A high significance level after analysis of covariance in a small-group study?

To the Editor: I read with great interest the article by Bellani et al. and praise their important work in the field of weaning from mechanical ventilation. More than anything, I question how the authors formed the hypothesis that oxygen consumption (V̇O2) increases more in patients unable to sustain decreasing ventilatory assistance. In a landmark article by Jubran et al., weaning failure was associated with increased oxygen extraction and decreased oxygen delivery. In the same article, the measured V̇O2 increased in both the success and failure from weaning groups, with a lower increase in the success group. In contrast, Zakynthinos et al. demonstrated that patients who cannot be weaned have one of two hemodynamic and oxygen use profiles. (1) Those who fail without increasing V̇O2 demonstrate increased oxygen extraction and decreased oxygen delivery. (2) In those who fail and increase their V̇O2, the increase mainly occurs secondary to increased oxygen extraction. Direct measures of mixed venous oxygen saturation are increased in the first group and decreased in the second group, supporting their findings. Given the complex physiologic nature of respiratory weaning and weaning failure, it is widely believed that failure to wean occurs secondary to decreased oxygen delivery and increased oxygen extraction. Given the proposal of 1870 by Fick, a decrease in cardiac output in combination with an increase in the arteriovenous oxygen content difference would yield a relatively stable V̇O2. Combining these data with those of Bellani et al., it is clear that the weaning process is complex and highly variable between patients. Overall, this work supports previous studies demonstrating that there are patients who fail weaning in the absence of increased V̇O2.