To the Editor: The revised recommendations of the Australasian Creatinine Consensus Working Group1 are improved with the recognition of an age-related reduction in glomerular filtration rate (GFR), but don’t deal with a number of other significant problems associated with an estimated GFR (eGFR).

When a plasma creatinine measurement is requested, an eGFR is commonly provided, increasing the sensitivity but reducing the specificity of diagnosis of kidney disease. The eGFR remains a substantially flawed estimate of GFR. It is associated with significant predictive error (up to 30% of individual eGFRs differ by more than 30% from the measured GFR at 60–90 mL/min)2 and with substantial false positive and false negative outcomes.

The flaws in the eGFR are, firstly, the limitations of creatinine clearance rate as a measure of GFR, and secondly (and more importantly), the use of age, sex and race as surrogates for muscle mass (the determining factor in creatinine production and, together with creatinine clearance, plasma creatinine level).

Age, sex and race are imperfect predictors of muscle mass, and this leads to underestimation of GFR in people who are fit and well muscled and overestimation in those who are wasted and disabled. While reporting eGFR values represents a worthwhile advance on using plasma creatinine levels to detect kidney disease, it could be considered, at best, the “least bad” readily available measure of GFR.

When no better test is readily available, how should we handle a suboptimal measure of GFR? Educating the medical profession about the limitations of eGFR is important, but, based on personal experience and anecdotal evidence, I believe that using conventional methods of informing doctors has not been uniformly effective. Providing “just in time” information support is likely to assist this process.

Thus, I support the recommendation that laboratories routinely report eGFRs, but suggest that, when they do so, they add a product warning along the following lines:

In patients over 70 years of age, an additional product warning, consistent with the Australasian Creatinine Consensus Working Group’s revised recommendations,1 could be as follows:

To the Editor: The revised recommendations for the use of the estimated glomerular filration rate (eGFR) in the clinical setting use the Modification of Diet in Renal Disease (MDRD) formula to adjust drug dosing in people with renal impairment (Recommendation 6).1 Although the utility of the MDRD-based eGFR as a screening tool to identify people with chronic kidney disease represents an important clinical opportunity, we, and others,2,3 remain concerned about this recommendation.

Certainly, adjusting the dose of renally excreted medicines based on a patient’s serum creatinine concentration alone is not appropriate. Accordingly, many drug monographs provide explicit dosing recommendations based on the estimated creatinine clearance (calculated using the Cockcroft–Gault formula) as a rational basis for dose adjustment using sound pharmacological principles. This assumes a significant correlation between the estimated creatinine clearance and the actual clearance of a drug (or metabolite). But such a correlation has not been established for the MDRD formula.

Indeed, an empirical study comparing the use of the Cockcroft–Gault formula to the MDRD formula in 1067 elderly patients found that the MDRD formula significantly overestimated renal function and would, if used, lead to significantly higher doses of two drugs in question (enoxaparin and gentamicin) being administered.2 This highlights the need for further research to rigorously characterise the relationship between MDRD estimates of renal function and drug clearance before this formula can be recommended to guide dose adjustment in the clinical setting.

Recommendation 6, that the MDRD-based eGFR should be used for dosing after considering body size,1 requires further clinical information about the patient — the same information needed to use the Cockcroft–Gault formula. Even limiting Recommendation 6 to drugs that are not “critical-dose drugs”1 is confusing. Many drugs would be considered critical-dose drugs when used in frail older people with some degree of renal impairment.

We reaffirm the statement, from the 2008 Australian medicines handbook, that “there is no evidence that [automatically reported eGFR] is suitable for adjusting drug doses in people with renal impairment”.4

In reply: We agree with Adam’s points that:

• The estimated glomerular filtration rate (eGFR) represents a worthwhile advance on the plasma creatinine level for detecting kidney disease, and is currently the best readily available measure of GFR;

• Educating the medical profession on the limitations of eGFR is important; and

• Further refinements of eGFR accuracy are highly desirable.

Since the advent of automated laboratory reporting of eGFR, there has been a concerted effort by the Kidney Check Australia Taskforce to educate medical professionals and other clinicians about the strengths and limitations of eGFR and about how best to approach the patient with a significantly reduced eGFR. Education has been via printed office materials, accredited workshops, medical journal articles, online education, and decision-support systems embedded in medical software. Adding statements to laboratory reports has also been considered by the Australasian Creatinine Consensus Working Group, and some laboratories do provide explanatory eGFR statements. However, many have not embarked on this because of the difficulty of providing adequately informative explanatory remarks in very limited space.

Individual laboratories balance the need for such supporting information against the space requirements on reports and the diminishing effect of excessive repetition. Given that routine reports are already complex, often containing over 20 result items, and that many millions of reports with serum creatinine results are produced annually in Australia, the value of any repetitive comment needs careful consideration.

The response of clinicians to the introduction of automatic reporting of eGFR together with the linked educational campaign has been strongly positive, with outcomes that have included easier identification of chronic kidney disease, better decision making for affected patients, and more appropriate referral patterns, both in Australia and overseas. Nevertheless, we agree with Adam that eGFR is only an approximation of actual GFR and is subject to error. There are ongoing attempts to further improve the accuracy and clinical utility of eGFR through such ventures as the universal standardisation of creatinine calibration (which has already taken place in Australia and New Zealand) and the Chronic Kidney Disease Epidemiology Collaboration.1

McLachlan, on behalf of the Editorial Advisory Board of the Australian medicines handbook, reaffirms the Board’s position that eGFR is not appropriate for use in dosage calculations. However, for the reasons stated in the consensus document,2 we remain convinced that eGFR is a useful tool for most drug dosing decisions.

We note with interest the recent change in the position taken by the British National Formulary (BNF 54) to one of support for the Modification of Diet in Renal Disease (MDRD)-based eGFR1 being used in place of creatinine clearance rate “for most drugs and for most patients of average build and height”3 — a stance similar to ours. In particular, we re-emphasise that eGFR, because of its ready availability, increases the rate of identification of renal failure.

We agree with McLachlan that prescribers should continue to follow specific published recommendations for drugs such as enoxaparin and gentamicin (these are good examples of “critical-dose drugs” in the hospital setting). We note additionally that there is increasing acceptance of the eGFR in the drug literature. In the case of enoxaparin, the MDRD equation for eGFR has been used to assess the effect of renal function on bleeding in elderly patients4 and has been found to provide the best relationship with enoxaparin clearance in this setting.5

The eGFR is now an established feature of pathology reports in Australia, and we believe it is important to integrate this information into routine practice for drug dosing decisions. We therefore offer to work with the authors of the Australian medicines handbook and other interested parties to develop guidelines for drug dosing decisions using all available information.