Letter by Delanaye et al Regarding Article, “Drawbacks and Prognostic Value of Formulas Estimating Renal Function in Patients With Chronic Heart Failure and Systolic Dysfunction”
We read with interest the study by Smilde et al1 about the formulae estimating renal function in patients with chronic heart failure (CHF). The topic is certainly important. Moreover, the population studied was large, and glomerular filtration rate (GFR) was measured by a reference method (iothalamate). However, we would like to know the opinion of the authors about some points of the methodology that we consider to be potential sources of bias.
First, the authors normalized GFR for body surface area. This is questionable, especially in a population with a high percentage of obesity.2 Secondly, in the Modification of Diet in Renal Disease study (MDRD),3 creatinine was measured with a kinetic Jaffé method from Beckman, whereas in the present paper, the authors used a different method, apparently without calibrating their creatinine results. Can the authors be sure that this is not a source of bias, especially in the range of low creatinine values?4 It is generally accepted that there is an underestimation of MDRD formulae in normal creatinine levels. Lack of calibration is not the only reason for the underestimation of MDRD formulae in normal creatinine levels, but the absolute values of the creatinine-based formulae given in the receiver operating characteristic curve analysis should not be interpretable. We would greatly appreciate the authors’ opinion regarding this issue.
Thirdly, the authors do not mention the relationship between serum creatinine and muscular mass as another potential source of bias in their population. This might explain some results in the subgroups analysis. In ill patients with decreased muscular mass, serum creatinine will not increase as much as expected for the GFR decrease. In these low–muscular mass patients with chronic renal failure, MDRD formula would thus overestimate true GFR.5 The authors state that in patients with severe CHF, formulae precision and accuracy were improved compared with patients with mild CHF. We propose another explanation: the most severe CHF patients probably would have the highest decreases in muscular mass. In such patients, the underestimation of the MDRD formulae, owing to most of the patients having GFR over 65 mL/min per 1.73 square meters, could have been thwarted by overestimation attributable to the decreased production of creatinine. If MDRD can be considered better, one should keep in mind that the explanation of this fact could not be intellectually satisfactory (ie, association of 2 compensating errors). Our last comment concerns the statement of the authors that in obese populations, the use of the Cockcroft–Gault formula might be recommended instead of the MDRD formula. This assertion does not seem to be supported by the literature, including the article by Cockcroft and Gault.6 In an obese, renal-failure population, the Cockcroft–Gault formula will usually overestimate GFR because weight is included in the numerator. Thus, in obese patients with normal GFR, the underestimation of GFR could again be thwarted by the overestimation linked to the weight in the numerator. Once again, this could be a phenomenon of 2 compensating errors. We would greatly appreciate the authors’ comments regarding the concerns we have expressed. Nevertheless, the study realized by Smilde et al1 is fundamental and gives relevant, clinically important information. Indeed, this study justifies the use of the MDRD equation in CHF patients, reminding cardiologists that these patients are at high risk of renal failure.
Smilde TD, van Veldhuisen DJ, Navis G, Voors AA, Hillege HL. Drawbacks and prognostic value of formulas estimating renal function in patients with chronic heart failure and systolic dysfunction. Circulation. 2006; 114: 1572–1580.
Delanaye P, Radermecker RP, Rorive M, Depas G, Krzesinski JM. Indexing glomerular filtration rate for body surface area in obese patients is misleading: concept and example. Nephrol Dial Transplant. 2005; 20: 2024–2028.
Coresh J, Eknoyan G, Levey AS. Estimating the prevalence of low glomerular filtration rate requires attention to the creatinine assay calibration. J Am Soc Nephrol. 2002; 13: 2811–2812.