Use of a High-Sensitivity Rapid Strep Test Without Culture Confirmation of Negative Results

Acute pharyngitis is one of the most common ill-nesses in children and adults in the United States and one of the leading reasons for outpatient office visits.¹ In both children and adults, the majority of cases are of viral etiology. Only 5% to 30% of pharyngitis patients are infected with group A b-hemolytic streptococci (GABHS), with the higher incidence being more common among pediatric populations during the winter months.^2-4 Accurate diagnosis of infections due to GABHS is important, since antibiotic therapy has been shown to shorten the clinical course,^5,6 reduce the rate of transmission,⁷ and prevent suppurative and nonsuppurative complications, such as peritonsillar abscess and acute rheumatic fever.^8,9

Although many clinical scoring systems have been developed to help with the diagnosis of pharyngitis due to GABHS, physicians vary greatly in their ability to diagnose streptococcal infections accurately on clinical grounds alone.^2,4,10,11 Because of this variability, both the American Academy of Pediatrics and the American Heart Association have recommended the use of either throat culture or an antigen detection test with culture confirmation of all negative results as the standard of care.^12,13 The rationale for culture confirmation is based on the historically low sensitivity of most antigen detection tests compared with blood agar plate (BAP) cultures.^14-16 Because of this low sensitivity, it has been feared that falsely negative results might lead to an increase in otherwise preventable suppurative and nonsuppurative complications. However, with the current commercial availability of a rapid antigen test with sensitivity equivalent or superior to BAP culture,^17,18 these recommendations have been challenged.^18,19

To date, there have been no clinical outcomes studies on the use of streptococcal antigen tests without culture confirmation. In our study, we examined the effect of a systemwide change in the management of pharyngitis on subsequent complication rates at the Lahey Clinic. Before April 1996, BAP culture was the sole test available for the diagnosis of streptococcal pharyngitis at the clinic. After demonstrating both high sensitivity and a beneficial effect on therapeutic decisions and appropriate antibiotic use associated with the use of a high-sensitivity antigen test¹⁹ (STREP A OIA, BioStar, Inc, Boulder, Colorado), the Department of Laboratory Medicine suggested that primary care physicians at the Lahey Clinic begin using this test without culture confirmation as the preferred method of diagnosis in April 1996. Although primary care physicians would still be able to order BAP cultures either alone or in combination with the high-sensitivity antigen test, reflexive culture confirmation of all negative antigen test results would no longer be performed. In this study, we compared the number of suppurative and nonsuppurative complications during 2 time periods: when throat culture was performed almost exclusively (period 1) and when the high-sensitivity antigen test without culture confirmation was the test performed in the majority of cases (period 2).

Methods

Setting

The Lahey Clinic is an integrated delivery system based in Burlington, Massachusetts, with a major satellite medical center in Peabody, Massachusetts. At these 2 locations, the Lahey Clinic medical staff includes more than 300 physicians in all disciplines. Primary, specialty, and hospital care are delivered to more than 650,000 patients each year at these 2 locations. Patients treated at the other 35 satellite offices of the clinic were not included in this study.

Subjects

The Lahey Clinic’s Decision Support System tracks the clinical care delivered at these 2 locations, and we used it to identify all eligible patients. All those seen for pharyngitis in the outpatient clinics were identified by the International Classification of Diseases, Ninth Revision (ICD-9) codes for acute pharyngitis (462), sore throat (784.1), acute tonsillitis (463), and streptococcal tonsillitis (034.0). All patients seen with suppurative complications were identified by the ICD-9 codes for peritonsillar abscess (475) and retropharyngeal abscess (478.24). All patients seen with nonsuppurative complications were identified by the ICD-9 codes for acute rheumatic fever (390) and poststreptococcal glomerulonephritis (580.0). Diagnostic tests were identified by Current Procedural Terminology (CPT) codes for throat culture (87081) and the high-sensitivity antigen test (83518 or 87880).

Patients were separated into 2 cohorts, those seen during period 1 (April 1, 1994, to March 31, 1996) and those seen during period 2 (April 1, 1996, to March 31, 1998). All available medical records of those patients identified as having suffered a complication of GABHS were then reviewed in an effort to assess both coding accuracy and whether a false laboratory test (culture or antigen test) might have preceded the complication.

Microbiologic Investigations

Both cultures and the high-sensitivity antigen test were performed in the clinical microbiology laboratory in Burlington, Massachusetts, using methods previously described.¹⁹

Data Analysis

We used chi-square analysis to compare the testing strategies used in the 2 periods. The Fisher exact test was used to compare the complication rates between the 2 periods.

Results

Patients and Diagnostic Tests Performed

The Decision Support System identified a total of 30,036 patient encounters with a primary or secondary diagnosis of pharyngitis during the 4 years studied. We identified a total of 15,399 patients in period 1 and 14,637 in period 2 Table 1.

There was a statistically significant change in the diagnostic strategy used in each of the 2 periods. During period 1, throat culture was used in 65.6% of patient encounters, and no test was used in the remaining encounters. During period 2, the high-sensitivity antigen test was the only test used in 50.8% of patient encounters; no test was used in 33.7% of the encounters; throat culture was the only test used in 14.3% of patient encounters; and the high-sensitivity antigen test with culture confirmation was used in 1.3% of patient encounters (P <.001).

Suppurative Complication rates

There were no statistically significant differences in the rates of suppurative complications observed during the 2 time periods (P = .92).

