involvement. All the aforementioned diseases could explain the intraabdominal lymphadenopathy, but my suspicion is highest for infection. I would biopsy and culture the skin lesions, repeat the RPR and/or send a treponemal‐specific test, place a PPD skin test, and send fungal studies (serum serologies and urine antigens) for evaluation. If the results of these noninvasive studies are unrevealing, I would consider a liver biopsy.

The patient's medications were discontinued, and a skin biopsy of the rash from his chest showed atypical lymphohistiocytic infiltrates without acute inflammatory cells and with negative Gomori methenamine silver (GMS), acid‐fast bacilli (AFB), and Fite (for Nocardia) stains. The infiltrates were predominantly T cells with a 1:1 CD4:CD8 ratio. This was read as suspicious for cytotoxic (CD8) mycosis fungoides.

I do not have reason to doubt the pathologist's impression of mycosis fungoides on histopathologic grounds, but from a clinical standpoint, I do not think mycosis fungoides is a disease that has a prolonged febrile prodrome or an explosive cutaneous onset. Rather, it is frequently preceded by nonspecific skin findings over a long period. Thinking broadly and pathophysiologically and noting that T cells are the predominant lymphocytes in skin, I wonder if they could represent a nonmalignant, immunological reaction in the skin. The stains, although not perfectly sensitive, make mycobacterial and fungal diseases less likely, although incubation of cultures is necessary.

Over the next 10 days (bringing the total duration of the patient's illness to 6 weeks), the skin lesions increased in number. In the physician's office at his next follow‐up, the patient had a temperature of 104.1F, was uncomfortable, shivering, and ill‐appearing. His blood pressure was 108/66 mm Hg, and his pulse 114 beats/min. He complained of severe shooting pains, predominantly in his pretibial regions and arms. Examination showed no other new findings, including no focal neurological findings. The results of the T‐cell rearrangement study from the skin biopsy showed evidence of a monoclonal T‐cell population. He was admitted to the hospital for further evaluation and treatment.

The extremity dysesthesias could represent a lesion of the spinal cord (including the CSF/meninges), a polyradiculopathy, or a polyneuropathy. Unfortunately, this does not add a tremendous amount of diagnostic resolution, as infection, malignancy, and autoimmune syndromes, such as vasculitis, may all involve the nervous system in these ways. In general, I associate monoclonal lymphocyte responses with hematological malignancies and polyclonal responses with the less specific inflammation that could accompany infection, autoimmunity, or solid malignancies. His age, fever, and rapid progression seem atypical for mycosis fungoides, but given the monoclonal T cells, this must now be considered. Adult T‐cell leukemia/lymphoma, with its prominent skin manifestations and its association with HLTV‐1, is an alternative T‐cell malignancy that could explain the fever, neurological symptoms, and possible visceral involvement (elevated alkaline phosphatase, which could reflect liver or bone). In cases that are diagnostic challenges, one of the highest‐yield maneuvers is to repeat the preceding evaluation, starting with the history, exam, and basic labs, and if necessary, to review or repeat the imaging or skin biopsy. Given the elevated alkaline phosphatase, disseminated rash, new neurological symptoms, and his HIV status, I remain particularly concerned about syphilis and would do further testing (accounting for the prozone phenomenon) before proceeding with the malignancy evaluation.

A lumbar puncture demonstrated clear cerebrospinal fluid, with 2 leukocytes and 195 erythrocytes/cm³, protein of 26 mg/dL, and glucose of 52 mg/dL. Bacterial and fungal cultures of the fluid were negative. The results of colonoscopy were normal. A bone marrow biopsy demonstrated ring granulomas. GMS, AFB, Fite, and Steiner (for spirochetes) stains were negative, cultures of the aspirate were negative for bacteria, and smears were negative for fungi and mycobacteria. Antibody tests for human T‐cell lymphotropic virus types I and II, Coxiella burnetii, and Bartonella henselae were negative. The dermatology consultant believed the absence of lymphadenopathy and the pruritic nature of the lesions was atypical for cytotoxic T‐cell lymphoma (CTCL). Before initiating therapy for CTCL, she suggested repeating the skin biopsy and RPR.

The repeat RPR was positive at 1:64 dilutions, and a confirmatory fluorescent treponemal antibody absorption test showed a positive result. He was prescribed intramuscular benzathine pencillin 2.4 million units weekly for 3 weeks, with almost immediate defervescence and slower resolution of his rash and shooting pains in his limbs. The repeat skin biopsy done during the hospitalization demonstrated lichenoid‐type dermatitis with interstitial and perivascular lymphohistiocytic infiltrates and granulomas. Steiner stains for spirochetes were positive. Immunohistochemical stains ruled out a lymphoproliferative process. One year later his RPR was nonreactive.

COMMENTARY

Fever of unknown origin (FUO) was first defined by Petersdorf and Beeson in 1961 as a temperature higher than 38.3C on several occasions lasting longer than 3 weeks and defying diagnosis despite 1 week of inpatient investigation.1 Dramatic changes in medical practice have rendered this definition outdated, with more recent proposals allowing thoughtful outpatient investigation to serve as a surrogate for hospitalization. Some have proposed that HIV‐associated FUO be considered a distinct entity, with the most complete North American series finding the etiology of the HIV‐associated FUO in 56 of 70 patients.2 The mean CD4+ count in this series was 58/cm³. Disseminated M. avium was the most frequently diagnosed cause, followed by P. jirovecii pneumonia, cytomegalovirus infection, disseminated histoplasmosis, and lymphoma. Of 14 patients with fever of no definable etiology, 12 eventually proved to have self‐limiting illness.

Despite numerous attempts to reduce the investigation of the patient with FUO to an algorithm, the approach must be individualized. A thorough history and careful, serial physical examinations are frequently and appropriately stressed as the foundation, followed by thoughtful selection of laboratory and imaging studies. Although FUO has a lengthy differential diagnosis, it often proves to be, as Mackowiak and Durack stress, an unusual manifestation of a common disease, rather than a typical presentation of a rare disease.3 A relatively uncommon disease in conjunction with an initially negative diagnostic test result, as was the case with this patient, may lead to a protracted diagnostic puzzle.

Syphilis is a rare cause of FUO. In 6 large studies of a total of 947 patients published over a 40‐year period, only 2 cases of syphilis (1 secondary and 1 neurosyphilis) were reported.1, 48 Syphilis as a cause of prolonged cryptic fever appears to have been seen with greater frequency in the preantibiotic era.9 In the first half of the 20th century, syphilis was known as the great imitator, with its unusual manifestations recognized and indeed expected. As a result of the dramatically lower incidence of syphilis in recent decades, these lessons have largely been forgotten, however, which may lead to diagnostic confusion when syphilis presents atypically. The manifestations of secondary syphilis are protean, including a variety of rashes, aphthous ulcers, arthralgias, pharyngitis, weight loss, fever, meningitis, ocular symptoms, cranial nerve palsies, glomerulonephritis, hepatitis, and periostitis (which afflicted this patient, who complained of severe shooting pains in his arms and shins).

After declining in the last decade of the 20th century, the rates of primary and secondary syphilis are rising in the United States.10 Oral sex is a clear risk factor for syphilis transmission, particularly for men who have sex with men.11 Because of the patient's exposure history and clinical picture, his outpatient physician considered the diagnosis of secondary syphilis early in the course of his illness. The diagnosis was not entertained further when an RPR test, highly sensitive at this stage of the disease, returned nonreactive. Likewise, when a rash subsequently appeared, the lack of palm and sole involvement dissuaded multiple clinicians from reconsidering the diagnosis of syphilis. A skin biopsy that appeared to lead in a distinctly different direction understandably confused the picture still further. Even at the time of the lumbar puncture, VDRL of the CSF was not ordered.

In retrospect, the chief confounder in the case was the false‐negative RPR test, as the discussant suspected early on. Although nontreponemal tests are generally accurate in individuals with HIV, delayed seropositivity and false‐negatives have been reported in this population.12 The false‐negative could have also been a result of the prozone phenomenon, an unusual event, occurring in fewer than 2% of cases of secondary syphilis and attributed to a mismatch between antibody and very high antigen level. The prozone reaction can be corrected for by requesting dilution of the serum prior to repeating the test. Simple lab error must be considered as well, but without access to this patient's serum from his original testing, the cause of his initial false‐negative test cannot be known with certainty.

An unusual presentation in conjunction with failure to recognize the causes of rare false‐negative testing for secondary syphilis led to a delayed diagnosis in this patient. Although syphilis and mycosis fungoides have previously been reported to mimic one another both clinically and histopathologically, the potential for secondary syphilis to be misdiagnosed in this fashion is not generally appreciated.1315 Recognition of the possibility of secondary syphilis occurred just in time to spare this patient the rash decision of treating him with cytotoxic therapy directed against CTCL.

A 38‐year‐old HIV+ Ohio man with a recent CD4+ count of 534 cells/mL presented to his physician with 3 weeks of fever as high as 102F. He noted mild myalgias, pruritus, and an occasional cough but no headache, sore throat, dyspnea, rash, or gastrointestinal or genitourinary complaints. He had been seen elsewhere 2 weeks previously, when he had reported a single episode of receptive oral sex with a male partner several weeks earlier. He had been prescribed ciprofloxacin and azithromycin, but a throat swab came back negative for Chlamydia and Neisseria gonorrhoeae, and he reported no change in his symptoms after the course of antibiotics. He denied smoking or using street drugs. His only medications were citalopram and trazodone for depression.

This is a HIV+ man with a mild degree of immunosuppression with a fever of unknown origin (FUO). It is not yet known if the requisite basic infectious evaluation has been completed to meet this definition, but the duration certainly qualifies, and regardless of semantics, the FUO framework is a helpful starting point. The primary considerations in FUO are infections, neoplasms, and autoimmune illnesses. Autoimmune diseases are relatively less common in HIV patients. Although pruritis is quite common in HIV alone, it may also herald renal failure, cholestasis, or a malignancy (usually hematologic). Drugs must also be considered as a cause of unexplained fever; the pruritis might suggest an allergic reaction, although I do not think of citalopram or trazodone as having this effect. The failure to respond to broad‐spectrum antimicrobials (along with the duration) lowers my suspicion for common infections such as pneumonia, urinary tract infection, or cellulitis. Among sexually transmitted diseases, syphilis can be protean and merits consideration.

On examination he appeared well. His temperature was 102.4F, pulse 111 beats/min, blood pressure 138/78 mm Hg. The head, neck, cardiovascular system, and lungs appeared normal on examination. The abdomen was soft and nontender without organomegaly; skin, extremities, and neurological system were unremarkable. Rectal examination showed small anal condylomata. Hemoglobin was 14.3 g/dL, white blood cell count 6200/cm³, and platelet count 230,000/cm³. Serum electrolytes and lactate dehydrogenase were normal. The results of his liver function tests (LFTs) demonstrated a serum aspartate transaminase of 60 U/L (normal, 7‐40 U/L), alanine transaminase of 125 U/L (normal, 5‐50 U/L), alkaline phosphatase 218 U/L (normal, 40‐150 U/L), and total bilirubin 2.1 mg/dL (normal, 0.0‐1.5 mg/dL). Urinalysis demonstrated 2+ bilirubin and was otherwise normal. His erythrocyte sedimentation rate was 32 mm/hr (normal, 0‐15 mm/hr).

After 3 weeks of illness, his CBC demonstrates no signs of chronic illness (such as anemia of a chronic disease or a reactive leukocytosis or thrombocytosis). The results of his liver function tests showed moderate elevation, slightly more cholestatic than hepatocellular. This finding may reflect a disease process involving the liver, but such abnormal findings are often nonspecific in acute and chronic illnesses. With an unremitting fever, infectious complications in the liver merit early consideration. The time course rules out common biliary disorders such as cholangitis or cholecystitis. Pyogenic or amoebic liver abscesses are possible (homosexual men are at increased risk for the latter), but the absence of pain or abdominal tenderness is atypical. This biochemical profile can also be seen in chronic (but not acute) viral infections of the liver. Chronic hepatitis B and C predispose to hepatocellular carcinoma (HCC), which can be associated with fever. Cancers that infiltrate the liver, such as lymphoma or carcinoma, could also account for this picture. Indolent infections such as tuberculosis (TB) and syphilis are also possible, so associated signs of these systemic diseases should be sought. I do not believe either of his antibiotics is commonly associated with LFT abnormalities, and his CD4 count is too high for HIV cholangiopathy. In sum, a host of liver diseases are possible, but an extrahepatic systemic disease deserves equal attention.

