Background Duloxetine is effective for chronic musculoskeletal and neuropathic pain, but there are insufficient data to recommend the use of antidepressants for postoperative pain. The authors hypothesized that administration of duloxetine for 15 days would reduce pain with ambulation at 2 weeks after total knee arthroplasty. Methods In this triple-blinded, randomized, placebo-controlled trial, patients received either duloxetine or placebo for 15 days, starting from the day of surgery. Patients also received a comprehensive multimodal analgesic regimen including neuraxial anesthesia, epidural analgesia, an adductor canal block, meloxicam, and oxycodone/acetaminophen as needed. The primary outcome was the pain score (0 to 10 numeric rating scale) with ambulation on postoperative day 14. Results One hundred six patients were randomized and analyzed. On day 14, duloxetine had no effect on pain with ambulation; mean pain was 3.8 (SD, 2.3) for placebo versus 3.5 (SD, 2.1) for duloxetine (difference in means [95% CI], 0.4 [−0.5 to 1.2]; P = 0.386). Symptoms potentially attributable to duloxetine discontinuation at study drug completion (nausea, anxiety) occurred among nine patients (duloxetine) and five patients (placebo); this was not statistically significant (P = 0.247). Statistically significant secondary outcomes included opioid consumption (difference in mean milligram oral morphine equivalents [95% CI], 8.7 [3.3 to 14.1], P = 0.002 by generalized estimating equation) over the postoperative period and nausea on day 1 (P = 0.040). There was no difference in other side effects or in anxiety and depression scores. Conclusions When included as a part of a multimodal analgesic regimen for knee arthroplasty, duloxetine does not reduce subacute pain with ambulation.

This study was registered at Clinicaltrials.gov, identifier NCT02506660.

Background: General anesthesia with neuromuscular blockade may facilitate total shoulder arthroplasty but appears to increase risk of cerebral oxygen desaturation. Cerebral desaturation is undesirable and is a proxy for risk of stroke. Purposes/Questions: This study tested the hypothesis that cerebral oxygen desaturation occurs frequently during general anesthesia with neuromuscular blockade and positivepressure ventilation but does not occur with spontaneous ventilation. Correlations were sought among cerebral oxygen saturation, blood pressure, and cardiac index. Methods: We designed a prospective, observational, cohort study to measure cerebral oxygenation in 25 patients during general anesthesia, both with and without positive-pressure ventilation. Patients undergoing elective shoulder arthroplasty in the sitting position received an arterial catheter, near-infrared spectroscopic measurement of cerebral oxygenation, and non-invasive cardiac output measurement. Moderate hypotension was allowed. Blood pressure was supported as needed with ephedrine or low-dose epinephrine (but avoiding phenylephrine). Hypercapnia (45 to 55 mmHg) was targeted during positive-pressure ventilation. Results: No cerebral oxygen desaturations occurred, regardless of ventilation mode. Under positive-pressure ventilation, the median (interquartile range: Q1, Q3) cerebral oxygenation was 110% of baseline (104, 113), the mean arterial pressure was 62% of baseline (59, 69), and the cardiac index was 82% of baseline (71, 104). Cerebral oxygenation did not correlate with blood pressure or cardiac index but had moderate correlation with end-tidal carbon dioxide. No strokes occurred. Conclusions: There were no signs of inadequate brain perfusion during general anesthesia using paralytic agents. Positive-pressure ventilation with moderate hypotension in the sitting position does not endanger patients, in the context of moderate hypercapnia and hemodynamic support using ephedrine or epinephrine.