Dear Editor

We read Zhou’s study [1] with great interest. This study aims to explore clinical and laboratory characteristics of primary Sjögren's syndrome (pSS) complicated with interstitial lung disease (ILD) and investigate the risk factors for respiratory infections in pSS-ILD. After the comparison of the clinical characteristics, treatments, and laboratory data between infection group with non infection group, and univariate analysis of respiratory infection, then this study found that higher levels of ESSDAI, CRP, ESR, and C3 may be correlated with increased infection risk. Pulmonary arterial hypertension (PAH) and reduction of diffusing capacity of the lung for carbon monoxide (DLCO) were identified as independent risk factors. Despite outstanding results, we mention that a great factor may be ignored that could greatly influence the result of risk factors in this study.

In Fig. 2, binary logistic regression analysis was performed to identify potential risk factors for respiratory infections in patients with pSS-ILD. However, there are 25 variables in this univariate analysis, but the outcome events were only 53 infection cases in this study. Most importantly, this statistical analysis method severely breaks the rule that 1 variable per 10 outcome events (i.e., infection case) for the univariate analysis [2, 3]. In other words, this study needs more than 250 infection cases in patients with pSS-ILD, which is the tenfold of 25 variables. Without obeying this basic rule of statistical analysis, it could yield unreliable statistical results, while the PAH and reduction of DLCO may not be independent risk factors. Thus, more respiratory infections in patients with pSS-ILD need to be included in this univariate analysis work. However, despite this statistical error comment, we show great gratitude to Zhou et al. outstanding retrospective work.