His CD4+ count was 537 cells/mL, and his HIV RNA viral load was 44,300 copies/mL. Radiographs of the chest were normal. Two sets of blood cultures were negative. The rapid plasma reagin (RPR) was nonreactive. The results of serologies for acute hepatitis A, B, C, and E, chronic hepatitis B and C, and toxoplasmosis were negative. Testing for both Epstein‐Barr virus and cytomegalovirus showed evidence of remote infection. Results of serologies for bartonella species, human herpesviruses 6 and 7, and parvovirus B19 were negative.

The negative RPR makes disseminated (secondary) syphilis improbable, provided the prozone phenomenon has been excluded. An extensive serological workup is common in the evaluation of fever of unknown origin, although the threat of false‐positive results always looms when many studies are sent simultaneously. This must be considered in advance here, as his relatively preserved CD4 count affords him significant protection against many opportunistic infections. His HIV infection, however, regardless of CD4 count, increases his risk for TB and lymphoma, which remain high on my list. Both may be residing primarily in the liver. In FUO, the abdominal CT is frequently a high‐yield test (primarily by demonstrating unsuspected tumors and abscesses), even in the absence of symptoms, and would certainly be of interest here given the liver function test results. Imaging could diagnose febrile tumors such as lymphoma, HCC, or renal cell carcinoma. In the event that imaging is unrevealing, causes of granulomatous hepatitis should be entertained. The constellation of cough, LFT abnormalities, and fever is compatible with Q fever. As with any FUO case, I would also carefully revisit this patient's history to discern where he was born, where he has been, and what activities or exposures he is engaged in.

He was seen 2 days later with fever of 104F and new papules over his sternal area. Over the next week, he had intermittent fevers and severe fatigue. The rash progressed, predominantly involving his chest and back, but also his legs, arms, and face (see Fig. 1). The lesions spared his palms and soles. The exanthem was intensely pruritic and maculopapular, consisting of lesions with a diameter of 0.5 cm or less, with some scaling. There were no vesicles or pustular lesions. There were no other new findings on examination. His transaminase and bilirubin had normalized, and his CBC and electrolytes were unchanged. Repeat blood cultures held for extended incubation were negative. Computerized tomography of the chest, abdomen, and pelvis demonstrated mild lymphadenopathy at the porta hepatis with increased portocaval and periaortic lymphadenopathy.

Figure 1

The only LFT abnormality that persists is the elevated alkaline phosphatase, which suggests (1) that liver involvement was not specific and that there is a disease process involving the bone, (2) that there is a persistent infiltrative disorder of the liver such as infection or malignancy or, less likely, amyloidosis or sarcoidosis, or (3) that the porta hepatis lymphadenopathy is causing biliary obstruction. The underlying diagnosis must explain the rash, intraabdominal lymphadenopathy, and fever. The time course does somewhat limit the extensive differential of fever and rash. After 3 weeks of illness, some of the most life‐threatening entities such as meningococcal disease, Rocky Mountain spotted fever, and toxic shock syndrome are unlikely. Concern remains for infections that are more indolent, such as mycobacteria, fungi, or spirochetes. The most striking elements of the rash are the extensive distribution, rapid progression, large number, and discreteness of the lesions, which collectively point more toward disseminated fungal (eg, histoplasmosis, as he lives in Ohio), spirochetal, rickettsial, or viral etiologies, rather than bacterial or mycobacterial entities. The absence of vesicles detracts from the diagnosis of a disseminated herpes virus such as herpes simplex or varicella. I believe that this rash is too disseminated to be caused by a common mycobacterial illness. This extent of cutaneous metastases would usually accompany a far more ill patient with an obvious primary cancer (none is seen on imaging, including the liver), and it appears too extensive to be caused by a paraneoplastic phenomenon such as Sweet's syndrome. A systemic vasculitis or another autoimmune disease remains possible, but there is minimal evidence of visceral organ involvement. All the aforementioned diseases could explain the intraabdominal lymphadenopathy, but my suspicion is highest for infection. I would biopsy and culture the skin lesions, repeat the RPR and/or send a treponemal‐specific test, place a PPD skin test, and send fungal studies (serum serologies and urine antigens) for evaluation. If the results of these noninvasive studies are unrevealing, I would consider a liver biopsy.

The patient's medications were discontinued, and a skin biopsy of the rash from his chest showed atypical lymphohistiocytic infiltrates without acute inflammatory cells and with negative Gomori methenamine silver (GMS), acid‐fast bacilli (AFB), and Fite (for Nocardia) stains. The infiltrates were predominantly T cells with a 1:1 CD4:CD8 ratio. This was read as suspicious for cytotoxic (CD8) mycosis fungoides.

I do not have reason to doubt the pathologist's impression of mycosis fungoides on histopathologic grounds, but from a clinical standpoint, I do not think mycosis fungoides is a disease that has a prolonged febrile prodrome or an explosive cutaneous onset. Rather, it is frequently preceded by nonspecific skin findings over a long period. Thinking broadly and pathophysiologically and noting that T cells are the predominant lymphocytes in skin, I wonder if they could represent a nonmalignant, immunological reaction in the skin. The stains, although not perfectly sensitive, make mycobacterial and fungal diseases less likely, although incubation of cultures is necessary.

Over the next 10 days (bringing the total duration of the patient's illness to 6 weeks), the skin lesions increased in number. In the physician's office at his next follow‐up, the patient had a temperature of 104.1F, was uncomfortable, shivering, and ill‐appearing. His blood pressure was 108/66 mm Hg, and his pulse 114 beats/min. He complained of severe shooting pains, predominantly in his pretibial regions and arms. Examination showed no other new findings, including no focal neurological findings. The results of the T‐cell rearrangement study from the skin biopsy showed evidence of a monoclonal T‐cell population. He was admitted to the hospital for further evaluation and treatment.

The extremity dysesthesias could represent a lesion of the spinal cord (including the CSF/meninges), a polyradiculopathy, or a polyneuropathy. Unfortunately, this does not add a tremendous amount of diagnostic resolution, as infection, malignancy, and autoimmune syndromes, such as vasculitis, may all involve the nervous system in these ways. In general, I associate monoclonal lymphocyte responses with hematological malignancies and polyclonal responses with the less specific inflammation that could accompany infection, autoimmunity, or solid malignancies. His age, fever, and rapid progression seem atypical for mycosis fungoides, but given the monoclonal T cells, this must now be considered. Adult T‐cell leukemia/lymphoma, with its prominent skin manifestations and its association with HLTV‐1, is an alternative T‐cell malignancy that could explain the fever, neurological symptoms, and possible visceral involvement (elevated alkaline phosphatase, which could reflect liver or bone). In cases that are diagnostic challenges, one of the highest‐yield maneuvers is to repeat the preceding evaluation, starting with the history, exam, and basic labs, and if necessary, to review or repeat the imaging or skin biopsy. Given the elevated alkaline phosphatase, disseminated rash, new neurological symptoms, and his HIV status, I remain particularly concerned about syphilis and would do further testing (accounting for the prozone phenomenon) before proceeding with the malignancy evaluation.

A lumbar puncture demonstrated clear cerebrospinal fluid, with 2 leukocytes and 195 erythrocytes/cm³, protein of 26 mg/dL, and glucose of 52 mg/dL. Bacterial and fungal cultures of the fluid were negative. The results of colonoscopy were normal. A bone marrow biopsy demonstrated ring granulomas. GMS, AFB, Fite, and Steiner (for spirochetes) stains were negative, cultures of the aspirate were negative for bacteria, and smears were negative for fungi and mycobacteria. Antibody tests for human T‐cell lymphotropic virus types I and II, Coxiella burnetii, and Bartonella henselae were negative. The dermatology consultant believed the absence of lymphadenopathy and the pruritic nature of the lesions was atypical for cytotoxic T‐cell lymphoma (CTCL). Before initiating therapy for CTCL, she suggested repeating the skin biopsy and RPR.

The repeat RPR was positive at 1:64 dilutions, and a confirmatory fluorescent treponemal antibody absorption test showed a positive result. He was prescribed intramuscular benzathine pencillin 2.4 million units weekly for 3 weeks, with almost immediate defervescence and slower resolution of his rash and shooting pains in his limbs. The repeat skin biopsy done during the hospitalization demonstrated lichenoid‐type dermatitis with interstitial and perivascular lymphohistiocytic infiltrates and granulomas. Steiner stains for spirochetes were positive. Immunohistochemical stains ruled out a lymphoproliferative process. One year later his RPR was nonreactive.

COMMENTARY

Fever of unknown origin (FUO) was first defined by Petersdorf and Beeson in 1961 as a temperature higher than 38.3C on several occasions lasting longer than 3 weeks and defying diagnosis despite 1 week of inpatient investigation.1 Dramatic changes in medical practice have rendered this definition outdated, with more recent proposals allowing thoughtful outpatient investigation to serve as a surrogate for hospitalization. Some have proposed that HIV‐associated FUO be considered a distinct entity, with the most complete North American series finding the etiology of the HIV‐associated FUO in 56 of 70 patients.2 The mean CD4+ count in this series was 58/cm³. Disseminated M. avium was the most frequently diagnosed cause, followed by P. jirovecii pneumonia, cytomegalovirus infection, disseminated histoplasmosis, and lymphoma. Of 14 patients with fever of no definable etiology, 12 eventually proved to have self‐limiting illness.

Despite numerous attempts to reduce the investigation of the patient with FUO to an algorithm, the approach must be individualized. A thorough history and careful, serial physical examinations are frequently and appropriately stressed as the foundation, followed by thoughtful selection of laboratory and imaging studies. Although FUO has a lengthy differential diagnosis, it often proves to be, as Mackowiak and Durack stress, an unusual manifestation of a common disease, rather than a typical presentation of a rare disease.3 A relatively uncommon disease in conjunction with an initially negative diagnostic test result, as was the case with this patient, may lead to a protracted diagnostic puzzle.

Syphilis is a rare cause of FUO. In 6 large studies of a total of 947 patients published over a 40‐year period, only 2 cases of syphilis (1 secondary and 1 neurosyphilis) were reported.1, 48 Syphilis as a cause of prolonged cryptic fever appears to have been seen with greater frequency in the preantibiotic era.9 In the first half of the 20th century, syphilis was known as the great imitator, with its unusual manifestations recognized and indeed expected. As a result of the dramatically lower incidence of syphilis in recent decades, these lessons have largely been forgotten, however, which may lead to diagnostic confusion when syphilis presents atypically. The manifestations of secondary syphilis are protean, including a variety of rashes, aphthous ulcers, arthralgias, pharyngitis, weight loss, fever, meningitis, ocular symptoms, cranial nerve palsies, glomerulonephritis, hepatitis, and periostitis (which afflicted this patient, who complained of severe shooting pains in his arms and shins).

After declining in the last decade of the 20th century, the rates of primary and secondary syphilis are rising in the United States.10 Oral sex is a clear risk factor for syphilis transmission, particularly for men who have sex with men.11 Because of the patient's exposure history and clinical picture, his outpatient physician considered the diagnosis of secondary syphilis early in the course of his illness. The diagnosis was not entertained further when an RPR test, highly sensitive at this stage of the disease, returned nonreactive. Likewise, when a rash subsequently appeared, the lack of palm and sole involvement dissuaded multiple clinicians from reconsidering the diagnosis of syphilis. A skin biopsy that appeared to lead in a distinctly different direction understandably confused the picture still further. Even at the time of the lumbar puncture, VDRL of the CSF was not ordered.

In retrospect, the chief confounder in the case was the false‐negative RPR test, as the discussant suspected early on. Although nontreponemal tests are generally accurate in individuals with HIV, delayed seropositivity and false‐negatives have been reported in this population.12 The false‐negative could have also been a result of the prozone phenomenon, an unusual event, occurring in fewer than 2% of cases of secondary syphilis and attributed to a mismatch between antibody and very high antigen level. The prozone reaction can be corrected for by requesting dilution of the serum prior to repeating the test. Simple lab error must be considered as well, but without access to this patient's serum from his original testing, the cause of his initial false‐negative test cannot be known with certainty.

