UC Irvine

Objectives

Serum creatinine-based estimated glomerular filtration rate equations and muscle mass are powerful markers of health and mortality risk. However, the serum creatinine-to-cystatin-C ratio may be a better indicator of health status. The objective of this study was to describe the relationship between creatinine-to-cystatin-C ratio and all-cause mortality when stratifying patients as per race and as per chronic kidney disease status.

Methods

This was a retrospective cohort study examining black and nonblack US veterans between October 2004 and September 2019, with baseline cystatin C and creatinine data from those not on dialysis during the study period. Veterans were divided into four creatinine-to-cystatin-C ratio groups: <0.75, 0.75-<1.00, 1.0-<1.25, and ≥1.25. The primary outcome of interest was all-cause mortality subsequent to the cystatin C laboratory measure.

Results

Among 22,316 US veterans, the mean (± standard deviation) age of the cohort was 67 ± 14 years, 5% were female, 82% were nonblack, and 18% were black. The proportion of black veterans increased across creatinine-to-cystatin-C ratio groups. In the fully adjusted model, compared with the reference (creatinine-to-cystatin-C ratio: 1.00-<1.25), a creatinine-to-cystatin-C ratio <0.75 had the highest mortality risk among both black and nonblack veterans (nonblack: hazard ratio [HR] [95% confidence interval {CI}]: 3.01 [2.78-3.26] and black: 4.17 [3.31-5.24]). A creatinine-to-cystatin-ratio ≥1.25 was associated with lower death risk than the referent in both groups (nonblack: HR [95% CI]: 0.89 [0.80-0.99] and black: HR [95% CI]: 0.55 [0.45-0.69]). However, there was a significant difference in the effect by race (Wald's P-value: <0.01).

Conclusions

Higher creatinine-to-cystatin-C ratios indicate better health status and are strongly associated with lower mortality risk regardless of the kidney function level, and the relation was similar for both black and nonblack veterans, but with different strengths of effect across racial groups. Thereby, use of a fixed race coefficient in estimating kidney function may be biased.