Baseline Characteristics of Participants
The study included 480 patients, of whom 328 (68.33%) were male, with an average age of 78.28 (ranging from 70.21 to 85.06) years. The average follow-up time was 21.41 ± 14.90 months, with 224 participants (46.67%) reaching the composite endpoint. Table 1 presents demographic data and laboratory indicators such as age, time, height, weight, BMI, diastolic blood pressure (DBP), mean arterial pressure (MAP), blood urea nitrogen (BUN), triglyceride (TG), high-density lipoprotein cholesterol (HDL-c), low-density lipoprotein cholesterol (LDL-c), total cholesterol (TC), potassium (K), sodium (Na), calcium (Ca), chloride (Cl), and phosphorus (P). There were no statistically significant differences in these indicators between the composite endpoint and without composite endpoint groups. Compared to the group without composite endpoints, the composite endpoint group had higher baseline estimated glomerular filtration rate (eGFR), hemoglobin (Hb); and higher systolic blood pressure (SBP), serum high-sensitivity C-reactive protein (hs-CRP), hemoglobin A1c (HbA1c), uric acid (UA), and albumin-to-creatinine ratio (ACR). Table 2 displays complications and medication status.
Table 1. Specific Classification of Different Statuses
The composite endpoint group had a higher prevalence of diabetes mellitus (DM) (104/224, 46.43% vs. 91/256, 35.55%), used fewer inhalers (LABA/LAMA/ICS) (15/224, 6.70% vs. 55/256, 21.48%), fewer uric acid-lowering drugs (21.88% vs. 33.98%), and more antiplatelet medications (61.16% vs. 48.05%). There were no significant differences in the proportions of smokers, those with hypertension, cardiac insufficiency (CI), coronary artery disease (CAD), and in the use of ACE inhibitors/ARBs, statins, or in the occurrence of edema.
Specific Incidents of Composite Endpoint
Table 3 details the specific outcomes within the composite endpoint group. Of the 224 individuals, the majority (202 people) experienced a decline of 40% from baseline eGFR, and 22 developed end-stage renal disease (ESRD). As of May 31, 2023, 77 individuals had died, including 12 (4.69%) from the group without composite endpoints and 65 (29.02%) from the composite endpoint group. Figure 2 shows survival curves stratified by age (≥60 or <60), sex (male/female), and the presence of hypertension or diabetes, comparing those who met the composite endpoint. There were no significant differences in log-rank tests among these groups in terms of age, sex, and hypertension, but individuals with diabetes were more likely to meet the composite endpoint.
The Results of Univariate and Multivariate Analyses Using Cox Proportional-Hazards Regression Model
Table 4 shows the Cox regression analysis based on the occurrence of the composite endpoint. Significant associations were found between the occurrence of composite endpoint events and factors such as SBP, Hb, albumin (Alb), creatinine, eGFR, albumin-to-creatinine ratio (ACR), uric acid (UA), potassium (K), calcium (Ca), and edema in univariate analysis. Variables with p<0.100 in univariate analysis were again included in the multivariate regression analysis. After adjustment, SBP, Hb, Alb, and edema were independently associated with the occurrence of composite endpoint events. Specifically, for each 10 mmHg increase in SBP, the risk of composite endpoint events increased by 10% (HR 1.10, 95% CI 1.04-1.18, p=0.002); for each 1 g/dl increase in Hb, the risk decreased by 11% (HR 0.89, 95% CI 0.83-0.96, p=0.002); for each 1 g/L increase in Alb, the risk decreased by 4% (HR 0.96, 95% CI 0.93-0.99, p=0.009); notably, the presence of edema significantly increased the risk by 73%.
Model Performance Analysis
Based on the Akaike Information Criterion (AIC), a stepwise regression approach was used to select SBP, Hb, Alb, and edema to construct the model with the lowest AIC. The resulting Cox regression model included these four variables. Figure 3.A shows the Kaplan-Meier curves at different time points (12 months, 24 months, 36 months). Figure 3.B displays Harrell's Concordance Index at various intervals (6 months, 12 months, 18 months, 24 months, 30 months, 36 months), with percentage values respectively being 72.7%, 70.4%, 71.5%, 69.3%, 68.6%, and 68.1%.
Internal Validation of the Model
The model's internal validation was conducted using the bootstrap method, assessing the model's stability. The Cox model built on SBP, Hb, Alb, and edema was validated through bootstrap sampling 1000 times, resulting in a 95% confidence interval for the C-index of (0.626-0.705).