An unusual presentation in conjunction with failure to recognize the causes of rare false‐negative testing for secondary syphilis led to a delayed diagnosis in this patient. Although syphilis and mycosis fungoides have previously been reported to mimic one another both clinically and histopathologically, the potential for secondary syphilis to be misdiagnosed in this fashion is not generally appreciated.1315 Recognition of the possibility of secondary syphilis occurred just in time to spare this patient the rash decision of treating him with cytotoxic therapy directed against CTCL.

A 45‐year‐old Vietnamese woman presented to the emergency department reporting 4 days of vomiting, diarrhea, and epigastric pain. Other than poliomyelitis as a child, she had an unremarkable medical history. A slight left‐sided limp was noted on exam. Although she was diagnosed with gastroenteritis, laboratory studies revealed her blood urea nitrogen was low, at 4 mg/dL, and her creatinine was 0.2 mg/dL, raising concern for an intercurrent abdominal malignancy. A CT scan of the abdomen and pelvis was obtained.

The CT scan did not reveal any abdominal pathology but, incidentally, showed near‐complete atrophy of the left pelvic girdle and proximal femur. Specifically noted was near‐complete fatty atrophy and involution of the left iliacus, piriformis, and gluteus muscles (Fig. 1). Atrophy also involved muscles surrounding the left proximal femur (Fig. 2). The CT and laboratory findings were attributed to childhood poliomyelitis. The patient was discharged home in good condition.

Figure 1

Figure 2

Poliovirus, a small RNA enterovirus, causes destruction of the motor neuron ganglia of the brain stem and anterior horn cells of the spine. Resultant Wallerian degeneration leads to muscle atrophy, as seen here.

A 45‐year‐old Vietnamese woman presented to the emergency department reporting 4 days of vomiting, diarrhea, and epigastric pain. Other than poliomyelitis as a child, she had an unremarkable medical history. A slight left‐sided limp was noted on exam. Although she was diagnosed with gastroenteritis, laboratory studies revealed her blood urea nitrogen was low, at 4 mg/dL, and her creatinine was 0.2 mg/dL, raising concern for an intercurrent abdominal malignancy. A CT scan of the abdomen and pelvis was obtained.

The CT scan did not reveal any abdominal pathology but, incidentally, showed near‐complete atrophy of the left pelvic girdle and proximal femur. Specifically noted was near‐complete fatty atrophy and involution of the left iliacus, piriformis, and gluteus muscles (Fig. 1). Atrophy also involved muscles surrounding the left proximal femur (Fig. 2). The CT and laboratory findings were attributed to childhood poliomyelitis. The patient was discharged home in good condition.

Figure 1

Figure 2

Poliovirus, a small RNA enterovirus, causes destruction of the motor neuron ganglia of the brain stem and anterior horn cells of the spine. Resultant Wallerian degeneration leads to muscle atrophy, as seen here.

A 45‐year‐old Vietnamese woman presented to the emergency department reporting 4 days of vomiting, diarrhea, and epigastric pain. Other than poliomyelitis as a child, she had an unremarkable medical history. A slight left‐sided limp was noted on exam. Although she was diagnosed with gastroenteritis, laboratory studies revealed her blood urea nitrogen was low, at 4 mg/dL, and her creatinine was 0.2 mg/dL, raising concern for an intercurrent abdominal malignancy. A CT scan of the abdomen and pelvis was obtained.

The CT scan did not reveal any abdominal pathology but, incidentally, showed near‐complete atrophy of the left pelvic girdle and proximal femur. Specifically noted was near‐complete fatty atrophy and involution of the left iliacus, piriformis, and gluteus muscles (Fig. 1). Atrophy also involved muscles surrounding the left proximal femur (Fig. 2). The CT and laboratory findings were attributed to childhood poliomyelitis. The patient was discharged home in good condition.

Figure 1

Figure 2

Poliovirus, a small RNA enterovirus, causes destruction of the motor neuron ganglia of the brain stem and anterior horn cells of the spine. Resultant Wallerian degeneration leads to muscle atrophy, as seen here.

There are currently limited data to guide perioperative management of antiplatelet therapy after drug‐eluting stent (DES) implantation. The clinician must balance the risk of excessive bleeding if antiplatelet agents are continued perioperatively with the risk of stent thrombosis if antiplatelet agents are discontinued for surgerya risk that may be amplified in the perioperative period because of the prothrombotic state that accompanies the stress of surgery.

Paclitaxel‐ and sirolimus‐eluting stents have supplanted bare‐metal stents as first‐line treatment for coronary stenosis because of their efficacy in preventing in‐stent restenosis by inhibiting neointimal proliferation. However, the antiproliferative effects of DESs may also delay endothelialization, rendering them vulnerable to stent thrombosis when antiplatelet therapy is prematurely discontinued.15 Some patients with DESs may be vulnerable to stent thrombosis when antiplatelet therapy is discontinued even after a year or more of treatment.6 Although stent thrombosis is uncommon, it is deadly, with a mortality rate approaching 50%.1 Generally, antiplatelet therapy is discontinued prior to surgery. This presents a clinical dilemma for patients with DES because guidelines recommend lifelong aspirin therapy and at least 36 months of clopidogrel for patients who have undergone DES placement.79

In the bare‐metal stent era, studies demonstrated an alarming risk of stent thrombosis in the setting of noncardiac surgery within 26 weeks of stent placement.10, 11 However, the appropriate interval before elective noncardiac surgery following DES placement has not been defined and may be longer. Case reports and case series have highlighted this risk12 and have even suggested that a DES may be susceptible to stent thrombosis as long as a year after its placement.6 More recently, pooled data from controlled trials have suggested that although the overall rate of DES thrombosis may not be consistently higher than that of bare‐metal stents, the risk appears to persist far longer (probably from delayed endothelialization of the target vessel) and may be more pronounced following discontinuation of antiplatelet agents.9, 1316 This has led to recent recommendations to continue dual antiplatelet therapy (with aspirin and clopidogrel) for at least a year following DES placement and possibly indefinitely, provided that the therapy is tolerated.9 Whether this risk is accentuated in the perioperative setting independent of discontinuation of antiplatelet therapy remains unknown. In 1 registry, the strongest predictor of DES thrombosis was premature discontinuation of antiplatelet therapy (hazard ratio 90, 95% confidence interval 30270, P < .001), and noncardiac surgery was the most frequent reason for discontinuation of antiplatelet therapy.1 However, the actual incidence of stent thrombosis in patients undergoing surgery was unavailable because the denominator was unknown (ie, number of patients with stents who underwent surgery). Although it is certainly plausible that the prothrombotic and proinflammatory postoperative state augments the risk of stent thrombosis independent of discontinuation of antiplatelet therapy alone, this remains unproven.

At the time of the present study, protocol‐based clinical practice at the Cleveland Clinic Foundation's Internal Medicine Preoperative Assessment Consultation and Treatment (IMPACT) Center included routine discontinuation of all antiplatelet agents (including aspirin and clopidogrel) at least 7 days prior to noncardiac surgery, including in patients with coronary stents. Exceptions to this policy were generally made only for very minor procedures. The purpose of this study was to systematically quantify the risk of adverse cardiovascular events in patients who had DES placement and subsequently underwent elective or semielective noncardiac surgery, most of whom had discontinued all antiplatelet agents at least 7 days before surgery.

Methods

We identified all patients who had DES placement at the Cleveland Clinic who subsequently underwent preoperative evaluation for noncardiac surgery at the IMPACT Center between July 2003 and July 2005. About half the patients undergoing surgery at the Cleveland Clinic were seen in the IMPACT Center prior to surgery during the study period. Preoperative evaluation at the IMPACT Center included a standardized assessment by a hospitalist with expertise in preoperative medicine. Clinical data for each patient were contemporaneously entered into an electronic medical record. Written preoperative medication instructions were provided to each patient and documented in the electronic record, indicating specific instructions to discontinue any antiplatelet agents 710 days preoperatively.

The IMPACT Center database was crosslinked to the Cleveland Clinic Foundation Heart Center Database, which contains records of all patients who have undergone coronary stenting at the Cleveland Clinic. Computerized and written medical records of all patients in both databases were reviewed using a standardized data collection instrument. All medical data generated up to 30 days postoperatively at the Cleveland Clinic were reviewed. Social Security numbers were linked with the Social Security Death Index to verify that no patients died within 30 days of surgery.

Predefined outcomes included catheterization‐confirmed DES thrombosis, any myocardial infarction, and major bleeding within 30 days of the surgical procedure. Myocardial infarction was defined as elevation of troponin T to more than twice the upper limit of normal (0.2 mg/mL) with or without associated electrocardiographic changes or symptoms. This biochemically based definition was used with the understanding that cardiac enzyme tests are consistently ordered for patients at the Cleveland Clinic with suspected coronary events and that postoperative myocardial infarction may be atypical in presentation (eg, delirium or hypotension without chest pain). Stent thrombosis was considered present if confirmed by catheterization or autopsy and considered possible if a patient suffered from a myocardial infarction but did not have a definitive diagnostic procedure performed. DES thrombosis was considered absent if a patient underwent postoperative catheterization and the DES appeared patent. Major bleeding was defined as any bleeding requiring unplanned reoperation or bleeding in a critical location (intracranial or retroperitoneal). Invasiveness of surgery was defined prospectively according to a Cleveland Clinic bleeding classification scheme based on that of Pasternak17, 18:

• Category 1. Minimal risk to patient; little or no anticipated blood loss (eg, breast biopsy, cystoscopy).

• Category 2. Mild risk to patient; minimal to moderately invasive procedure; estimated blood loss < 500 cc (eg, laparoscopy, arthroscopy, hernia repair).

• Category 3. Moderate risk to patient and moderate to significantly invasive; blood loss potential 5001000 cc (eg, laminectomy, total hip or knee replacement).

• Category 4. Major risk to patient; highly invasive procedure; anticipated blood loss > 1500 cc (eg, major spinal reconstruction, major reconstruction of GI tract, major vascular repair without intensive care unit stay).

• Category 5. Critical risk to patient; highly invasive procedure; anticipated blood loss > 1500 cc with anticipated postoperative intensive care unit stay (eg, cardiac procedure, major vascular repair with anticipated intensive care unit stay).

Statistical analyses were descriptive. We determined the rate of adverse outcomes with 95% confidence intervals (CIs) in the entire patient cohort and among prespecified patient subsets, based on timing of discontinuation of antiplatelet therapy. Predefined subsets included those who had clopidogrel and aspirin discontinued less than 3 months and less than 6 months following DES implantation. The ² test was used to test the hypothesis that discontinuation of antiplatelet therapy was a function of the type of surgery or timing of stent placement.

The study was approved by the Cleveland Clinic Foundation's institutional review board. The requirement for informed consent was waived.

RESULTS

In total, 114 patients were evaluated in the IMPACT Center following DES placement. Baseline patient characteristics are shown in Table 1. The median age was 71 years (interquartile range 6476 years), and 66% were male. Patients had a moderate degree of comorbidity: 41% had diabetes, 12% had an ejection fraction < 45%, 34% had undergone coronary bypass, 17% had atrial fibrillation or flutter, and 20% had chronic renal insufficiency (creatinine 2.0 or end‐stage renal disease). Most patients received ‐adrenergic blockers (97%), statins (95%), and either angiotensin‐converting enzyme (ACE) inhibitors or angiotensin receptor blockers (77%) preoperatively. Patients underwent a variety of surgeries (Table 1).

