Do gastrointestinal symptoms accompanying sore throat predict streptococcal pharyngitis? An UPRNet study. Upper Peninsula Research Network.

BACKGROUND

The purpose of this study was to determine whether gastrointestinal (GI) symptoms are more common in streptococcal than nonstreptococcal pharyngitis, and, if so, whether these symptoms are useful diagnostic predictors.

METHODS

Patients aged 4 and older presenting consecutively to one of three family practice clinics and one emergency department with the chief complaint of sore throat were invited to participate in the study. A nurse administered a brief symptom checklist; after documenting clinical signs, the clinician assessed and treated the patient. All patients were screened for group A streptococcus using the Abbott Test Pack Plus. Patients were enrolled from January 1996 through March 1996. Significant associations of signs and symptoms with streptococcal pharyngitis were determined by chi square, likelihood ratios were calculated, and logistic regression was used to compare diagnostic prediction models with and without GI symptoms.

RESULTS

Six hundred fifty-seven consecutive patients with the presenting complaint of sore throat were enrolled in the study. The mean age of the patients enrolled was 19 years; the median age was 14. Thirty-two percent of the children (ages 4 to 18), 23% of the adults (ages 19 to 74), and 29% of all patients had streptococcal pharyngitis. Symptom frequencies for streptococcal and nonstreptococcal pharyngitis, respectively, were: nausea (39% vs 31%, P = .14); vomiting (14% vs 7%, P = .004); abdominal pain (27% vs 26%, P = .621); and any GI symptom (47% vs 41%, P = .45). When included in a predictive model with other significant predictors of streptococcal pharyngitis including age, palatal petechiae, absence of cough, and anterior cervical adenopathy, the addition of nausea or vomiting added slight predictive power to the models, but abdominal pain and "any GI symptom" did not.

CONCLUSIONS

Nausea and vomiting are somewhat more common in streptococcal than in nonstreptococcal pharyngitis, but appear to have limited usefulness as clinical predictors of streptococcal pharyngitis.