INTRODUCTION

Chronic kidney disease (CKD) is currently recognised as a leading health problem worldwide, not only because of the need for significant healthcare resources in patients reaching end-stage CKD, requiring dialysis and transplant, but also because of the significant burden of cardiovascular disease, hospitalisation, and early death inherent to CKD diagnosis.1

There is increasing evidence that these adverse events may be prevented or, at least, delayed.2 In 2002, the US National Kidney Foundation (NKF) Kidney Disease Outcome Quality Initiative (K/DOQI) published clinical guidelines to define and to classify CKD in different stages.3 In 2005, another initiative, Kidney Disease: Improving Global Outcomes (KDIGO), international in character, accepted with minor clarifications the definition and classification initially proposed by the K/DOQI.4

There is currently general agreement in that CKD classification is based on the estimated glomerular filtration rate (GFR) as the parameter to measure kidney function, particularly at low GFR ranges. For GFR values > 60 mL/min/1.73 m2, other renal «damage» markers (albuminuria, haematuria, changes in imaging tests, etc¿) are required due to the imprecision of GFR estimation in higher values, amongst other reasons.3,4 Despite the existence of some limitations, the abbreviated equation from the MDRD study (Modification of Diet in Renal Disease)5 has been validated in multiple studies and conditions, and is now the formula to estimate GFR in adults most commonly recommended4 not only by many nephrological societies,3,6 but also in other clinical guidelines, such as those from the Joint National Committee 7 (JNC7)7, the American Heart Association, and others.8,9 We recently published, on behalf of a work group of the Spanish Society of Nephrology (SEN) and the Spanish Society of Clinical Chemistry (SEQC), a consensus document on «Recommendations about use of equations for estimating the glomerular filtration rate in adults» in which automatic calculation of GFR using the MDRD-4 formula is advised whenever plasma creatinine levels are requested.6 On the other hand, for laboratories using a creatinine measuring procedure traceable to the reference method (isotopic dilution mass spectrophotometry), use of the MDRD-IDMS variant is recommended, as advised by the National Kidney Disease Education Program.10

However, these recommendations have not bee still implemented by all laboratories, and international recommendations have not been followed yet in many countries. There is thus a need for tools that allow for rapid translation of serum creatinine into an estimated GFR at different ages and for both sexes. We therefore calculated and designed numerical tables with different clinically significant cut-off points, as well as a table globally integrating an estimation of GFR from plasma creatinine levels, age and sex using the MDRD-IDMS method, emphasising the need for standardisation of plasma creatinine measurement at the different laboratories.

METHODS

Results shown in the tables were obtained using a spreadsheet prepared with Excel software (Microsoft, USA). The creatinine value corresponding to the different GFRs at each of the 5-year intervals between 20 and 90 years of age was calculated for both sexes by first finding serum creatinine raised to its exponent in each formula (MDRD-4/ MDRD- IDMS) (table I). Then, by fixing the GFR value, the spreadsheet is used to obtain the value of creatinine raised to its exponent for each of the ages of interest. By a linear estimation consisting of the practice of 100 iterations by the Newton method, with a precision of 0.000001, the serum creatinine value is finally obtained. Because of the characteristics of the average population in our country, the correction factor for the black race was obviated.

Data in table V were obtained by applying the MDRDIDMS formula to each of the creatinine values and ages.

RESULTS

Table II shows the serum creatinine value corresponding to an estimated GFR of 60 mL/min/1.73 m2 according to the MDRD-4 and IDMS formulas. The MDRD-IDMS formula should be used when creatinine was measured using a method with spectrophotometric traceability, as previously mentioned. 10 This GFR maintained for > 3 months normally defines the presence of CKD without the need for kidney damage markers (stage 3).3,4

Table III shows the serum creatinine value corresponding to an estimated GFR of 30 mL/min/1.73 m2 according to the MDRD-4 and IDMS formulas. This GFR maintained for > 3 months normally defines the presence of stage 4 CKD without the need for kidney damage markers.3,4 This stage is generally considered per se, among others, a criterion for referral to specialised care, although some divide stage 3 into two substages (3A and 3B) partially for this purpose.11

Table IV shows the serum creatinine value corresponding to an estimated GFR of 15 mL/min/1.73 m2 according to the MDRD-4 and IDMS formulas. This GFR maintained for > 3 months normally defines the presence of stage 5 CKD without the need for kidney damage markers.3,4 This stage not only represents a criterion for mandatory (though late) referral to specialised care, but may in itself be an indication for early entry in a dialysis programme in some patients (i.e. CKD in diabetic patients, refractory heart failure, etc.).12-14

Table V shows the integrated estimation of GFR by plasma creatinine values, age and sex, using mean creatinine values and age for each interval and the MDRD-IDMS formula, stratified into the different CKD stages