Patients had received both paclitaxel and sirolimus stents (28% and 73% of patients, respectively); 33% of patients had had more than 1 DES (Table 2). Most patients underwent surgery within 1 year of stent placement (77%), but only 40% had surgery within 180 days of stenting and only 13% within 90 days of stenting. Most patients (77%) had antiplatelet therapy completely discontinued a median of 10 days before surgery and remained off antiplatelet therapy for a median of 14 days total. Ten of the 15 patients (67%) who underwent surgery within 90 days of stenting had all antiplatelet agents discontinued preoperatively, 24 of the 30 patients (80%) who had surgery between 91 and 180 days after stenting had antiplatelet therapy completely discontinued, and 54 of the 69 patients (78%) who had surgery more than 180 days after stenting had antiplatelet therapy completely discontinued. There was no significant relationship between timing of stent placement relative to surgery (<90, 91180, or >180 days) and decision about whether to discontinue antiplatelet therapy (P = .59). However, invasiveness of the surgery was associated with antiplatelet management: 85% of those who continued antiplatelet therapy (aspirin or aspirin and clopidogrel) during the perioperative period were patients who underwent minimally invasive surgery (P < .0001).

The outcome events are presented in Table 3. Two patients (1.8%, 95% CI 0.5%6.2%) suffered a non‐ST‐elevation myocardial infarction (NSTEMI) postoperatively, and another patient (0.9%, 95% CI 0.2%4.8%) developed major bleeding, a retroperitoneal hemorrhage following kidney transplantation. This patient had been taking both aspirin and clopidogrel until 7 days prior to surgery and began to hemorrhage the day after surgery; antiplatelet agents were resumed 12 days postoperatively. No patients died (0%, 95% CI 0%3.3%). One of the 2 patients who suffered an MI was a 72‐year‐old man who had had placement of a single sirolimus‐eluting stent in the posterior descending artery 284 days prior to elective hip arthroplasty. He had no history of myocardial infarction but had undergone coronary bypass surgery 4 years earlier. Echocardiography showed he had aortic stenosis, with a calculated valve area of 0.9 cm². He had a baseline left ventricular ejection fraction of 45%. His preoperative cardiac medications included lovastatin, lisinopril, and atenolol; he discontinued both aspirin and clopidogrel 7 days before the surgery. His NSTEMI occurred on the day of his operation, presenting with hypotension and anterolateral ST depressions. His troponin T peaked at 0.48 mg/mL, with a peak creatinine kinase of 795 U/L (MB fraction 6%). His left ventricular ejection fraction was 45% on postoperative day 2 (unchanged from baseline). He was discharged on postoperative day 8 and returned for catheterization 3 weeks later, at which time he was found to have a 70% ostial lesion in a saphenous vein graft to an obtuse marginal, which was stented. The previously placed DES was widely patent. The other patient who suffered a postoperative NSTEMI was a 68‐year‐old man with a history of carotid artery stenting and renal artery stenosis who had undergone placement of 3 sirolimus‐eluting stents in the right coronary artery 50 days prior to cervical laminectomy. He had had elective placement of the stents following a positive pharmacologic stress test. He was taking 50 mg of atenolol daily and had been taking aspirin and clopidogrel until 17 days before surgery. On postoperative day 3 he developed dyspnea, and leads V4 and V5 showed ST depressions. His troponin T peaked at 1.24 mg/mL, with a peak creatinine kinase of 879 U/L (MB fraction 6%). The patient underwent left‐heart catheterization on hospital day 10. All 3 DESs were widely patent. His left ventricular ejection fraction was estimated at 65%. He was discharged on postoperative day 15. Because neither of the patients who had a postoperative NSTEMI showed evidence of stent thrombosis on catheterization, the overall rate of stent thrombosis was 0% (95% CI 0%3.3%).

DISCUSSION

Although 2 patients in our study cohort suffered a postoperative myocardial infarction and underwent postoperative catheterization, neither was found to have stent thrombosis, and the MIs of both patients were NSTEMIs with modest cardiac enzyme elevations only. No patients died. A rate of myocardial infarction of less than 2% is well within that expected for patients with established coronary disease undergoing noncardiac surgery.19 That most of our patients discontinued both aspirin and clopidogrel and did not receive antiplatelet agents for a median of 14 days suggests that transient termination of antiplatelet agents in the perioperative setting is not associated with high morbidity or mortality in patients with DES, even when patients have had their stents implanted in the preceding 36 months.

Our study builds on the limited data on this topic. One small case series examined outcomes in 38 patients who had had DES placement and subsequently underwent noncardiac surgery a median of 297 days after stenting.20 None of the patients in this series suffered from stent thrombosis or myocardial infarction, but most underwent surgery without discontinuing aspirin, and 41% underwent surgery without discontinuing clopidogrel. Another recent study demonstrated a high rate of adverse cardiovascular events in patients with coronary stents who underwent noncardiac surgery up to a year after stenting, but the authors of this study did not differentiate between drug‐eluting and bare‐metal stents, and all patients were continued on antiplatelet agents and received parenteral antithrombotic treatment.21

The major strength of our study was its systematic approach. Using a computerized and comprehensive search strategy, we identified all patients who had undergone DES placement at the Cleveland Clinic who subsequently had a preoperative evaluation at the IMPACT Center. Therefore, we are confident that the number of patients in our cohort truly reflects a well‐defined at‐risk population, allowing for an accurate calculation of event rates. This approach contrasts sharply with prior case reports and case series, in which the number of patients at risk was unknown. Nevertheless, these previous reports demonstrate that DES thromboses do occur and can be devastating, so even a small risk of DES thrombosis should be taken seriously. The upper bound of the 95% confidence interval of our estimate of the rate of DES thrombosis was 3.3%, so it is entirely plausible that sampling error contributed to the low rate of thrombosis that we observed.

One major limitation of our study is its sample size. Although our cohort was more than 3 times larger than the only other published cohort of DES patients undergoing noncardiac surgery,20 we had only limited precision to quantify the risk of DES thrombosis. This limitation is particularly relevant for patients who have undergone stent implantation within 36 months of surgery, as they are the patients most likely to have incomplete reendothelialization of the stented artery. We believe that when possible it remains prudent to delay noncardiac surgery for at least 36 months and perhaps up to 12 months following DES implantation, in keeping with recent guidelines.7, 8 However, for patients with conditions such as cancer whose surgery is semielective or patients with nonsurgical bleeding problems (such as gastrointestinal bleeding), our study provides at least some reassurance that short‐term discontinuation of antiplatelet agents may not be as dangerous as some authors have suggested,1 even within 36 months of DES placement. Another important limitation of our study is potential referral bias. At the Cleveland Clinic, most patients undergoing vascular and thoracic procedures are not evaluated at the IMPACT Center. Similarly, some of the patients with severe cardiovascular disease may also have bypassed the IMPACT Center and gone to a cardiologist for preoperative evaluation. As such, we believe our findings should not be generalized to high‐risk cardiac patients or to those undergoing high‐risk procedures.

A noteworthy distinction between our cohort and the cohort reported by Compton and colleagues is that in the perioperative period, most of our patients underwent complete discontinuation of antiplatelet therapy and remained off both aspirin and clopidogrel for an average of 2 weeks, whereas most patients in the other cohort were continued on antiplatelet therapy.20 This highlights the continued controversy surrounding management of antiplatelet therapy in perioperative patients with established coronary disease, who are at substantial risk for both bleeding and myocardial infarction because of the surgery.22 Our data offer little guidance on the optimal management of antiplatelet agents perioperatively because the incidence of both bleeding and thrombosis was low and whether or not patients were continued on antiplatelet agents was not random. We advocate individualized management strategies of perioperative patients with DES. Patients undergoing procedures that carry a high risk of outcome‐affecting bleeding (such as brain surgery) should probably have their antiplatelet agents discontinued preoperatively, whereas those undergoing minor surgery may have their antiplatelet agents continued, provided the surgeon and the anesthesiologist are in agreement with this approach. The timing of DES placement should also be factored into this decision because recently placed stents carry a higher risk of thrombosis.

In summary, our findings clarify the risks of stent thrombosis and postoperative myocardial infarction in clinically stable patients with DES who undergo low‐ and intermediate‐risk noncardiac surgery. Because it is unlikely to ever be ethically appropriate or logistically feasible to conduct a randomized study of patients with DES having early versus delayed noncardiac surgery, observational cohorts will have to suffice. Additional similar studies will help to validate (or refute) our findings and to more precisely quantify the risk of adverse cardiac events when patients with DES undergo surgery, which is real, feared, and potentially catastrophic but may be overestimated.

There are currently limited data to guide perioperative management of antiplatelet therapy after drug‐eluting stent (DES) implantation. The clinician must balance the risk of excessive bleeding if antiplatelet agents are continued perioperatively with the risk of stent thrombosis if antiplatelet agents are discontinued for surgerya risk that may be amplified in the perioperative period because of the prothrombotic state that accompanies the stress of surgery.

Paclitaxel‐ and sirolimus‐eluting stents have supplanted bare‐metal stents as first‐line treatment for coronary stenosis because of their efficacy in preventing in‐stent restenosis by inhibiting neointimal proliferation. However, the antiproliferative effects of DESs may also delay endothelialization, rendering them vulnerable to stent thrombosis when antiplatelet therapy is prematurely discontinued.15 Some patients with DESs may be vulnerable to stent thrombosis when antiplatelet therapy is discontinued even after a year or more of treatment.6 Although stent thrombosis is uncommon, it is deadly, with a mortality rate approaching 50%.1 Generally, antiplatelet therapy is discontinued prior to surgery. This presents a clinical dilemma for patients with DES because guidelines recommend lifelong aspirin therapy and at least 36 months of clopidogrel for patients who have undergone DES placement.79

In the bare‐metal stent era, studies demonstrated an alarming risk of stent thrombosis in the setting of noncardiac surgery within 26 weeks of stent placement.10, 11 However, the appropriate interval before elective noncardiac surgery following DES placement has not been defined and may be longer. Case reports and case series have highlighted this risk12 and have even suggested that a DES may be susceptible to stent thrombosis as long as a year after its placement.6 More recently, pooled data from controlled trials have suggested that although the overall rate of DES thrombosis may not be consistently higher than that of bare‐metal stents, the risk appears to persist far longer (probably from delayed endothelialization of the target vessel) and may be more pronounced following discontinuation of antiplatelet agents.9, 1316 This has led to recent recommendations to continue dual antiplatelet therapy (with aspirin and clopidogrel) for at least a year following DES placement and possibly indefinitely, provided that the therapy is tolerated.9 Whether this risk is accentuated in the perioperative setting independent of discontinuation of antiplatelet therapy remains unknown. In 1 registry, the strongest predictor of DES thrombosis was premature discontinuation of antiplatelet therapy (hazard ratio 90, 95% confidence interval 30270, P < .001), and noncardiac surgery was the most frequent reason for discontinuation of antiplatelet therapy.1 However, the actual incidence of stent thrombosis in patients undergoing surgery was unavailable because the denominator was unknown (ie, number of patients with stents who underwent surgery). Although it is certainly plausible that the prothrombotic and proinflammatory postoperative state augments the risk of stent thrombosis independent of discontinuation of antiplatelet therapy alone, this remains unproven.

At the time of the present study, protocol‐based clinical practice at the Cleveland Clinic Foundation's Internal Medicine Preoperative Assessment Consultation and Treatment (IMPACT) Center included routine discontinuation of all antiplatelet agents (including aspirin and clopidogrel) at least 7 days prior to noncardiac surgery, including in patients with coronary stents. Exceptions to this policy were generally made only for very minor procedures. The purpose of this study was to systematically quantify the risk of adverse cardiovascular events in patients who had DES placement and subsequently underwent elective or semielective noncardiac surgery, most of whom had discontinued all antiplatelet agents at least 7 days before surgery.

Methods

We identified all patients who had DES placement at the Cleveland Clinic who subsequently underwent preoperative evaluation for noncardiac surgery at the IMPACT Center between July 2003 and July 2005. About half the patients undergoing surgery at the Cleveland Clinic were seen in the IMPACT Center prior to surgery during the study period. Preoperative evaluation at the IMPACT Center included a standardized assessment by a hospitalist with expertise in preoperative medicine. Clinical data for each patient were contemporaneously entered into an electronic medical record. Written preoperative medication instructions were provided to each patient and documented in the electronic record, indicating specific instructions to discontinue any antiplatelet agents 710 days preoperatively.

The IMPACT Center database was crosslinked to the Cleveland Clinic Foundation Heart Center Database, which contains records of all patients who have undergone coronary stenting at the Cleveland Clinic. Computerized and written medical records of all patients in both databases were reviewed using a standardized data collection instrument. All medical data generated up to 30 days postoperatively at the Cleveland Clinic were reviewed. Social Security numbers were linked with the Social Security Death Index to verify that no patients died within 30 days of surgery.

