Predictors and nomogram models for postoperative headache in patients undergoing heart valve surgery

Background Headache is a frequent complication after cardiac surgery. However, studies on the risk factors of postoperative headache (POH) are rare. The purpose of this study was to identify independent risk factors for POH in patients undergoing heart valve surgery (HVS) and to develop and validate risk prediction models. Methods Consecutive patients undergoing open HVS from 2016 to 2019 were enrolled in this study. Patients were randomly assigned to training and validation sets at a 2:1 ratio. Univariate and multivariate analysis were applied to identify independent predictors for POH in the training set. A nomogram predicting POH was developed based on these factors, and was validated in the independent validation set. Results POH developed in 1,061 of the 3,853 patients (27.5%). The overall mortality was 2.9%, and it was significantly higher in patients with POH (4.3% versus 2.4%, P<0.001). In the training set, six independent predictors were identified by multivariate analysis, including female, smoking history, hypertension, headache history, left ventricular ejection fraction, and cardiopulmonary bypass time. The model demonstrated good discrimination in both the training (c-index: 0.811) and validation sets (c-index: 0.814), and calibration was assessed by visual inspection. A second nomogram was also constructed including only preoperative predictors, with good discrimination (c-index: 0.792) and calibration. The decision and clinical impact curves of the models showed good clinical utility. Conclusions We developed and validated two risk prediction models for POH in patients undergoing HVS. The models may have clinical utility in individualized risk assessment and preventive interventions.