Development of a prognostic model for in-hospital mortality in idiopathic membranous nephropathy patients with Pneumocystis jirovecii pneumonia based on persistent lymphopenia.

objectivePneumocystis jirovecii pneumonia (PJP) represents a potentially life-threatening complication in patients with idiopathic membranous nephropathy (IMN). The objective of this study is to identify the risk factors for in-hospital mortality in IMN-PJP patients and to construct an early warning model aimed at assisting clinicians in promptly recognizing high-risk cases and formulating more aggressive treatment strategies.

methodsWe retrospectively analyzed IMN patients diagnosed with PJP at the National Clinical Research Center of Kidney Diseases between December 2014 and December 2024. Group-based trajectory modeling (GBTM) was employed to characterize the dynamic patterns of absolute lymphocyte count (ALC) during the first week of hospitalization. Based on the analysis results, the characteristics and diagnostic criteria of the persistent lymphopenia (PL) population were determined. Logistic regression was performed to screen for risk factors of in-hospital mortality in IMN-PJP patients, and a prognostic prediction model was constructed. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA).

resultsA total of 63 IMN-PJP patients were included, of whom 10 died during hospitalization (mortality rate 15.9%). Based on model fit criteria and ALC trends, we selected a three-trajectory model: persistently low-level group, slow increase group, and rapid increase group. In the persistently low-level group, patients’ ALC were consistently below 0.5 × 10⁹/L during the first week of hospitalization. Accordingly, we classified PL as an ALC level < 0.5 × 10⁹/L within the initial week post-admission. Notably, patients with PL exhibited a significantly higher in-hospital mortality rate compared to their non-PL counterparts. Univariate and multivariate logistic regression analyses identified PL, complement C3 levels, and the use of mechanical ventilation after admission as significant independent risk factors for in-hospital mortality in IMN-PJP patients. These predictors were integrated into a nomogram to develop an in-hospital mortality risk prediction model. The model showed outstanding discrimination (AUC = 0.945, 95% CI 0.890-1.000), satisfactory calibration (Hosmer-Lemeshow test: χ² = 2.418, p = 0.965), and clinically meaningful net benefit on decision curve analysis (DCA), demonstrating its robust predictive performance and practical applicability.

conclusionA subset of IMN-PJP patients exhibit PL, which is strongly associated with adverse outcomes. Our prognostic model, integrating PL, demonstrated robust predictive performance and practical value for risk stratification in this population. CLINICAL TRIAL: Not applicable.