Predefined outcomes included catheterization‐confirmed DES thrombosis, any myocardial infarction, and major bleeding within 30 days of the surgical procedure. Myocardial infarction was defined as elevation of troponin T to more than twice the upper limit of normal (0.2 mg/mL) with or without associated electrocardiographic changes or symptoms. This biochemically based definition was used with the understanding that cardiac enzyme tests are consistently ordered for patients at the Cleveland Clinic with suspected coronary events and that postoperative myocardial infarction may be atypical in presentation (eg, delirium or hypotension without chest pain). Stent thrombosis was considered present if confirmed by catheterization or autopsy and considered possible if a patient suffered from a myocardial infarction but did not have a definitive diagnostic procedure performed. DES thrombosis was considered absent if a patient underwent postoperative catheterization and the DES appeared patent. Major bleeding was defined as any bleeding requiring unplanned reoperation or bleeding in a critical location (intracranial or retroperitoneal). Invasiveness of surgery was defined prospectively according to a Cleveland Clinic bleeding classification scheme based on that of Pasternak17, 18:

• Category 1. Minimal risk to patient; little or no anticipated blood loss (eg, breast biopsy, cystoscopy).

• Category 2. Mild risk to patient; minimal to moderately invasive procedure; estimated blood loss < 500 cc (eg, laparoscopy, arthroscopy, hernia repair).

• Category 3. Moderate risk to patient and moderate to significantly invasive; blood loss potential 5001000 cc (eg, laminectomy, total hip or knee replacement).

• Category 4. Major risk to patient; highly invasive procedure; anticipated blood loss > 1500 cc (eg, major spinal reconstruction, major reconstruction of GI tract, major vascular repair without intensive care unit stay).

• Category 5. Critical risk to patient; highly invasive procedure; anticipated blood loss > 1500 cc with anticipated postoperative intensive care unit stay (eg, cardiac procedure, major vascular repair with anticipated intensive care unit stay).

Statistical analyses were descriptive. We determined the rate of adverse outcomes with 95% confidence intervals (CIs) in the entire patient cohort and among prespecified patient subsets, based on timing of discontinuation of antiplatelet therapy. Predefined subsets included those who had clopidogrel and aspirin discontinued less than 3 months and less than 6 months following DES implantation. The ² test was used to test the hypothesis that discontinuation of antiplatelet therapy was a function of the type of surgery or timing of stent placement.

The study was approved by the Cleveland Clinic Foundation's institutional review board. The requirement for informed consent was waived.

RESULTS

In total, 114 patients were evaluated in the IMPACT Center following DES placement. Baseline patient characteristics are shown in Table 1. The median age was 71 years (interquartile range 6476 years), and 66% were male. Patients had a moderate degree of comorbidity: 41% had diabetes, 12% had an ejection fraction < 45%, 34% had undergone coronary bypass, 17% had atrial fibrillation or flutter, and 20% had chronic renal insufficiency (creatinine 2.0 or end‐stage renal disease). Most patients received ‐adrenergic blockers (97%), statins (95%), and either angiotensin‐converting enzyme (ACE) inhibitors or angiotensin receptor blockers (77%) preoperatively. Patients underwent a variety of surgeries (Table 1).

Patients had received both paclitaxel and sirolimus stents (28% and 73% of patients, respectively); 33% of patients had had more than 1 DES (Table 2). Most patients underwent surgery within 1 year of stent placement (77%), but only 40% had surgery within 180 days of stenting and only 13% within 90 days of stenting. Most patients (77%) had antiplatelet therapy completely discontinued a median of 10 days before surgery and remained off antiplatelet therapy for a median of 14 days total. Ten of the 15 patients (67%) who underwent surgery within 90 days of stenting had all antiplatelet agents discontinued preoperatively, 24 of the 30 patients (80%) who had surgery between 91 and 180 days after stenting had antiplatelet therapy completely discontinued, and 54 of the 69 patients (78%) who had surgery more than 180 days after stenting had antiplatelet therapy completely discontinued. There was no significant relationship between timing of stent placement relative to surgery (<90, 91180, or >180 days) and decision about whether to discontinue antiplatelet therapy (P = .59). However, invasiveness of the surgery was associated with antiplatelet management: 85% of those who continued antiplatelet therapy (aspirin or aspirin and clopidogrel) during the perioperative period were patients who underwent minimally invasive surgery (P < .0001).

The outcome events are presented in Table 3. Two patients (1.8%, 95% CI 0.5%6.2%) suffered a non‐ST‐elevation myocardial infarction (NSTEMI) postoperatively, and another patient (0.9%, 95% CI 0.2%4.8%) developed major bleeding, a retroperitoneal hemorrhage following kidney transplantation. This patient had been taking both aspirin and clopidogrel until 7 days prior to surgery and began to hemorrhage the day after surgery; antiplatelet agents were resumed 12 days postoperatively. No patients died (0%, 95% CI 0%3.3%). One of the 2 patients who suffered an MI was a 72‐year‐old man who had had placement of a single sirolimus‐eluting stent in the posterior descending artery 284 days prior to elective hip arthroplasty. He had no history of myocardial infarction but had undergone coronary bypass surgery 4 years earlier. Echocardiography showed he had aortic stenosis, with a calculated valve area of 0.9 cm². He had a baseline left ventricular ejection fraction of 45%. His preoperative cardiac medications included lovastatin, lisinopril, and atenolol; he discontinued both aspirin and clopidogrel 7 days before the surgery. His NSTEMI occurred on the day of his operation, presenting with hypotension and anterolateral ST depressions. His troponin T peaked at 0.48 mg/mL, with a peak creatinine kinase of 795 U/L (MB fraction 6%). His left ventricular ejection fraction was 45% on postoperative day 2 (unchanged from baseline). He was discharged on postoperative day 8 and returned for catheterization 3 weeks later, at which time he was found to have a 70% ostial lesion in a saphenous vein graft to an obtuse marginal, which was stented. The previously placed DES was widely patent. The other patient who suffered a postoperative NSTEMI was a 68‐year‐old man with a history of carotid artery stenting and renal artery stenosis who had undergone placement of 3 sirolimus‐eluting stents in the right coronary artery 50 days prior to cervical laminectomy. He had had elective placement of the stents following a positive pharmacologic stress test. He was taking 50 mg of atenolol daily and had been taking aspirin and clopidogrel until 17 days before surgery. On postoperative day 3 he developed dyspnea, and leads V4 and V5 showed ST depressions. His troponin T peaked at 1.24 mg/mL, with a peak creatinine kinase of 879 U/L (MB fraction 6%). The patient underwent left‐heart catheterization on hospital day 10. All 3 DESs were widely patent. His left ventricular ejection fraction was estimated at 65%. He was discharged on postoperative day 15. Because neither of the patients who had a postoperative NSTEMI showed evidence of stent thrombosis on catheterization, the overall rate of stent thrombosis was 0% (95% CI 0%3.3%).

DISCUSSION

Although 2 patients in our study cohort suffered a postoperative myocardial infarction and underwent postoperative catheterization, neither was found to have stent thrombosis, and the MIs of both patients were NSTEMIs with modest cardiac enzyme elevations only. No patients died. A rate of myocardial infarction of less than 2% is well within that expected for patients with established coronary disease undergoing noncardiac surgery.19 That most of our patients discontinued both aspirin and clopidogrel and did not receive antiplatelet agents for a median of 14 days suggests that transient termination of antiplatelet agents in the perioperative setting is not associated with high morbidity or mortality in patients with DES, even when patients have had their stents implanted in the preceding 36 months.

Our study builds on the limited data on this topic. One small case series examined outcomes in 38 patients who had had DES placement and subsequently underwent noncardiac surgery a median of 297 days after stenting.20 None of the patients in this series suffered from stent thrombosis or myocardial infarction, but most underwent surgery without discontinuing aspirin, and 41% underwent surgery without discontinuing clopidogrel. Another recent study demonstrated a high rate of adverse cardiovascular events in patients with coronary stents who underwent noncardiac surgery up to a year after stenting, but the authors of this study did not differentiate between drug‐eluting and bare‐metal stents, and all patients were continued on antiplatelet agents and received parenteral antithrombotic treatment.21

The major strength of our study was its systematic approach. Using a computerized and comprehensive search strategy, we identified all patients who had undergone DES placement at the Cleveland Clinic who subsequently had a preoperative evaluation at the IMPACT Center. Therefore, we are confident that the number of patients in our cohort truly reflects a well‐defined at‐risk population, allowing for an accurate calculation of event rates. This approach contrasts sharply with prior case reports and case series, in which the number of patients at risk was unknown. Nevertheless, these previous reports demonstrate that DES thromboses do occur and can be devastating, so even a small risk of DES thrombosis should be taken seriously. The upper bound of the 95% confidence interval of our estimate of the rate of DES thrombosis was 3.3%, so it is entirely plausible that sampling error contributed to the low rate of thrombosis that we observed.

One major limitation of our study is its sample size. Although our cohort was more than 3 times larger than the only other published cohort of DES patients undergoing noncardiac surgery,20 we had only limited precision to quantify the risk of DES thrombosis. This limitation is particularly relevant for patients who have undergone stent implantation within 36 months of surgery, as they are the patients most likely to have incomplete reendothelialization of the stented artery. We believe that when possible it remains prudent to delay noncardiac surgery for at least 36 months and perhaps up to 12 months following DES implantation, in keeping with recent guidelines.7, 8 However, for patients with conditions such as cancer whose surgery is semielective or patients with nonsurgical bleeding problems (such as gastrointestinal bleeding), our study provides at least some reassurance that short‐term discontinuation of antiplatelet agents may not be as dangerous as some authors have suggested,1 even within 36 months of DES placement. Another important limitation of our study is potential referral bias. At the Cleveland Clinic, most patients undergoing vascular and thoracic procedures are not evaluated at the IMPACT Center. Similarly, some of the patients with severe cardiovascular disease may also have bypassed the IMPACT Center and gone to a cardiologist for preoperative evaluation. As such, we believe our findings should not be generalized to high‐risk cardiac patients or to those undergoing high‐risk procedures.

A noteworthy distinction between our cohort and the cohort reported by Compton and colleagues is that in the perioperative period, most of our patients underwent complete discontinuation of antiplatelet therapy and remained off both aspirin and clopidogrel for an average of 2 weeks, whereas most patients in the other cohort were continued on antiplatelet therapy.20 This highlights the continued controversy surrounding management of antiplatelet therapy in perioperative patients with established coronary disease, who are at substantial risk for both bleeding and myocardial infarction because of the surgery.22 Our data offer little guidance on the optimal management of antiplatelet agents perioperatively because the incidence of both bleeding and thrombosis was low and whether or not patients were continued on antiplatelet agents was not random. We advocate individualized management strategies of perioperative patients with DES. Patients undergoing procedures that carry a high risk of outcome‐affecting bleeding (such as brain surgery) should probably have their antiplatelet agents discontinued preoperatively, whereas those undergoing minor surgery may have their antiplatelet agents continued, provided the surgeon and the anesthesiologist are in agreement with this approach. The timing of DES placement should also be factored into this decision because recently placed stents carry a higher risk of thrombosis.

In summary, our findings clarify the risks of stent thrombosis and postoperative myocardial infarction in clinically stable patients with DES who undergo low‐ and intermediate‐risk noncardiac surgery. Because it is unlikely to ever be ethically appropriate or logistically feasible to conduct a randomized study of patients with DES having early versus delayed noncardiac surgery, observational cohorts will have to suffice. Additional similar studies will help to validate (or refute) our findings and to more precisely quantify the risk of adverse cardiac events when patients with DES undergo surgery, which is real, feared, and potentially catastrophic but may be overestimated.

There are currently limited data to guide perioperative management of antiplatelet therapy after drug‐eluting stent (DES) implantation. The clinician must balance the risk of excessive bleeding if antiplatelet agents are continued perioperatively with the risk of stent thrombosis if antiplatelet agents are discontinued for surgerya risk that may be amplified in the perioperative period because of the prothrombotic state that accompanies the stress of surgery.

