Determinants of Kidney Function Decline and Rapid Progression in Diabetic Kidney Disease

Dear Editor,

Diabetic kidney disease (DKD), a leading cause of ESKD worldwide, is characterized by diverse clinical presentations and varying rates of progression. Despite implementation of kidney-protective therapies such as renin-angiotensin system inhibitors (RASi) and sodium-glucose cotransporter-2 (SGLT2) inhibitors, kidney function measured as estimated glomerular filtration rate (eGFR) slope, continues to decline over time, culminating in ESKD.^1-3 Clinical trials report an annual eGFR decline of 2.8 to 3.5 mL/min/year with these treatments^1-3 and observational studies suggest a similar decline of 2.7 - 4 mL/min/year in DKD patients.^4-6 While these estimates represent the average eGFR decline, a subset of patients with DKD have rapid progression,^S1 defined by Kidney Disease: Improving Global Outcomes (KDIGO) as a decrease in eGFR of >5 mL/min/year.^S2

Rapid progression is prevalent in DKD with estimates ranging from 60% to 80%.^S3,S4 Factors such as age, sex, hypertension, proteinuria, baseline eGFR, and nephrotoxic drug exposure are known to influence eGFR decline.^4,S4,S5 There is limited data from developing countries on the determinants of kidney function decline in DKD in the real- world context.

This was a single-center retrospective longitudinal study that measured mean annual eGFR decline in an Indian cohort of patients with type 2 diabetes mellitus and DKD. It also identified novel predictors of eGFR decline and rapid progression in DKD. We measured DKD progression by eGFR slope using linear mixed-effects regression. Rapid progression was defined as eGFR-slope <-5 mL/min/year. eGFR slope predictors were identified by multivariable regression. The methods have been detailed in the Supplementary File.^S6-S31

This study included 274 patients. The mean age at initial diagnosis was 60.04 ± 9.9 years; there were 210 (77%) males. The median eGFR at initial DKD diagnosis was 43.8 (IQR: 31-57) mL/min/1.73 m². Most patients (59%) initially presented with stage III CKD (29% and 30% with stage IIIa and stage IIIb, respectively). The mean follow-up from the initial diagnosis was 3.3 years. The baseline characteristics of the sample at initial diagnosis have been listed in Table 1.

The cohort’s mean eGFR slope was -6.43 (95% CI: -7.02 to -5.84) mL/min/year. The unadjusted eGFR slopes in different DKD subsets have been presented in Table 2. Patients with non-proteinuric DKD showed the least tendency for progression (eGFR slope -3.72 mL/min/year). Those with heart failure with mildly reduced or preserved ejection fraction (HFmrEF/HFpEF) had the greatest propensity for progression (eGFR slope -10.84 mL/min/year). The eGFR-slope according to baseline CKD stage has been summarized in Supplementary Table 1. eGFR-slope exhibited a gradient according to the baseline CKD stage, with fastest decline observed in stage I (-12.87ml/min/year) and slowest decline in stage IV (-4.25 ml/min/year). Supplementary Figure 1 illustrates the unadjusted eGFR-slope in different DKD subsets. Crossover between curves (interaction) implies a significant difference in eGFR slope between groups.

Supplementary Table 2 shows univariate and multivariate associations between predictors and eGFR slope. Multivariate regression revealed that rural dwellers lost 1.13 mL/min/year eGFR (p = 0.02) more than urban residents. Patients with HFmrEF/HFpEF lost an excess 3.26 mL/min/year eGFR (p < 0.001) than those without heart failure. Patients with resistant hypertension lost 1.74 mL/min/year (p = 0.001) eGFR more than those with well-controlled hypertension. Intravitreal anti-VEGF exposure was linked to a 1.51 mL/min/year (p = 0.01) greater eGFR loss than in unexposed individuals. Intercurrent AKI was associated with 1.83 mL/min/year (p = 0.002) excess eGFR loss. Compared to patients with proteinuric DKD, those with non-proteinuric DKD showed a lesser eGFR decline (2.73 mL/min/year) (p < 0.001). Age and sex did not predict eGFR slope.

