Diabetes
Early post-screening diabetes care and 10-year risk of hemodialysis initiation: a target trial emulation Kiyomitsu Fukaguchi* Kiyomitsu Fukaguchi Fukaguchi Fukaguchi Fukaguchi 1. Department of Health Data Science, Yokohama City University Graduate School of Data Science, Yokohama, Japan. 2. Department of Emergency Medicine, Shonan Kamakura General Hospital, Kamakura, Japan
Background: Early engagement in diabetes care after screening could prevent kidney failure, but evidence on hard renal endpoints such as hemodialysis is limited.
Methods: We emulated a target trial using the nationwide JMDC Claims Database (>12 million beneficiaries) in Japan (2005–2021). Adults aged 40–74 years with newly screen-detected diabetes (HbA1c ≥6.5% or fasting glucose ≥126 mg/dL) and no history of renal failure or dialysis were eligible. Participants were assigned to: (1) initiating ≥1 diabetes-related visit <1 year after detection (Strategy 1), or (2) no diabetes-related visit <1 year (Strategy 2). To address immortal time bias and selection bias due to censoring, we used a clone–censor–weight approach. Stabilized inverse-probability weights were estimated using >30 baseline confounders (demographics, lifestyle, metabolic parameters including eGFR and proteinuria, and health-seeking behaviors) and time-varying confounders (cumulative healthcare visits, medications). Follow-up started at diabetes detection. Outcome was maintenance hemodialysis initiation. Weighted pooled logistic regression estimated 10-year risk differences (RD) and risk ratios (RR).
Results: Among 153,455 participants (mean age 53 years; 78% men), 54,278 initiated care <1 year and 71,659 did not; the remainder were censored. Hemodialysis occurred in 74 (Strategy 1) and 54 (Strategy 2) participants (total=128). Early care was associated with RD −0.12 percentage points (95% CI −0.74 to 0.32) and RR 0.94 (95% CI 0.51 to 1.51) at 10 years. Risk curves showed slight separation after approximately 7 years, but confidence intervals substantially overlapped throughout the follow-up.
Conclusions: Among adults with screen-detected diabetes, initiating care <1 year showed no clear association with 10-year hemodialysis risk. Given renal failure’s latency and slight later divergence, longer follow-up is warranted to clarify early engagement’s impact on hard renal outcomes.
