We thank Mukherjee et al. for their comments on our study, in which there was no association between cardiac dysfunction and reintubation rate in patients following application of a postextubation high-flow nasal cannula (HFNC) [
1]. In their letter, the authors questioned whether we considered several factors relevant to extubation failure. Perhaps they were concerned about the possibility that these factors acted as confounding variables.
We were able to calculate the respiratory rate-oxygenation-heart rate (ROX-HR) index from our raw data [
2]. We only calculated the ROX-HR index after 2 hours of HFNC because the number of patients from whom the ROX-HR index could be obtained decreased over time. In the normal function group, the ROX-HR index was 9.99 (range, 7.23 to 14.80). However, the value was not available in five patients. In the cardiac dysfunction group, the ROX-HR index was 10.60 (range, 8.18 to 13.58). There was no significant difference in the ROX-HR index between the two groups (p=0.517). Thus, it was not a confounding variable.
Unfortunately, we did not have data regarding N-terminal pro-brain natriuretic peptide (NT-proBNP), the Simplified Acute Physiology Score 2 score, or the secretion burden. It was not feasible to analyze all factors known to predict extubation failure. The possibility of hidden confounding variables is an inherent limitation of a retrospective study, which is why a randomized controlled trial is needed.
With regard to NT-proBNP, 11% of patients were treated with renal replacement therapy on the day of extubation. Therefore, prediction of extubation failure using NT-proBNP would not be reliable in our study population. Patients with renal impairment were also excluded in the study [
3] cited by Mukherjee et al. Although secretion burden was not documented objectively, it was taken into consideration when assessing weaning readiness.
We do not understand the meaning of the third question. We have already addressed the issue of dobutamine. Moreover, the study cited by Mukherjee et al. focused on a treatment strategy, rather than the diagnosis, for heart failure [
4]. Thus, it is not related to our study.
The fifth question is also difficult to understand. Tachypnea, thoracoabdominal asynchrony, and lack of improvement in oxygenation are signs of extubation failure rather than predictors thereof. They are not mentioned at all in the cited paper [
5].
Finally, we have also mentioned the study of Roca et al. [
6], which measured inferior vena cava (IVC) collapsibility to show a potential benefit of HFNC. However, measuring physiological parameters such as IVC collapsibility and confirming beneficial effects of HFNC were not the focus of our study.
Our study started with the recognition that the definition of high-risk extubation patients is broad, heterogeneous, and inconsistent and that it is not practical to apply noninvasive ventilation (NIV) in all cases. A recent study has shown that the beneficial effect of NIV in high-risk patients is dependent on obesity status [
7]. Further studies are warranted to determine the appropriate level of postextubation respiratory support according to patient characteristics, which will improve patient outcomes and enable efficient use of intensive care resources.