Diagnostic value of blood culture growth patterns in distinguishing contaminants from pathogens

TL;DR

This study shows that blood culture growth patterns can help distinguish contaminants from true pathogens, improving antibiotic use.

Contribution

The study demonstrates that discordant growth of coagulase-negative staphylococci in blood culture sets has a high negative predictive value for true bacteremia.

Findings

Discordant CoNS-positive cultures have a 98.1% negative predictive value for true CoNS bacteremia.

Contaminants are more likely to grow in aerobic bottles compared to anaerobic bottles in discordant sets.

Discordant growth patterns can help clinicians make timely decisions about antibiotic use.

Abstract

Blood culture contamination is common, causing diagnostic uncertainty and unnecessary antibiotic use. Analyzing growth patterns within culture sets might offer diagnostic value. We retrospectively analyzed peripheral blood culture sets from 2019 and 2024. Growth pattern (one bottle [discordant] vs both bottles [concordant]) was analyzed according to clinical significance (contaminant vs true pathogen). Overall, 38,216 blood culture sets were analyzed, including 1,491 (3.9%) discordant and 1,938 (5.1%) concordant sets (remaining 34,787 [91.0%] sets were sterile). Discordant sets grew 1,060/1,491 (71.1%) contaminants and 431/1,491 (28.9%) true pathogens. Concordant sets grew 222/1,938 (11.4%) contaminants and 1,716/1,938 (88.5%) true pathogens (P < 0.001). Examining coagulase-negative staphylococci (CoNS) only (2019 data set), 629/642 (98.0%) discordant sets grew contaminants, while 13/642 (2.0%) grew true pathogens. In contrast, Staphylococcus aureus grew in only 82/270 (30.4%) discordant sets. Among 858 first CoNS-positive cultures per patient, 624/636 (98.1%) discordant sets grew contaminants, and 12/636 (1.9%) grew CoNS defined as a true pathogen. The negative predictive value of a discordant first CoNS set to exclude true CoNS bacteremia was 98.1% (95% confidence interval 96.7%–98.9%). Examining aerobic vs anaerobic bottles in 356 discordant sets, contaminants were found more frequently in aerobic bottles (135/356, 37.9% vs 73/356, 20.5%, P = 0.04). The proportion of true pathogens was similar in both (79/356, 22.2% vs 69/356, 19.4%, P = 0.4). Discordant CoNS-positive cultures were strongly associated with contamination. This could assist in interpreting blood culture results and supporting antimicrobial stewardship. Discordance might result from a diversion effect, the aerobic bottle acting as a diversion device for the anaerobic bottle.

Rapidly distinguishing blood culture contaminants from true pathogens is essential for optimizing antimicrobial stewardship and avoiding unnecessary antibiotic therapy. In this large, two-period study, we demonstrate that discordant growth of coagulase-negative staphylococci in a two-bottle set has a negative predictive value of 98.1% for true bacteremia. This finding remained robust across both study years and when restricted to first positive cultures, highlighting its reliability. Incorporating simple growth pattern analysis into early blood culture interpretation can provide clinicians with reliable and timely information within 24 hours, supporting more targeted and judicious antibiotic use.

Figures