In Response

    loading  Checking for direct PDF access through Ovid


We thank Grocott1 for his comments regarding our article on aerosolized vasodilators for the treatment of pulmonary hypertension.2 Our original primary outcome and our research question were indeed mortality and the extent to which the use of inhaled agents were associated with a reduction in mortality. However, when a meta-analysis is performed, it is difficult to know in advance if there will be sufficient studies and patients to answer the question. Power analysis is mostly performed for clinical trials not meta-analysis because we have limited control over what we will find in the literature. Therefore, we could only confirm that there were insufficient data to answer this important question. Consequently, we could analyze only those data that were made available from the studies we found suitable for the analysis. Those data were mainly hemodynamic and echocardiographic.
The Bonferroni correction for multiple comparisons did not include effects on mortality and length of stay in hospital and in intensive care unit as we only considered correcting for the various hemodynamic parameters that were grouped under the broader secondary outcome assessing the hemodynamic profile; hence, we controlled for the family-wise error rate of these effects only. Furthermore, the Bonferroni correction was not adjusted by a multiple of 2 for the comparisons against placebo and intravenous agents because these comparisons were treated as separate studies, and therefore, the meta-analyses were performed separately. Combining studies comparing against placebo and intravenous agents would have introduced a high level of heterogeneity in the analysis, and for this reason, they were analyzed as 2 distinct studies. Nevertheless, as mentioned by Grocott,1 the overall results of our analyses would remain the same had we additionally adjusted the Bonferroni corrections as suggested. We agree that although pulmonary vascular resistance and mean arterial pressure were statistically significant at a P level of significance <.05, when correcting for multiple testing errors using a corrected P value, only results for right ventricular ejection fraction remained significant. As mentioned in our discussion, mean arterial pressure and pulmonary vascular resistance are not statistically significant at this corrected P value.
Our conclusion is somehow similar to the literature and the experience on nitric oxide alone that has definitively favorable hemodynamic effect but no clear effect on clinical outcome and mortality.3 One of the explanations that we observed in the multicenter clinical trial on inhaled milrinone is the interindividual response to inhaled agents.4 Despite a modest but favorable hemodynamic effect of inhaled milrinone compared to placebo, we observed variable drug levels for the same amount of drug administered. There was also a direct correlation between the hemodynamic effect with the observed blood concentration. We are currently exploring other therapeutic alternatives in pulmonary hypertension and right heart failure such as combined inhaled agents because we find it more effective.5,6
In addition, the benefit of inhaled agents might be offset with other interventions such as fluid management, which in our study and many others reported, has not been controlled. Therefore, it is unlikely that the use of a single strategy, in a situation as complex as right ventricular dysfunction, will be the final answer. We agree that the information we provided is just a small step, as mentioned by Grocott,1 “in our understanding of its significance.” We fully agree that “it does remain a potentially fruitful subject for further investigation.
    loading  Loading Related Articles