(ID12) Multi-institutional Retrospective Study Investigating Blood Culture Protocols and Test Positivity in 701 Dogs

Abstract:

Background: Blood cultures (BCs) are used to identify bacteremia. No standardized protocols exist for the timing or number of samples.

Hypothesis/

Objectives: Evaluate if BC positivity is impacted by i) the number or timing of blood samples, ii) antibiotic or immunosuppressive therapy, and iii) primary disease process. Secondary objectives were to document bacterial isolate and susceptibility data, and correlate clinicopathologic variables to culture positivity.

Animals: 701 dogs.

Methods: Multi-institutional retrospective study. Mixed-effect logistic regression was used to determine whether primary disease process, the number of BCs or the timing of sample collection influenced BC positivity. Prediction plots were generated, and correlations between minimum database and culture results were performed using two-sided t-tests.

Results: Dogs that had 3 samples had the greatest likelihood of obtaining a positive BC result (adjusted predictive value: 0.44, 95% CI: 0.21-0.70). The timing between sample collection had no significant effect on test positivity. Dogs with cardiac or urogenital disease had the greatest likelihood of a positive culture (adjusted predictive values: 0.33, 95% CI: 0.15-0.56 and CI: 0.16-0.56). Dogs with a positive BC were more likely to have a positive urine culture (OR: 4.36, 95% CI: 2.12-8.97, P = .003), and had higher mean cholesterol concentrations (279.5 mg/dL vs 247.0 mg/dL, P = .008), and lower mean magnesium concentrations (1.88 mg/dL vs 2.40 mg/dL, P = .003) than those with negative cultures.

Conclusions and Clinical Importance: Three BCs might be the optimal protocol. Urine culture, cholesterol, and magnesium results were associated with BC positivity in our population.