Improvement of microbiological diagnostics in sepsis

Abstract

Sepsis, a life-threatening condition characterized by a dysregulated host response to infection, poses significant diagnostic and therapeutic challenges, contributing to high morbidity and mortality rates worldwide. The overall objective of this thesis is to provide an insight in how to improve the microbiological diagnosis of sepsis both by optimizing current blood culture (BC) practices and by evaluating novel diagnostic methods. The thesis also explores the impact on microbiological diagnosis imposed by two major recent events; the change of clinical sepsis criteria and the COVID-19 pandemic. The thesis consists of the four following projects:

Study I was a prospective non-inferiority study, assessing the utility of implementing a single-sampling strategy (SSS) versus multi-sampling strategy (MSS) for BCs in sepsis. The study group consisted of 549 suspected BSI episodes. The outcomes were detection of pathogens and occurrence of sample contamination. We found no significant difference in pathogen detection rates, and a tendency toward less contamination using SSS, thereby suggesting a potential simplification of the BC sampling protocol without compromising diagnostic yield.

In Study II, we described microbiological findings when using different sepsis criteria (Sepsis-2 versus Sepsis-3) in 514 patients with suspected sepsis. There were 357/514 (79.5%) Sepsis-3 and 411/514 (80.0%) Sepsis-2 episodes, with twothirds of patients fulfilling both definitions. The BC positivity rate was 130/357 (36.1%) and 145/411 (35.3%) in Sepsis-2 and Sepsis-3, respectively. The study shows that even with the improved definition, the frequency of BC positivity remains largely consistent.

Study III explored the changes in bacteremia rates during the COVID-19 pandemic, comparing BCs from patients with COVID-19 (n = 3,027) with two control groups without COVID-19, consisting of a contemporary group (n = 6,663) during the same period, as well as historical group (n = 8,175) from a year prior. Clinically relevant growth was found in 6.5% of the BC episodes in the COVID-19 group compared to 10.8% and 10.4% in the control groups respectively. Contamination was present in 8.4% in the COVID-19 group compared to 5.0% and 4.3% in the control groups. The study concluded that in COVID-19 patients, frequency of relevant growth was lower and contamination rates were higher.

Study IV retrospectively evaluated the diagnostic performance of T2Bacteria, a rapid molecular-based tool, against traditional BCs. We compared 640 T2Bacteria results to all BCs sampled within ± 72 hours of the T2Bacteria sampling. In total, 46/101 (45.5%) episodes were T2Bacteria positive/BC negative and 26/101 (25.7%) were T2Bacteria negative/BC positive. Only 29 (28.7%) episodes were positive using both methods. Total turn-around time was 27 hours and 33 minutes for T2Bacteria and 36 hours and 48 minutes for BCs (p < 0.001). We concluded that identification using molecular methods directly from blood provide a valuable complement to BCs.

Collectively, the findings from these studies adds to the understanding of microbiological diagnosis in sepsis, and highlights the need of both improving established methods, as well as the judicious implementation of novel techniques.