2370 Creatinine, cystatin C, muscle mass, and mortality: findings from a primary and replication population-based cohort

Abstract Background and Aims Creatinine measures are used as initial test to derive eGFR and confirmatory testing with circulating cystatin C is recommended when creatinine-based eGFR is considered less accurate due to deviant muscle mass. Low muscle mass is associated with increased risk of premature mortality. However, the associations of circulating creatinine and cystatin C with muscle mass and mortality remain unclear and require further investigation to better inform clinical decision-making. This study aimed to disentangle these complex relationships and better inform clinical decisions on marker reliability in varying settings by investigating whether cystatin C adjustment could clarify the associations of circulating creatinine with muscle mass and mortality. Method We conducted a prospective, observational cohort study including 8, 437 community-dwelling adults enrolled in the Dutch PREVEND study and 5, 033 in the U.S. NHANES. Associations of serum creatinine and/or cystatin C with muscle mass surrogates and mortality were quantified with linear and Cox proportional-hazards regression, respectively. Missing observations in covariates were multiply imputed using Substantive Model Compatible Fully Conditional Specification. Results Mean (SD) age of PREVEND and NHANES participants (50% and 48% male) were 49.8 (12.6) and 48.7 (18.7) years, respectively. Median (Q1-Q3) serum creatinine and cystatin C were 0.79 (0.69-0.90) and 0.90 (0.70-1.00) mg/dl and 0.87 (0.78-0.98) and 0.91 (0.80-1.10) mg/l, respectively. Higher serum creatinine was associated to greater muscle mass, while serum cystatin C was not associated with muscle mass. Adjusting both markers for each other strengthened the positive relationship between serum creatinine and muscle mass and revealed an inverse association between serum cystatin C and muscle mass. In the PREVEND cohort, 1, 636 (19%) deaths were registered over a median follow-up of 12.9 (5.8-16.3) years with a 10-year mortality rate (95% CI) of 7.6% (7.1-8.2%). In the NHANES, 1, 273 (25%) deaths were registered over a median follow-up of 17.9 (17.3-18.5) years with a 10-year mortality rate of 13.8% (12.8-14.7%). Both markers were associated with increased mortality (Fig. panels A and B). Notably, when adjusted for each other, higher serum creatinine was associated with decreased mortality (Fig. panel C), while the association between serum cystatin C and increased mortality strengthened (Fig. panel D). The shapes of the associ