Development and Validation of a Nomogram Prediction Model for Pulmonary Infection in Cerebral Infarction Patients During the Recovery Phase After Acupuncture Treatment

Objective To analyze the risk factors for pulmonary infection in convalescent patients with cerebral infarction under acupuncture and moxibustion intervention, and to develop a nomogram model.

Methods According to a modeling-to- verification set ratio of 7∶3, 350 patients with cerebral infarction treated in the hospital from January 2022 to January 2024 were selected as the modeling set. At the same time, 150 patients with cerebral infarction admitted from February 2024 to December 2024 were selected as the verification set. General patient information, laboratory indicators, and acupuncture intervention variables were collected. After screening variables using LASSO regression in the modeling set, multivariate logistic regression was performed. Based on the results of the multivariate regression, a nomogram prediction model was established, and the models were evaluated in the modeling and validation set, respectively.

Results The incidence of pulmonary infection in patients with cerebral infarction during convalescence was 26.20%(131/500), including 95 patients in the model group. A total of 147 pathogen strains were detected, mainly Gram- negative bacteria. Modeling concentration, LASSO regression, and multivariate stepwise logistic regression were used to identify factors influencing pulmonary infection: increased age (OR = 1.043, 95%CI: 1.010-1.077), higher Eating Assessment Tool 10 (EAT-10) score (OR = 1.249, 95%CI: 1.125-1.387), invasive procedures (OR = 2.715, 95%CI: 1.340-5.501), higher National Institutes of Health Stroke Scale (NIHSS) score (OR = 2.477, 95%CI: 1.951-3.144), lower albumin (OR = 0.856, 95%CI: 0.786-0.931), lower prealbumin (OR = 0.992, 95%CI: 0.987-0.997), needle retention time of 30 min (OR = 2.439,95%CI: 1.146-5.191), acupuncture course less than 14 d (OR = 2.463,95%CI: 1.157-5.243), and initiation of intervention acupuncture treatment more than >3 d after onset (OR = 2.660, 95%CI: 1.190-5.942). Based on these factors, a nomogram model was constructed, with areas under the curve (AUC) of 0.916 (95%CI: 0.877-0.954) and 0.923 (95%CI: 0.874-0.973). The calibration curve showed that the predicted probabilities of the model were in close agreement with the observed probabilities, and its application could yield positive net clinical benefits.

Conclusion The nomogram model constructed used Lasso-logistic regression can identify the risk of pulmonary infection in patients with cerebral infarction during convalescence after acupuncture treatment. Quantifying risk factors can guide patients in optimizing acupuncture treatment plans, achieving precise intervention, and reducing the risk of infection.