Paclitaxel‐ and sirolimus‐eluting stents have supplanted bare‐metal stents as first‐line treatment for coronary stenosis because of their efficacy in preventing in‐stent restenosis by inhibiting neointimal proliferation. However, the antiproliferative effects of DESs may also delay endothelialization, rendering them vulnerable to stent thrombosis when antiplatelet therapy is prematurely discontinued.15 Some patients with DESs may be vulnerable to stent thrombosis when antiplatelet therapy is discontinued even after a year or more of treatment.6 Although stent thrombosis is uncommon, it is deadly, with a mortality rate approaching 50%.1 Generally, antiplatelet therapy is discontinued prior to surgery. This presents a clinical dilemma for patients with DES because guidelines recommend lifelong aspirin therapy and at least 36 months of clopidogrel for patients who have undergone DES placement.79

In the bare‐metal stent era, studies demonstrated an alarming risk of stent thrombosis in the setting of noncardiac surgery within 26 weeks of stent placement.10, 11 However, the appropriate interval before elective noncardiac surgery following DES placement has not been defined and may be longer. Case reports and case series have highlighted this risk12 and have even suggested that a DES may be susceptible to stent thrombosis as long as a year after its placement.6 More recently, pooled data from controlled trials have suggested that although the overall rate of DES thrombosis may not be consistently higher than that of bare‐metal stents, the risk appears to persist far longer (probably from delayed endothelialization of the target vessel) and may be more pronounced following discontinuation of antiplatelet agents.9, 1316 This has led to recent recommendations to continue dual antiplatelet therapy (with aspirin and clopidogrel) for at least a year following DES placement and possibly indefinitely, provided that the therapy is tolerated.9 Whether this risk is accentuated in the perioperative setting independent of discontinuation of antiplatelet therapy remains unknown. In 1 registry, the strongest predictor of DES thrombosis was premature discontinuation of antiplatelet therapy (hazard ratio 90, 95% confidence interval 30270, P < .001), and noncardiac surgery was the most frequent reason for discontinuation of antiplatelet therapy.1 However, the actual incidence of stent thrombosis in patients undergoing surgery was unavailable because the denominator was unknown (ie, number of patients with stents who underwent surgery). Although it is certainly plausible that the prothrombotic and proinflammatory postoperative state augments the risk of stent thrombosis independent of discontinuation of antiplatelet therapy alone, this remains unproven.

At the time of the present study, protocol‐based clinical practice at the Cleveland Clinic Foundation's Internal Medicine Preoperative Assessment Consultation and Treatment (IMPACT) Center included routine discontinuation of all antiplatelet agents (including aspirin and clopidogrel) at least 7 days prior to noncardiac surgery, including in patients with coronary stents. Exceptions to this policy were generally made only for very minor procedures. The purpose of this study was to systematically quantify the risk of adverse cardiovascular events in patients who had DES placement and subsequently underwent elective or semielective noncardiac surgery, most of whom had discontinued all antiplatelet agents at least 7 days before surgery.

Methods

We identified all patients who had DES placement at the Cleveland Clinic who subsequently underwent preoperative evaluation for noncardiac surgery at the IMPACT Center between July 2003 and July 2005. About half the patients undergoing surgery at the Cleveland Clinic were seen in the IMPACT Center prior to surgery during the study period. Preoperative evaluation at the IMPACT Center included a standardized assessment by a hospitalist with expertise in preoperative medicine. Clinical data for each patient were contemporaneously entered into an electronic medical record. Written preoperative medication instructions were provided to each patient and documented in the electronic record, indicating specific instructions to discontinue any antiplatelet agents 710 days preoperatively.

The IMPACT Center database was crosslinked to the Cleveland Clinic Foundation Heart Center Database, which contains records of all patients who have undergone coronary stenting at the Cleveland Clinic. Computerized and written medical records of all patients in both databases were reviewed using a standardized data collection instrument. All medical data generated up to 30 days postoperatively at the Cleveland Clinic were reviewed. Social Security numbers were linked with the Social Security Death Index to verify that no patients died within 30 days of surgery.

Predefined outcomes included catheterization‐confirmed DES thrombosis, any myocardial infarction, and major bleeding within 30 days of the surgical procedure. Myocardial infarction was defined as elevation of troponin T to more than twice the upper limit of normal (0.2 mg/mL) with or without associated electrocardiographic changes or symptoms. This biochemically based definition was used with the understanding that cardiac enzyme tests are consistently ordered for patients at the Cleveland Clinic with suspected coronary events and that postoperative myocardial infarction may be atypical in presentation (eg, delirium or hypotension without chest pain). Stent thrombosis was considered present if confirmed by catheterization or autopsy and considered possible if a patient suffered from a myocardial infarction but did not have a definitive diagnostic procedure performed. DES thrombosis was considered absent if a patient underwent postoperative catheterization and the DES appeared patent. Major bleeding was defined as any bleeding requiring unplanned reoperation or bleeding in a critical location (intracranial or retroperitoneal). Invasiveness of surgery was defined prospectively according to a Cleveland Clinic bleeding classification scheme based on that of Pasternak17, 18:

• Category 1. Minimal risk to patient; little or no anticipated blood loss (eg, breast biopsy, cystoscopy).

• Category 2. Mild risk to patient; minimal to moderately invasive procedure; estimated blood loss < 500 cc (eg, laparoscopy, arthroscopy, hernia repair).

• Category 3. Moderate risk to patient and moderate to significantly invasive; blood loss potential 5001000 cc (eg, laminectomy, total hip or knee replacement).

• Category 4. Major risk to patient; highly invasive procedure; anticipated blood loss > 1500 cc (eg, major spinal reconstruction, major reconstruction of GI tract, major vascular repair without intensive care unit stay).

• Category 5. Critical risk to patient; highly invasive procedure; anticipated blood loss > 1500 cc with anticipated postoperative intensive care unit stay (eg, cardiac procedure, major vascular repair with anticipated intensive care unit stay).

Statistical analyses were descriptive. We determined the rate of adverse outcomes with 95% confidence intervals (CIs) in the entire patient cohort and among prespecified patient subsets, based on timing of discontinuation of antiplatelet therapy. Predefined subsets included those who had clopidogrel and aspirin discontinued less than 3 months and less than 6 months following DES implantation. The ² test was used to test the hypothesis that discontinuation of antiplatelet therapy was a function of the type of surgery or timing of stent placement.

The study was approved by the Cleveland Clinic Foundation's institutional review board. The requirement for informed consent was waived.

RESULTS

In total, 114 patients were evaluated in the IMPACT Center following DES placement. Baseline patient characteristics are shown in Table 1. The median age was 71 years (interquartile range 6476 years), and 66% were male. Patients had a moderate degree of comorbidity: 41% had diabetes, 12% had an ejection fraction < 45%, 34% had undergone coronary bypass, 17% had atrial fibrillation or flutter, and 20% had chronic renal insufficiency (creatinine 2.0 or end‐stage renal disease). Most patients received ‐adrenergic blockers (97%), statins (95%), and either angiotensin‐converting enzyme (ACE) inhibitors or angiotensin receptor blockers (77%) preoperatively. Patients underwent a variety of surgeries (Table 1).

Patients had received both paclitaxel and sirolimus stents (28% and 73% of patients, respectively); 33% of patients had had more than 1 DES (Table 2). Most patients underwent surgery within 1 year of stent placement (77%), but only 40% had surgery within 180 days of stenting and only 13% within 90 days of stenting. Most patients (77%) had antiplatelet therapy completely discontinued a median of 10 days before surgery and remained off antiplatelet therapy for a median of 14 days total. Ten of the 15 patients (67%) who underwent surgery within 90 days of stenting had all antiplatelet agents discontinued preoperatively, 24 of the 30 patients (80%) who had surgery between 91 and 180 days after stenting had antiplatelet therapy completely discontinued, and 54 of the 69 patients (78%) who had surgery more than 180 days after stenting had antiplatelet therapy completely discontinued. There was no significant relationship between timing of stent placement relative to surgery (<90, 91180, or >180 days) and decision about whether to discontinue antiplatelet therapy (P = .59). However, invasiveness of the surgery was associated with antiplatelet management: 85% of those who continued antiplatelet therapy (aspirin or aspirin and clopidogrel) during the perioperative period were patients who underwent minimally invasive surgery (P < .0001).

The outcome events are presented in Table 3. Two patients (1.8%, 95% CI 0.5%6.2%) suffered a non‐ST‐elevation myocardial infarction (NSTEMI) postoperatively, and another patient (0.9%, 95% CI 0.2%4.8%) developed major bleeding, a retroperitoneal hemorrhage following kidney transplantation. This patient had been taking both aspirin and clopidogrel until 7 days prior to surgery and began to hemorrhage the day after surgery; antiplatelet agents were resumed 12 days postoperatively. No patients died (0%, 95% CI 0%3.3%). One of the 2 patients who suffered an MI was a 72‐year‐old man who had had placement of a single sirolimus‐eluting stent in the posterior descending artery 284 days prior to elective hip arthroplasty. He had no history of myocardial infarction but had undergone coronary bypass surgery 4 years earlier. Echocardiography showed he had aortic stenosis, with a calculated valve area of 0.9 cm². He had a baseline left ventricular ejection fraction of 45%. His preoperative cardiac medications included lovastatin, lisinopril, and atenolol; he discontinued both aspirin and clopidogrel 7 days before the surgery. His NSTEMI occurred on the day of his operation, presenting with hypotension and anterolateral ST depressions. His troponin T peaked at 0.48 mg/mL, with a peak creatinine kinase of 795 U/L (MB fraction 6%). His left ventricular ejection fraction was 45% on postoperative day 2 (unchanged from baseline). He was discharged on postoperative day 8 and returned for catheterization 3 weeks later, at which time he was found to have a 70% ostial lesion in a saphenous vein graft to an obtuse marginal, which was stented. The previously placed DES was widely patent. The other patient who suffered a postoperative NSTEMI was a 68‐year‐old man with a history of carotid artery stenting and renal artery stenosis who had undergone placement of 3 sirolimus‐eluting stents in the right coronary artery 50 days prior to cervical laminectomy. He had had elective placement of the stents following a positive pharmacologic stress test. He was taking 50 mg of atenolol daily and had been taking aspirin and clopidogrel until 17 days before surgery. On postoperative day 3 he developed dyspnea, and leads V4 and V5 showed ST depressions. His troponin T peaked at 1.24 mg/mL, with a peak creatinine kinase of 879 U/L (MB fraction 6%). The patient underwent left‐heart catheterization on hospital day 10. All 3 DESs were widely patent. His left ventricular ejection fraction was estimated at 65%. He was discharged on postoperative day 15. Because neither of the patients who had a postoperative NSTEMI showed evidence of stent thrombosis on catheterization, the overall rate of stent thrombosis was 0% (95% CI 0%3.3%).

DISCUSSION

Although 2 patients in our study cohort suffered a postoperative myocardial infarction and underwent postoperative catheterization, neither was found to have stent thrombosis, and the MIs of both patients were NSTEMIs with modest cardiac enzyme elevations only. No patients died. A rate of myocardial infarction of less than 2% is well within that expected for patients with established coronary disease undergoing noncardiac surgery.19 That most of our patients discontinued both aspirin and clopidogrel and did not receive antiplatelet agents for a median of 14 days suggests that transient termination of antiplatelet agents in the perioperative setting is not associated with high morbidity or mortality in patients with DES, even when patients have had their stents implanted in the preceding 36 months.

Our study builds on the limited data on this topic. One small case series examined outcomes in 38 patients who had had DES placement and subsequently underwent noncardiac surgery a median of 297 days after stenting.20 None of the patients in this series suffered from stent thrombosis or myocardial infarction, but most underwent surgery without discontinuing aspirin, and 41% underwent surgery without discontinuing clopidogrel. Another recent study demonstrated a high rate of adverse cardiovascular events in patients with coronary stents who underwent noncardiac surgery up to a year after stenting, but the authors of this study did not differentiate between drug‐eluting and bare‐metal stents, and all patients were continued on antiplatelet agents and received parenteral antithrombotic treatment.21

The major strength of our study was its systematic approach. Using a computerized and comprehensive search strategy, we identified all patients who had undergone DES placement at the Cleveland Clinic who subsequently had a preoperative evaluation at the IMPACT Center. Therefore, we are confident that the number of patients in our cohort truly reflects a well‐defined at‐risk population, allowing for an accurate calculation of event rates. This approach contrasts sharply with prior case reports and case series, in which the number of patients at risk was unknown. Nevertheless, these previous reports demonstrate that DES thromboses do occur and can be devastating, so even a small risk of DES thrombosis should be taken seriously. The upper bound of the 95% confidence interval of our estimate of the rate of DES thrombosis was 3.3%, so it is entirely plausible that sampling error contributed to the low rate of thrombosis that we observed.

