Inhalation NO in the HFNC group may result in a meaningless extension of survival time

Dear Editor,

We read the article by Isha et al. published in Critical Care with great interest [1]. The authors conducted a retrospective cohort study to explore the correlation between inhaled nitric oxide (iNO) and mortality in COVID-19 patients. This study had an important cohort, and one of the key conclusions drawn from their research is that COVID-19 patients supported by HFNC/NIV may benefit from the iNO. However, we believe that some biases have not been adequately addressed.

First, this study lacks an analysis of immortal time bias [2]. It seems unlikely that patients in the iNO group would receive iNO immediately after being placed on HFNC. It is necessary to clarify the timing from patient enrollment (i.e., starting HFNC) to the initiation of iNO treatment. During this time period, patients in the iNO group are considered immortal, which can artificially extend the survival days of patients in the iNO group. Using the initiation of iNO as the starting point for patient enrollment to analyze the 28-day mortality rate is a viable approach.

Second, we observe that although the 28-day survival curves for the iNO group and the No-iNO group in the HFNC cohort show significant separation, the in-hospital mortality rates for the two groups are nearly identical (24.2% vs. 22.3% after matching, SMD = 0.046, p = 0.85 according to our calculations). In the Methods section, the authors describe using a Cox univariate regression model for survival analysis, making it unlikely that there would be a difference in in-hospital mortality rates. Based on this, we would like to further inquire whether the definition of mortality in Table 4 refers to the 28-day mortality rate (as described in Fig. 3) or in-hospital mortality, and whether adjustments were made for additional covariates.

Finally, we would like to point out that the discrepancy between the 28-day mortality rate in Fig. 3 and the matched in-hospital mortality rate in Table 1 may be due to the inhalation of NO delaying the progression of death in the spontaneously breathing group. However, this extension of survival time may not be meaningful, similar to how we occasionally observe patients receiving futile ECMO treatment, resulting in increased days of survival. If we analyze the survival curves at 90 days or 180 days, the curves may overlap again. Additionally, incorporating extubation or HFNC withdrawal as a competing risk for death would also be a viable approach to assess the differences in mortality due to respiratory failure between the two groups [3].

This cohort’s data is important; however, the authors must interpret the conclusion that COVID-19 patients supported by HFNC/NIV may benefit from the inhalation of NO with caution.