Role of Non-Group A Streptococci in Acute Pharyngitis

Methods

We conducted a retrospective case-control study of all adult patients who presented to our 2 family medicine clinics with sore throat and had a rapid strep test (RST) during a 12-month period. Our clinics serve a mixed urban population. Approximately one third of our patient population is the student body of a large urban university. The remainder are primarily managed care and privately insured patients and their families, with <10% covered by Medicare and <10% covered by Medicaid or uninsured.

Our clinics use the QuickVue Inline Strep A test (Quidel Corp., San Diego, CA). Per our medical center policy all patients with a negative RST must have a throat culture done. Cultures are inoculated to a streptococcal-selective agar and the plates are incubated for 24 to 48 hours in 5% CO₂-enriched atmosphere at 35°. The Hardy Diagnostics Strep PRO Grouping Kit (Santa Maria, CA) was used for typing; this kit uses a rapid latex agglutination method for the serologic identification of Lancefield's group A, B, C, F, and G from isolated colonies of B hemolytic streptococci.

All patients with non-GAS identified on culture were included as cases. Control group 1 consisted of all patients with GAS infection identified by RST or culture; control group 2 included all patients with a negative RST and culture (presumed viral pharyngitis). Each chart was reviewed for documentation of 5 clinical criteria traditionally associated with GAS infection: headache, fever (by history or documented in the office >99.5°), pharyngeal exudates, cervical lymphadenopathy, and absence of cough. The incidence of these criteria was compared by multivariate analysis using SPSS software (SPSS, Inc., Chicago, IL).

Results

Nine hundred fifteen charts were reviewed; 6 patients, who had negative RSTs but no cultures sent, were excluded. The mean age was 26 years for cases, 25 years for control group 1, and 27 years for control group 2 (P > .05). The population was primarily young adults (Figure 1). The prevalence of GAS by culture or RST was 16.0%, the prevalence of non-GAS was 19.8%, and the presence of viral (culture negative) was 64.2% (Figure 2). Of the 145 GAS patients, RST was positive for 75% of patients; the remainder were diagnosed by culture. Prevalence rates for non-GAS by type were 8.5% (group C), 6.0% (group B), 3.6% (group G), 1% (group F), and 1% (nontypeable) (Figure 3).

• Download figure
• Open in new tab

Figure 1.

Age distribution of patients.

• Download figure
• Open in new tab

Figure 2.

Study flow diagram. *75% by rapid strep test and 25% by culture. GAS, group A streptococci.

• Download figure
• Open in new tab

Figure 3.

Prevalence rates by type of streptococci.

The incidence of each of the clinical criteria are shown in Table 1. The presence of headache and fever was significantly associated with streptococcal infection, with no difference between GAS and non-GAS infection. Exudates and lymphadenopathy were also significantly associated with both GAS and non-GAS infection. Although both of these criteria were significantly more common in GAS infection than non-GAS infection, the magnitude of the difference was much smaller than between either streptococcal infection versus viral infection. The absence of cough did not differ between non-GAS infection and viral infection but was significantly more common in GAS infection (Figure 4). Analysis was also performed for individual non-GAS types. For group B and group C alone or combined the results were not significantly different; for groups F, G, and nontypeable, the numbers were too small for meaningful analysis (Figure 5).

• Download figure
• Open in new tab

Figure 4.

Comparison of the clinical features by pharyngitis type.

• Download figure
• Open in new tab

Figure 5.

Clincial features for groups A, B, and C versus viral.

Clinical guidelines are based on the number of criteria present; Table 2 shows the culture results by the number of criteria present for each patient. Patients with 0 or 1 criteria had an 85% and 72% rate of viral infection, respectively. When 2 criteria were present, the risk of any streptococcal infection rose to 55% (27% for non-GAS or GAS), and when 3 or more criteria were present the rate of any streptococcal infection rose to 81% (non-GAS infection, 34%; GAS infection, 47%; Figure 6). The frequencies of these clinical features for patients with GAS is consistent with frequencies reported in other studies.

• Download figure
• Open in new tab

Figure 6.

Pharyngitis type by the number of clinical criteria present.

Discussion

As a retrospective study we could not ensure capture of all patients who presented to the clinics with a chief complaint of sore throat during this time frame. Billing code data were reviewed but ultimately this was not useful because patients with negative tests could be billed under multiple diagnoses, such as upper respiratory infection, cough, or viral syndrome. However, because our rates of GAS infection were similar to those reported in other studies, and because 589 patients tested had only 0 or 1 criteria present, it seems likely that we captured a representative majority of patients presenting with sore throat. It is possible that some patients with 0 or 1 criteria were not tested, but it is also possible that patients with 4 or 5 criteria were presumptively treated without testing. Given these factors it seems unlikely that missed patients would skew the results significantly in one direction.

In this retrospective chart review we relied on clinician documentation for the presence of clinical criteria. Clinicians may under-report clinical findings when they have a positive RST or when they are relying primarily on the RST result. Fever, exudates, and adenopathy were probably most reliably documented and notably had the strongest association with streptococcal infection. As a “negative” finding, the absence of cough was probably the least reliably documented and was found to have the weakest association with streptococcal infection. Because our clinics are teaching practices for medical students, residents, and nurse practitioner students documentation tends to be thorough. These clinics also use an electronic medical record that significantly streamlines documentation. Given the reported rates of these findings in our study underdocumentation does not seem to be a major confounder.

Our rapid strep test sensitivity of 75% may seem lower than expected. Although the manufacturer reports a 92% to 96% sensitivity, published studies of rapid antigen detection of GAS show rates ranging from 70% to 90%, which is more consistent with our results.¹³

In this study population of primarily young adults, non-GAS was as common as GAS in those with acute pharyngitis. Non-GAS infection was significantly associated with headache, fever, exudates, and cervical lymphadenopathy compared with patients a negative culture, and with frequencies approaching those of patients with GAS infection, which suggests that non-GAS are likely to be pathogens in some of these cases. The presence of 3 clinical criteria predicts non-GAS or GAS infection with equal likelihood, although the presence of 4 or 5 criteria more reliably predicts GAS infection than non-GAS infection. Although we do not have a control group of asymptomatic patients with which to compare the frequency of non-GAS isolates, the association of a symptom complex recognized as bacterial in origin is strong evidence for non-GAS as pathogens in these patients.

If the goal of treatment is to target GAS infection exclusively, then the IDSA guidelines recommending treatment for only patients who test positive seem reasonable; in this study the Centor criteria predict at best only a 50% chance of GAS infection. If the goal is to treat any streptococcal infection, then treating adult patients with 3 or more criteria is reasonable because they have an 80% chance of any streptococcal infection. At this time the benefits of treating non-GAS pharyngitis in terms of either symptomatic relief or prevention of sequelae are unproven. Clinicians may consider treating patients with proven or presumptive non-GAS pharyngitis who fail to respond to symptomatic therapy or who are at increased risk for sequelae of group B or group C streptococcal infections, such as patients who are or have close contact with pregnant women, neonates, or elderly or immunocompromised patients. Further study is needed to determine whether patients with non-GAS pharyngitis benefit from targeted antibiotic treatment.