Diagnostic Use
Estimated glomerular filtration rate (eGFR) is calculated using the CKD-EPI Creatinine 2009 equation, from plasma creatinine (in umol/L), age (in years, ≥18years) and gender. Overall accuracy of the equation: about 87% of the eGFR are within 30% of the measured GFR (ref 1). This equation is endorsed by the Australasian Creatinine Consensus Working Group.
Routine eGFR calculation is not offered alongside all plasma creatinine reports. However, eGFR calculation can be ordered specifically on the request form or test add by ringing 09-2760044 ext 58519.
eGFR is UNRELIABLE in the following situations:
- When meat or fish is ingested during the 2 hours period preceding collection of the blood sample used for the creatinine measurement
- Acute changes in renal function. eGFR is only valid in subjects in a steady state with respect to plasma creatinine
- Dialysis-dependent patients
- Patients with unusually high or low muscle mass. Glomerular function may be better assessed by e.g. 24 hour urine creatinine clearance test.
- Creatine supplements
- Children less than 18 years of age
- Pregnancy
- Severe liver disease
In most out-of-hospital settings, where an eGFR result is on hand and no other measure of GFR is known or readily accessible, eGFR (corrected back for the patient’s specific body size/body surface area especially if extreme body habitus) can be used to assist drug-dosing decision making. However, if critical dose adjustments is required for some narrow therapeutic index medications affected by reduced GFR, then therapeutic drug monitoring or a valid marker of drug effect should be used to individualise dosing. Direct measurement of creatinine clearance (via timed 8 hours or 24 hours urine collection) may be more appropriate – consult your team pharmacist.
eGFR and Definition of Chronic Kidney Disease
Chronic Kidney Disease (CKD) is defined as abnormalities of kidney structure or function, present for >3 months, with implications for health (ref 2). CKD involves either of the following 2 criteria:
A. Evidence of kidney damage (with or without decrease in GFR) for ≥ 3months, irrespective of underlying cause of the kidney disease:
- Persisting albuminuria (>3mg/mmol creatinine ratio. First void urine preferable)
- Haematuria after exclusion of urological causes
- Structural abnormalities e.g. from renal imaging
- Pathological abnormalities e.g. from renal biopsy
- Urine sediment abnormalities
- Electrolyte or other abnormalities indicative of tubular disorders
- History of kidney transplantation
The first 4 criteria are described in Kidney Health NZ, BPAC and Kidney Health Australia guidelines (refs. 4-6). The KDIGO 2012 guideline (ref. 2) also included the latter 3 criteria as markers of kidney damage.
B. Decreased GFR (with or without evidence of kidney damage) for ≥ 3months:
Estimated or measured GFR <60ml/min/1.73m2 (i.e. GFR categories Grade 3a – Grade 5)
Although GFR of 60-89 ml/min/1.73m2 (G2) is described as “mildly decreased” kidney function, in the absence of evidence of kidney damage a GFR of ≥ 60ml/min/1.73m2 (i.e. both G1 and G2) does not fulfill criteria for CKD.
On the other hand, in some patients although their eGFR are still within the G1 or G2 category (i.e. ≥ 60ml/min/1.73m2), a significant decline (fall of ≥ 20%, ref. 5) from a baseline higher value may reflect a pathologic process even with no other evidence of kidney damage. It will then require a high index of suspicion by the clinician to embark on additional testing or close follow up to detect early onset of CKD (ref. 2).
What is a significant change in serum creatinine and eGFR?
Minor fluctuations in eGFR are common and are not necessarily indicative of progression. Temporary elevation in creatinine (and fall in eGFR) may occur due to factors such as dehydration or recent intake of cooked red meat.
There are unavoidable biological and analytical variations in creatinine and calculated eGFR. The intra-individual biological variation of creatinine is around 4.5% (ref. 7). Coupled with the week-to-week analytical variation of creatinine assays of around 3.2%, this means 2 consecutive creatinine results from our laboratory will have to differ by more than ±15% before one can say with 95% confidence that the change cannot be solely explained by biological and analytical variation (this minimum statistical difference is called ‘Reference Change Value’, or RCV).
With creatinine as one of the variables in the CKD-EPI equation, the RCV between two consecutive eGFR results is around 23%. For example a female with creatinine of 100umol/L and an initial eGFR of 51ml/min/1.73m2, then the difference from the second eGFR will have to be > ±12ml/min/1.73m2 (or outside the 39-63 ml/min/1.73m2 range) before one can be 95% confident that the change cannot be explained by biological and analytical variation alone (refs. 8,9).
“Drop in eGFR” or “Progression” is indicated by:
- A fall in eGFR of ≥25% with a drop in eGFR category (by convention)
- A sustained decline in GFR by >5ml/min/1.73m2 per year (preferably monitored by the same method in the same laboratory). The confidence in assessing progression increases with increasing number of plasma creatinine measurements and duration of follow up (ref. 2).
Note that some medications such as ACE inhibitor and angiotensin receptor blockers can cause a predictable fall in eGFR of up to 25%, due to changes in renal blood flow. Such a fall is not regarded as pathological if the patient then reaches and stabilized at a new baseline.
Age and eGFR
Age (>60 years) is a risk factor for CKD , but is among numerous other risk factors such as (refs. 3-6):
- Māori /Pacific/Indo-Asian ethnicity
- Obesity
- Smoking
- Diabetes
- Hypertension
- Cardiovascular disease history
- Systemic diseases that may affect the kidney
- Family history of renal disease/CKD
Age is one of the variables used to calculate eGFR, noting that eGFR does fall slowly in the general population. However, age-related decision points for eGFR are not recommended in adults (ref 10). The reason is that eGFR slowly falls with ‘healthy’ aging, with many elderly falling into the CKD stage 3a (45-59ml/min/1.73m 2 ) without evidence of active or structural kidney diseases, as defined by the criteria above . An eGFR in this range therefore does not necessarily of itself indicate pathology in the absence of one or more of the above factors (ref. 3).
Conversely, in population studies CKD stage 3a (compared with those with eGFR ≥60ml/min/1.73m 2 ) is associated with a higher cardiovascular (CVD) risk at all ages (ref10). Thus, while eGFR does fall slowly in healthy older individuals it is arguable whether it should be considered purely “physiological” (refs. 2,10).
Regardless, eGFR <45ml/min/1.73m2 (CKD3b) is pathological whether other factors are present or not, and is associated with increased risk of renal and cardiovascular complications irrespective of age, particularly if persisting albuminuria is also noted (Refs. 2,11).
Testing for CKD
Targeted testing for CKD should generally be linked to routine cardiovascular disease assessment or diabetes checks. It is recommended to match the eGFR with the degree of albuminuria and assess them against a 2 dimensional grid table (e.g. at ref 2) to evaluate risk of future deterioration of kidney function as well as cardiovascular and all cause mortality (Refs. 2-6, 9).
In a stable patient without acute symptoms, an initial eGFR of 45-59ml/min/1.73m2 should prompt review of the above other possible risk factors, and should be followed up with a repeat in 3 months, to check that it is stable. Caution should also be made in prescribing certain potentially nephrotoxic or predominantly renal cleared medications. Earlier follow up may be justified depending on clinical circumstances, e.g. if the patient has been unwell or there are other reasons to suspect acute renal injury.
For further guidance, please refer to Auckland Regional Health Pathways (Medical-Nephrology – Chronic Kidney Disease (CKD) in adults)
Online eGFR calculator is available e.g. from Kidney Health Australia, at https://kidney.org.au/health-professionals/egfr-calculator/ (last accessed 15 4 2026)