Reply to Letter: Alvimopan Is Associated With Improved Outcomes and Cost Savings in Enhanced Recovery Colorectal Surgery Protocols
We appreciate the interest that Dr. Kehlet had in our study on impact of alvimopan on short-term outcomes and costs in the setting of enhanced recovery colorectal surgery.1 The study included data from 660 patients who underwent major colorectal resections within a rigorous enhanced recovery protocol. After adjustment for competing factors, alvimopan was associated with improved outcomes, shorter length of hospitalization, and cost savings. Although the study is retrospective in nature with the potential for selection bias, the large sample size permitted for adequate adjustment for known confounders, addressing many of the concerns raised by the author. In our adjusted analyses, we accounted for the effects of patients’ demographics, ASA score, indication for surgery, year of the procedure, extent of surgical resection, use of laparoscopy and epidural analgesia, and the operating surgeon. This list of cofounders included in the multivariable adjustment is comprehensive and represents known confounders. Given the adequate power of our study, we believe that the beneficial effect of alvimopan reported in the study is independent of these competing factors. In addition, our study findings are in line with randomized clinical trials that examined the effect of alvimopan after bowel resection in general.2–4
As acknowledged and discussed in the limitations section of our article, the control group included patients who were treated a year before implementation of alvimopan use in our institution. However, the Enhanced Recovery Program at our institution was well established from the start time of the study. Although we acknowledge that the continuous process improvement at engaged enhanced recovery centers does result in regular auditing and updating of practice management, our protocol changes from 2010 to 2013 were evenly distributed and impacted the entire dataset. All patients had their procedures between 2010 and 2013 with equal distribution and under the same protocol. In addition, we adjusted for year of the procedure to control for the potential for any bias related to this concern.
All patients in our cohort received perioperative opioid analgesia with or without thoracic epidural placement. The competing effect of epidural analgesia was accounted for in the multivariable models. We did not report on the independent effect of epidural analgesia, because it was not the primary or secondary question of the study. However, we found that epidural analgesia alone did not confer any improvement in outcomes or costs. This finding is corroborated by the recent clinical trial comparing epidural analgesia to intravenous opioids.5 The alvimopan group included more patients who underwent segmental and laparoscopic resections; however, these factors were controlled for in the clinical outcomes and costs analyses.
Our study represents the largest single-institution series that has examined the effect of alvimopan in the setting of an enhanced recovery protocol, permitting us to overcome limitations of previous studies that attempted to answer this question. The retrospective nature of our study does not automatically invalidate our findings. Appropriately designed retrospective studies are extremely useful and informative, especially in the absence of randomized data. We, therefore, believe that our study provides important information regarding the role of alvimopan in the setting of enhanced recovery colorectal surgery protocols and argues for its consideration.