During period 1, ICD-9 codes identified 34 patients with a principal diagnosis of peritonsillar abscess and 3 patients with a principal diagnosis of retropharyngeal abscess. During period 2, ICD-9 codes identified 35 patients with a principal diagnosis of peritonsillar abscess and 1 patient with a principal diagnosis of retropharyngeal abscess.

Charts were available for review for 72 of the 73 patients with suppurative complications (99%). The principal diagnosis was correctly coded on 71 of 72 of these patients (99%), with the one miscode being a patient who was initially thought to have a peritonsillar abscess but ultimately found to have a parotid tumor.

Of the 71 patients with suppurative complications whose charts were available, the median age of patients with a suppurative complication was 28.6 years (range = 8 to 64 years). Thirty-one of the 71 (44%) initially presented with the complication, so no antecedent diagnostic test was performed.

Of the 40 patients who had been seen before the development of the suppurative complication, 23 had received a streptococcal diagnostic test in a visit before the one during which the suppurative diagnosis was made (12, culture; 11, high-sensitivity antigen test). Of the 12 throat cultures performed, 3 (25%) were positive. Of 11 high-sensitivity antigen tests performed, 3 (27%) were positive.

Of the 40 patients who had been seen before the development of the suppurative complication, 27 had been previously treated with an antibiotic that would have been expected to eradicate GABHS: 12 with penicillin, 5 with amoxicillin/clavulanic acid, 4 with amoxicillin, 3 with clindamycin, 2 with erythromycin, and 1 with cephalexin.

Nonsuppurative Complication Rates

There were no statistically significant differences in the rates of nonsuppurative complication rates observed during the 2 time periods (P = .30).

The combination of ICD-9 codes and chart review revealed no new or recurrent cases of acute rheumatic fever during the 4 years studied. During period 1, ICD-9 codes identified 6 patients with a principal diagnosis of rheumatic fever. All 6 of these charts were available for review. After chart review, we found that 5 of these patients had a medical history of rheumatic fever, and one was a woman aged 65 years with rheumatoid arthritis. During period 2, ICD-9 codes identified 4 patients with a principal diagnosis of rheumatic fever. All 4 of these charts were available for review. After chart review, we found that 3 of these patients had a medical history of rheumatic fever. One was a boy aged 5 years who presented with arthritis in both knees, normal results on cardiac and electrocardiograph examinations, and was given a diagnosis of serum sickness following treatment with cefixime.

ICD-9 codes identified one patient with a principal diagnosis of poststreptococcal glomerulonephritis in period 2.

Discussion

Rapid accurate diagnosis of pharyngeal infections attributable to GABHS can be difficult. Diagnosis based on clinical signs and symptoms alone is often inaccurate. Although BAP throat culture is widely held as the gold standard of diagnosis, the sensitivity of this method varies from 75% to 98% when compared with a more rigorous gold standard, such as Todd-Hewitt broth culture.^17,18 Reasons for the variable performance of BAP cultures include the absence of a universally accepted procedure and variability in the number and type of collection swabs, the type of culture media used, the duration of incubation, the atmosphere of incubation, and the specific Lancefield grouping procedure used.²⁰ Also, because throat cultures require 24 to 48 hours to achieve a definitive result, their use may lead to incomplete follow-up, missed opportunity to treat those infected, increased morbidity, delayed return to normal activities for both child and caretaker, and increased spread of disease.¹⁸

During the 1980s a number of antigen detection tests for GABHS became available, with most of these tests relying on latex agglutination or membrane-based immunoassay techniques. The advantages of these tests were rapid diagnosis and initiation of therapy. The main disadvantage was their sensitivity, which has been demonstrated to be as low as 50% and 70% with the latex agglutination tests14 and 70% to 90% with the more recently developed immunoassay tests.^15-18 Because of their historically low sensitivity, the American Academy of Pediatrics (AAP) and the American Heart Association (AHA) have always recommended that a confirmatory culture should be obtained for antigen test–negative patients suspected of having GABHS pharyngitis.^12,13

Policies advocating routine culture confirmation have been challenged.^17,18 Although the AAP and AHA concerns are well-grounded in theory, there have been no outcomes studies that have examined pharyngitis complication rates in detail. Despite the widespread use of antigen tests without culture confirmation of negative results,²¹ we have not witnessed a rise in the all-time low rates of acute rheumatic fever currently being seen in this country.²² Although there have been several local epidemics of acute rheumatic fever reported in this country during the past 2 decades,^23,24 falsely negative antigen tests have not been mentioned as a contributing factor for any of them.

Limitations

This is the first clinical outcomes study in which complication rates following use of a high-sensitivity antigen test without culture confirmation of negative results was assessed. Our study has many limitations. Because it is a retrospective case series, we cannot definitively dismiss the possibility that there was a change in the virulence of the GABHS strains encountered during the 4 years studied, although there was no change in the rates of invasive GABHS seen in the clinical microbiology laboratory at the Lahey Clinic during each of the periods studied.

Although chart review revealed the ICD-9 and CPT codes used to be accurate, we also cannot dismiss the possibility that some codes other than those specified were used to document complications of pharyngitis, such that those cases could not be included in our analysis. In an effort to maximize patient inclusion before beginning this study, we reviewed all potential CPT and ICD-9 codes with coding personnel and subsequently used all those codes in our analysis. We also included all patients with either a primary or secondary diagnosis code matching one of the ICD-9 codes selected.

