BACKGROUND Postoperative nausea and vomiting (PONV) is a key driver of unplanned admission and patient satisfaction after surgery. Because traditional risk factors do not completely explain variability in risk, this study hypothesized that genetics may contribute to the overall risk for this complication. The objective of this research is to perform a genome-wide association study of PONV, derive a polygenic risk score for PONV, assess associations between the risk score and PONV in a validation cohort, and compare any genetic contributions to known clinical risks for PONV. METHODS Surgeries with integrated genetic and perioperative data performed under general anesthesia at Michigan Medicine (Ann Arbor, Michigan) and Vanderbilt University Medical Center (Nashville, Tennessee) were studied. PONV was defined as nausea or emesis occurring and documented in the postanesthesia care unit. In the discovery phase, genome-wide association studies were performed on each genetic cohort, and the results were meta-analyzed. Next, the polygenic phase assessed whether a polygenic score, derived from genome-wide association study in a derivation cohort from Vanderbilt University Medical Center, improved prediction within a validation cohort from Michigan Medicine, as quantified by discrimination (c-statistic) and net reclassification index. RESULTS Of 64,523 total patients, 5,703 developed PONV (8.8%). The study identified 46 genetic variants exceeding the threshold of P < 1 × 10-5, occurring with minor allele frequency greater than 1%, and demonstrating concordant effects in both cohorts. Standardized polygenic score was associated with PONV in a basic model, controlling for age and sex (adjusted odds ratio, 1.027 per SD increase in overall genetic risk; 95% CI, 1.001 to 1.053; P = 0.044), a model based on known clinical risks (adjusted odds ratio, 1.029; 95% CI, 1.003 to 1.055; P = 0.030), and a full clinical regression, controlling for 21 demographic, surgical, and anesthetic factors, (adjusted odds ratio, 1.029; 95% CI, 1.002 to 1.056; P = 0.033). The addition of polygenic score improved overall discrimination in models based on known clinical risk factors (c-statistic, 0.616 compared to 0.613; P = 0.028) and improved net reclassification of 4.6% of cases. CONCLUSIONS Standardized polygenic risk was associated with PONV in all three of the study's models, but the genetic influence was smaller than exerted by clinical risk factors. Specifically, a patient with a polygenic risk score greater than 1 SD above the mean has 2 to 3% greater odds of developing PONV when compared to the baseline population, which is at least an order of magnitude smaller than the increase associated with having prior PONV or motion sickness (55%), having a history of migraines (17%), or being female (83%) and is not clinically significant. Furthermore, the use of a polygenic risk score does not meaningfully improve discrimination compared to clinical risk factors and is not clinically useful. EDITOR’S PERSPECTIVE

中文翻译：

---

预测术后恶心和呕吐的多基因评分：一项回顾性推导和验证队列研究。


背景 术后恶心呕吐 （PONV） 是术后意外入院和患者满意度的关键驱动因素。由于传统风险因素并不能完全解释风险的可变性，因此本研究假设遗传学可能导致这种并发症的总体风险。本研究的目的是进行 PONV 的全基因组关联研究，得出 PONV 的多基因风险评分，评估验证队列中风险评分与 PONV 之间的关联，并将任何遗传贡献与 PONV 的已知临床风险进行比较。方法 研究了在密歇根医学院（密歇根州安娜堡）和范德堡大学医学中心（田纳西州纳什维尔）在全身麻醉下进行的具有综合遗传和围手术期数据的手术。PONV 被定义为发生恶心或呕吐，并在麻醉后监护病房记录。在发现阶段，对每个遗传队列进行了全基因组关联研究，并对结果进行了荟萃分析。接下来，多基因阶段评估了来自范德堡大学医学中心衍生队列的全基因组关联研究的多基因评分是否改善了密歇根医学院验证队列内的预测，如通过区分 （c 统计量） 和净重分类指数量化的那样。结果 在总共 64,523 名患者中，5,703 名 （8.8%） 发生了 PONV。该研究确定了 46 个超过 P < 1 × 10-5 阈值的遗传变异，以大于 1% 的次要等位基因频率发生，并在两个队列中显示出一致的效果。在基本模型中，标准化多基因评分与 PONV 相关，控制年龄和性别（调整后的比值比，总体遗传风险每增加 1.027;95% CI，1.001 比 1。053;P = 0.044），一个基于已知临床风险的模型（调整后的比值比，1.029;95% CI，1.003 至 1.055;P = 0.030），以及完整的临床回归，控制 21 个人口统计学、手术和麻醉因素（调整后的比值比，1.029;95% CI，1.002 至 1.056;P = 0.033）。增加多基因评分提高了基于已知临床风险因素的模型的总体区分度（c 统计量，0.616 对 0.613;P = 0.028），并改善了 4.6% 的病例的净重新分类。结论 在该研究的所有三个模型中，标准化多基因风险都与 PONV 相关，但遗传影响小于临床危险因素施加的影响。具体来说，与基线人群相比，多基因风险评分高于平均值 1 SD 的患者患 PONV 的几率高 2% 至 3%，这至少比与既往 PONV 或晕动病相关的增加 （55%）、有偏头痛病史 （17%） 小一个数量级， 或女性 （83%） 且无临床意义。此外，与临床风险因素相比，使用多基因风险评分并不能有意义地提高鉴别力，并且在临床上没有用处。编辑观点