Participant characteristics

Data from 403,677 participants who underwent health check-ups between January 2002 and December 2018 were extracted from an anonymized dataset, and 113,813 participants aged ≥ 18 years with baseline data on total lymphocyte count, serum albumin, and serum creatinine were included in the study cohort. After excluding subjects with missing data on serum creatinine at 5 years (n = 98,375) and those on dialysis at baseline (n = 1), data from 15,437 subjects were finally used for analysis.

The baseline characteristics for the participants across PNI categories are presented in Table 1. The participants’ age (mean ± SD) was 43.7 ± 7.9 years, and 45.9% were male. The mean ± SD PNI and baseline eGFR were 56.9 ± 4.2 and 95.4 ± 15.0 mL/min/1.73 m2, respectively. Participants in the higher PNI groups were younger, and these groups included a larger proportion of male, smokers, diabetics, individuals with hypertension, and individuals with coronary artery disease. In the higher PNI groups, levels of BMI, SBP, DBP, hemoglobin, total cholesterol, CRP, uric acid, HbA1c, and serum glucose were also higher, while the baseline eGFR was lower.

Table 1 Baseline characteristics of 15,437 participants across PNI categoriesAssociation between PNI quintile and 25% decline in eGFR after 5 years from baseline

A total of 2,272 (14.7%) participants experienced a 25% decline in eGFR after 5 years from baseline, including 18.7%, 15.7%, 14.5%, 13.5%, and 11.1% of the first, second, third, fourth, and fifth PNI groups, respectively. In the logistic regression analysis with full adjustment, lower PNI groups had a greater risk of a 25% decline in eGFR compared to the highest PNI group; odds ratios and 95% confidence intervals (CIs) were 1.42 (1.20, 1.68), 1.23 (1.04, 1.45), 1.21 (1.03, 1.43), and 1.19 (1.01, 1.40) for the first, second, third, and fourth quintiles of PNI, respectively (Table 2). When PNI was introduced as a continuous variable in the fully adjusted model, the odds ratio and 95% CIs for 25% decline in eGFR associated with a decrease of 1 unit in PNI level were 1.03 (1.01, 1.03). In subgroup analyses, the association between PNI quintiles and a 25% decline in eGFR after 5 years was not modified by age (Pinteraction = 0.99), sex (Pinteraction = 0.20), diabetes (Pinteraction = 0.22), hypertension (Pinteraction = 0.17), CAD (Pinteraction = 0.55), smoking status (Pinteraction = 0.49), SBP (Pinteraction = 0.16), DBP (Pinteraction = 0.28), BMI (Pinteraction = 0.37), hemoglobin (Pinteraction = 0.22), total cholesterol (Pinteraction = 0.54), serum glucose (Pinteraction = 0.57), CRP (Pinteraction = 0.05), uric acid (Pinteraction = 0.09), or HbA1c (Pinteraction = 0.71) (Supplemental Fig. 1). In contrast, the association between PNI quintile and a 25% decline in eGFR after 5 years was modified by eGFR (Pinteraction = 0.01); the association was more evident in participants with an eGFR ≥ 90 mL/min/1.73 m2 compared to those with an eGFR < 90 mL/min/1.73 m2, but showed similar trends between the two groups (Fig. 1).

Table 2 Odds ratios for 25% decline in eGFR after 5 year according to PNI quintileFig. 1

Association between PNI quintile and risk of a 25% decline in eGFR after 5 years, stratified by baseline eGFR (≥ 90 and < 90 mL/min/1.73 m2) among 15,437 participants. Points and bars depict odds ratios and 95% confidence intervals, respectively. Abbreviations: eGFR, estimated glomerular filtration rate; PNI, prognostic nutritional index

Association between PNI quintile and 5-year change in renal function

The mean (SD) eGFR levels after 5 years were 85.8 (15.0), 85.6 (14.4), 85.2 (14.4), 84.8 (14.2), and 84.6 (14.3) mL/min/1.73 m2 from the first to fifth quintiles of PNI, respectively. The median (IQR) 5-year change in eGFR was − 5.2 (− 18.8, − 3.3) mL/min/1.73 m2. In the linear regression analysis with full adjustment, lower PNI groups showed a greater decline in the eGFR over 5 years compared to the highest PNI group (the reference group); beta (β) coefficients and 95% CIs were − 2.57 (− 3.18, − 1.97), − 1.20 (− 1.77, − 0.63), − 1.03 (− 1.59, − 0.48), and − 0.86 (− 1.40, − 0.31) for the first, second, third, and fourth quintiles of PNI, respectively (Fig. 2). When PNI was introduced as a continuous variable in the fully adjusted model, the β coefficient and 95% CIs for 5-year change in eGFR associated with a decrease of 1 unit in PNI level were − 0.19 (− 0.24, − 0.15). The restricted cubic splines showed an incremental association between PNI as a continuous variable and 5-year change in eGFR (Fig. 3).

Fig. 2

Association between PNI quintile and 5-year change in eGFR among 15,437 participants. Lower PNI groups showed a larger decline in the eGFR over 5 years compared to the highest PNI group. Points and bars depict beta coefficients and 95% confidence intervals, respectively. Abbreviations: eGFR, estimated glomerular filtration rate; PNI, prognostic nutritional index

Fig. 3

A linear regression model with restricted cubic splines showing an incremental association between PNI as a continuous variable and 5-year change in eGFR: unadjusted model (A); fully adjusted model (B). Abbreviations: eGFR, estimated glomerular filtration rate; PNI, prognostic nutritional index

Sensitivity analysis

We investigated the 5-year change in PNI assessed by PNI slope, considering the influence of PNI change on changes in renal function. PNI increased more in the lower baseline PNI groups over a 5-year period. The median (IQR) 5-year slope of PNI were 0.16 (0.10–0.22), 0.11 (0.06–0.17), 0.09 (0.03–0.13), 0.05 (0.01–0.10), and 0.01 (− 0.05–0.04) for the first to fifth quintiles of PNI, respectively (P for trend < 0.001). When we further adjusted the PNI slope in a fully adjusted logistic regression model to assess the association between PNI and renal function decline, the results were still robust; odds ratios and 95% CIs were 1.37 (1.14, 1.63), 1.20 (1.01, 1.42), 1.19 (1.01, 1.40), and 1.18 (1.00, 1.39) for the first, second, third, and fourth quintiles of PNI, respectively (reference: the highest PNI quintile) (Supplemental Fig. 2).

Considering that the definition of proteinuria in this study cannot rule out the possibility of temporary proteinuria or false positives, we also investigated the association between PNI quintile and 25% decline in eGFR among participants without proteinuria, the results remained similar; odds ratios and 95% CIs were 1.41 (1.19, 1.67), 1.22 (1.04, 1.45), 1.21 (1.02, 1.42), and 1.19 (1.01, 1.40) for the first, second, third, and fourth quintiles of PNI, respectively (Supplemental Table 1).

To address the potential for selection bias when excluding participants without data on serum creatinine at 5 years, we analyzed the association between the PNI quintile and the 5-year eGFR slope by including all participants with at least one follow-up eGFR data, and the results remained robust; β coefficients and 95% CIs were − 0.28 (− 0.32, − 0.24), − 0.15 (− 0.19, − 0.11), − 0.10 (− 0.13, − 0.06), and − 0.02 (− 0.06, 0.01) for the first, second, third, and fourth quintiles of PNI, respectively (Supplemental Table 2).