Constructing and Validating a Dynamic Nomogram to Predict Response to Bariatric Surgery: A Multicenter Retrospective Study.

PURPOSE

Suboptimal response is one of the major problems for bariatric surgery, and constructing an individualized model for predicting outcomes of bariatric surgery is essential. Thus, the aim of this study is to develop a nomogram to predict the response to bariatric surgery.

MATERIALS AND METHODS

509 patients who underwent bariatric surgery between 2019 to 2020 from 6 centers were retrieved and assessed. Multiple Imputation was used to replace missing data. Patients with %TWL ≥ 20% 1 year after bariatric surgery were classified as patients with optimal response, while the others were patients with suboptimal response. A web-based nomogram was constructed and validated. ROC curve and calibration curve were used to determine the predictive ability of our model.

RESULTS

56 (11.0%) patients were classified as patients with suboptimal response, and they showed advanced age, lower pre-operative BMI, smaller waist circumference, higher fasting glucose, higher HbA1c and lower fasting insulin compared to patients with optimal response. A forward likelihood ratio logistic regression analysis indicated that age (OR = 0.943, 95% CI: 0.915-0.971, p < 0.001), pre-operative BMI (OR = 1.109, 95% CI: 1.002-1.228, p = 0.046) and waist circumference (OR = 1.043, 95% CI: 1.000-1.088, p = 0.048) were essential factors contributing to the response to bariatric surgery. Lastly, a web-based nomogram was constructed to predict the response to bariatric surgery and demonstrated an AUC of 0.829 and 0.798 upon internal and external validation.

CONCLUSION

Age, BMI and fasting glucose were proved to be essential factors influencing the response to bariatric surgery. The nomogram constructed in this study demonstrated good adaptivity.