Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Estimated glomerular filtration rate (eGFR)

The problems
eGFR and age
eGFR and mortality
eGFR and diabetes
eGFR, heart failure, and MI
Useful websites:

For Bandolier eGFR is another case of angels fearing to tread. The estimation of GFR from serum creatinine and demographic characteristics (age, sex, ethnicity) is usually done using equations derived from blood samples taken during a baseline period of the Modification of Diet in Renal Disease study [1]. The methodology used was excellent, with 1,600 patients split between derivation and validation cohorts and using an independent measure of GFR. Most samples were taken in the morning following a period of fasting, and so provide a somewhat idealised population.

Since about 2002, following general consensus that measuring eGFR was a good thing, many laboratories have been providing an eGFR result routinely when biochemistry has been ordered on a blood sample. The problem many doctors have is how to use this information, and whom to treat, especially when eGFR is linked to the staging classification of chronic kidney disease (Table 1). For classes 2 and 3 the advice is about observation, with control of blood pressure and risk factors. Despite caveats in guidance this has the potential to cause confusion.

The problems

The kidney has much redundancy. The concern is that without age and sex related “normal” ranges (and perhaps ethnic ranges, too), there could be a huge increase in unnecessary investigations and inappropriate referrals. The problem is not whether eGFR is a useful result, but rather about the judgement of when and in whom it is useful.

Most would agree that a GFR above about 60 mL/minute was pretty good in everyone, and that one of below 30 mL/minute was probably not so good, with more concern with even lower results than this. It is the grey area between 30 and 60 mL/minute that is the problem. At its crudest, the argument would be that a GFR value of 50 mL/minute would be worrying in a young man, while one of 40 mL/minute might not be in an otherwise healthy woman of 80 years. It should be noted, though, that others would not recognise this argument at all.

A further element of the problem relates to the circumstances in which blood samples are taken. After a meat meal (hospital canteen rather than trencherman heroics) serum creatinine is increased and eGFR decreased (Figure 1) [2]. In the 32 participants in this little experiment, 12 would have fallen by at least one stage shown in Table 1, and in 11 from stages 1 or 2 to 3.

Figure 1: Changes in serum creatinine and eGFR after eating cooked meat

Table 1: Classification of kidney disease according to eGFR

Stage Description
(mL/minute/1.73 sq m)
0 Normal
≥ 90
1 Kidney damage with normal or raised GFR
≥ 90
2 Kidney damage with mild reduced GFR
3 Moderate reduced GFR
4 Severe reduced GFR
5 Kidney failure
< 15 or dialysis

It is also the case that the original eGFR equation required calibration of the serum creatinine measurement to the laboratory that developed the equation. The calculations have recently been re-expressed using a standardised assay [3].

eGFR and age

GFR in healthy young men and women is about 130 and 120 mL/minute/1.73 sq m respectively (following this, for clarity, we drop the 1.73 sq m, so it has to be assumed). After the age of 40 years GFR drops by 1 mL/minute per year. So the incidence of low GFR should rise with age.

Figure 2 shows the percentage of men and women with eGFR below 60 mL/minute in an Italian population of 4,500 people from a single town [4]. Surveys in the USA tend to show somewhat higher rates, and the largest survey there involved 2.5 million veterans [5].

Figure 2: Low eGFR prevalence in Italy

In Italians with eGFR <60 mL/minute, disorders associated with kidney disease were more common than in those with higher eGFR. In men, hypertension, high serum urate, high potassium and low calcium, and cardiovascular disease were significantly more common, while in women significantly more common were high serum urate, high potassium and low calcium, cardiovascular disease, and anaemia. With lower eGFR, the number of these associated conditions increased, to over two per patient when eGFR was below 40 mL/minute.

In US veterans [5], 20% of the whole cohort had eGFR of <60 mL/minute, with half of these having moderate reductions in the range of 50-59 mL/minute. Very low eGFR of <30 mL/minute was rare, and usually below 5% even in the cohort of 53,000 people aged 85 to 100 years old.

eGFR and mortality

Two large US studies relate eGFR and mortality or hospital admission. The earliest [6] of these examined eGFR in 1.1 million persons in an integrated system of healthcare delivery between 1996 and 2000, with a median follow up of 2.8 years. Figures 3-5 show the age-standardised rates of death from any cause, cardiovascular events, and hospital admission. Major increases in all these rates began when eGFR was below 45 mL/minute, and especially below 30 mL/minute. Fully adjusted hazard ratios are shown in Table 2.

Figures 3-5: Mortality, CV event rate and hospital admission rate according to eGFR from a large US database

Table 2: Fully-adjusted hazard ratio for low eGFRs, compared with eGFR ≥60 mL/minute

(mL/minute/1.73 sq m)
Death from any cause
Any cardiovascular event
Any hospital admission
≥ 60
< 15
All results were statistically significantly greater than for eGFR ≥60 mL/minute

An even larger US Veterans study [5] used information on 2.5 million patients (predominantly men) aged 18-100 years and with at least one serum creatinine measurement in 2000-2002, with the outcome of death by mid-2005 (average 3.2 years of follow up). Annual risk of mortality for any cause was higher with higher age and lower eGFR (Figure 6 uses best resolution possible for the axes).

