Sepsis research is hampered by the lack of a clear definition of suspected infection.

Introduction: A prevalent approach in sepsis research is pairing obtained cultures with antibiotic treatment to identify suspected infections. However, cultures are insensitive and nonspecific. Therefore, the present study aimed to examine the proportion of patients with infections admitted to an emergency department (ED) with and without having cultures obtained and to estimate 28-day mortality and prognostic factors of mortality according to culture status.

Methods: We conducted a secondary analysis of prospectively collected data from adult ED patients with suspected or documented infections (1 October 2017 - 31 March 2018). Patients receiving both cultures and antibiotics were compared to those treated solely with antibiotics. Logistic regression analyses assessed mortality differences.

Results: Among 2,055 patients, 1,441 (70.1%) had at least one culture obtained in addition to antibiotic treatment. Among patients without cultures, 163 (26.6%) had a Sequential Organ Failure Assessment score (SOFA) ≥ 2 on admission, compared to 528 (36.6%) among patients with cultures obtained (difference: 5.7-14.3). The 28-day mortality was 7.3% and 7.7%, respectively (difference: -2.1-2.9). Age, SOFA and the Charlson Comorbidity Index were the most important prognostic factors in both groups.

Conclusions: Defining suspected infections using cultures and antibiotics may introduce bias in sepsis research. Data sources relying on these criteria should be validated to examine their applicability.

Funding: The Region Zealand Health Research Foundation and the Naestved, Slagelse and Ringsted Hospital Research Fund.

Trial registration: Not relevant.