First, Bhakta asserts that we used an insufficient dose of oral clonidine. Investigators designing a clinical trial need to balance selecting a high dose that maximizes efficacy and a lower dose that reduces the risk of complications. Many previous (small) studies suggest that the dose we used, and even lower ones, reduce pain and opioid consumption.2–4 Our very much larger trial clearly shows that clonidine (0.2 mg oral clonidine followed by 0.2-mg/d transdermal clonidine patch) does not produce meaningful analgesia. Given that the dose we used produced clinically important hypotension and bradycardia (ie, requiring intervention), we are amazed that anyone would seriously propose using a higher dose just for its putative analgesic effect.
Second, Bhakta asserts that the oral route has erratic absorption and the transdermal route is unreliable. Per protocol, an initial oral dose was accompanied by simultaneous application of a transdermal clonidine patch. Oral clonidine is rapidly and reliably absorbed, especially in patients who have not recently eaten, and provided a therapeutic plasma concentration until the transdermal drug became available.5 Specifically, oral clonidine reaches peak serum concentrations within 2 to 4 hours with therapeutic concentrations persisting 24 hours.5 Transdermal clonidine demonstrates physiologic effects at 24 hours and thereafter provides steady plasma concentrations so long as the patch remains in place.6 The combination of oral and transdermal clonidine thus provided therapeutic concentrations of clonidine from several hours after administration until 3 to 4 days thereafter—and thus throughout the period of greatest pain (and cardiac risk).
Third, Bhakta asserts that we did not clearly identify which transdermal patch was used. In fact, the type was fully specified in our separate methods article that was referenced in the primary result publications.7 The patch was Catapres-TTS, a transdermal therapeutic system by Boehringer Ingelheim. The pharmacokinetics of this type of clonidine patch are well established.
Fourth, Bhakta asserts that clonidine should have been given intravenously. For reasons documented, IV administration was unnecessary and certainly would have complicated the study.
Fifth, Bhakta notes that clonidine is not a first-line analgesic and that using it alone for immediate postoperative pain is not a good idea. Of course, that is the same as saying that the drug is not particularly effective—which is perfectly consistent with our conclusion that the drug provides no meaningful analgesia. In fact, clonidine was not used as the sole analgesic. Instead, clonidine administration was blinded and standard institutional acute pain management was provided to all participating patients. Opioid doses were presented in the published article.
Sixth, Bhakta asserts that heterogeneous types of surgery led to marked variation in the amount of postoperative pain. We remind him that the primary outcome of POISE-2 was myocardial infarction.8 Patients were thus selected for cardiovascular risk rather than expected postoperative pain. However, in a study of 10,000 patients, we can be sure that any clinically important analgesic effect of clonidine would be easily detected. None was.
Seventh, Bhakta cites a Cochrane Review by Lambert et al.9 That review was restricted to a pediatric patient population and included only 11 studies (742 patients), only 4 of which compared clonidine with a placebo or no treatment. Lambert et al9 simply do not support the assertion of Bhakta. He might better have cited our article that included an update to the relevant 2012 meta-analysis by Blaudszun et al.10 The meta-analysis results, Figure 3 in our article, clearly show that clonidine is not analgesic.
Eighth, Bhakta asserts that our conclusions are overreaching. In fact, our very large study was extremely well powered to detect even tiny analgesic effects of clonidine. There was none.