Bidirectional association between sarcopenia and diabetic kidney disease: a prospective cohort study.

BACKGROUND

Sarcopenia and diabetic kidney disease (DKD) commonly coexist in people with type 2 diabetes, but the temporal direction of their association and potential subgroup differences are not fully understood.

METHODS

In this hospital-based prospective cohort, we performed bidirectional analyses. We evaluated baseline sarcopenia as a predictor of incident DKD among participants without DKD at baseline (n= 4,826), and baseline DKD as a predictor of incident sarcopenia among participants without sarcopenia at baseline (n= 4,798). Sarcopenia was defined using the AWGS 2019 criteria based on appendicular skeletal muscle mass index (ASMI), handgrip strength, and 4 m gait speed, requiring low muscle mass plus low muscle strength or low physical performance. DKD was defined as eGFR < 60 mL/min/1.73 m² and or persistent UACR ≥ 30 mg/g, supported by physician diagnosis after excluding primary non-diabetic kidney diseases. Associations were examined using stepwise multivariable Cox models, restricted cubic splines to assess nonlinearity, and prespecified subgroup analyses with interaction testing.

RESULTS

Baseline sarcopenia was associated with a higher risk of incident DKD, with incidence rates of 69.49 and 40.39 per 1,000 person-years in participants with and without sarcopenia, respectively. The fully adjusted hazard ratio was 1.587 (95% CI, 1.432 to 1.755; P < 0.001). Conversely, baseline DKD was associated with a higher risk of incident sarcopenia, with incidence rates of 104.69 and 40.61 per 1,000 person-years in participants with and without DKD, respectively. The fully adjusted hazard ratio was 2.036 (95% CI, 1.741 to 2.384; P < 0.001). In prespecified subgroup analyses, hazard ratios were greater than 1 across strata. Significant interactions were observed for the sarcopenia to DKD association by age group and low muscle mass, and for the DKD to sarcopenia association by sex, low muscle mass, and diabetes medication use (all P for interaction < 0.05). Restricted cubic splines showed nonlinear relationships of eGFR and log-transformed UACR with incident sarcopenia risk, and inverse associations of ASMI, handgrip strength, and gait speed with incident DKD risk.

CONCLUSIONS

Sarcopenia and DKD were bidirectionally associated in this cohort of adults with type 2 diabetes. Interactions were present for selected stratifying factors, but the direction of association remained consistent across subgroups, supporting mutual screening and integrated risk-reduction strategies.