Abstract / Summary
Abstract Background: Idiopathic pulmonary fibrosis (IPF) is a progressive fibrotic interstitial lung disease characterized by poor prognosis and substantial clinical heterogeneity. Type 2 diabetes mellitus (T2DM) is a common comorbidity in IPF and may exacerbate outcomes through metabolic dysfunction, chronic low-grade inflammation, and microvascular injury. However, prognostic tools specifically designed for patients with both IPF and T2DM remain limited. Methods: We conducted a single-center retrospective cohort study of consecutive adult inpatients with physician-diagnosed IPF and T2DM at the Second Affiliated Hospital of Zhengzhou University between December 20, 2017 and January 20, 2023. The primary outcome was overall survival (OS), defined as the time from the baseline inpatient assessment (when predictors were collected) to death or last follow-up (January 20, 2026) . Candidate predictors included demographics, vital signs, pulmonary function, HRCT fibrosis grading, inflammatory markers, metabolic indices, and routine laboratory variables. Variables with more than 20% missingness were excluded, and missing data in the remaining variables were handled by multiple imputation. Predictor selection was performed via LASSO–Cox regression with 10-fold cross-validation, followed by multivariable Cox modeling. Internal validation was performed using 1000 bootstrap resamples. Model performance was assessed using Harrell's C-index, time-dependent area under the curve (AUC), calibration plots, and decision-curve analysis (DCA). Reporting followed TRIPOD. Results: Of 196 screened patients, 102 were included in the final analysis, of whom 74 died during follow-up. The LASSO procedure identified a parsimonious five-predictor model comprising age, pulmonary function grading, HRCT fibrosis grading, C-reactive protein (CRP), and interleukin-6 (IL-6). In the multivariable Cox model, age (HR 1.05, 95% CI 1.02–1.08; P<0.001), pulmonary function grading (HR 1.72, 95% CI 1.42–2.08; P<0.001), and HRCT fibrosis grading (HR 1.99, 95% CI 1.28–3.09; P=0.002) were independently associated with mortality. Although CRP and IL-6 were retained by the penalized selection process, they did not reach statistical significance in the multivariable model. The apparent C-index was 0.822 (SE 0.022), and the optimism-corrected C-index after 1000 bootstraps was 0.809 (95% CI 0.790–0.828). Time-dependent AUCs were 0.882 (95% CI 0.814–0.951) at 1 year and 0.856 (95% CI 0.778–0.923) at 3 years. Calibration plots demonstrated reasonable agreement between predicted and observed survival probabilities. DCA suggested potential net benefit across a range of threshold probabilities within this internally validated dataset. Conclusions: In hospitalized adults with coexisting IPF and T2DM, older age, worse pulmonary function, and greater HRCT fibrosis burden were independently associated with higher mortality. An exploratory LASSO–Cox model based on routine clinical variables showed good internal performance, however it should be regarded as hypothesis-generating, and external validation is required before clinical application.