Despite these flaws, this study is important because it illustrates a number of important facts about the management of pharyngitis in the current era. First, acute rheumatic fever is a rare disease in this country. During the 4 years of this study, there were more than 2 million outpatient visits to the Lahey Clinic and more than 30,000 visits for pharyngitis. There were no new or recurrent cases of acute rheumatic fever seen at the clinic during the 4 years studied. Moreover, during the first 5 years of this decade there was only one case of acute rheumatic fever from the entire Commonwealth of Massachusetts reported to the Centers for Disease Control and Prevention.²² Because of its low incidence, the absence of acute rheumatic fever we found in our study may have been due to chance alone and not to a change in diagnostic strategy used. Although this may be the case, it does not seem to be logical or cost-effective to recommend routine culture confirmation of all negative antigen tests unless a local epidemic of acute rheumatic fever warrants such a strategy.

Second, although there was one case of poststreptococcal glomerulonephritis during period 2, antibiotics are of little or doubtful benefit in preventing this complication.^25,26 Because diagnosis and therapy have little, if any, impact on the development of this complication, culture confirmation of negative antigen tests in an effort to prevent this rare complication cannot be justified.

Third, suppurative complications of pharyngitis are a great deal more common than nonsuppurative complications. Prevention of these complications might better serve as the rationale for culture confirmation of negative antigen tests were it not for the fact that most of these infections typically occur in young adults, a population in which group A streptococcal pharyngitis is relatively uncommon, and are typically polymicrobial in origin, with anaerobic organisms commonly identified.^27-30

Although cultures of the abscessed fluid were not routinely performed on patients reviewed in our study, there is circumstantial evidence that bacteria other than GABHS may have been responsible for many of the suppurative complications seen. In this study, 23 patients with a suppurative complication had undergone a previous test for GABHS, only 26% of which were positive (25% by culture, 27% by high-sensitivity antigen test). Given the high sensitivity that we have previously demonstrated with both culture and the STREP A OIA,¹⁹ this finding is more readily explained by the presence of other abscess-causing bacteria than by false-negative GABHS test results. There were also 27 patients included in this study with a suppurative complication despite concurrent use of an antibiotic that would be expected to effectively treat GABHS. Although poor compliance may be the explanation for the observed treatment failure, this finding may also have been due to infection with other bacteria that were less responsive to these agents. Finally, culture was performed on 3 of the abscesses reviewed in this study, and all 3 grew species other than GABHS. Therefore, because species other than GABHS are most commonly responsible for suppurative complications of pharyngitis, it seems unlikely that the use of diagnostic strategies designed to identify the presence of GABHS—whether streptococcal antigen tests, cultures, or both—have much chance of preventing these complications.

Conclusions

This study recounts our 2-year experience using a high-sensitivity GABHS antigen test without culture confirmation of the negative results and compares it with 2 years of culture-only experience. Using this strategy, we did not see an increase in either suppurative or nonsuppurative complication rates, as had been feared by the AHA and the AAP. Our study demonstrates that streptococcus testing by culture and antigen detection are of equal value in the prevention of suppurative and nonsuppurative complications of pharyngitis.

Selection of the appropriate diagnostic test should be based on other factors, such as patient satisfaction, cost-effectiveness, improved work flow, and so forth. Although not specifically studied in this cohort, use of the high-sensitivity rapid test has already been associated with more appropriate use of antibiotics when managing patients with pharyngitis at the Lahey Clinic.¹⁹ Use of a high-sensitivity antigen test may also offer the additional advantage of immediate diagnosis, improved treatment rates, more rapid clinical improvement and return to normal activities, and reduced spread of GABHS infections.

Acknowledgments

Dr Webb is the Medical Director at BioStar, Inc, the manufacturer of the STREP A OIA antigen test. He received salary support during the performance of this study. Drs Needham and Kurtz received no financial support.

Acute pharyngitis is one of the most common ill-nesses in children and adults in the United States and one of the leading reasons for outpatient office visits.¹ In both children and adults, the majority of cases are of viral etiology. Only 5% to 30% of pharyngitis patients are infected with group A b-hemolytic streptococci (GABHS), with the higher incidence being more common among pediatric populations during the winter months.^2-4 Accurate diagnosis of infections due to GABHS is important, since antibiotic therapy has been shown to shorten the clinical course,^5,6 reduce the rate of transmission,⁷ and prevent suppurative and nonsuppurative complications, such as peritonsillar abscess and acute rheumatic fever.^8,9

Although many clinical scoring systems have been developed to help with the diagnosis of pharyngitis due to GABHS, physicians vary greatly in their ability to diagnose streptococcal infections accurately on clinical grounds alone.^2,4,10,11 Because of this variability, both the American Academy of Pediatrics and the American Heart Association have recommended the use of either throat culture or an antigen detection test with culture confirmation of all negative results as the standard of care.^12,13 The rationale for culture confirmation is based on the historically low sensitivity of most antigen detection tests compared with blood agar plate (BAP) cultures.^14-16 Because of this low sensitivity, it has been feared that falsely negative results might lead to an increase in otherwise preventable suppurative and nonsuppurative complications. However, with the current commercial availability of a rapid antigen test with sensitivity equivalent or superior to BAP culture,^17,18 these recommendations have been challenged.^18,19

To date, there have been no clinical outcomes studies on the use of streptococcal antigen tests without culture confirmation. In our study, we examined the effect of a systemwide change in the management of pharyngitis on subsequent complication rates at the Lahey Clinic. Before April 1996, BAP culture was the sole test available for the diagnosis of streptococcal pharyngitis at the clinic. After demonstrating both high sensitivity and a beneficial effect on therapeutic decisions and appropriate antibiotic use associated with the use of a high-sensitivity antigen test¹⁹ (STREP A OIA, BioStar, Inc, Boulder, Colorado), the Department of Laboratory Medicine suggested that primary care physicians at the Lahey Clinic begin using this test without culture confirmation as the preferred method of diagnosis in April 1996. Although primary care physicians would still be able to order BAP cultures either alone or in combination with the high-sensitivity antigen test, reflexive culture confirmation of all negative antigen test results would no longer be performed. In this study, we compared the number of suppurative and nonsuppurative complications during 2 time periods: when throat culture was performed almost exclusively (period 1) and when the high-sensitivity antigen test without culture confirmation was the test performed in the majority of cases (period 2).