One major limitation of our study is its sample size. Although our cohort was more than 3 times larger than the only other published cohort of DES patients undergoing noncardiac surgery,20 we had only limited precision to quantify the risk of DES thrombosis. This limitation is particularly relevant for patients who have undergone stent implantation within 36 months of surgery, as they are the patients most likely to have incomplete reendothelialization of the stented artery. We believe that when possible it remains prudent to delay noncardiac surgery for at least 36 months and perhaps up to 12 months following DES implantation, in keeping with recent guidelines.7, 8 However, for patients with conditions such as cancer whose surgery is semielective or patients with nonsurgical bleeding problems (such as gastrointestinal bleeding), our study provides at least some reassurance that short‐term discontinuation of antiplatelet agents may not be as dangerous as some authors have suggested,1 even within 36 months of DES placement. Another important limitation of our study is potential referral bias. At the Cleveland Clinic, most patients undergoing vascular and thoracic procedures are not evaluated at the IMPACT Center. Similarly, some of the patients with severe cardiovascular disease may also have bypassed the IMPACT Center and gone to a cardiologist for preoperative evaluation. As such, we believe our findings should not be generalized to high‐risk cardiac patients or to those undergoing high‐risk procedures.

A noteworthy distinction between our cohort and the cohort reported by Compton and colleagues is that in the perioperative period, most of our patients underwent complete discontinuation of antiplatelet therapy and remained off both aspirin and clopidogrel for an average of 2 weeks, whereas most patients in the other cohort were continued on antiplatelet therapy.20 This highlights the continued controversy surrounding management of antiplatelet therapy in perioperative patients with established coronary disease, who are at substantial risk for both bleeding and myocardial infarction because of the surgery.22 Our data offer little guidance on the optimal management of antiplatelet agents perioperatively because the incidence of both bleeding and thrombosis was low and whether or not patients were continued on antiplatelet agents was not random. We advocate individualized management strategies of perioperative patients with DES. Patients undergoing procedures that carry a high risk of outcome‐affecting bleeding (such as brain surgery) should probably have their antiplatelet agents discontinued preoperatively, whereas those undergoing minor surgery may have their antiplatelet agents continued, provided the surgeon and the anesthesiologist are in agreement with this approach. The timing of DES placement should also be factored into this decision because recently placed stents carry a higher risk of thrombosis.

In summary, our findings clarify the risks of stent thrombosis and postoperative myocardial infarction in clinically stable patients with DES who undergo low‐ and intermediate‐risk noncardiac surgery. Because it is unlikely to ever be ethically appropriate or logistically feasible to conduct a randomized study of patients with DES having early versus delayed noncardiac surgery, observational cohorts will have to suffice. Additional similar studies will help to validate (or refute) our findings and to more precisely quantify the risk of adverse cardiac events when patients with DES undergo surgery, which is real, feared, and potentially catastrophic but may be overestimated.

Physical restraints, such as bed rails, Posey vests, and 2‐point and 4‐point soft or hard restraints, are commonly used in acute care hospitals to protect agitated patients from harming themselves or others.1 Yet restraints are viewed by patient advocates and health care practitioners as inhumane and overly restrictive. Furthermore, currently used physical restraints have been linked to minor injuries such as sores and abrasions, intensification of agitation, and even death.2, 3 Hospitals and nursing homes are therefore required to try alternative and less severe means of alleviating agitation and delirium among patients before resorting to physical restraints. However, despite a general dislike of restraints and stricter federal guidelines governing their use, the application of restraints is often unavoidable for some patients. It is estimated that between 4% and 25% of in‐patients will have physical restraints applied at some point during their hospital stay.4

Given these numbers, it is surprising that newer and potentially safer restraint systems have not been explored. Safe enclosures may provide health care facilities with an alternative option. This type of restraint consists of a nylon net canopy that safely surrounds both the patient and the mattress. The potential for safe enclosures to provide a safe, humane, and acceptable method of restraint for both hospital staff and patients warranted investigation. In addition, because this system does not restrict a person's ability to move within the enclosure, the many potential hazards of immobility associated with standard restraints may be reduced or eliminated. However, to our knowledge, there have been no reports published of randomized trials comparing standard restraints to newer and possibly safer restraint systems.

We report a randomized controlled trial that compared the use of safe enclosures with standard restraints among agitated, hospitalized patients. Compared with patients in standard restraints, we hypothesized that safe enclosures would: (1) be perceived as more acceptable and humane by family members, physicians, and nurses; (2) lead to improved health outcomes such as decreased duration of restraint use, decreased agitation, shorter length of stay, decreased need to administer medication to treat agitation, and fewer injuries to the patient.

METHODS

Subjects

Male and female hospitalized patients at least 18 years old in the general medicine in‐patient services at a community hospital in Connecticut were assessed for eligibility to participate in this study if they had been put in restraints by the health care team independent of the study for one of these acute conditions: (1) delirium from any cause, including drug or alcohol withdrawal, or other medical conditions resulting in acute delirium; (2) confusional state from any cause; (3) agitation and disruptive behavior requiring restraints; (4) psychosis, hallucinations, or delusions requiring acute intervention (such as medication, restraints, or sitter); or (5) suicidality. Once in restraints, patients were screened for eligibility to participate in this study.

Exclusion criteria included: (1) need for acute respiratory or hemodynamic support or cardiac or septic shock; (2) terminal illness; (3) documented history of claustrophobia; (4) refusal by the family to give consent; (5) hospital stay < 24 hours; and (6) need for intravenous vasopressors, intubation, or ventilatory support. We also excluded patients who had been in restraints for more than 48 hours prior to potential study enrollment. Because safe enclosure would be a redundant system of restraint for patients requiring more than 1 limb in restraints, those patients were also excluded from the study. Figure 1 shows participation flow.

Figure 1

A number of screened patients were excluded for not being appropriate candidates for the safe enclosure. Of these, 20% required more than 1 limb in restraints, 13% required restraint not for agitation but for IV or catheter protection only, 10% were in critical care or on ventilators, and 13% were not appropriate for various other reasons including claustrophobia. The remaining excluded patients may have been eligible but either were preparing for discharge on screening (26%), were in restraints for more than 48 hours on screening (11%), stayed in the hospital less than 24 hours (4%), or had previously been a study participant (3%).

A stratified permuted block randomization was used to control for age (65 vs. >65 years) and sex (male vs. female) to ensure equal representation in both study arms. The study was approved by the institutional review board of the study site, and written informed consent was obtained by the study coordinator from patients' families. Because eligible patients suffered from acute delirium or agitation, most were not sufficiently cognizant to participate in the consent process. As a result, consent was largely obtained from patients' family members. Although the intent of this trial was to recruit 60 patients over an 18‐month period, the study was closed at 49 after 2 years because of slow recruitment and a lack of remaining funds.

RESULTS

DISCUSSION

We have demonstrated that the SOMA Safe Enclosure may be a more acceptable alternative to the restraints currently in use. Our results show that the safe enclosure was rated as more acceptable by family members, physicians, and secondary nurses in our composite perception scores. The results from the primary nurses did not show a significant difference between the 2 groups. An analysis of the individual perception variables found that family members viewed the safe enclosure as more acceptable for 10 of the 11 variables examined. Furthermore, in this small‐scale study, safe enclosures appeared to be safe, as there were no injuries reported in the intervention group. As stated above, there was 1 minor injury reported in the standard restraint groups.

Restraints are commonly used to protect agitated hospitalized patients from harming themselves or others. Despite the significant reluctance of hospital staff members to use restraints, they continue to be necessary in certain situations. Factors such as a general nursing shortage and the expense required to allocate nursing or other ancillary health care workers as sitters contribute to the use of restraints. Therefore, it is reasonable to conclude that restraint use in some form or fashion will continue into the foreseeable future. There are no clear estimates of the prevalence of restraint use in acute care hospitals. A chart review study from Canada reported physical restraints in about 7.7% of in‐patients.12 Other studies have reported the use of restraints on patients in the range of 4%‐25%.2 Given the prevalence of restraint use in acute care hospitals, surprisingly little innovative research has been undertaken to develop more effective and humane systems of restraint. Furthermore, no research has examined how restraint use may affect important clinical outcomes such as length of stay. To our knowledge, this is the first clinical trial to compare currently used restraints to a newer method of restraint using the SOMA Safe Enclosure.

The idea that restraint use can lead to further agitation is not supported by our data. We observed a decrease in agitated behavior scale scores from 9.6 to 7.4 from the 24‐ and to the 72‐hour assessments; however, these results were not significant and appeared to be more dramatic for the safe enclosure group because of higher baseline levels. Our adjusted analyses of length of stayrelated outcomes indicated an association with age. Total length of stay and time from restraint application until discharge were significantly reduced for those subjects younger than 80 years of age in the safe enclosure group. The basis for this finding is not entirely clear. It may be a chance finding, or there may have been a complex combination of factors at work.

There was a reduction in overall length of stay by 1.5 days among those in the safe enclosure group when compared with the standard restraint group. Similarly, total duration of restraint use of the safe enclosure group was 551 minutes (9 hours) shorter. Although these findings were not significant, they warrant further investigation in a larger trial. If safe enclosure use truly reduces length of stay and duration of restraint use, it is an important finding, for it could translate into meaningful cost savings for acute care hospitals. It is possible, however, that any potential cost savings could be tempered by the additional time required to set up the enclosure. Ethically, if restraints are to be used, their use should be minimized, and in that sense, safe enclosures may help acute care hospitals achieve this goal more effectively.

Limitations of this trial include its small sample size and inadequate power to determine certain outcomes. Although we saw encouraging trends in several outcomes, they failed to reach statistical significance because of the limited power. For instance, the observed power for total length of stay difference was only 17%. It is conceivable that a larger trial powered specifically for length of stayrelated outcomes may show significant results. Because subjects in this study were patients in a single midsize community teaching hospital, the results may not be generalizable to patients in, for example, tertiary‐care centers or nursing homes. However, these results may apply to a large proportion of patients in the United States, as most are treated in community hospitals. We found that many patients required 2 wrist restraints in order to protect IV lines, and this resulted in exclusion of a large proportion of potential subjects. Therefore, safe enclosures may not be appropriate for all agitated patients. They may be an ideal method of restraining patients who are not at risk of pulling out their IV line or catheters but require restraints for other reasons. This could include patients in nursing homes or rehabilitation centers.

It is also important to discuss the issue of practitioner acceptability of a newer method of restraint in acute care hospitals. As expected, we found the nursing staff was originally reluctant to use the safe enclosure, even as part of a trial. This may have been because of fear of change and having a high level of comfort with the restraint systems already in use. The setup of safe enclosures can take 10‐15 minutes, whereas the use of 2‐point soft restraints or Posey vests can be accomplished in as little as a minute. However, we found that after initial use of the safe enclosure, resistance among nurses declined. In fact, in our hospital, nurses began using safe enclosures for confused and agitated patients not enrolled in the study in order to prevent wandering and falls at night. Another difficulty reported by the nursing staff was feeling somewhat limited in their access to patients by a safe enclosure. Nurses had to open a zipped flap to access the patient to administer medication or provide food. Health care providers must remember to close the flap to avoid potential falls.

In summary, safe enclosures seem to be a safe and more acceptable alternative to the restraints currently in use in acute care hospitals. These findings should be replicated in a larger trial.