Figure 6: Annual risk of mortality from any cause by age and eGFR

In patients with stable repeat creatinine measurements, and adjusting the risk of death for age, race, sex, and various chronic diseases, there was no increased risk of death at any age when eGFR was 50-59 mL/minute compared with patients in whom it was ≥60 mL/minute.

When eGFR was 40-49 mL/minute there was a significant increased risk of death in patients below 75 years, but not older. When eGFR was below 40 mL/minute, risk of death was significantly increased at all ages.

eGFR and diabetes

About 27% of 7,600 diabetic subjects in Manchester had eGFR of below 60 mL/minute (Figure 7), but only a tenth of these were below 30 mL/minute [7]. A similar pattern was seen in a survey of 3,300 diabetics in Teeside [8] and 4,300 in Wolverhampton [9]. In all these surveys only 2% or less of diabetics had eGFR below 30 mL/minute.

Figure 7: Prevalence of eGFR at different levels (log scale) in diabetic patients in Manchester, Teeside, and Wolverhampton

The Teeside study [8] had a 10-year follow up in which 36% of the cohort died. It was able to examine death rates over that time according to eGFR. The 10-year chance of death from any cause was much higher in those with the lowest eGFR (Table 3).

Table 3: Mortality by eGFR in diabetic patients over 10 years in Teeside

(mL/minute/1.73 sq m)
Death rate
≥ 90
≥ 29

The presence of low eGFR in patients with diabetes may be important. In 269 diabetic patients who were asymptomatic for coronary heart disease, all cardiac events were recorded over a 2.3 year period [10]. It transpired that 29% (77) had an eGFR below 60 mL/minute at baseline; these patients were older, had a longer duration of diabetes, and a higher prevalence of albuminuria (though 35% had normal albumin excretion).

In these 77 patients, 25% had some cardiac event (nonfatal MI, unstable angina, percutaneous coronary intervention, or coronary artery bypass surgery) That was significantly higher than the 13% rate in the 192 with eGFR of 60 mL/minute or more.

eGFR, heart failure, and MI

A systematic review of the prevalence of renal impairment in heart failure patients [11] assembled information on over 80,000 patients. Renal function was determined by a number of methods, but showed that any renal impairment was present in 63% of these patients, and moderate or severe impairment in 29%.

With greater impairment in renal function comes greater risk of mortality, especially in older patients. In a retrospective cohort of 57,000 patients (mean age 78 years) admitted to hospital with heart failure and 44,000 patients admitted with myocardial infarction who were followed up for one year [12], mortality was higher with lower eGFR, at about two or more times greater at lowest level of eGFR (60+%) compared with highest level (about 20-25%).


Using stimated GFR is complex. There are concerns over different equations for the calculations, methods of creatinine measurement, muscle mass, alcohol, and cooked meat ingestion that all serve to muddy the waters to an extent. Most of these are important, but represent for the most part the sort of squabbles that academics undertake for valid reasons, though most of the time we ordinary mortals don't need to trouble our neurone over it.

Estimated GFR will pick up impaired renal function. Most of what it does pick up tends to be in older patients with comorbidities. In most cases, the response is control of blood pressure and risk factors. The proportion of patients picked up with very low eGFR is usually tiny, 2% and below, for instance, in diabetic populations. The secret is to take note, but not over-react, and examine the result in context. eGFR is a hot topic about which there will be more in future.


  1. AS Levey et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Annals of Internal Medicine 1999 130: 461-470.
  2. DJ Preiss et al. The influence of a cooked-meat meal on estimated glomerular filtration rate. Annals of Clinical Biochemistry 2007 44: 35-42.
  3. AS Levey et al. Using standardised serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Annals of Internal Medicine 2006 145: 247-254.
  4. M Cirillo et al. Low glomerular filtration in the population: prevalence, associated disorders, and awareness. International Society of Nephrology 2006 70:800-806.
  5. AM O'Hare et al. Mortality risk stratified in chronic kidney disease: one size for all ages? Journal of the American Society of Nephrology 2006 17: 846-853.
  6. AS Go et al. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. New England Journal of Medicine 2004 351: 1296-1305.
  7. RJ Middleton et al. The unrecognized prevalence of chronic kidney disease in diabetes. Nephrology Dialysis Transplantation 2006 21: 88-92.
  8. S Nag et al. All-cause and cardiovascular mortality in diabetic subjects increases significantly with reduced estimated glomerular filtration rate (eGFR): 10 years' data from the South Tees diabetes mortality study. Diabetic Medicine 2007 24: 10-17.
  9. V Baskar et al. Clinical utility of estimated glomerular filtration rates in predicting renal risk in a district diabetes population. Diabetes Medicine 2006 23: 1057-1060.
  10. H Knobler et al. Reduced glomerular filtration rate in asymptomatic diabetic patients: predictor of increased risk of cardiac events independent of albuminuria. Journal of the American College of Cardiology 2004 44: 2142-2148.
  11. GL Smith et al. Renal impairment and outcomes in heart failure. Systematic review and meta-analysis. Journal of the American College of Cardiology 2006 47: 1987-1996.
  12. GL Smith. Serum urea nitrogen, creatinine, and estimators of renal function. Mortality in older patients with cardiovascular disease. Archives of Internal Medicine 2006 166: 1134-1142.

Useful websites:

previous or next story