Significant attention, time, and research have been devoted to the reduction of opioid consumption, both in the context of chronic pain treatment and for perioperative use. The 2023 Lancet Commission was dedicated to understanding the causes of the opioid crisis and headlining efforts to curb opioid-related mortality.1 Similarly, a consensus report from the Perioperative Quality Initiative and the Enhanced Recovery After Surgery (ERAS) Cardiac Society provided opioid stewardship guidelines geared specifically to reduce perioperative opioid administration to cardiac surgery patients.2 Despite published ERAS protocols and increased awareness of the adverse effects of overly generous opioid administration in North America, there continues to be significant variation in hospital opioid utilization across the United States. In this month’s issue of Anesthesia & Analgesia, Fisher and colleagues3 deliver insights from an extensive observational analysis of intraoperative opioid administration in adult cardiac surgeries across 30 unique US hospitals from 2014 to 2021. The authors highlight significant variation in intraoperative opioid use for cardiac surgery among institutions and between anesthesiologists, concordant with findings from a prior Multicenter Perioperative Outcomes Group (MPOG) analysis that included all types of surgical procedures.4
Fisher and colleagues3 identified a sample of 59,463 unique nonemergency cardiac cases (isolated coronary artery bypass grafts [CABGs], isolated valves, and combined CABG/valves) that reliably followed a similar anesthetic trajectory punctuated by a pre- and postbypass period. A total of 794 different anesthesiologists provided clinical care to the included patients. Exclusion criteria included American Society of Anesthesiologists (ASA) physical status classification of 5 or greater, extremes of weight (<20kg or >150 kg; BMI <10 or >50) or height (<140 cm or >200 cm), preexisting endotracheal tube in place suggesting a repeat surgical encounter, cases requiring mechanical circulatory support or hypothermic circulatory arrest, and cases in which methadone was administered. The authors found that in 82.5% of cases, fentanyl was the primary opioid used, with a mean dose of 1139 fentanyl microgram equivalents (95% confidence interval [CI], 1132–1146) after adjusting for anesthetic duration, year of surgery, weight, and patient demographics. Every additional 10 minutes in the operating room was associated with an additional 5.1 to 5.9 average fentanyl microgram equivalents administered. In addition to the strikingly high mean fentanyl dose, the observed range of intraoperative opioid administration varied from 0 to >10,000 fentanyl microgram equivalents for all case types and durations. Clinical factors associated with more intraoperative opioid use were any use of sufentanil (r = 0.64), the specific use of a sufentanil infusion (r = 0.81), or any implementation of an opioid infusion (r = 0.65).
Traditionally, opioid-heavy cardiac anesthetics were valued for their combined hemodynamic stability and analgesic effect. More recently, the drawbacks related to opioid side effects, especially in the context of accelerated clinical recovery, have gained greater appreciation.5 Accelerated recovery protocols, commonly referred to as “fast-track cardiac anesthesia” (FTCA), aim to provide analgesia while limiting recovery-prolonging side effects, especially respiratory depression, which can prolong postoperative mechanical ventilation. In cardiac surgical patients for whom early extubation and mobility are prioritized, high-dose intraoperative opioid administration may counteract this “fast” recovery. On the other hand, eliminating or drastically minimizing intraoperative opioids can lead to increased pain, especially with respiration, paradoxically increasing postoperative opioid use and possibly chronic postoperative pain.6 Lastly, a one-size-fits-all approach that dictates a single dose or approach for all patients is untenable, given the variability in postoperative pain seen among patients. All this sets the stage for the need for better evidence to personalize analgesic regimens for cardiac surgery patients.