Physical restraints, such as bed rails, Posey vests, and 2‐point and 4‐point soft or hard restraints, are commonly used in acute care hospitals to protect agitated patients from harming themselves or others.1 Yet restraints are viewed by patient advocates and health care practitioners as inhumane and overly restrictive. Furthermore, currently used physical restraints have been linked to minor injuries such as sores and abrasions, intensification of agitation, and even death.2, 3 Hospitals and nursing homes are therefore required to try alternative and less severe means of alleviating agitation and delirium among patients before resorting to physical restraints. However, despite a general dislike of restraints and stricter federal guidelines governing their use, the application of restraints is often unavoidable for some patients. It is estimated that between 4% and 25% of in‐patients will have physical restraints applied at some point during their hospital stay.4

Given these numbers, it is surprising that newer and potentially safer restraint systems have not been explored. Safe enclosures may provide health care facilities with an alternative option. This type of restraint consists of a nylon net canopy that safely surrounds both the patient and the mattress. The potential for safe enclosures to provide a safe, humane, and acceptable method of restraint for both hospital staff and patients warranted investigation. In addition, because this system does not restrict a person's ability to move within the enclosure, the many potential hazards of immobility associated with standard restraints may be reduced or eliminated. However, to our knowledge, there have been no reports published of randomized trials comparing standard restraints to newer and possibly safer restraint systems.

We report a randomized controlled trial that compared the use of safe enclosures with standard restraints among agitated, hospitalized patients. Compared with patients in standard restraints, we hypothesized that safe enclosures would: (1) be perceived as more acceptable and humane by family members, physicians, and nurses; (2) lead to improved health outcomes such as decreased duration of restraint use, decreased agitation, shorter length of stay, decreased need to administer medication to treat agitation, and fewer injuries to the patient.

METHODS

Subjects

Male and female hospitalized patients at least 18 years old in the general medicine in‐patient services at a community hospital in Connecticut were assessed for eligibility to participate in this study if they had been put in restraints by the health care team independent of the study for one of these acute conditions: (1) delirium from any cause, including drug or alcohol withdrawal, or other medical conditions resulting in acute delirium; (2) confusional state from any cause; (3) agitation and disruptive behavior requiring restraints; (4) psychosis, hallucinations, or delusions requiring acute intervention (such as medication, restraints, or sitter); or (5) suicidality. Once in restraints, patients were screened for eligibility to participate in this study.

Exclusion criteria included: (1) need for acute respiratory or hemodynamic support or cardiac or septic shock; (2) terminal illness; (3) documented history of claustrophobia; (4) refusal by the family to give consent; (5) hospital stay < 24 hours; and (6) need for intravenous vasopressors, intubation, or ventilatory support. We also excluded patients who had been in restraints for more than 48 hours prior to potential study enrollment. Because safe enclosure would be a redundant system of restraint for patients requiring more than 1 limb in restraints, those patients were also excluded from the study. Figure 1 shows participation flow.

Figure 1

A number of screened patients were excluded for not being appropriate candidates for the safe enclosure. Of these, 20% required more than 1 limb in restraints, 13% required restraint not for agitation but for IV or catheter protection only, 10% were in critical care or on ventilators, and 13% were not appropriate for various other reasons including claustrophobia. The remaining excluded patients may have been eligible but either were preparing for discharge on screening (26%), were in restraints for more than 48 hours on screening (11%), stayed in the hospital less than 24 hours (4%), or had previously been a study participant (3%).

A stratified permuted block randomization was used to control for age (65 vs. >65 years) and sex (male vs. female) to ensure equal representation in both study arms. The study was approved by the institutional review board of the study site, and written informed consent was obtained by the study coordinator from patients' families. Because eligible patients suffered from acute delirium or agitation, most were not sufficiently cognizant to participate in the consent process. As a result, consent was largely obtained from patients' family members. Although the intent of this trial was to recruit 60 patients over an 18‐month period, the study was closed at 49 after 2 years because of slow recruitment and a lack of remaining funds.

RESULTS

DISCUSSION

We have demonstrated that the SOMA Safe Enclosure may be a more acceptable alternative to the restraints currently in use. Our results show that the safe enclosure was rated as more acceptable by family members, physicians, and secondary nurses in our composite perception scores. The results from the primary nurses did not show a significant difference between the 2 groups. An analysis of the individual perception variables found that family members viewed the safe enclosure as more acceptable for 10 of the 11 variables examined. Furthermore, in this small‐scale study, safe enclosures appeared to be safe, as there were no injuries reported in the intervention group. As stated above, there was 1 minor injury reported in the standard restraint groups.

Restraints are commonly used to protect agitated hospitalized patients from harming themselves or others. Despite the significant reluctance of hospital staff members to use restraints, they continue to be necessary in certain situations. Factors such as a general nursing shortage and the expense required to allocate nursing or other ancillary health care workers as sitters contribute to the use of restraints. Therefore, it is reasonable to conclude that restraint use in some form or fashion will continue into the foreseeable future. There are no clear estimates of the prevalence of restraint use in acute care hospitals. A chart review study from Canada reported physical restraints in about 7.7% of in‐patients.12 Other studies have reported the use of restraints on patients in the range of 4%‐25%.2 Given the prevalence of restraint use in acute care hospitals, surprisingly little innovative research has been undertaken to develop more effective and humane systems of restraint. Furthermore, no research has examined how restraint use may affect important clinical outcomes such as length of stay. To our knowledge, this is the first clinical trial to compare currently used restraints to a newer method of restraint using the SOMA Safe Enclosure.

The idea that restraint use can lead to further agitation is not supported by our data. We observed a decrease in agitated behavior scale scores from 9.6 to 7.4 from the 24‐ and to the 72‐hour assessments; however, these results were not significant and appeared to be more dramatic for the safe enclosure group because of higher baseline levels. Our adjusted analyses of length of stayrelated outcomes indicated an association with age. Total length of stay and time from restraint application until discharge were significantly reduced for those subjects younger than 80 years of age in the safe enclosure group. The basis for this finding is not entirely clear. It may be a chance finding, or there may have been a complex combination of factors at work.

There was a reduction in overall length of stay by 1.5 days among those in the safe enclosure group when compared with the standard restraint group. Similarly, total duration of restraint use of the safe enclosure group was 551 minutes (9 hours) shorter. Although these findings were not significant, they warrant further investigation in a larger trial. If safe enclosure use truly reduces length of stay and duration of restraint use, it is an important finding, for it could translate into meaningful cost savings for acute care hospitals. It is possible, however, that any potential cost savings could be tempered by the additional time required to set up the enclosure. Ethically, if restraints are to be used, their use should be minimized, and in that sense, safe enclosures may help acute care hospitals achieve this goal more effectively.

Limitations of this trial include its small sample size and inadequate power to determine certain outcomes. Although we saw encouraging trends in several outcomes, they failed to reach statistical significance because of the limited power. For instance, the observed power for total length of stay difference was only 17%. It is conceivable that a larger trial powered specifically for length of stayrelated outcomes may show significant results. Because subjects in this study were patients in a single midsize community teaching hospital, the results may not be generalizable to patients in, for example, tertiary‐care centers or nursing homes. However, these results may apply to a large proportion of patients in the United States, as most are treated in community hospitals. We found that many patients required 2 wrist restraints in order to protect IV lines, and this resulted in exclusion of a large proportion of potential subjects. Therefore, safe enclosures may not be appropriate for all agitated patients. They may be an ideal method of restraining patients who are not at risk of pulling out their IV line or catheters but require restraints for other reasons. This could include patients in nursing homes or rehabilitation centers.

It is also important to discuss the issue of practitioner acceptability of a newer method of restraint in acute care hospitals. As expected, we found the nursing staff was originally reluctant to use the safe enclosure, even as part of a trial. This may have been because of fear of change and having a high level of comfort with the restraint systems already in use. The setup of safe enclosures can take 10‐15 minutes, whereas the use of 2‐point soft restraints or Posey vests can be accomplished in as little as a minute. However, we found that after initial use of the safe enclosure, resistance among nurses declined. In fact, in our hospital, nurses began using safe enclosures for confused and agitated patients not enrolled in the study in order to prevent wandering and falls at night. Another difficulty reported by the nursing staff was feeling somewhat limited in their access to patients by a safe enclosure. Nurses had to open a zipped flap to access the patient to administer medication or provide food. Health care providers must remember to close the flap to avoid potential falls.

In summary, safe enclosures seem to be a safe and more acceptable alternative to the restraints currently in use in acute care hospitals. These findings should be replicated in a larger trial.

Physical restraints, such as bed rails, Posey vests, and 2‐point and 4‐point soft or hard restraints, are commonly used in acute care hospitals to protect agitated patients from harming themselves or others.1 Yet restraints are viewed by patient advocates and health care practitioners as inhumane and overly restrictive. Furthermore, currently used physical restraints have been linked to minor injuries such as sores and abrasions, intensification of agitation, and even death.2, 3 Hospitals and nursing homes are therefore required to try alternative and less severe means of alleviating agitation and delirium among patients before resorting to physical restraints. However, despite a general dislike of restraints and stricter federal guidelines governing their use, the application of restraints is often unavoidable for some patients. It is estimated that between 4% and 25% of in‐patients will have physical restraints applied at some point during their hospital stay.4

Given these numbers, it is surprising that newer and potentially safer restraint systems have not been explored. Safe enclosures may provide health care facilities with an alternative option. This type of restraint consists of a nylon net canopy that safely surrounds both the patient and the mattress. The potential for safe enclosures to provide a safe, humane, and acceptable method of restraint for both hospital staff and patients warranted investigation. In addition, because this system does not restrict a person's ability to move within the enclosure, the many potential hazards of immobility associated with standard restraints may be reduced or eliminated. However, to our knowledge, there have been no reports published of randomized trials comparing standard restraints to newer and possibly safer restraint systems.

We report a randomized controlled trial that compared the use of safe enclosures with standard restraints among agitated, hospitalized patients. Compared with patients in standard restraints, we hypothesized that safe enclosures would: (1) be perceived as more acceptable and humane by family members, physicians, and nurses; (2) lead to improved health outcomes such as decreased duration of restraint use, decreased agitation, shorter length of stay, decreased need to administer medication to treat agitation, and fewer injuries to the patient.

METHODS

Subjects

Male and female hospitalized patients at least 18 years old in the general medicine in‐patient services at a community hospital in Connecticut were assessed for eligibility to participate in this study if they had been put in restraints by the health care team independent of the study for one of these acute conditions: (1) delirium from any cause, including drug or alcohol withdrawal, or other medical conditions resulting in acute delirium; (2) confusional state from any cause; (3) agitation and disruptive behavior requiring restraints; (4) psychosis, hallucinations, or delusions requiring acute intervention (such as medication, restraints, or sitter); or (5) suicidality. Once in restraints, patients were screened for eligibility to participate in this study.

Exclusion criteria included: (1) need for acute respiratory or hemodynamic support or cardiac or septic shock; (2) terminal illness; (3) documented history of claustrophobia; (4) refusal by the family to give consent; (5) hospital stay < 24 hours; and (6) need for intravenous vasopressors, intubation, or ventilatory support. We also excluded patients who had been in restraints for more than 48 hours prior to potential study enrollment. Because safe enclosure would be a redundant system of restraint for patients requiring more than 1 limb in restraints, those patients were also excluded from the study. Figure 1 shows participation flow.

Figure 1

A number of screened patients were excluded for not being appropriate candidates for the safe enclosure. Of these, 20% required more than 1 limb in restraints, 13% required restraint not for agitation but for IV or catheter protection only, 10% were in critical care or on ventilators, and 13% were not appropriate for various other reasons including claustrophobia. The remaining excluded patients may have been eligible but either were preparing for discharge on screening (26%), were in restraints for more than 48 hours on screening (11%), stayed in the hospital less than 24 hours (4%), or had previously been a study participant (3%).

A stratified permuted block randomization was used to control for age (65 vs. >65 years) and sex (male vs. female) to ensure equal representation in both study arms. The study was approved by the institutional review board of the study site, and written informed consent was obtained by the study coordinator from patients' families. Because eligible patients suffered from acute delirium or agitation, most were not sufficiently cognizant to participate in the consent process. As a result, consent was largely obtained from patients' family members. Although the intent of this trial was to recruit 60 patients over an 18‐month period, the study was closed at 49 after 2 years because of slow recruitment and a lack of remaining funds.

RESULTS