ABSTRACTPersistent opioid use after surgery is a common morbidity outcome associated with subsequent opioid use disorder, overdose, and death. While phenotypic associations have been described, genetic associations remain unidentified. Here, we conducted the largest genetic study of persistent opioid use after surgery, comprising ~40,000 non‐Hispanic, European‐ancestry Michigan Genomics Initiative participants (3198 cases and 36,321 surgically exposed controls). Our study primarily focused on the reproducibility and reliability of 72 genetic studies of opioid use disorder phenotypes. Nominal associations (p < 0.05) occurred at 12 of 80 unique (r2 < 0.8) signals from these studies. Six occurred in OPRM1 (most significant: rs79704991‐T, OR = 1.17, p = 8.7 × 10−5), with two surviving multiple testing correction. Other associations were rs640561‐LRRIQ3 (p = 0.015), rs4680‐COMT (p = 0.016), rs9478495 (p = 0.017, intergenic), rs10886472‐GRK5 (p = 0.028), rs9291211‐SLC30A9/BEND4 (p = 0.043), and rs112068658‐KCNN1 (p = 0.048). Two highly referenced genes, OPRD1 and DRD2/ANKK1, had no signals in MGI. Associations at previously identified OPRM1 variants suggest common biology between persistent opioid use and opioid use disorder, further demonstrating connections between opioid dependence and addiction phenotypes. Lack of significant associations at other variants challenges previous studies' reliability.