Reply to Topeli et al. and to Akinosoglou et al.

From the Authors:

We would like to thank Topeli and colleagues and Akinosoglou and colleagues for their interest in our manuscript (1).

We carefully read the discussion by Topeli and colleagues on our data and their own results regarding quick sepsis-related organ dysfunction (qSOFA) and other scores for sepsis’ mortality prediction in Turkey. We congratulate the authors for their initiative, as we believe it is very important to have data from low- and middle-income countries (LMICs). These countries represent more than 80% of world population, and the data from these settings on sepsis epidemiology and mortality are scarce (2).

There are important similarities between the authors’ results and ours. They found that qSOFA score has the worst sensitivity to predict mortality in septic patients, which adds to previous LMICs’ studies showing that qSOFA has low sensitivity to predict sepsis mortality in this population (3, 4). However, there are also major differences comparing both results. Their study is a single-center retrospective cohort with a limited number of patients, as can be suggested by the large confidence intervals of the data. Additionally, they collected qSOFA variables from patients at 48 hours before ICU admission, whereas we collected qSOFA data considering only the worst values prior to the suspicion of infection or sepsis, which may have contributed to more accurate findings in our study. The time window is crucial in assessing the sensitivity of a screening tool, as it is expected that if the interval of data collection is increased, more patients that deteriorate and eventually die will have a qSOFA ≥2. It would also be important to evaluate in Topeli’s data whether the use of a single qSOFA variable would increase the sensitivity of the score, as we demonstrated in our study. This modified score could be suggested as an alternative to improve its accuracy in determining mortality in LMICs.

Akinosoglou and colleagues assessed the role of qSOFA according to site of infection in a cohort of 614 septic patients from their institution. They identified that qSOFA accuracy to predict survival is dependent on the focus of infection. Because mortality rates are variable with the site of sepsis, and qSOFA variables may also be affected by the disease itself, their data are very reasonable. It would be interesting to assess data from other series to confirm if qSOFA can have adequate performance in all sepsis sites, or if we should modify the score according to the probable site of infection.