Two-wave modified cardiometabolic index change patterns and incident edentulism in Chinese adults: a longitudinal study from CHARLS

Abstract

Abstract Objectives Complete tooth loss is a late-stage oral health endpoint that may reflect cumulative oral, systemic, behavioral, and social risk processes. We examined whether two-wave patterns of the modified cardiometabolic index (MCMI) from 2011 to 2015 were associated with incident edentulism from 2015 to 2018 among Chinese adults. Methods We analyzed participants aged 45 years or older from the China Health and Retirement Longitudinal Study who reported having natural teeth in 2015 and had MCMI measured in 2011 and 2015. Using wave-specific medians, we grouped participants into four 2011–2015 MCMI patterns: stable high, high-to-low, low-to-high, and stable low. Incident edentulism was defined as self-reported complete tooth loss in 2018 among those at risk in 2015. Risk ratios were estimated using modified Poisson regression with robust variance, with stepwise adjustment for sociodemographic and behavioral factors; fully adjusted models additionally included baseline MCMI. Cumulative MCMI and simple change in MCMI were also examined as tertiles. Results Among 3,626 adults, 231 developed edentulism by 2018, yielding a 6.4% cumulative incidence. Compared with the stable high pattern, participants who moved from high to low MCMI had a lower risk of incident edentulism in the fully adjusted model (adjusted RR, 0.57; 95% CI, 0.34 to 0.97). The low-to-high and stable low patterns showed no clear associations after adjustment. In parallel analyses, neither cumulative MCMI nor simple change tertiles demonstrated a consistent gradient in edentulism risk. Conclusions In this longitudinal CHARLS analysis, a high-to-low MCMI change pattern was associated with lower subsequent risk of incident edentulism compared with persistently high MCMI, whereas cumulative MCMI and absolute change showed no consistent gradient. Because edentulism is a late-stage cumulative endpoint and baseline oral health measures were unavailable, these findings should be interpreted as observational associations rather than evidence of a direct cardiometabolic effect on tooth loss.