Impact of switching from manual to automated aerobic blood culture on bacteremia diagnosis in Lao PDR

Funding for antimicrobial resistance surveillance has led to more laboratories in low-resource settings switching to automated blood culture. There is limited evidence for the impact of switching from manual to automated blood culture in low-resource tropical settings. The objective of this study was to compare the yield of clinically significant blood culture isolates between manual and automated aerobic culture methods in adults and children in a diagnostic laboratory in Lao PDR. This retrospective study analyzed blood culture data from Mahosot Hospital, Vientiane. The switch from manual (two bottles in children and adults) to automated blood culture (one pediatric or one standard aerobic bottle) took place on 8 October 2018. Blood culture results, total blood volume, time to a positive culture of a significant isolate, and method used (manual or automated) were analyzed by multiple logistic regression and survival analysis. We analyzed 34,142 blood culture episodes from 1 January 2016 to 31 December 2023, of which 12,709 (37.22%) were processed by manual methods. The median blood volumes cultured per episode in babies, children, and adults were 1.9, 4.5, and 11.2 mL for manual vs 0.9, 1.8, and 10.6 mL for automated culture. The percentage of significant pathogens detected did not differ between manual and automated methods (6.7% vs 6.3%, 95% confidence interval [CI] [0.882-1.057]). However, presumed contaminants were detected at different rates between the two methods (4.1% vs 3.9%, 95% CI [0.684-0.864]). The odds ratio (OR) of detecting a significant pathogen increased by about 2.3% per 1 mL increase in blood volume (OR = 1.023, 95% CI [1.013-1.033], <i>P</i> < 0.001), while the odds of detecting a contaminant decreased by about 13% per 1 mL increase in blood volume (OR = 0.870, 95% CI [0.854-0.886], <i>P</i> < 0.001). The automated process led to reporting of significant growth approximately 1.3 times faster than the manual process, with a hazard ratio of 1.32 and a 95% CI of 1.20-1.46. In Laos, automated processing led to reporting of significant organisms from blood cultures more quickly than manual processing, with similar percentages of significant organisms and contaminants between the two methods. Switching to one automated bottle from two manual bottles for collection had the unintended consequence of reducing the volume of blood cultured from children.IMPORTANCEIn a primary to tertiary hospital in Laos, automated processing of blood cultures did not impact on yield of significant growth but did reduce the time to growth with a significant pathogen compared to manual processing; however, switching to a single bottle led to a reduction in the total volume of blood being sent to the laboratory for culture in all age groups. Monitoring and feedback to clinicians on blood volume received may help optimize the use of blood culture. Adding another aerobic or anaerobic culture bottle for adults would be desirable but would double the cost.