Investigating the time to blood culture positivity: why does it take so long?

Abstract

Introduction. Bloodstream infections (BSIs) are one of the most serious infections investigated by microbiologists. However, the time to detect a BSI fails to meet the rapidity required to inform clinical decisions in real time.Gap Statement. Blood culture (BC) is considered the gold standard for diagnosing bloodstream infections. However, the time to blood culture positivity can be lengthy. Underpinning this is the reliance on bacteria replicating to a high concentration, which is necessary for the detection using routine blood culture systems. To improve the diagnosis and management of patients with BSIs, more sensitive detection methods are required.Aim. The study aimed to answer key questions addressing the delay in BSI detection and whether the time to BSI detection could be expedited using a Scattered Light Integrated Collection (SLIC) device.Methodology. A proof-of-concept study was conducted to compare the time to positivity (TTP) of Gram-negative BCs flagging positive on BacT/ALERT with an SLIC device. An SLIC device was utilized to compare the TTP of the most prevalent BSI pathogens derived from nutrient broth and BC, the influence of bacterial load on TTP and the TTP directly from whole blood. Additionally, the overall turnaround time (TAT) of SLIC was compared with that of a standard hospital workflow.Results. Most pathogens tested took significantly longer to replicate when derived from BC than from nutrient medium. The median TTP of Gram-negative BC on BacT/ALERT was 13.56 h with a median bacterial load of 6.4×10⁹ c.f.u. ml^-1. All pathogens (7/7) derived from BC at a concentration of 10⁵ c.f.u. ml^-1 were detectable in under 70 min on SLIC. Decreasing Escherichia coli BC concentration from 10⁵ to 10² c.f.u. ml^-1 increased the TTP of SLIC from 15 to 85 min. Direct BSI detection from whole blood on SLIC demonstrated a 76% reduction in TAT when compared with the standard hospital workflow.Conclusion. An SLIC device significantly reduced the TTP of common BSI pathogens. The application of this technology could have a major impact on the detection and management of BSI.

Keywords: blood culture; bloodstream infection; rapid diagnostics.

Fig. 1.. The detection time (hours) of Gram-negative BSI using BacT/ALERT 3D system. A total of 111 Gram-negative BCs flagged positive on the BacT/ALERT system over the 5-month study with a median TTP of 13.56 h (IQR, 10.57–19.66). The TTP varied with bacterial spp. For bacteria spp. detected more than once, the number of cases is given [n] with the mean±sem.

Fig. 2.. Quantifying the bacterial concentration of BCs that had flagged positive with Gram-negative bacteria. The median bacterial concentration for 109 Gram-negative BCs was 6.4×10⁹ c.f.u. ml⁻¹ (IQR, 6.6×10⁸–7.26×10¹⁰). Two BCs, one positive for P. aeruginosa and another for Pantoea agglomerans, could not be quantified due to insufficient growth. For bacteria spp. detected more than once, the number of cases is given [n] with the mean±sem.

Fig. 3.. Comparison of TTP of different bacterial species grown in BC compared with BHI medium. TTP was defined as the time taken to achieve a clear differentiation between the positive and negative control on SLIC. The time to detection on SLIC was normalized by previously published generation times (Table S1, available in the online Supplementary Material). The TTP on SLIC was noticeably longer when bacteria were grown in BC compared with media except for S. aureus. All BSI pathogens were detectable on SLIC in under 60 min apart from S. agalactiae (70 min). The starting bacterial concentration was 10⁵ c.f.u. ml⁻¹.

Fig. 4.. E. coli BCs spiked with four different concentrations were examined on SLIC over 120 min. The LS-SST method was used to recover bacteria from BC, and the bacteria were grown in pre-warmed TSB media for SLIC analysis. The TTP was determined when the mV output deviated from the baseline consistently for two timepoints (indicated in green). The baseline was defined by the negative control mV output (not shown on graph). A total of three biological repeats were performed at each BC concentration, the mean±sem is plotted. A concentration time-dependent relationship was established for all BCs. BCs spiked with E. coli at 10⁵, 10⁴ and 10³ c.f.u. ml⁻¹ were detectable within 15 to 30 min, respectively. In comparison, BC spiked with E. coli at 10² c.f.u. ml⁻¹ was detectable within 85 min.

Fig. 5.. The TTP of E. coli on SLIC directly from whole blood. The mean TTP for a bacterial concentration range of 10–10³ was 8.54±0.99 h, 6.48±0.81 h and 6.48±0.67 h for the blood volumes of 2 ml, 4 ml, and 8 ml respectively. The overall mean TTP across all bacterial concentrations and blood volumes was 7.16±0.82 h.

Fig. 6.. Time saving of the SLIC workflow against the BacT/ALERT workflow. The SLIC workflow involves the direct processing of whole blood with a 10-min serum separation step before the sample is ready for use in SLIC. The analysis of TTP on SLIC directly from whole blood was comparable to BacT/ALERT for the detection of E. coli. However, by directly sampling from whole blood (patient sample), the mean transfer time of 12.54±1.28 h was nullified and could potentially enable BC positivity to be confirmed within the same day as sampling. The slowest meantime for detection on SLIC (E. coli at <10 c.f.u. ml⁻¹) was taken as an illustrative example. Depending on blood volume and starting bacterial concentration, the SLIC workflow may be quicker than indicated.