BACKGROUND: The study aimed to investigate the associations of baseline serum albumin level and its dynamic change with type 2 diabetes mellitus (T2DM) risk in a large Chinese cohort study. METHODS: This cohort study included 30 442 adults without T2DM at first entry, who completed at least one follow-up of annual examinations between 2009 and 2016. Serum albumin level was measured at baseline and at every annual check-up. The dynamic change in serum albumin level (∆ALB) was calculated by subtracting serum albumin level at baseline from that at the last follow-up. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated with Cox regression models. RESULTS: During 7 years of follow-up, we identified 1634 T2DM events. From the lowest to the highest quartile of serum albumin level, adjusted HRs (95% CI) were 1.00 (reference), 0.96 (0.94, 1.01), 0.98 (0.95, 1.02) and 0.88 (0.85, 0.98), respectively. As compared with stable change in serum albumin (-0.2 ≤ ∆ALB <1.0 g/L), the risk of T2DM increased for ∆ALB < -2.0 g/L (HR 1.44, 95% CI 1.24-1.68) and decreased for ∆ALB ≥3.0 g/L (0.81, 0.68-0.97) after adjusting for potential confounding factors. Restricted cubic splines showed a linear dose-response association between baseline serum albumin level and T2DM risk (Pnonlinearity 0.715) and a nonlinear dose-response association between ∆ALB and T2DM risk (Pnonlinearity 0.011). CONCLUSIONS: Baseline serum albumin level appears to be inversely associated with T2DM risk. Adults with reduced serum albumin level could be early identified for diabetes risk in clinical practice.
BACKGROUND: The study aimed to investigate the associations of baseline serum albumin level and its dynamic change with type 2 diabetes mellitus (T2DM) risk in a large Chinese cohort study. METHODS: This cohort study included 30 442 adults without T2DM at first entry, who completed at least one follow-up of annual examinations between 2009 and 2016. Serum albumin level was measured at baseline and at every annual check-up. The dynamic change in serum albumin level (∆ALB) was calculated by subtracting serum albumin level at baseline from that at the last follow-up. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated with Cox regression models. RESULTS: During 7 years of follow-up, we identified 1634 T2DM events. From the lowest to the highest quartile of serum albumin level, adjusted HRs (95% CI) were 1.00 (reference), 0.96 (0.94, 1.01), 0.98 (0.95, 1.02) and 0.88 (0.85, 0.98), respectively. As compared with stable change in serum albumin (-0.2 ≤ ∆ALB <1.0 g/L), the risk of T2DM increased for ∆ALB < -2.0 g/L (HR 1.44, 95% CI 1.24-1.68) and decreased for ∆ALB ≥3.0 g/L (0.81, 0.68-0.97) after adjusting for potential confounding factors. Restricted cubic splines showed a linear dose-response association between baseline serum albumin level and T2DM risk (Pnonlinearity 0.715) and a nonlinear dose-response association between ∆ALB and T2DM risk (Pnonlinearity 0.011). CONCLUSIONS: Baseline serum albumin level appears to be inversely associated with T2DM risk. Adults with reduced serum albumin level could be early identified for diabetes risk in clinical practice.