Article Text
Statistics from Altmetric.com
Introduction
There is currently no standard timing method used to quantify postoperative opioid consumption. Some studies define postoperative day (POD) 0 as anesthesia stop time until 23:59:59 of the date of surgery and POD 1 as beginning at midnight following the date of surgery (POD@MN).1 While this may be practical for administrative reasons, the variability in surgical procedure end time on POD 0 is a source of inconsistency in calculating postoperative opioid consumption. Other studies report opioid consumption in 24-hour intervals after anesthesia stop time (24i).2 Although this method provides more consistency, the need to track patients based on varying procedure end times is a labor-intensive endeavor. Figure 1 illustrates the POD@MN and 24i nomenclature used in this report. Lack of standardization complicates study comparisons in meta-analyses. Consequently, previous meta-analyses include summative data only from studies using the 24i method, thus excluding a substantial body of experience.3 The goal of this study was to use a single spine surgery dataset to evaluate these two methods and their impact on interpretation of postoperative opioid consumption.
Methods
This is a retrospective study of consecutive cancer-related laminectomy cases from January to November 2019 at our institution (online supplemental figure 1). The home medication list was used to determine whether a patient was opioid tolerant or naïve. Postoperative opioid drug, dosage, and time administered were collected from patient-controlled analgesia (PCA) pump records and the electronic medication administration record system. All dosages were converted to milligram morphine equivalents (MME).4 Inpatient postoperative opioid consumption was calculated in Microsoft Excel using POD@MN and 24i methods. Four comparison scenarios (POD 0 to 1 vs 0 to 24 hours, POD 0 to 2 vs 0 to 48 hours, POD 1 vs 0 to 24 hours, and POD 1 to 2 vs 0 to 48 hours) were chosen to represent common time points at which opioid consumption is measured.
Supplemental material
Statistical analyses were conducted using GraphPad Prism V.7 (GraphPad Software, San Diego, California, USA). Bland-Altman analysis5 was used to measure the bias, which is the average of values computed by one method (POD@MN) minus the value computed by another method (24i), in postoperative opioid consumption. Within-patient differences were analyzed using Wilcoxon signed-rank test. Analyses were also performed separately for opioid-tolerant and opioid-naïve cohorts.
Results
Thirty-nine of the 180 analyzed cases were determined to be opioid tolerant (online supplemental table 1). POD 0 ranged from 31 min to 13.4 hours. If POD 0 was not included in postoperative opioid consumption calculations, the average percentage of opioid missed by the end of POD 1 was 43.8% (86.5±89.0 MME). Bland-Altman analysis (figure 2) revealed a persistent bias between timing methods which increased as total opioid dose increased. There was a larger bias and wider 95% limits of agreement in the opioid-tolerant cohort, with the exception of POD 1 to 2 vs 0 to 48 hours (online supplemental table 2).
Supplemental material
Wilcoxon signed-rank test showed significant within-patient differences between the POD@MN and 24i methods at all time points tested. When analyzed according to chronic pain status, opioid-tolerant patients had larger median of differences compared with opioid-naïve patients (table 1).
Discussion
These findings suggest that POD@MN and 24i yield varying values for postoperative opioid consumption. Differences are magnified at higher opioid doses. Establishment of a standardized methodology for reporting postoperative opioid consumption increases the validity of interstudy comparisons and is especially important for studies involving high opioid use, such as arduous surgeries or opioid-tolerant patients. Our findings were limited by inconsistencies in times at which opioid administration was recorded from PCA pumps. However, these inconsistencies would be expected to have little influence as the analyses were focused on differences between POD@MN and 24i rather than total opioid consumption.
Ethics statements
Ethics approval
This retrospective study was approved by our local Institutional Review Board (Memorial Sloan Kettering Cancer Center, New York, New York, USA). The requirement for written informed consent was waived.
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @StephLam_med, @patrickmdnet
Contributors SL: This author wrote the manuscript, developed the statistical plan, performed statistical analyses, analyzed data, created the figures and tables, completed revisions, and reviewed the final manuscript. RV: This author helped to write the manuscript, developed the study concept, critically revised the manuscript, and reviewed the final manuscript. VM: This author helped with data acquisition, critically revised the manuscript, and reviewed the final manuscript. PMC: This author acquired data, helped to write the manuscript, revised the manuscript, and reviewed the final manuscript.
Funding The research was supported by the National Cancer Institute of the National Institutes of Health under Award Numbers R25CA020449 and P30CA008748. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Competing interests Dr McCormick’s spouse holds stock in Johnson & Johnson.
Provenance and peer review Not commissioned; externally peer reviewed.