Methods

Setting

The Lahey Clinic is an integrated delivery system based in Burlington, Massachusetts, with a major satellite medical center in Peabody, Massachusetts. At these 2 locations, the Lahey Clinic medical staff includes more than 300 physicians in all disciplines. Primary, specialty, and hospital care are delivered to more than 650,000 patients each year at these 2 locations. Patients treated at the other 35 satellite offices of the clinic were not included in this study.

Subjects

The Lahey Clinic’s Decision Support System tracks the clinical care delivered at these 2 locations, and we used it to identify all eligible patients. All those seen for pharyngitis in the outpatient clinics were identified by the International Classification of Diseases, Ninth Revision (ICD-9) codes for acute pharyngitis (462), sore throat (784.1), acute tonsillitis (463), and streptococcal tonsillitis (034.0). All patients seen with suppurative complications were identified by the ICD-9 codes for peritonsillar abscess (475) and retropharyngeal abscess (478.24). All patients seen with nonsuppurative complications were identified by the ICD-9 codes for acute rheumatic fever (390) and poststreptococcal glomerulonephritis (580.0). Diagnostic tests were identified by Current Procedural Terminology (CPT) codes for throat culture (87081) and the high-sensitivity antigen test (83518 or 87880).

Patients were separated into 2 cohorts, those seen during period 1 (April 1, 1994, to March 31, 1996) and those seen during period 2 (April 1, 1996, to March 31, 1998). All available medical records of those patients identified as having suffered a complication of GABHS were then reviewed in an effort to assess both coding accuracy and whether a false laboratory test (culture or antigen test) might have preceded the complication.

Microbiologic Investigations

Both cultures and the high-sensitivity antigen test were performed in the clinical microbiology laboratory in Burlington, Massachusetts, using methods previously described.¹⁹

Data Analysis

We used chi-square analysis to compare the testing strategies used in the 2 periods. The Fisher exact test was used to compare the complication rates between the 2 periods.

Results

Patients and Diagnostic Tests Performed

The Decision Support System identified a total of 30,036 patient encounters with a primary or secondary diagnosis of pharyngitis during the 4 years studied. We identified a total of 15,399 patients in period 1 and 14,637 in period 2 Table 1.

There was a statistically significant change in the diagnostic strategy used in each of the 2 periods. During period 1, throat culture was used in 65.6% of patient encounters, and no test was used in the remaining encounters. During period 2, the high-sensitivity antigen test was the only test used in 50.8% of patient encounters; no test was used in 33.7% of the encounters; throat culture was the only test used in 14.3% of patient encounters; and the high-sensitivity antigen test with culture confirmation was used in 1.3% of patient encounters (P <.001).

Suppurative Complication rates

There were no statistically significant differences in the rates of suppurative complications observed during the 2 time periods (P = .92).

During period 1, ICD-9 codes identified 34 patients with a principal diagnosis of peritonsillar abscess and 3 patients with a principal diagnosis of retropharyngeal abscess. During period 2, ICD-9 codes identified 35 patients with a principal diagnosis of peritonsillar abscess and 1 patient with a principal diagnosis of retropharyngeal abscess.

Charts were available for review for 72 of the 73 patients with suppurative complications (99%). The principal diagnosis was correctly coded on 71 of 72 of these patients (99%), with the one miscode being a patient who was initially thought to have a peritonsillar abscess but ultimately found to have a parotid tumor.

Of the 71 patients with suppurative complications whose charts were available, the median age of patients with a suppurative complication was 28.6 years (range = 8 to 64 years). Thirty-one of the 71 (44%) initially presented with the complication, so no antecedent diagnostic test was performed.

Of the 40 patients who had been seen before the development of the suppurative complication, 23 had received a streptococcal diagnostic test in a visit before the one during which the suppurative diagnosis was made (12, culture; 11, high-sensitivity antigen test). Of the 12 throat cultures performed, 3 (25%) were positive. Of 11 high-sensitivity antigen tests performed, 3 (27%) were positive.

Of the 40 patients who had been seen before the development of the suppurative complication, 27 had been previously treated with an antibiotic that would have been expected to eradicate GABHS: 12 with penicillin, 5 with amoxicillin/clavulanic acid, 4 with amoxicillin, 3 with clindamycin, 2 with erythromycin, and 1 with cephalexin.

Nonsuppurative Complication Rates

There were no statistically significant differences in the rates of nonsuppurative complication rates observed during the 2 time periods (P = .30).

