Narcotic sparing postoperative analgesic strategy reduces length of stay after pancreatoduodenectomy: Analysis of practice patterns for 1004 patients

Abstract

Background: Improved post-operative outcomes have been demonstrated in gastrointestinal (GI) procedures where a narcotic sparing strategy has been utilized. Data for pancreaticoduodenectomy (PD) patients is limited, and currently a clinical trial (PAKMAN) underway in Germany is still recruiting patients to determine the impact of intravenous versus epidural analgesia to reduce the incidence of GI complications after elective pancreatoduodenectomy. The surgeon preference driven approach to post-operative analgesia has not been studied extensively for PD. Our center has historically utilized epidurals liberally in the postoperative period and we reviewed our postoperative outcomes over the past decade based on initial analgesic strategy employed. Methods: Demographic, operative, and post-operative data for patients at Emory University between 2010-2017 was reviewed. 1004 consecutive patients who underwent elective PD at our center in the study period were included. Patients were divided into groups based on primary analgesic strategy (opiate based-patient controlled analgesia first (PCA), thoracic epidural analgesic administration first (EPI), simultaneous-PCA/EPI (dual-PCA/EPI) use and neither (non-PCA/EPI). Data for all analgesic doses was extracted from medication administration record for each patients' entire length of stay (LOS). All oral and parenteral opiate doses were converted to parenteral morphine equivalent doses for comparative analysis. Postoperative outcomes for each group were analyzed and multivariate linear regression analysis was performed generate beta-coefficients for predictors of LOS specifically. Results: 448 (44.6%) patients were treated with EPI as primary modality of pain control compared to 300 (29.9%) with PCA, 176 (17.7%) with non-PCA/EPI and 78 (7.8%) with dual-PCA/EPI. Patients with chronic pain and chronic pancreatitis were more likely to have EPI (82%, n = 89 and 69%, n = 88) as part of primary modality of analgesia with dual-PCA/EPI use being 3.36 and 1.44-fold higher. The EPI patients received on average 16.8 mg/day (±28.9) of parenteral morphine equivalent compared to 13.7 mg/day (±17.0) for non-PCA/EPI, 30.7 mg/day (±28.4) for PCA and 65.3 mg/day (±98.1) for dual-PCA/EPI (p < 0.001). Non-PCA/EPI had significantly higher number of patients receiving ketorolac (57.3%) and oral NSAIDs (17.4%) (p < 0.001) as well as a higher ketorolac dose (128.4mg, ±69.4) compared to the other three groups (p < 0.001). 66 (14.7%) patients had failure of pain control with EPI necessitating a PCA, compared to 8 (2.6%) and 2 (1.1%) failures for the PCA and non-PCA/EPI patients (p < 0.001). Mean LOS for PCA group was significantly longer (11.44 days, ±12.3) compared to EPI (9.1, ±6.8), non-PCA/EPI (9.2, ±8.6) and dual-PCA/EPI (9.2, ±4.5) (p < 0.001). On univariate analysis increased BMI (p = 0.014), male gender (0.008), advanced age (p = 0.025), lack of EPI (p = 0.022), PCA use (p < 0.001), VTE (p < 0.001), POPF (p < 0.001) and Ileus/DGE (p < 0.001) were all correlated with increased LOS. On multivariate linear regression, VTE (b-coefficient 9.07, p = 0.004) POPF (8.846, p = 0.001), Ileus/DGE (4.464, p = 0.004), PCA (1.75, p = 0.003) and chronic pain (0.434, p = 0.019) were associated with a significantly increased LOS. When patients with POPF, chronic pain, initial analgesic failures and dual-EPI/PCA are excluded from this model, development of VTE (10.13, p = 0.010), Ileus/DGE (5.11, p = 0.011) and PCA use (2.097, p = 0.025) continued to be directly correlated with increased LOS. Conclusion: Our results suggest a primary narcotic sparing strategy (utilizing epidural or predominant combination of non-opiate parenteral-oral analgesia) led to significantly reduced LOS. The overall effect on increased LOS is predominantly driven by POPF, Ileus/DGE and VTE in our dataset with PCA use standing out as the only non-traditional measure. At the very least PCA use added 1.75 days to LOS based on our unadjusted linear modelling. On an adjusted linear regression excluding patients who have conventionally required high opiate use (POPF, chronic pain, initial analgesic failure, dual-EPI/PCA), an opiate PCA analgesic strategy continued to remain vastly inferior to epidural or non-PCA/EPI strategy and was associated with an additional 2.1 days to LOS. These data provide further evidence for utilizing alternatives to opiate-based PCAs after PD.