Currently, intraoperative opioid utilization is often neither standardized across all surgeries nor individualized based on individual patient factors. Instead, it is predominantly influenced by institutional culture, as the authors point out.3 Interestingly, major aspects of care controlled by the anesthesiologist, ranging from peripheral nerve block administration7 to preoperative consultation,8 are driven by institutional practice patterns or surgical clinician preference rather than patient-specific factors. Certain substance use best practices, including alcohol withdrawal management protocols or personalized discharge opioid prescriptions based on in-hospital requirements, are arguably suitable for universal widescale implementation.9 However, intraoperative opioid administration evades standardization, as many factors influence it, and unexplained variability still exists in previous models.3
A significant limitation of any retrospective analysis is the limited indexing of the potentially influential factors to explain the variability observed. Fisher and colleagues observed considerable variability in intraoperative opioid utilization and, remarkably, found that 60% of it could be explained by institutional and individual anesthesiologists, calculating the median absolute difference derived from the institution and anesthesiologist as random Gaussian variables.3 However, this left a substantial portion of the variability unaccounted for by the institution, anesthesiologist, or other case-level factors. The hierarchical modeling enabled a comprehensive analysis that also accounted for some aspects of patient health status, case year, surgical duration, and demographic variables that might inform a clinician when implementing a fast-track protocol. However, these factors only accounted for 6% of the observed variation in intraoperative opioid administration. The causal diagram (Supplemental Figure 3) in their paper enumerates the complex interrelationships between an anesthesiologist nested within an institution and the resultant amount of opioid administration, with anesthetic technique and involvement of trainees as mediators, and a multitude of other measured (year, duration, patient factors) and unmeasured variables (background and surgeon) as confounders or colliders. Importantly, clinician and hospital case volumes were not considered. These may be pertinent variables, as prior multicenter analyses have found associations between low clinician cardiac case volume and a decreased likelihood of benzodiazepine administration10 and more significant inotrope utilization variability.11 In the authors’ analysis, institution cardiac case volumes varied from 29 to 4000 during the period evaluated, reflecting significant hospital-level case variability.3
An advantage of including multiple centers in a systematic approach, though, is that it allows assessment of variation in practice, adding to the knowledge about anesthesiology practice variation in the past decade.4,7,8 Historically, the lack of widespread data integration of relevant intraoperative variables limited such acquisition and analysis. Now, multisite collaborative registries, such as MPOG, provide the opportunity to more granularly assess a multitude of anesthesiologists’ practices when facing similar clinical tasks. Well-indexed, high-volume data of this nature has the potential to serve as the basis for tailored intraoperative treatment algorithms and may iteratively inform their improvement through ongoing machine-learning feedback.12 In the face of a rapidly shifting opioid prescription landscape, such real-time approaches could provide prompt actionable changes to clinical practice. In the future, the changes in prescribing and administration patterns could be leveraged as a natural experiment to assess meaningful clinical outcomes.
Interestingly, the link between intraoperative opioid regimens and postoperative pain is not always intuitive or straightforward. One might expect that greater opioid use would be associated with lower pain scores. However, in a prospective mixed surgical cohort study evaluating 360 patients, intraoperative opioid dose (median [interquartile range {IQR}, 3–1] 60 [70.5]) was not associated with postoperative pain trajectories.13 More commonly, pain and opioid use are actually positively associated, as seen in several prospective observational studies in noncardiac postoperative patients.14,15 Decreasing opioid utilization in cardiac cases may, therefore, have benefits, as demonstrated in a Cochrane review of 28 randomized controlled trials comparing fast-track cardiac care with conventional treatment, where the authors found that lower-dose opioid anesthesia (≤20 µg/kg fentanyl) was associated with reduced time to extubation and shorter length of intensive care unit stay but had no effect on total hospital length of stay.16
The opioid epidemic and its public health implications impact the “opioid ecosystem” and are relevant to the context of this study. As the authors address, intraoperative opioid use for cardiac surgery has decreased over the past decade by an average of 478 fewer fentanyl microgram equivalents than those performed in 2014. Even so, 12.6% of the cases in this analysis received >20 fentanyl microgram equivalents per kilogram, a well-cited threshold for high-dose opioid anesthesia.2 Use of non-opioid analgesics for cardiac surgery, including novel regional techniques (eg, parasternal block), is welcome and likely beneficial to patients, independent of their opioid-lowering potential. Future investigation of the benefits of non-opioid adjuvant analgesic approaches should include determination of optimal timing, dosing, and cost.2 The authors examined how regional anesthetics and multimodal pain strategies influenced opioid administration variation. Interestingly, coanalgesics were frequently administered, but they did not appreciably affect the amounts of opioids administered.
Fundamentally, opioids are used for their antinociceptive effect to decrease stimulation incited by surgical incision, sternal retraction, and chest tube placement and suppress the accompanying sympathetic activation. Preemptively blocking this activation intraoperatively presumably also impacts postoperative pain, which is most pronounced in the first 2 postoperative days after cardiac surgery. Notably, the extent and persistence of pain likely varies based on unique patient characteristics. Previous literature has identified younger age, female sex, anxiety, depression, negative affect, and greater preoperative temporal summation of pain as risk factors for postoperative pain and opioid use after both general and cardiac surgery.15,17 Previous studies suggest that postoperative pain trajectories are related more to patient characteristics, including various pain-relevant modulatory factors, and less defined by surgical factors.13 Hence, intraoperative opioid administration based on person-level variables may be better suited to address intraoperative nociception and postoperative pain rather than protocols based on surgery type, institutional culture, or clinician preference.
REFERENCES