The combination of ICD-9 codes and chart review revealed no new or recurrent cases of acute rheumatic fever during the 4 years studied. During period 1, ICD-9 codes identified 6 patients with a principal diagnosis of rheumatic fever. All 6 of these charts were available for review. After chart review, we found that 5 of these patients had a medical history of rheumatic fever, and one was a woman aged 65 years with rheumatoid arthritis. During period 2, ICD-9 codes identified 4 patients with a principal diagnosis of rheumatic fever. All 4 of these charts were available for review. After chart review, we found that 3 of these patients had a medical history of rheumatic fever. One was a boy aged 5 years who presented with arthritis in both knees, normal results on cardiac and electrocardiograph examinations, and was given a diagnosis of serum sickness following treatment with cefixime.

ICD-9 codes identified one patient with a principal diagnosis of poststreptococcal glomerulonephritis in period 2.

Discussion

Rapid accurate diagnosis of pharyngeal infections attributable to GABHS can be difficult. Diagnosis based on clinical signs and symptoms alone is often inaccurate. Although BAP throat culture is widely held as the gold standard of diagnosis, the sensitivity of this method varies from 75% to 98% when compared with a more rigorous gold standard, such as Todd-Hewitt broth culture.^17,18 Reasons for the variable performance of BAP cultures include the absence of a universally accepted procedure and variability in the number and type of collection swabs, the type of culture media used, the duration of incubation, the atmosphere of incubation, and the specific Lancefield grouping procedure used.²⁰ Also, because throat cultures require 24 to 48 hours to achieve a definitive result, their use may lead to incomplete follow-up, missed opportunity to treat those infected, increased morbidity, delayed return to normal activities for both child and caretaker, and increased spread of disease.¹⁸

During the 1980s a number of antigen detection tests for GABHS became available, with most of these tests relying on latex agglutination or membrane-based immunoassay techniques. The advantages of these tests were rapid diagnosis and initiation of therapy. The main disadvantage was their sensitivity, which has been demonstrated to be as low as 50% and 70% with the latex agglutination tests14 and 70% to 90% with the more recently developed immunoassay tests.^15-18 Because of their historically low sensitivity, the American Academy of Pediatrics (AAP) and the American Heart Association (AHA) have always recommended that a confirmatory culture should be obtained for antigen test–negative patients suspected of having GABHS pharyngitis.^12,13

Policies advocating routine culture confirmation have been challenged.^17,18 Although the AAP and AHA concerns are well-grounded in theory, there have been no outcomes studies that have examined pharyngitis complication rates in detail. Despite the widespread use of antigen tests without culture confirmation of negative results,²¹ we have not witnessed a rise in the all-time low rates of acute rheumatic fever currently being seen in this country.²² Although there have been several local epidemics of acute rheumatic fever reported in this country during the past 2 decades,^23,24 falsely negative antigen tests have not been mentioned as a contributing factor for any of them.

Limitations

This is the first clinical outcomes study in which complication rates following use of a high-sensitivity antigen test without culture confirmation of negative results was assessed. Our study has many limitations. Because it is a retrospective case series, we cannot definitively dismiss the possibility that there was a change in the virulence of the GABHS strains encountered during the 4 years studied, although there was no change in the rates of invasive GABHS seen in the clinical microbiology laboratory at the Lahey Clinic during each of the periods studied.

Although chart review revealed the ICD-9 and CPT codes used to be accurate, we also cannot dismiss the possibility that some codes other than those specified were used to document complications of pharyngitis, such that those cases could not be included in our analysis. In an effort to maximize patient inclusion before beginning this study, we reviewed all potential CPT and ICD-9 codes with coding personnel and subsequently used all those codes in our analysis. We also included all patients with either a primary or secondary diagnosis code matching one of the ICD-9 codes selected.

Despite these flaws, this study is important because it illustrates a number of important facts about the management of pharyngitis in the current era. First, acute rheumatic fever is a rare disease in this country. During the 4 years of this study, there were more than 2 million outpatient visits to the Lahey Clinic and more than 30,000 visits for pharyngitis. There were no new or recurrent cases of acute rheumatic fever seen at the clinic during the 4 years studied. Moreover, during the first 5 years of this decade there was only one case of acute rheumatic fever from the entire Commonwealth of Massachusetts reported to the Centers for Disease Control and Prevention.²² Because of its low incidence, the absence of acute rheumatic fever we found in our study may have been due to chance alone and not to a change in diagnostic strategy used. Although this may be the case, it does not seem to be logical or cost-effective to recommend routine culture confirmation of all negative antigen tests unless a local epidemic of acute rheumatic fever warrants such a strategy.

Second, although there was one case of poststreptococcal glomerulonephritis during period 2, antibiotics are of little or doubtful benefit in preventing this complication.^25,26 Because diagnosis and therapy have little, if any, impact on the development of this complication, culture confirmation of negative antigen tests in an effort to prevent this rare complication cannot be justified.

Third, suppurative complications of pharyngitis are a great deal more common than nonsuppurative complications. Prevention of these complications might better serve as the rationale for culture confirmation of negative antigen tests were it not for the fact that most of these infections typically occur in young adults, a population in which group A streptococcal pharyngitis is relatively uncommon, and are typically polymicrobial in origin, with anaerobic organisms commonly identified.^27-30

Although cultures of the abscessed fluid were not routinely performed on patients reviewed in our study, there is circumstantial evidence that bacteria other than GABHS may have been responsible for many of the suppurative complications seen. In this study, 23 patients with a suppurative complication had undergone a previous test for GABHS, only 26% of which were positive (25% by culture, 27% by high-sensitivity antigen test). Given the high sensitivity that we have previously demonstrated with both culture and the STREP A OIA,¹⁹ this finding is more readily explained by the presence of other abscess-causing bacteria than by false-negative GABHS test results. There were also 27 patients included in this study with a suppurative complication despite concurrent use of an antibiotic that would be expected to effectively treat GABHS. Although poor compliance may be the explanation for the observed treatment failure, this finding may also have been due to infection with other bacteria that were less responsive to these agents. Finally, culture was performed on 3 of the abscesses reviewed in this study, and all 3 grew species other than GABHS. Therefore, because species other than GABHS are most commonly responsible for suppurative complications of pharyngitis, it seems unlikely that the use of diagnostic strategies designed to identify the presence of GABHS—whether streptococcal antigen tests, cultures, or both—have much chance of preventing these complications.