DISCUSSION

Current recommendations by different national and international societies to estimate GFR using formulas have not still been automatically implemented by some laboratories, and such initiative has not been taken in many countries because of a lack of leadership or willingness. As a result, there is a need for tools permitting rapid translation of serum creatinine into an estimated GFR for the different ages and weights, thus adapting clinical practice to current guidelines. Plasma creatinine levels have been used as a measure of kidney function because they are simply tested, but are affected by many other parameters that go beyond the GFR itself. Hence, a same plasma creatinine value does not reflect the same grade of kidney function in all patients. Thus, the tables shown allow for converting a parameter as little sensitive and precise as plasma creatinine into an estimated GFR that, being inaccurate, is currently considered to be the most practical parameter for assessing kidney function without requiring 24-hour urine collection with their attendant inaccuracies.15,16 In addition, early detection of CKD also allows for detection of the associated complications (anaemia, secondary hyperparathyroidism, etc.),17 modification of treatments or therapeutic objectives18,19 (e.g. blood pressure values, start of angiotensin converting inhibitors or angiotensin II receptor blockers, plasma cholesterol or LDL goals, among others) for preventing progression and the associated cardiovascular complications,20 or awareness that once CKD is detected, nephrotoxics or dangerous drugs in this setting should be avoided (i.e. potassium sparing agents).

The tables provided therefore not only allow for seeing that methodological differences exist in creatinine measurement with use of both formulas, but also very visually show (e.g. Table II) that plasma creatinine values as low as 0.88 to 1.15 mg/dL (78 to 102 μmol/L) may correspond to a diagnosis of CKD in women at various ages. On the other hand, at the other side of the spectrum, plasma creatinine levels of 3 mg/dL (270 μmol/L) could indicate a GFR of 15 mL/min/1.73 m2 in a 70-year old woman, suggesting the nephrogenic origin of a potential associated anaemia, or the need to start dialysis in a diabetic patient.

While any formula (i.e. the Cockcroft-Gault formula) is probably better than serum creatinine, the formulas derived from the MDRD study are the ones currently recommended by most clinical guidelines and societies.3,4,11,21-23 However, if they are not automatically reported by the clinical laboratories, they cannot be calculated without use of programmed computers. This is one of the reasons why the Cockcroft and Gault formula24 continues to be the most widely used in clinical practice because of its simplicity. However, it is well known that this formula is highly imprecise, particularly as the GFR decreases, and only reports creatinine clearance (mL/min), not GFR (mL/min/1.73 m2). While this latter aspect is controversial,25,26 the European Best Practice Guidelines (EBPG), for instance, do not consider the Cockcroft-Gault equation as an adequate method for deciding entry of a patient into a dialysis programme.12 In fact, many studies comparing both equations in different population groups have been published in recent years. Results reported by the different studies have varied, depending not only on the characteristics of the populations analysed and their mean GFRs, but especially on the gold standard used to assess GFR, and particularly on the creatinine measurement method, which makes comparison of the results obtained difficult.6 Moreover, the Cockcroft-Gault equation has not been re-expressed for use with standardised creatinine tests, and is therefore likely to routinely overestimate the actual GFR and be even less useful in the future.27

Our comprehensive review of international guidelines found that only the CARI guidelines (Caring for Australasians with Renal Impairment) recommend use of the Cockcroft-Gault formula for calculating kidney function in patients with CKD. The Australian organisation published similar tables based in this other formula.28 On the other hand, the British Columbia Health Service recently published tables based on the MDRD-4 equation only where the age range for which the GFR for each creatinine value is analysed is wider, and which are therefore less precise.29 In this study, however, Table V overall shows the GFRs as a function of serum creatinine and age, using mean creatinine values and age for each of the intervals according to the MDRD-IDMS formula. It should be noted that the main limitation for use of estimating equations stems from the lack of standardisation of methods to measure serum creatinine and their different degrees of inaccuracy, imprecision, and susceptibility to interference. Use of calibration materials with traceability to the accepted reference method (isotopic dilution mass spectrophotometry or IDMS) is therefore currently recommended.6,10 This is why this study provides tables corrected by the MDRD-IDMS formula and integrates GFRs in a single table in accordance with these recommendations, thus calling clinician attention to this significant aspect of methodology. In fact, diagnostic laboratories are revising their tests to be in line with this method, that also appears to have a greater precision in different ranges.21,30

To summarise, while the procedure currently recommended for estimating GFR would be automatic calculation using the MDRD-IDMS formula, with adequate creatinine measurement by a method with traceabililty, the availability of tables with the MDRD-4 and MDRD-IDMS formulas allows clinicians for visualisation and conversion of plasma creatinine, not only to detect the presence of CKD with creatinine values even in the normal range, but also to rapidly and simply transform plasma creatinine into a more clinically significant parameter such as GFR.