Chronic Kidney Disease
was measured on the same day as GFR in only 14 % of patients. Interestingly, the authors demonstrated that bias was influenced by age and GFR level, but to the same extent with both equations. GFR estimations were not influenced by BMI with either equation.35
This is of interest because, in these subjects, bias and precision were only influenced by changes in mGFR, not by changes in age or other anthropometric variables. The authors confirmed that the CKD-EPI equation underestimated GFR less than the MDRD study equation. The performance of both equations improved after donation, and this was particularly the case for the MDRD study equation. Importantly, the decrease in mGFR was underestimated to the same extent with the two equations.36
Lane et al. also studied patients before and after nephrectomy, but these patients had urological diseases.32
Their study is somewhat less
informative because, even if the authors also demonstrated a better performance of the CKD-EPI equation, they analysed in the same set results before and after nephrectomy. Moreover, the patients are not compared to themselves.32
Four studies have focused on the estimation of GFR in diabetic patients.31,37–39
Results have been somewhat controversial, as some methodological errors have occurred in some studies. However, all studies concluded that the performance of the CKD-EPI equation was, at best, comparable to the performance of the MDRD study equation.
Several studies compared the ability of the CKD-EPI and MDRD study equations to estimate GFR in renal transplant patients.40–45 authors came to the same conclusion as the CKD-EPI group,40,41
superiority of the CKD-EPI equation was far from being confirmed by all the studies, even in the higher GFR levels.42–45
The most interesting study on the relative performances of the MDRD study and CKD-EPI equations is certainly the one published in 2011 by Murata et al. from the Mayo Clinic.45
The sample was
impressive, with 5,238 patients included. The authors analysed the results both globally and according to health status. Therefore, results were analysed globally but also according to five subgroups: potential kidney donors; donors after nephrectomy; native CKD patients; renal transplant patients; and non-kidney transplant patients. The vast majority were Caucasian (only 98 were AA). Analysing bias and accuracy, the authors confirmed that the CKD-EPI equation underestimated mGFR in kidney donors before and after donation to a lesser degree. They thus confirmed prior data from Tent et al.36
Data about precision were not clearly given,
but were probably comparable between the two equations.45 However, the performance of the CKD-EPI equation was not better, and actually slightly worse, compared with the MDRD study equation in the three other types of patients (‘the CKD patients’), at least for eGFR below 90 ml/min/1.73 m2. In CKD stages 3A, 3B and 4, the CKD-EPI equation overestimated the mGFR more than the MDRD study equation. Among the total patient population, 20 % were over 70 years of age, but the majority of these patients were in the three CKD subgroups, making definitive conclusions difficult. However, it seems that the performance of the CKD-EPI equation was globally identical to that of the MDRD study equation, but that it was worse in patients aged over 70 years compared with patients aged between 40 and 69 years. The authors also underlined the relative
18
If some the
Tent et al. studied the same kidney donors before and after donation.36
poor performance of both equations when they studied the concordance in determining CKD stages. For example, the CKD-EPI equation classified in the same CKD stage as mGFR only 50 % of potential kidney donors for CKD stage 2 (38 % for the MDRD study equation) and only 27 % of potential kidney donors for CKD stage 3A (15 % for the MDRD study equation). In the three CKD patient groups (native CKD, renal transplant and non-kidney transplant patients), the capacity of the CKD-EPI and MDRD study equations to correctly classify patients as CKD stage 2, 3A and 3B was between 40 and 62 %, which was slightly disappointing.45
Epidemiological Impact
The choice of both the biomarker and the equation to estimate GFR can strongly impact the results of epidemiological studies evaluating the prevalence of CKD in the general population (most of the time defined by an eGFR below 60 ml/min/1.73 m2, which is the limit of stage 3 CKD).25,46
Because the CKD-EPI equation is supposed to underestimate mGFR less, it can be expected that the prevalence of CKD will be lower when this equation is used compared with the MDRD study equation. This was indeed observed in the princeps study analysing data from the National Health and Nutrition Examination Survey (NHANES) study.21
For example, the prevalence of stage 3 CKD
in the NHANES population changed from 7.8 % with the MDRD study equation to 6.3 % with the CKD-EPI equation.21
Such differences have
been confirmed by other authors, including ourselves, even if the difference in prevalence seen with the two equations may vary according to the population studied.25,47–56
Interestingly, White et al.
and Matsushita et al. analysed data from the Australian diabetes, obesity and lifestyle (AusDiab) study and the Atherosclerosis risk in communities (ARIC) study, respectively.47,49
Both groups showed that
patients reclassified to upward CKD stage (mostly from stage 3 to stage 2) with the CKD-EPI equation had a lower mortality risk than those not reclassified, suggesting a better categorisation. These results have, however, not been confirmed by all authors.52
Some points merit recall and discussion. First, the differences in prevalence seem particularly important in women and in younger subjects. It has already been demonstrated that the underestimation of mGFR by the MDRD study equation was particularly important in young women, and the CKD-EPI equation seems better in this specific population, from an epidemiologic point of view as well as for a better CKD screening policy.9,21,25,51
The second point concerns the prevalence of stage 3 CKD in the elderly population. As already mentioned, there is still doubt regarding the true performance of the two equations in elderly subjects. Moreover, there is now a hot debate in the literature regarding whether or not the current definition of stage 3 CKD (set at 60 ml/min/1.73 m2) is applicable to elderly patients.51,57–59
This topic
is outside the scope of our article, but it is interesting to stress that differences in eGFR between the MDRD study and the CKD-EPI equations will decrease with increasing age. We can reasonably say that the impact of the CKD-EPI equation on the prevalence of CKD stage 3 in subjects over 70 years of age is negligible, and that this prevalence measured by the CKD-EPI equation is not very different from the results obtained with the MDRD study equation.21,51,53,56
Once
again, it must not be concluded from this observation that the CKD-EPI equation is unbiased and precise in the elderly subjects. Data are lacking34 are needed.
EUROPEAN NEPHROLOGY and additional studies on this specific population
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76