Conclusions

This study recounts our 2-year experience using a high-sensitivity GABHS antigen test without culture confirmation of the negative results and compares it with 2 years of culture-only experience. Using this strategy, we did not see an increase in either suppurative or nonsuppurative complication rates, as had been feared by the AHA and the AAP. Our study demonstrates that streptococcus testing by culture and antigen detection are of equal value in the prevention of suppurative and nonsuppurative complications of pharyngitis.

Selection of the appropriate diagnostic test should be based on other factors, such as patient satisfaction, cost-effectiveness, improved work flow, and so forth. Although not specifically studied in this cohort, use of the high-sensitivity rapid test has already been associated with more appropriate use of antibiotics when managing patients with pharyngitis at the Lahey Clinic.¹⁹ Use of a high-sensitivity antigen test may also offer the additional advantage of immediate diagnosis, improved treatment rates, more rapid clinical improvement and return to normal activities, and reduced spread of GABHS infections.

Acknowledgments

Dr Webb is the Medical Director at BioStar, Inc, the manufacturer of the STREP A OIA antigen test. He received salary support during the performance of this study. Drs Needham and Kurtz received no financial support.

Acute pharyngitis is one of the most common ill-nesses in children and adults in the United States and one of the leading reasons for outpatient office visits.¹ In both children and adults, the majority of cases are of viral etiology. Only 5% to 30% of pharyngitis patients are infected with group A b-hemolytic streptococci (GABHS), with the higher incidence being more common among pediatric populations during the winter months.^2-4 Accurate diagnosis of infections due to GABHS is important, since antibiotic therapy has been shown to shorten the clinical course,^5,6 reduce the rate of transmission,⁷ and prevent suppurative and nonsuppurative complications, such as peritonsillar abscess and acute rheumatic fever.^8,9

Although many clinical scoring systems have been developed to help with the diagnosis of pharyngitis due to GABHS, physicians vary greatly in their ability to diagnose streptococcal infections accurately on clinical grounds alone.^2,4,10,11 Because of this variability, both the American Academy of Pediatrics and the American Heart Association have recommended the use of either throat culture or an antigen detection test with culture confirmation of all negative results as the standard of care.^12,13 The rationale for culture confirmation is based on the historically low sensitivity of most antigen detection tests compared with blood agar plate (BAP) cultures.^14-16 Because of this low sensitivity, it has been feared that falsely negative results might lead to an increase in otherwise preventable suppurative and nonsuppurative complications. However, with the current commercial availability of a rapid antigen test with sensitivity equivalent or superior to BAP culture,^17,18 these recommendations have been challenged.^18,19

To date, there have been no clinical outcomes studies on the use of streptococcal antigen tests without culture confirmation. In our study, we examined the effect of a systemwide change in the management of pharyngitis on subsequent complication rates at the Lahey Clinic. Before April 1996, BAP culture was the sole test available for the diagnosis of streptococcal pharyngitis at the clinic. After demonstrating both high sensitivity and a beneficial effect on therapeutic decisions and appropriate antibiotic use associated with the use of a high-sensitivity antigen test¹⁹ (STREP A OIA, BioStar, Inc, Boulder, Colorado), the Department of Laboratory Medicine suggested that primary care physicians at the Lahey Clinic begin using this test without culture confirmation as the preferred method of diagnosis in April 1996. Although primary care physicians would still be able to order BAP cultures either alone or in combination with the high-sensitivity antigen test, reflexive culture confirmation of all negative antigen test results would no longer be performed. In this study, we compared the number of suppurative and nonsuppurative complications during 2 time periods: when throat culture was performed almost exclusively (period 1) and when the high-sensitivity antigen test without culture confirmation was the test performed in the majority of cases (period 2).

Methods

Setting

The Lahey Clinic is an integrated delivery system based in Burlington, Massachusetts, with a major satellite medical center in Peabody, Massachusetts. At these 2 locations, the Lahey Clinic medical staff includes more than 300 physicians in all disciplines. Primary, specialty, and hospital care are delivered to more than 650,000 patients each year at these 2 locations. Patients treated at the other 35 satellite offices of the clinic were not included in this study.

Subjects

The Lahey Clinic’s Decision Support System tracks the clinical care delivered at these 2 locations, and we used it to identify all eligible patients. All those seen for pharyngitis in the outpatient clinics were identified by the International Classification of Diseases, Ninth Revision (ICD-9) codes for acute pharyngitis (462), sore throat (784.1), acute tonsillitis (463), and streptococcal tonsillitis (034.0). All patients seen with suppurative complications were identified by the ICD-9 codes for peritonsillar abscess (475) and retropharyngeal abscess (478.24). All patients seen with nonsuppurative complications were identified by the ICD-9 codes for acute rheumatic fever (390) and poststreptococcal glomerulonephritis (580.0). Diagnostic tests were identified by Current Procedural Terminology (CPT) codes for throat culture (87081) and the high-sensitivity antigen test (83518 or 87880).

