This site is intended for health professionals only

At the heart of general practice since 1960

CKD targets: good intentions, bad policy

The QOF targets for CKD amount to a screening programme that may do more harm than good, argues Dr Paul Giles

The QOF targets for CKD amount to a screening programme that may do more harm than good, argues Dr Paul Giles

By following the quality framework targets for chronic kidney disease, GPs are effectively operating an eGFR screening programme. After all, a patient's eGFR is now reported with every new creatinine result, while GPs' information systems seek out old creatinine results too. There's no doubt eGFR is valuable for monitoring patients with known chronic kidney disease, but is it a valid screening test?

We should be asking some hard questions about its use for screening without evaluation against any of the established screening criteria. Serum creatinine is insensitive in early renal impairment, and the hope was that the MDRD formula for eGFR, which includes age, sex and racial group, as well as creatinine, would overcome this. The problem is that whereas eGFR agrees well with reference methods in patients with known CKD, in individuals with normal or near-normal renal function it is seriously inaccurate – underestimating true GFR on average by 10-27%. Worse still, the degree of error varies hugely between individuals.

False positives

In a patient with normal serum creatinine but an eGFR of 60ml/min/1.73m2, the true GFR could lie anywhere between about 40 and more than 100. Sensitivity has been enhanced at the expense of specificity. If you apply low specificity tests to populations with low prevalence you generate many false positives – and with eGFR screening there could be more false positives than true positives. Moreover, the MDRD formula is not well validated in important sub-groups, including Asian people and the elderly.

Some of the errors with eGFR will be reduced by standardising the ways labs measure creatinine, but there are serious problems besides. Creatinine can be affected by non-renal factors, which can translate into significant errors in eGFR. And what proof is there that patient outcomes can be improved by eGFR screening?

Some CKD patients develop kidney failure; many succumb to cardiovascular disease. The hypothesis is that early detection of CKD might help GPs predict and modify these risks. But a recent evaluation of the original MDRD trial published in the Annals of Internal Medicine found no graded association between baseline eGFR and subsequent cardiovascular death or all-cause mortality. If eGFR performed so disappointingly in CKD patients in the trial in which the MDRD formula originated, why suppose it will do any better in screening a mixed population, most of whom will not have CKD?

Even if early CKD could be reliably detected by screening, what evidence is there outcomes could be improved beyond what is achievable by treatment of other risk factors such as blood pressure, diabetes and cholesterol? Nephrology suffers from a paucity of trials of interventions, and patients with renal impairment have been excluded from many of the cardiovascular trials that underpin current practice.


It's hard to gauge the effect of being labelled as having a chronic disease on patient anxiety, but it may distort therapeutic decisions. Consider hypertension in the elderly. Under the QOF, there is no upper age limit for inclusion on the CKD register, and most will be elderly. There will clearly be tension between the QOF target, to treat up to 80% of hypertensive patients on the CKD register with an ACE inhibitor or ARB, and the 2006 NICE guidance on hypertension, which recommends other agents as first-line treatment in patients over 55.

GPs can take a number of practical steps to minimise problems, including checking whether the local lab participates in the National Quality Assurance Scheme to harmonise eGFR results, confirming borderline results on fasting samples and seeking evidence of renal disease from other sources. They can also pass the Scottish verdict of 'not proven' on patients with an eGFR below 60 but normal serum creatinine and urine, and no other evidence of CKD.

But we need researchers and policymakers to do more. We need better ways to predict risk of renal and cardiovascular complications in patients with early CKD. The underlying assumption that if only we could measure GFR properly we could predict these problems may itself be wrong. In the meantime, we need objective measures of the balance of benefit and harm of eGFR screening, both to individuals and to health economies.

In my view, it is based on wishful thinking and extrapolation far beyond the evidence. Good intentions are not enough to determine public policy.

Dr Paul Giles is a consultant chemical pathologist at Walsall Hospitals NHS Trust and co-author of an analysis of eGFR published in the BMJ last month

Rate this article 

Click to rate

  • 1 star out of 5
  • 2 stars out of 5
  • 3 stars out of 5
  • 4 stars out of 5
  • 5 stars out of 5

0 out of 5 stars

Have your say