Interchangeability of blood culture bottles and automats for blood culture systems: an in-vitro performance evaluation.

Abstract

Blood cultures are central to diagnosing bloodstream infections and guiding appropriate treatment. Because microbial concentrations in blood are low, culturing is necessary for any downstream analysis, and rapid detection of microbial growth is essential. Automated systems, such as those for blood culture (ie, blood culture automats) facilitate rapid detection, but are not equally available worldwide; their implementation in low-income and middle-income countries (LMICs) is hindered by complex supply chains, low availability of equipment and consumables, such as blood culture bottles (BCBs), and higher prices. In this study, we aimed to evaluate the interchangeability of BCBs and blood culture automats from different brands to improve accessibility to blood culture diagnostics in resource-limited settings by reducing brand dependency. This in-vitro evaluation study was conducted at the Institute of Tropical Medicine, Antwerp, Belgium. We evaluated the interchangeability of paediatric BCBs and blood culture automats from four different systems: bioMérieux BacT/ALERT 3D, Autobio BC60, DL-Biotech DL60, and Mindray TDR60. The combination of Autobio BCBs with the Mindray automat was excluded, resulting in 11 BCB-automat combinations. A panel of 20 bloodstream infection-causing microbial species was assembled with seven isolates (one reference and six clinical) for each species. Simulated blood cultures were prepared by spiking defibrinated horse blood with microbial isolates at approximately 5-50 colony-forming units per mL, with 2 mL inoculated per BCB. Non-spiked blood cultures were also included to assess specificity. Interchangeability was assessed based on feasibility (ie, physical fit) and performance, which comprised of yield, specificity, and time to positivity (TTP). We evaluated interchangeability across 4620 spiked blood cultures (11 BCB-automat combinations, 20 species, seven isolates, in triplicate) and 301 non-spiked cultures. All BCB-automat combinations were shown to have physical compatibility. The overall yield ranged from 97·6-100%, with 100% yield (420 of 420) for bioMérieux BCBs with Autobio or Mindray automats and for Autobio BCBs with bioMérieux automats. Specificity was 100% (301 of 301) for all non-spiked combinations. Median TTP was shortest with Autobio automats and showed some species-specific differences, but no differences were observed between same-brand and different-brand BCB-automat combinations. False negative detections were rare, occurring mainly with Cryptococcus neoformans in DL-Biotech and Autobio BCBs, and six false signals were detected, four of which occurred in DL-Biotech BCBs with the Autobio automat. This study shows that BCBs can be used with blood culture automats from different brands. For LMICs facing complex supply chains and higher prices, such flexibility could substantially improve accessibility to blood culture diagnostics. These findings warrant verification in a randomised controlled trial with clinical samples. The study was financially supported by the Health Innovations for All programme at the Institute of Tropical Medicine, funded by the Flemish Government, Department of Economy, Science and Innovation.