Patients were separated into 2 cohorts, those seen during period 1 (April 1, 1994, to March 31, 1996) and those seen during period 2 (April 1, 1996, to March 31, 1998). All available medical records of those patients identified as having suffered a complication of GABHS were then reviewed in an effort to assess both coding accuracy and whether a false laboratory test (culture or antigen test) might have preceded the complication.

Microbiologic Investigations

Both cultures and the high-sensitivity antigen test were performed in the clinical microbiology laboratory in Burlington, Massachusetts, using methods previously described.¹⁹

Data Analysis

We used chi-square analysis to compare the testing strategies used in the 2 periods. The Fisher exact test was used to compare the complication rates between the 2 periods.

Results

Patients and Diagnostic Tests Performed

The Decision Support System identified a total of 30,036 patient encounters with a primary or secondary diagnosis of pharyngitis during the 4 years studied. We identified a total of 15,399 patients in period 1 and 14,637 in period 2 Table 1.

There was a statistically significant change in the diagnostic strategy used in each of the 2 periods. During period 1, throat culture was used in 65.6% of patient encounters, and no test was used in the remaining encounters. During period 2, the high-sensitivity antigen test was the only test used in 50.8% of patient encounters; no test was used in 33.7% of the encounters; throat culture was the only test used in 14.3% of patient encounters; and the high-sensitivity antigen test with culture confirmation was used in 1.3% of patient encounters (P <.001).

Suppurative Complication rates

There were no statistically significant differences in the rates of suppurative complications observed during the 2 time periods (P = .92).

During period 1, ICD-9 codes identified 34 patients with a principal diagnosis of peritonsillar abscess and 3 patients with a principal diagnosis of retropharyngeal abscess. During period 2, ICD-9 codes identified 35 patients with a principal diagnosis of peritonsillar abscess and 1 patient with a principal diagnosis of retropharyngeal abscess.

Charts were available for review for 72 of the 73 patients with suppurative complications (99%). The principal diagnosis was correctly coded on 71 of 72 of these patients (99%), with the one miscode being a patient who was initially thought to have a peritonsillar abscess but ultimately found to have a parotid tumor.

Of the 71 patients with suppurative complications whose charts were available, the median age of patients with a suppurative complication was 28.6 years (range = 8 to 64 years). Thirty-one of the 71 (44%) initially presented with the complication, so no antecedent diagnostic test was performed.

Of the 40 patients who had been seen before the development of the suppurative complication, 23 had received a streptococcal diagnostic test in a visit before the one during which the suppurative diagnosis was made (12, culture; 11, high-sensitivity antigen test). Of the 12 throat cultures performed, 3 (25%) were positive. Of 11 high-sensitivity antigen tests performed, 3 (27%) were positive.

Of the 40 patients who had been seen before the development of the suppurative complication, 27 had been previously treated with an antibiotic that would have been expected to eradicate GABHS: 12 with penicillin, 5 with amoxicillin/clavulanic acid, 4 with amoxicillin, 3 with clindamycin, 2 with erythromycin, and 1 with cephalexin.

Nonsuppurative Complication Rates

There were no statistically significant differences in the rates of nonsuppurative complication rates observed during the 2 time periods (P = .30).

The combination of ICD-9 codes and chart review revealed no new or recurrent cases of acute rheumatic fever during the 4 years studied. During period 1, ICD-9 codes identified 6 patients with a principal diagnosis of rheumatic fever. All 6 of these charts were available for review. After chart review, we found that 5 of these patients had a medical history of rheumatic fever, and one was a woman aged 65 years with rheumatoid arthritis. During period 2, ICD-9 codes identified 4 patients with a principal diagnosis of rheumatic fever. All 4 of these charts were available for review. After chart review, we found that 3 of these patients had a medical history of rheumatic fever. One was a boy aged 5 years who presented with arthritis in both knees, normal results on cardiac and electrocardiograph examinations, and was given a diagnosis of serum sickness following treatment with cefixime.

ICD-9 codes identified one patient with a principal diagnosis of poststreptococcal glomerulonephritis in period 2.

Discussion

Rapid accurate diagnosis of pharyngeal infections attributable to GABHS can be difficult. Diagnosis based on clinical signs and symptoms alone is often inaccurate. Although BAP throat culture is widely held as the gold standard of diagnosis, the sensitivity of this method varies from 75% to 98% when compared with a more rigorous gold standard, such as Todd-Hewitt broth culture.^17,18 Reasons for the variable performance of BAP cultures include the absence of a universally accepted procedure and variability in the number and type of collection swabs, the type of culture media used, the duration of incubation, the atmosphere of incubation, and the specific Lancefield grouping procedure used.²⁰ Also, because throat cultures require 24 to 48 hours to achieve a definitive result, their use may lead to incomplete follow-up, missed opportunity to treat those infected, increased morbidity, delayed return to normal activities for both child and caretaker, and increased spread of disease.¹⁸

During the 1980s a number of antigen detection tests for GABHS became available, with most of these tests relying on latex agglutination or membrane-based immunoassay techniques. The advantages of these tests were rapid diagnosis and initiation of therapy. The main disadvantage was their sensitivity, which has been demonstrated to be as low as 50% and 70% with the latex agglutination tests14 and 70% to 90% with the more recently developed immunoassay tests.^15-18 Because of their historically low sensitivity, the American Academy of Pediatrics (AAP) and the American Heart Association (AHA) have always recommended that a confirmatory culture should be obtained for antigen test–negative patients suspected of having GABHS pharyngitis.^12,13

Policies advocating routine culture confirmation have been challenged.^17,18 Although the AAP and AHA concerns are well-grounded in theory, there have been no outcomes studies that have examined pharyngitis complication rates in detail. Despite the widespread use of antigen tests without culture confirmation of negative results,²¹ we have not witnessed a rise in the all-time low rates of acute rheumatic fever currently being seen in this country.²² Although there have been several local epidemics of acute rheumatic fever reported in this country during the past 2 decades,^23,24 falsely negative antigen tests have not been mentioned as a contributing factor for any of them.

