The Role of Source Control in Septic Patients
However, several factors that might have affected the reported results should be addressed. First, as indicated in Table 1, almost all patient baseline characteristics were significantly different between patients requiring source control and those not requiring source control. This indicates that the two groups were fundamentally incomparable even after adjusting for available confounders using a sophisticated statistical approach, leading to the conclusion that the results in Tables 3 and 4 and Table S1 are not acceptable. Data analysis aiming at revealing associated factors, after stratification by performance with or without surgical source control (2), would be feasible.
Second, the authors showed hospital mortality was lower in patients with source control. However, the large sample size of this study might lead to statistically significant differences regardless of clinically small differences. Information regarding the predefined sample size calculation for hospital mortality analysis should be provided.
Third, this study consisted of septic shock patients with various types of infectious foci. The procedure and associated benefit and risks are not equivalent among these type of infections or source control procedures. Specifically, lung infections, the most frequent foci of septic shock, infrequently necessitate source control; that is, pleural drainage, and the impact on overall outcome is assumed to not be high, suggesting that the inclusion of lung infection in this study is fundamentally inappropriate. Furthermore, abdominal infection consists of various types of infections (cholangitis, peritonitis caused by upper gastrointestinal or colorectal perforation) and procedures (percutaneous drainage to open-abdominal surgery), which should be analyzed separately.
Finally, the analysis of timing to source control in skin and soft tissue foci found no statistical difference in mortality (early source control: 18.8% vs late source control: 27.3%; p = 0.314), although the absolute difference was about 9%. The lower mortality in skin and soft tissue foci in this study (17.7%) in comparison to that of a previous study (40.6%) (3) suggests the inclusion of patients with less severe infection/illness, which might have biased the therapeutic effect of source control. Conducting further studies by including a more severe population with sample size calculation based on a difference of about 10% mortality might be considered.