HOW SHOULD WE MEASURE RENAL FUNCTION IN OLDER HOSPITAL PATIENTS?

Abstract

To the Editor: With interest we have read the article by Péquignot and colleagues.1 The important question the authors address is how best to estimate renal function in older hospital patients. Their conclusion is that, in patients admitted to a geriatric ward, the Modification Diet in Renal Disease (MDRD) formula overestimates renal function and that the Cockcroft-Gault (CG) formula is a better predictor in these patients, but we believe that some of their statements must be put into perspective. Péquignot and colleagues1 used measured creatinine clearance (ClCr) as the gold standard for renal function, but the real measure of interest for renal function is not ClCr per se but glomerular filtration rate (GFR). The CG formula is an estimate for ClCr,2 whereas the MDRD formula is an estimate for GFR.3 Measured ClCr according to 24-hour urine collection is only an approximation of GFR, the real measure of interest. Although CG performed better than MDRD in estimating measured ClCr in Péquignot and colleagues' study,1 this does not necessarily make CG a better estimate for GFR. Cockcroft and Gault observed that older people have lower creatinine production.4 The population of older patients in Péquignot and colleagues' study1 can be expected to have a lower creatinine production. Under these conditions, estimates of GFR based on measured ClCr will be lower than the MDRD estimate.5 Whether this then truly reflects overestimation of GFR according to the MDRD formula or underestimation according to measured ClCr is not clear. Another point of concern is whether these older patients were in steady state not only regarding GFR but also in creatinine production. Creatinine production may be even lower in this population of older hospitalized patients because of concomitant acute illness such as infection. Sepsis, for example, lowers creatinine production.6 It can be seriously doubted whether measured ClCr is a good estimate of GFR if serum creatinine is not in steady state. Tubular secretion of creatinine is another possible source of error when using measured ClCr, because active tubular secretion of creatinine will result in an overestimation of the GFR of creatinine. In addition to tubular secretion of creatinine, there might also be creatinine reabsorption by older kidneys, which further complicates the interpretation of ClCr.7 Finally, collection errors during 24-hour urine collection are a ready source of bias. Collecting urine for longer than 24 hours will result in an overestimation, whereas missing urine samples will underestimate creatine clearance. In conclusion, there are many reasons why estimating renal function in hospitalized geriatric patients is difficult and prone to errors. Further research is necessary, preferably using gold standard methods to measure GFR (for example, using inulin, 51Cr-labeled ethylenediaminetetraacetic acid, or 125I-iothalamate), before the conclusion is drawn that the CG formula is a better estimate of renal function than the MDRD formula. Conflict of Interest: The editor in chief has reviewed the conflict of interest checklist provided by the authors and has determined that the authors have no financial or any other kind of personal conflicts with this paper. Author Contributions: Joep Lagro designed and wrote this letter with concept and editing support from Jurgen A. Claassen. Both authors participated in revising the letter. Sponsor's Role: None.