Influence of referral bias on the apparent clinical spectrum of infective endocarditis.

PURPOSE: To assess the effect of referral bias on the clinical spectrum of infective endocarditis. PATIENTS AND METHODS: We performed a retrospective study comparing a population-based cohort of incidence cases from Olmsted County, Minnesota, with a cohort of referred cases from the practice of the Mayo Clinic during the period from 1970 to 1987. RESULTS: In the community cohort, age was an important risk factor for acquiring endocarditis (incidence rate ratio 8.8:1 for age 65 years or older versus age less than 65 years), but episodes in elderly patients were underrepresented in the referral practice. The proportion of cases due to Staphylococcus aureus was greater in the community than in the referral practice (p less than 0.02), while a trend toward overrepresentation of enterococcal endocarditis was seen in the referral population (p = 0.057). Symptom duration prior to diagnosis was significantly shorter in the community. Overall, measures of in-hospital morbidity and mortality were similar in the two populations, but advanced age was associated with adverse outcome in the community cohort. CONCLUSION: The clinical spectrum of infective endocarditis may be distorted by referral. The increased risk of endocarditis in the elderly underscores the importance of adherence to recommendations for prophylaxis in this patient population.