Slide 1

Well, the question is, how to assess glomerular filtration rate or GFR to be short?

Slide 2

Before addressing this question I would like to address another one which is why to assess GFR? I have 3 answers to this question. The first one is that classification of CKD is based on GFR values and not on creatinine values.

Slide 3

As you know now, CKD is classified into 5 stages. Stage 1, which is known as kidney damage with normal or increased GFR.

Slide 4

It’s defined by a GFR value above 90ml/min normalised by body surface area together with at least one sign of renal damage, albuminuria, proteinuria or hematuria. Stage 2. Kidney damage with mild decrease in GFR. It’s defined by a GFR value between 60 and 89 with at least one sign of renal damage. Stage 3 or moderate decrease in GFR is defined by a GFR value between 30 and 59. Stage 4 or severe reduction in GFR is defined by a GFR value between 15 and 29. Finally, stage 5 or kidney failure is defined by a GFR value below 15 ml/min normalised by body surface area.

The other reason is that the risk of death increases with the stage of chronic kidney disease. This is one of the results of a well known study released 2 years ago that shows that the risk of death from any cause increases with a decrease in GFR. Also the risk of cardiovascular events increases with the decrease of GFR. The risk of hospitalisation increases with the decrease in GFR.
The third reason for assessing GFR is that the action plan is based on the CKD stage. For patients in CKD stage 1 it is advised to diagnose and treat the underlying disease, to treat comorbid conditions, to slow progression and to reduce CVD risk.

Slide 5

So, because we need to measure or to know the GFR value the question is how to assess GFR? What we know now is that creatinine concentration does not provide a reliable estimation of GFR.

Slide 6

Probably you are very familiar with this cartoon showing a plot of serum creatinine concentration against measured GFR. What you see with such a cartoon is that most of the patients with CKD stage 2 and many patients with CKD stage 3 have a creatinine concentration below 120 µmol/L.

Slide 7

Another way to put the question is to try to answer the question which could be the GFR value when serum creatinine concentration is equal to 80 µmol/L?
In patients younger than 65 the 29% confidence interval for measured GFR with a serum creatinine of 80 µmol/L is 50-120, so patients with such a creatinine value may well have CKD stage 3. In patients over 65 the 95% confidence interval for measured GFR is 40-95 and about 1/3 of patients with a serum creatinine at 80 have CKD stage 3.

Slide 8

Another estimator of GFR is 24-hour creatinine clearance and as you can see here, 24-hour creatinine clearance is not a good estimator of GFR. On this plot you can see that 24-hour creatinine clearance overestimates GFR and you can see also that many patients with a normal value of creatinine clearance, in fact, have stage 2 or even stage 3 CKD. You can also see that patients with a totally normal GFR have reduced creatinine clearance due to the poor quality of urine collection.

Slide 9

Creatinine based formulas have been developed to estimate GFR and one of the oldest is the Cockroft and Gault formula or Gault and Cockroft formula as you wish.
On the left part of the slide you can see the plot of estimated GFR value against measured GFR value and what appears is that this estimator provides a quite good correlation but despite the fact that the correlation is quite good that means that in the whole population the estimator is not biased, you can see using a Bland-Altman plot that the estimator is not very sensitive because of the scattering of individual values around the mean.

Slide 10

A more recent estimator is the MDRD formula and you can see here that the correlation between GFR estimated with the MDRD formula and measured GFR is quite good and once again, in the whole population the estimator is not biased because the mean difference between measured and estimated GFR is very close to 0. But once again, the estimator is not very sensitive and the scattering of individual values around the mean indicates it.

Slide 11

We tried to identify some reasons for this lack of sensitivity and we have seen, for example, that in patients younger than 65 the Cockroft and Gault formula seriously overestimates the GFR both in males and in females and that in patients older than 65 the Cockroft and Gault formula underestimates seriously GFR. The MDRD formula was less sensitive to the influence of age than was the Cockroft and Gault formula.

Slide 12

Another source of bias is the weight of the patient and more precisely the body mass index of the patient. As you can see here, in patients with low body mass index the MDRD formula very seriously overestimates GFR. In using the Cockroft and Gault formula we can observe that GFR is overestimated in patients with low BMI but also in patients with high BMI over 30.

Slide 13

So the current recommendations to assess GFR are to use creatinine based formulas but in selected subpopulations it is advised to measure GFR with an exogenous tracer because of the bias. The subpopulations that are considered for GFR measurement are patients at extremes of age and body size, pregnant patients, patients with severe malnutrition or obesity, patients with diseases of skeletal muscle, paraplegia or quadriplegia. Patients on a vegetarian diet or patients with rapidly changing kidney function because these formulas are not appropriate for acutely changing renal function. It is also advised to measure GFR prior to using drugs with a significant toxicity that are actually excreted by the kidney. It is advised to measure GFR prior to kidney donation and to measure GFR in clinical research projects in which GFR is the primary outcome. Thank you