Limitations

This is the first clinical outcomes study in which complication rates following use of a high-sensitivity antigen test without culture confirmation of negative results was assessed. Our study has many limitations. Because it is a retrospective case series, we cannot definitively dismiss the possibility that there was a change in the virulence of the GABHS strains encountered during the 4 years studied, although there was no change in the rates of invasive GABHS seen in the clinical microbiology laboratory at the Lahey Clinic during each of the periods studied.

Although chart review revealed the ICD-9 and CPT codes used to be accurate, we also cannot dismiss the possibility that some codes other than those specified were used to document complications of pharyngitis, such that those cases could not be included in our analysis. In an effort to maximize patient inclusion before beginning this study, we reviewed all potential CPT and ICD-9 codes with coding personnel and subsequently used all those codes in our analysis. We also included all patients with either a primary or secondary diagnosis code matching one of the ICD-9 codes selected.

Despite these flaws, this study is important because it illustrates a number of important facts about the management of pharyngitis in the current era. First, acute rheumatic fever is a rare disease in this country. During the 4 years of this study, there were more than 2 million outpatient visits to the Lahey Clinic and more than 30,000 visits for pharyngitis. There were no new or recurrent cases of acute rheumatic fever seen at the clinic during the 4 years studied. Moreover, during the first 5 years of this decade there was only one case of acute rheumatic fever from the entire Commonwealth of Massachusetts reported to the Centers for Disease Control and Prevention.²² Because of its low incidence, the absence of acute rheumatic fever we found in our study may have been due to chance alone and not to a change in diagnostic strategy used. Although this may be the case, it does not seem to be logical or cost-effective to recommend routine culture confirmation of all negative antigen tests unless a local epidemic of acute rheumatic fever warrants such a strategy.

Second, although there was one case of poststreptococcal glomerulonephritis during period 2, antibiotics are of little or doubtful benefit in preventing this complication.^25,26 Because diagnosis and therapy have little, if any, impact on the development of this complication, culture confirmation of negative antigen tests in an effort to prevent this rare complication cannot be justified.

Third, suppurative complications of pharyngitis are a great deal more common than nonsuppurative complications. Prevention of these complications might better serve as the rationale for culture confirmation of negative antigen tests were it not for the fact that most of these infections typically occur in young adults, a population in which group A streptococcal pharyngitis is relatively uncommon, and are typically polymicrobial in origin, with anaerobic organisms commonly identified.^27-30

Although cultures of the abscessed fluid were not routinely performed on patients reviewed in our study, there is circumstantial evidence that bacteria other than GABHS may have been responsible for many of the suppurative complications seen. In this study, 23 patients with a suppurative complication had undergone a previous test for GABHS, only 26% of which were positive (25% by culture, 27% by high-sensitivity antigen test). Given the high sensitivity that we have previously demonstrated with both culture and the STREP A OIA,¹⁹ this finding is more readily explained by the presence of other abscess-causing bacteria than by false-negative GABHS test results. There were also 27 patients included in this study with a suppurative complication despite concurrent use of an antibiotic that would be expected to effectively treat GABHS. Although poor compliance may be the explanation for the observed treatment failure, this finding may also have been due to infection with other bacteria that were less responsive to these agents. Finally, culture was performed on 3 of the abscesses reviewed in this study, and all 3 grew species other than GABHS. Therefore, because species other than GABHS are most commonly responsible for suppurative complications of pharyngitis, it seems unlikely that the use of diagnostic strategies designed to identify the presence of GABHS—whether streptococcal antigen tests, cultures, or both—have much chance of preventing these complications.

Conclusions

This study recounts our 2-year experience using a high-sensitivity GABHS antigen test without culture confirmation of the negative results and compares it with 2 years of culture-only experience. Using this strategy, we did not see an increase in either suppurative or nonsuppurative complication rates, as had been feared by the AHA and the AAP. Our study demonstrates that streptococcus testing by culture and antigen detection are of equal value in the prevention of suppurative and nonsuppurative complications of pharyngitis.

Selection of the appropriate diagnostic test should be based on other factors, such as patient satisfaction, cost-effectiveness, improved work flow, and so forth. Although not specifically studied in this cohort, use of the high-sensitivity rapid test has already been associated with more appropriate use of antibiotics when managing patients with pharyngitis at the Lahey Clinic.¹⁹ Use of a high-sensitivity antigen test may also offer the additional advantage of immediate diagnosis, improved treatment rates, more rapid clinical improvement and return to normal activities, and reduced spread of GABHS infections.

Acknowledgments

Dr Webb is the Medical Director at BioStar, Inc, the manufacturer of the STREP A OIA antigen test. He received salary support during the performance of this study. Drs Needham and Kurtz received no financial support.