Predictive characteristics and model development for acute heart failure preceding hip fracture surgery in elderly hypertensive patients: a retrospective machine learning approach

Abstract

Background: Hip fractures are a serious health concern among the elderly, particularly in patients with hypertension, where the incidence of acute heart failure preoperatively is high, significantly affecting surgical outcomes and prognosis. This study aims to assess the risk of preoperative acute heart failure in elderly patients with hypertension and hip fractures by constructing a predictive model using machine learning on potential risk factors.

Methods: A retrospective study design was employed, collecting preoperative data from January 2018 to December 2019 of elderly hypertensive patients with hip fractures at the Third Hospital of Hebei Medical University. Using SPSS 24.0 and R software, predictive models were established through LASSO regression and multivariable logistic regression analysis. The models' predictive performance was evaluated using metrics such as the concordance index (C-index), receiver operating characteristic curve (ROC curve), and decision curve analysis (DCA), providing insights into the nomogram's predictive accuracy and clinical utility.

Results: Out of 1038 patients screened, factors such as gender, age, history of stroke, arrhythmias, anemia, and complications were identified as independent risk factors for preoperative acute heart failure in the study population. Notable predictors included Sex (OR 0.463, 95% CI 0.299-0.7184, P = 0.001), Age (OR 1.737, 95% CI 1.213-2.488, P = 0.003), Stroke (OR 1.627, 95% CI 1.137-2.327, P = 0.008), Arrhythmia (OR 2.727, 95% CI 1.490-4.990, P = 0.001), Complications (OR 2.733, 95% CI 1.850-4.036, P < 0.001), and Anemia (OR 3.258, 95% CI 2.180-4.867, P < 0.001). The prediction model of acute heart failure was Logit(P) = -2.091-0.770 × Sex + 0.552 × Age + 0.487 × Stroke + 1.003 × Arrhythmia + 1.005 × Complications + 1.181 × Anemia, and the prediction model nomogram was established. The model's AUC was 0.785 (95% CI, 0.754-0.815), Decision curve analysis (DCA) further validated the nomogram's excellent performance, identifying an optimal cutoff value probability range of 3% to 58% for predicting preoperative acute heart failure in elderly patients with hypertension and hip fractures.

Conclusion: The predictive model developed in this study is highly accurate and serves as a powerful tool for the clinical assessment of the risk of preoperative acute heart failure in elderly hypertensive patients with hip fractures, aiding in the optimization of preoperative risk assessment and patient management.

Keywords: Heart failure; Hip fracture; Hypertension; Nomogram; Prediction model; Preoperative.

Fig. 1

The patient flow chart in our study

Fig. 2

Data statistics and clinical feature selection using the LASSO binary logistic regression model. A Optimal parameter (lambda) selection in the LASSO model used fivefold cross-validation via minimum criteria. The partial likelihood deviance (binomial deviance) curve was plotted versus log(lambda). B LASSO coefficient profiles of the 14 features. A coefficient profile plot was produced against the log(lambda) sequence

Fig. 3

Forest plot showing the relationship between risk factors and the occurrence of preoperative acute heart failure in elderly patients with hypertension combined with hip fracture

Fig. 4

A nomogram model for predicting the occurrence of preoperative acute heart failure in elderly patients with hypertension combined with hip fractures

Fig. 5

Calibration curves of the acute heart failure nomogram prediction in the cohort. The x-axis represents the predicted acute heart failure risk. The y-axis represents the actual diagnosed acute heart failure. The diagonal dotted line represents a perfect prediction by an ideal model. The solid line represents the performance of the nomogram, of which a closer fit to the diagonal dotted line represents a better prediction

Fig. 6

Analysis of ROC curve for the predictive values of preoperative acute heart failure

Fig. 7

Decision curve analysis for the acute heart failure nomogram. The y-axis measures the net benefit. The blue line represents the acute heart failure risk nomogram. The thin solid line represents the assumption that all patients have acute heart failure, while the thin thick solid line represents the assumption that no patients have acute heart failure. This indicates that using the nomogram to guide clinical intervention can provide a positive net benefit to patients within a cutoff value probability range of 3% to 58%