About 37% of patients in this cohort received SGLT2 inhibitor treatment, and 43% received RASi for at least half of the follow-up duration [Table 1]. Baseline or intercurrent hyperkalemia precluded RASi usage in 37% of patients. SGLT2 inhibitor and RASi exposure [Supplementary Table 2] were associated with 1.44 mL/min/year (p = 0.003) and 0.83 mL/min/year (p = 0.09) lesser decline in eGFR, respectively. There was no significant interaction between RASi or SGLT2 inhibitor treatment and proteinuric DKD categories [Supplementary Table 3].

Of 274, 159 (58%) had rapidly progressive DKD. The eGFR slopes among rapid (-9.50 mL/min/year) and non-rapid progressors (-2.19 mL/min/min) were significantly different (p < 0.001). Both groups were systematically different at initial diagnosis [Table 1].

Rapid progression predictors in DKD have been shown in Supplementary Table 4. After multivariate regression analysis, baseline eGFR (odds ratio [OR]- 1.03 per mL of baseline eGFR, p = 0.001), HFmrEF/HFpEF (OR 3.46, p = 0.02), and intercurrent AKI (OR 2.25, p = 0.02) significantly increased the odds of rapid progression in DKD. Non-proteinuric DKD (OR 0.19, p < 0.001) reduced the odds of rapid progression. SGLT2 inhibitor exposure (OR 0.56, p = 0.06) lowered the odds of rapid progression; however, not statistically significantly.

Our study quantitatively assesses multiple eGFR slope predictors in the real-world setting of a developing country. Social factors are known to influence kidney health.^S32,S33 In this study, rural residence caused significantly faster eGFR decline [Supplementary Table 2], likely due to limited healthcare access and environmental toxins. Heart failure, particularly HFmrEF/HFpEF, has been linked to accelerated kidney function decline.^S16,S34 Our findings align with these studies, as HFmrEF/HFpEF was associated with the most significant eGFR decline in DKD [Supplementary Table 2].

There is emerging evidence that intravitreal anti-VEGF agents promote rapid DKD progression.S35 In our study, patients receiving anti-VEGF treatment experienced additional 1.5 mL/min/year eGFR decline [Supplementary Table 2], emphasizing this therapy’s long-term impact on kidney function.

Baseline eGFR emerged as a strong eGFR slope predictor in DKD [Supplementary Table 2]. High baseline eGFR translated to rapid eGFR decline, implying that interventions intended to retard DKD should target the initial vulnerable stage of DKD.

This research also quantifies the impact of recognized factors on eGFR slope. Blood pressure control is an established strategy for kidney function preservation in DKD. In our study, 31% of the cohort had resistant hypertension, which led to an excess 1.74 mL/min/year eGFR loss. Patients with proteinuric DKD lost more eGFR than those with non-proteinuric DKD. These findings support blood pressure optimization and proteinuria reduction in DKD.

In clinical trials, SGLT2 inhibitor exposure^2,3,S36 led to a 1.13 to 2.26 mL/min/year lesser eGFR decline compared to the placebo group. In our study, the real-world benefit of SGLT2 inhibitor treatment in DKD was evident as 1.44 mL/min/year lesser decline in eGFR. The benefit of SGLT2 inhibitor treatment persisted across proteinuric DKD categories, as there was no significant interaction between the two. This finding concurs with a meta-analysis of clinical trials, which showed similar benefits of SGLT2 inhibitor therapy across varying CKD proteinuria categories.^S37 These real-world findings reinforce the motivation for SGLT2 inhibitor prescription in DKD. Considering the myriad factors that influence DKD progression, a holistic approach to retardation is needed with a special focus on heart failure.