Prevalence of kidney disease in anaemia differs by GFR-estimating method: the Third National Health and Nutrition Examination Survey (1988-94)

Abstract

Background: Anaemia worsens as kidney function declines. Both conditions are associated with increased mortality. Serum cystatin C is purportedly a more sensitive marker of kidney disease and a better predictor of mortality than serum creatinine. However, studies suggest that extrarenal factors also influence cystatin C levels.

Methods: We determined whether estimates of glomerular filtration rate [estimated glomerular filtration rate (eGFR)] based on serum cystatin C alone or in combination with serum creatinine were superior to those based on serum creatinine in recognizing impaired kidney function in the setting of anaemia in a sub-sample of the Third National Health and Nutrition Examination Survey of the USA consisting of 6734 participants, 20 years or older.

Results: The prevalence of moderate to severe kidney disease (eGFR 15-59 mL/min/1.73 m(2)) among anaemic persons was 15-16% when based on serum creatinine alone (eGFR(SCR)) or combined with cystatin C (eGFR(SCR) (+) (CYSC)); this estimate increased to nearly 25% when kidney function was estimated by cystatin C (eGFR(CYSC)). The adjusted odds ratios of kidney disease in anaemic versus non-anaemic persons were slightly higher with eGFR(CYSC) than eGFR(SCR) and eGFR(SCR) (+) (CYSC) in younger adults [odds ratio (OR) = 5.22, 95% confidence interval (CI): 2.23, 12.17], women (OR = 5.34, 95% CI: 2.36, 12.06) and those with elevated C-reactive protein (CRP) (OR = 7.36, 95% CI: 1.98-27.36).

Conclusions: Impaired kidney function was common in individuals with anaemia. Among anaemic individuals, the prevalence estimate for kidney disease was notably higher when kidney function was estimated by cystatin C alone compared with the estimations by serum creatinine alone or in combination with serum cystatin C. eGFR(CYSC) may be particularly helpful in identifying kidney disease in the setting of anaemia among younger persons, women and those with elevated CRP. Regardless of which renal biomarker is used, our study suggests that an evaluation for underlying kidney disease should be considered in the standard workup of anaemia.

Fig. 1

Predicted prevalence of kidney disease (defined as eGFR 15–59 mL/min/1.73 m²) using different GFR-estimating methods in US adults aged 20 years or older by haemoglobin levels. Estimated GFR are based separately on serum creatinine (SCr), serum cystatin C (CysC) and combined serum creatinine and cystatin C (SCr & CysC). Prevalence curves are truncated when the number of relevant participants is <30.