Prolonged Opioid Use and Pain after Surgery: Reply

We are grateful to Drs. Patel and Al-Shehab for their insightful response1 and Dr. Chelly’s thoughtful commentary2 regarding our recent publication on the multicenter study of prolonged opioid use and pain outcome after surgery.3 We appreciate their thoughtful and constructive critique and welcome this opportunity to address the issues they have raised.Dr. Patel et al.’s first concern pertains to the variety of surgical interventions involved in our study, which they point out may cast doubts on the validity of our findings. They suggest subgroup analyses to help clarify the correlation between the type of surgical intervention and persistent opioid use. We would like to point out that we have indeed factored in the type of surgery within our logistic multivariate models. No identifiable association was found between the surgery type and opioid use at 3 months postsurgery. This observation is further substantiated by the posterior distribution, illustrated in figure 2 of our article, which reveals a low probability direction. Given these findings, we believe that additional subgroup analyses would not provide further or clearer insights.Second, the authors point out that our article does not explore the impact of preoperative opioid dosage on continued opioid use. We agree that such an exploration would have been valuable. In fact, we incorporated dosage-related questions into our questionnaires for this very reason. Regrettably, despite our confidence in detecting whether opioids were used or not, the limitations in the quality of dosage data, coupled with high variability, did not allow us to extract quantitative, comparable, or standardized data regarding the level of opioid use. Hence, the question of whether those patients who were still using opioids 3 months postsurgery were using them at lower amounts cannot be answered with the data we were able to collect. In future research, the use of state-level prescription monitoring data might help in getting information on detailed preoperative opioid prescriptions.Last, the relationship between pain severity and continued opioid use was highlighted by the authors as an area in need of assessment. We concur that a more comprehensive quantitative analysis of pain severity in correlation with postoperative opioid use would have been beneficial. However, due to limitations in sample size and difficulties in extracting detailed opioid dose data, we opted to limit the pain analysis to a binary variable, assessing only whether pain was present or absent.Dr. Chelly suggests that focusing on a population with Patient-Reported Outcomes Measurement Information System scores greater than 1 SD above the mean could have afforded a more optimal analysis of the relationship between psychosocial factors and prolonged opioid use. Following that suggestion, we have taken another look at our data and found that only 85 of the 680 subjects exhibited a Patient-Reported Outcomes Measurement Information System depression score of one SD above the mean (i.e., of 60 or above), with a median (interquartile range) in these subjects of 63.4 (63.4 to 67.3). While Dr. Chelly’s suggestion might enable a more robust assessment in general, in this particular instance, the significant reduction in sample size and the low variability would likely hinder, rather than enhance, our ability to assess the relationship between psychosocial factors and prolonged opioid use.Dr. Chelly also comments on the omission of catastrophizing and sleep disturbance data from tables 1 and 2. In planning our study, we carefully selected the variables to be included based on existing literature and guided by discussions with clinicians and researchers experienced in opioids, pain management, and surgical anesthesia. The goal was to incorporate factors with a potential to impact the outcomes we were investigating, bearing in mind the limitations of our anticipated sample size. Unfortunately, due to constraints in our achieved sample size, we needed to limit the number of variables included in our analysis to avoid overfitting and the risk of modeling random noise in the data. As a result, we did not include catastrophizing and sleep disturbance data in our analysis; hence, their absence from tables 1 and 2. Future studies with larger sample sizes could accommodate these variables for a more comprehensive understanding. Unfortunately, in this instance, we had to prioritize and could not include them.We once again thank Drs. Patel, Al-Shebab, and Chelly for their constructive comments and hope that our responses further elucidate the rationale behind our methodologic decisions and analysis approach.The reply was supported in part through computational resources and services provided by Advanced Research Computing, a division of Information and Technology Services at the University of Michigan, Ann Arbor, Michigan. The crafting of this response involved the use of generative artificial intelligence. The authors of the response carefully drafted a detailed bullet-pointed outline of the response and then instructed ChatGPT 4.0 to create the reply based on that outline (i.e., generative artificial intelligence was not used for creating content, but solely for writing assistance). The draft created by ChatGPT was edited, reviewed by all authors for content and style, and checked for plagiarism using Scribbr (result: “Low Probability of Plagiarism”).Support was provided solely from institutional and/or departmental sources.Dr. Domino declares no competing interests. She reports research funding to the institution provided by Mathematica (Princeton, New Jersey) and Edwards Lifesciences (Irvine, California), unrelated to this project. Dr. Kuck declares no competing interests. He reports research funding, unrelated to this work, from the National Science Foundation (Alexandria, Virginia), Foundation for Anesthesia Education and Research (Schaumburg, Illinois), National Institute of General Medical Sciences (Bethesda, Maryland), Medtronic (Minneapolis, Minnesota), and Dynasthetics LLC (West Valley City, Utah). He also reports equity interest, unrelated to this work in KSCube, LLC (Park City, Utah), CKC Medical, LLC (Park City, Utah), and Atapir Inc. (San Rafael, California), and financial interest, unrelated to this work, from patents assigned to Drägerwerk (Lübeck, Germany) and University of Utah Research Foundation (Salt Lake City, Utah). Dr. Posner declares no competing interests. She reports research funding, unrelated to this work, to the institution provided by Mathematica, Microsoft Research (Redmond, Washington), American Society of Anesthesiologists (Schaumburg, Illinois), Society for Anesthesia and Sleep Medicine (Lisle, Illinois), Anesthesia Patient Safety Foundation (Rochester, Minnesota), and Anesthesia Quality Institute (Schaumburg, Illinois); and travel and speaker funding, unrelated to this work, to the author from The Doctors Company (Napa County, California). Dr. Saager declares no competing interests. He reports funding, unrelated to this work, from The Surgical Company (Amersfoort, The Netherlands). Dr. Pace declares no competing interests. He reports consulting fees, unrelated to this work, from Elute, Inc. (Salt Lake City, Utah). Dr. Naik declares no competing interests.