Meier et al.1 contend that our trial poorly tested the effect of multimodal analgesia because we did not control intraoperative local anesthetic use or postoperative analgesia. Epidural analgesia is rarely used for spine surgery in our setting. As presented in Table 1, fewer than 1% of our patients had epidural analgesia.2 Furthermore, local wound infiltration was used only in one quarter of the patients in each group: analgesic pathway 39 (26%) versus 32 (22%) in placebo.2 In a robust double-blind randomized trial (n = 299), there is no reason to expect substantive differences between treatment and placebo groups.
Meier et al. write that “measuring the Quality of Recovery score 3 days after surgery, as well as other postsurgical pain measures, is confounded when both the experimental and control groups received multimodal analgesics during and after surgery.” Confounding—by definition—is restricted to factors that influence both exposure and outcome. Randomization usually prevents confounding; but in any case, an intervention after a blinded exposure cannot be a confounder. What Meier et al. presumably mean is that postrandomization treatments might influence outcomes. We agree, but the fact that patients in each group consumed nearly the same amounts of various analgesics during the initial postoperative days is not a limitation; instead, it confirms that the four combined treatments we tested are ineffective.
A reasonable question is whether Quality of Recovery 3 days after surgery is a suitable primary outcome. Granted, 3 days is distal to the interventions which were largely intraoperative. But Quality of Recovery is a well-validated, patient-centered outcome3 focused on assessing patient pain and comfort level. Proponents of enhanced recovery or multimodal analgesic pathways presumably believe that their interventions noticeably improve outcomes that patients can detect. Clearly, the four drugs we tested did not. In any case, our predefined secondary outcomes were proximal, namely pain scores and opioid consumption, and neither improved.
Postoperative analgesia certainly matters, and presumably some approaches are better than others. We look forward to trials comparing various approaches, but evaluating postoperative analgesic technique was not our goal. We can, though, conclude that immediate perioperative use of acetaminophen, gabapentin, lidocaine, and ketamine—all of which act via separate pathways—contributes little. Future trials might therefore concentrate more on postoperative treatments or different analgesic regimens for perioperative analgesia.
Our results, while robustly equivocal, are nonetheless valuable since all drugs impose risk and cost and should therefore only be used if they are actually effective in a given context. We stand by our conclusion that “an analgesic pathway based on preoperative acetaminophen and gabapentin, combined with intraoperative infusions of lidocaine and ketamine, does not improve recovery in patients recovering from multilevel spine surgery.”
The authors declare no competing interests.