Associations of physiologic subtypes based on HOMA2 indices of β-cell function and insulin sensitivity with the risk of kidney function decline, cardiovascular disease, and all-cause mortality from the 4C study

Abstract

Background: Previous studies have been limited by their inability to differentiate between the effects of insulin sensitivity and β-cell function on the risk of kidney function decline, cardiovascular disease (CVD), and all-cause mortality. To address this knowledge gap, we aimed to investigate whether the physiological subtypes based on homeostasis model assessment-2 (HOMA2) indices of β-cell function (HOMA2-B) and insulin sensitivity (HOMA2-S) could be used to identify individuals with subsequently high or low of clinical outcome risk.

Methods: This retrospective cohort study included 7,317 participants with a follow-up of up to 5 years. Based on HOMA2 indices, participants were categorized into four physiologic subtypes: the normal phenotype (high insulin sensitivity and high β-cell function), the insulinopenic phenotype (high insulin sensitivity and low β-cell function), the hyperinsulinaemic phenotype (low insulin sensitivity and high β-cell function), and the classical phenotype (low insulin sensitivity and low β-cell function). The outcomes included kidney function decline, CVD events (fatal and nonfatal), and all-cause mortality. Cox regression models were used to calculate hazard ratios (HRs) for outcomes, and spline models were used to examine the dose-dependent associations of HOMA2-B and HOMA2-S with outcomes.

Results: A total of 1,488 (20.3%) were classified as normal, 2,179 (29.8%) as insulinopenic, 2,173 (29.7%) as hyperinsulinemic, and 1,477 (20.2%) as classical subtypes. Compared with other physiological subtypes, the classical subtype presented the highest risk of kidney function decline (classical vs. normal HR 11.50, 95% CI 4.31-30.67). The hyperinsulinemic subtype had the highest risk of CVD and all-cause mortality (hyperinsulinemic vs. normal: fatal CVD, HR 6.56, 95% CI 3.09-13.92; all-cause mortality, HR 4.56, 95% CI 2.97-7.00). Spline analyses indicated the dose-dependent associations of HOMA2-B and HOMA2-S with outcomes.

Conclusions: The classical subtype had the strongest correlation with the risk of kidney function decline, and the hyperinsulinemic subtype had the highest risk of CVD and all-cause mortality, which should be considered for interventions with precision medicine.

Keywords: All-cause mortality; CVD events; HOMA2-B; HOMA2-S; Kidney function decline.

Fig. 1

Flowchart of the cohort study CVD cardiovascular disease, CKD chronic kidney disease, HOMA2 homeostatic model assessment-2

Fig. 2

Distributions of HOMA2-B, HOMA2-S, and corresponding physiological subtypes based on HOMA2 indices of β-cell function and insulin sensitivity. The HOMA2 computational model was used to calculate insulin sensitivity (HOMA2-S, %) and beta-cell function (HOMA2-B, %) based on fasting glucose and insulin values. Each dot represents a person. Participants were grouped into four subtypes including normal, insulinopenic, hyperinsulinemic, and classical subtypes. High and low HOMA2-S and HOMA2-B were defined as being above or below the median values (HOMA2-S: 165.0 and HOMA2-B: 63.6).A: Distribution of HOMA2-B, HOMA2-S, and corresponding physiological subtypes based on HOMA2 indices of β-cell function and insulin sensitivity was presented in subgroups of different glucose statuses (A). B–D: Distributions of HOMA2-B, HOMA2-S, and corresponding physiological subtypes based on HOMA2 indices of β-cell function and insulin sensitivity were presented in all participants (B), NGT (C), and prediabetes (D) respectively.HOMA2 homeostatic model assessment-2, NGT normal glucose status-2

Fig. 3

Frequency of outcomes according to physiologic subtypes based on HOMA2 indices of β-cell function and insulin sensitivity in all participants

Fig. 4

Forest plot showing the adjusted HRs of physiological subtypes based on HOMA2 indices of β-cell function and insulin sensitivity for kidney function decline in all participants and participants with NGT and prediabetes

Fig. 5

Forest plot showing the adjusted HRs of physiological subtypes based on HOMA2 indices of β-cell function and insulin sensitivity for CVD events and all-cause mortality in all participants and participants with NGT and prediabetes. The model was adjusted for age, sex, BMI, SBP, FPG, HbA1c, TC, LDL-C, TG, ALT, GGT, eGFR, smoking status, drinking status, hypotensive drug, and lipid-lowering drug. HOMA2 homeostatic model assessment-2, HR hazard ratio, CI confidence interval, NGT normal glucose status, CVD cardiovascular disease

Fig. 6

RCS showing the dose-response relationships of HOMA2-B and HOMA2-S with the risk of kidney function decline, fatal CVD events, nonfatal CVD events, and all-cause mortality The dose-response relationships of HOMA2-B with the risk of kidney function decline (A for the MDRD method,B for the CKD-EPI method), fatal CVD events (C), nonfatal CVD events (D), and all-cause mortality (E). The dose-response relationships of HOMA2-B with the risk of kidney function decline (F for the MDRD method,G for the CKD-EPI method), fatal CVD events (H), nonfatal CVD events (I), and all-cause mortality (J). RCS restricted cubic spline, HOMA2 homeostatic model assessment-2, MDRD Modification of Diet in Renal Disease, CKD-EPI the Chronic Kidney Disease Epidemiology Collaboration, CVD cardiovascular disease