Development of a nomogram prediction model for pulmonary infection risk in elderly diabetic patients

Wang Jingjing; Fang Shencun; Zhang Weiwei; Xu Ting; Wang Xiaoyue; Ye Lei; Chen Yan; Miao Chou

doi:10.11986/j.issn.1673-873X.2025.07.03

Abstract：Objective To develop a nomogram prediction model for pulmonary infection risk in elderly diabetic patients,screen high-risk populations of pulmonary infection,and assist early clinical intervention.Methods Clinical data from 200 elderly dia-betic patients without pulmonary infection(non-infection group)and 200 with pulmonary infection(infection group)admitted to Nanjing Chest Hospital(January 2020-December 2023)were collected in a 1∶1 ratio.Logistic regression was used to i-dentify risk factors;R software was employed to construct the nomogram model of elderly diabetes complicated by pulmonary infection.receiver operator characteristic curve(ROC)curves,calibration curves,and decision curve analysis(DCA)were a-dopted to validate the nomogram model.Results No significant differences were noted in gender,body mass index(BMI),al-cohol history,or total cholesterol levels between groups(P＞0.05).The infection group had longer disease duration and hos-pital stay;older age,smoking history,invasive procedures,abnormal pulmonary function,combined chronic bronchitis,com-bined emphysema,combined chronic obstructive pulmonary disease(COPD),and combined hypoalbuminemia;and elevated u-ric acid(UA),D-dimer,fasting plasma glucose(FPG),glycosylated hemoglobin A1e(HbA1c),systolic blood pressure(SBP),diastolic blood pressure(DBP),CD8+,interleukin-6(IL-6),tumor necrosis factor-α(TNF-α),and C-reactive protein(CRP)levels but lower CD3+,CD4+,and CD4+/CD8+levels than the non-infection group,with differences being statistically significant(P＜0.05).Multivariate analysis identified disease duration(OR=1.341,95％CI:1.212-1.484),smoking his-tory(OR=1.975,95％CI:1.167-3.342),invasive procedures(OR=2.614,95％CI:1.368-4.994),hypoalbuminemia(OR=2.385,95％CI:1.327-4.287),and FPG(OR=1.509,95％CI:1.349-1.688)as independent risk factors of eld-erly diabetes complicated by pulmonary infection(P＜0.05).The nomogram assigned 12.6 points per 2-year disease duration increase,8.4 points per 1 μmol/L FPG increase,14.3 points for smoking history,20.2 points for invasive procedures,and 17.7 points for hypoalbuminemia.The area under the curve(AUC)was 0.826(95％CI:0.786-0.866).The calibration curve slope approached 1(Hosmer-Lemeshow test:x2=6.262,P=0.618).DCA showed high predictive value at threshold probabilities of 0.04-0.96 for elderly diabetes complicated by pulmonary infection.Conclusion The nomogram model based on disease duration,smoking history,invasive procedures,hypoalbuminemia,and FPG effectively predicts pulmonary infection risk in elderly diabetic patients,demonstrating good discrimination and calibration for individualized risk assessment and intervention.