| 摘要 |
Dear Editor,
We have read with great interest an original article entitled, “Predictors of mortality in necrotizing fasciitis” by Kumar and Lal in a recent issue at Formosan Journal of Surgery (2021; 54: 52–60).[1] We would like to congratulate the authors for their incredible work and contributions. Necrotizing fasciitis (NF) is a diagnosis involving skin and soft tissue infections that commonly requires immediate definitive surgical intervention.[2,3] As a consequence,
recognizing risk factors that may attribute to poor outcomes in an early encounter is important to acknowledge physicians to understand the disease progressivity. Several prior studies have been conducted to extensively analyze early signs that can be obtained from history‑taking, physical examination,
and laboratory results. Apart from the importance of this study – providing the study results from another study setting’s perspective, there are other components of predictors that should be further discussed in this study.
The authors discussed several physical and laboratory examination findings and their influence on mortality rate. However, the study findings did not include hemorrhagic bullae in its analysis. The presence of bullae differentiates
NF from other skin and soft tissue infections.[4] A previous meta‑analysis involving 23 studies indicates that hemorrhagic bullae was associated with NF with a sensitivity of 25.2% (95% confidence interval [CI] 12.8%–43.7%) and specificity of 95.8% (95% CI 87.3%–98.7%).[2] Therefore, that the presence
of hemorrhagic bullae should be included in the analysis to understand whether it influences NF mortality in the study.
In addition, choosing the effect measure in the study is essential to provide clinicians with an accurate statistical interpretation that will be translated into the clinical practice. As this study provides important findings to take the results into account in better predicting patient mortality, presenting the value of sensitivity and specificity of several most influencing factors is
needed. Further, although the authors have stated the P value of univariate and multivariate regression analyses, the value of odd ratio as an effect measure ought to be included to provide another value for researchers who are conducting a secondary study to investigate these factors. |
