Opioid tolerance or opioid withdrawal?

To the Editor: We thank Kim et al. for their recent work examining the effect of intraoperative remifentanil on postoperative analgesia in children.1 Certainly, the pharmacokinetic profile of this potent opioid is attractive in modern anesthetic practice, and a thorough understanding of the principles behind tolerance, hyperalgesia, and withdrawal of remifentanil are critical for its safe administration. The authors correctly define acute tolerance as a “progressive decrease in response to a drug which can be overcome by increasing the dose of the drug.” They conclude that remifentanil doses of 0.6–0.9 μg·kg−1·min−1 induce acute tolerance for 24 h after surgery based on evidence of increased postoperative fentanyl consumption. This conclusion does not seem entirely accurate based on the data presented. The study methodology called for fixed dosing of remifentanil with abrupt discontinuation of the infusion when the surgical procedure ended. Despite this fixed dosing, their figure 1 demonstrates no hemodynamic evidence of waning remifentanil effect, and the article provides no evidence of increased volatile anesthetic requirements with time. Therefore, the requisite decreased response to or increased drug requirement for remifentanil remains unproven, and the occurrence of acute tolerance cannot be confirmed. Another explanation for the authors’ observations would be withdrawal from remifentanil, especially considering the study methodology called for abrupt discontinuation of the experimental infusion at the end of surgery. Administration of a bridging dose of opioid, an established technique for mitigating the effects of remifentanil withdrawal, was not allowed in this protocol.2 In fact, patients did not receive any other analgesia until emergence from their anesthetic. Unfortunately, the authors did not collect clinical endpoints specific to opioid withdrawal including agitation or hemodynamic changes in the immediate postoperative period, and no consideration of this diagnosis is obvious in the article. Furthermore, figure 2 of the article demonstrates that pain scores were only significantly different between the control and intervention groups during the first postoperative hour. Unfortunately, the authors only presented total fentanyl consumption at the 24and 48-h time points; therefore evaluation of fentanyl consumption at earlier time points cannot be derived from the article. It is entirely feasible that opioid requirements were really increased only early in the postoperative course for those receiving remifentanil until acute withdrawal had resolved via infusion of fentanyl. Although the authors argue that subanesthetic doses of volatile anesthetic have little effect on analgesia or hyperalgesia, they overlook the potential for continued postoperative sedation in the control group as a result of higher volatile anesthetic exposure during the intraoperative period. Data regarding postoperative sedation are not presented in the article. This potential confounder could also account for the finding of lower pain scores in the control group during the first postoperative hour. In 2010, this same group published results of parentor nurse-controlled analgesia in the same surgical population and used these data as the basis for their power analysis in the current article.3 For patients receiving only fentanyl for postoperative analgesia, typical fentanyl consumption was 18.1 ± 4.6 μg·kg−1·day−1 at 24 h and 16.6 ± 5.5 μg·kg−1·day−1 at 48 h. In the current study, subjects randomized to receive the highest remifentanil dosing strategy (0.9 μg·kg−1·min−1) actually consumed less fentanyl than historic controls (17.8 ± 3.6 μg·kg−1·day−1 at 24 h, 9.2 ± 2.1 μg·kg−1·day−1 at 48 h). This inconsistency further highlights that the observed differences in postoperative opioid consumption were probably related more to the abrupt discontinuation of and subsequent withdrawal from remifentanil rather than acute tolerance to μ-agonists. Finally, we question the approval of such a research protocol in a population of subjects that cannot themselves consent for participation. Although we understand the need to conduct research in children, this protocol seems to put both control and experimental subjects at risk of significant pain. The authors acknowledge that ureteroneocystostomy is associated with “moderate to severe pain,” yet they allowed the control group to effectively emerge with no analgesic component to their anesthetic. Although their pain scores were superior in the first postoperative hour, the potential for pain should have mandated some analgesia at emergence within the experimental protocol as these authors have reported previously.3 Furthermore, the phenomenon of withdrawal and hyperalgesia from remifentanil is well-documented, yet the absence of a bridging dose of opioid coinciding with remifentanil discontinuation placed those children in the experimental group at risk of both significant pain and the physiologic effects of acute opioid withdrawal.4 Future evaluations of the effects of remifentanil should abandon this methodology as it is inappropriate for those unable to personally consent to experience significant pain.