Over the past few years, there has been controversy in the editorial pages of nephrology journals over the definition and classification of chronic kidney disease (CKD). While the debate occurred primarily among nephrologists, the controversy had implications for the care of CKD across all disciplines of medicine. This month, a Controversies Conference on “Chronic Kidney Disease: Definition, Classification and Prognosis” sponsored by Kidney Disease Improving Global Outcomes (KDIGO) took place in London and reached a consensus on revisions to the classification of CKD based on prognosis, but did not propose to change the definition of CKD. KDIGO is an international non-profit organization whose purpose is to improve the care and outcomes of kidney disease patients worldwide by promoting coordination, collaboration, and integration of initiatives to develop and implement clinical practice guidelines.

Prior to this conference, widespread agreement existed that kidney failure (Stage 5 CKD) is a life-threatening condition, with increasing prevalence around the world, high cost and poor outcomes. In the US, the prevalence of kidney failure treated by dialysis and transplantation is approximately 0.2% of the population (500,000 people), with an annual cost of $35 billion. Kidney disease is silent in its early stages, but can be detected by commonly available laboratory tests, such as serum creatinine to estimate glomerular filtration rate (GFR) and urinary albumin-to-creatinine ratio (ACR) as a marker of kidney damage. Earlier detection and treatment could potentially reduce disease complications and the risk of developing kidney failure.

The controversies aired at the conference centered on the current definition and classification of kidney disease, proposed by the National Kidney Foundation (NKF) Kidney Disease Outcomes Quality Initiative (KDOQI) in 2002 and subsequently adopted, with minor modifications, by KDIGO in 2005. The guidelines define CKD as either GFR <60 ml/min/1.73 m2 (less than half of the normal level in young adults) or kidney damage for >3 months, irrespective of cause of disease. A urine albumin-to-creatinine ratio >30 mg/g is defined as a marker of kidney damage. In people with CKD, the disease is further classified by the level of GFR (known as stages). Population surveys of estimated GFR and urinary ACR identify around 11% of the US adult population as having CKD using this definition (23 million people). The prevalence of CKD is as high as 40% among people over aged 70 years, primarily because of the large number of people with GFR 30-59 ml/min/1.73 m2 (CKD stage 3), many of whom have no other features of kidney damage. The prognosis of earlier stages of CKD is highly variable, with more people dying of cardiovascular disease (CVD) than kidney failure. Based on similar findings around the world, the International Society of Nephrology and International Federation of Kidney Foundations adopted the message for World Kidney Day in 2008 that “CKD is common, harmful and treatable.” One of the purposes of the KDIGO Conference was to identify absolute and relative risks of complications of CKD, including all-cause mortality, CVD mortality, kidney failure, acute kidney injury and progressive kidney disease.

The main concern about the current definition and classification was the possibility of overdiagnosis of CKD and overuse of resources in the investigation and management of CKD, without appropriate modification for variation in prognosis. The specific issues raised were the appropriateness of the GFR thresholds, albuminuria thresholds and absence of age-modifications, since lower GFR levels and higher albumin excretion rates are commonly observed in the apparently “healthy” elderly. Underlying these controversies was concern regarding the methods for assessment of eGFR and albuminuria, and the discomfort with the term “disease” for labeling a large number of people, mostly elderly, with lower levels of GFR and albuminuria.

In response to this debate, the KDIGO Board of Directors convened a Controversies Conference for the purpose of review and revision of the definition and classification in light of current knowledge of prognosis in CKD, with the goal for improving patient outcomes. KDIGO appointed a Planning Committee chaired by Andrew S. Levey, MD (US) and co-chaired by Meguid El Nahas, MD (UK), Paul de Jong, MD (NL), and Josef Coresh, MD, PhD (US). The KDIGO Controversies Conference was planned to address five questions; namely, i) What are the key outcomes of CKD?; ii) What progress has been made in CKD testing (eGFR and albuminuria)?; iii) What are the key factors determining prognosis of CKD (eGFR, albuminuria, etc)?; iv) Should the current CKD classification (based on eGFR) be modified to include additional factors associated with prognosis?; v) Should the current CKD definition be modified?

The planning committee invited representatives of studies to contribute data on outcomes of CKD in clinical or research populations in which eGFR and albuminuria had been determined at baseline. Outcomes considered included all-cause mortality, CVD mortality, kidney failure treated by dialysis or transplantation (end-stage renal disease), acute kidney injury, and decline in eGFR (progressive CKD). An analytical committee provided a uniform analysis plan for systematic evaluation of the data for each cohort and in addition, performed meta-analyses of results provided by the studies. Altogether, more than 50 cohorts submitted data and participated in the conference and metaanalyses on 1.5 million study participants on a range of outcomes were performed and reviewed at the Conference. A databook consisting of 1704 pages of cohort data and an additional 464 pages of results of meta-analyses was distributed to all conference participants. The conference consisted of plenary sessions during which KDIGO Co-Chairs Bertram Kasiske, MD (US) and Kai-Uwe Eckardt, MD (GER), members of the Planning Committee, Richard Glassock, MD (US), a noted critic of the current definition and classification, and other experts on CKD outlined the background and objectives of the conference. Following plenary sessions, the conference participants broke out into smaller groups for in-depth discussions of data and a proposal for revisions, and then reconvened in a plenary session for expression of viewpoints on a number of subjects, including a non-binding vote on questions prepared by the organizers.

The data reviewed showed a strong, consistent gradation in risk for all outcomes of CKD according to the level of estimated GFR and urine ACR across a wide range of study populations. Interestingly, the gradation was linear for all levels of albumin excretion and non-linear for GFR. In general, increased risk was noted below a level of GFR around 60 ml/min/1.73 m2 and at urinary ACR at all levels above 10 mg/g (the lowest value examined). The risk for cardiovascular mortality and kidney disease outcomes tended to be elevated at a higher eGFR than all-cause mortality. Results were adjusted for age, CVD risk factors diabetes, hypertension, smoking and hypercholesterolemia, and history of CVD.

A strong consensus reached by those present was that the current classification did not adequately describe the severity of CKD, and that predicting prognosis could be improved by the following modifications to the classification:

1. Emphasize classification by cause, if known, in addition to stage
2. Add albuminuria stages, in addition to GFR stages (ACR <30mg/g, 30-300 mg/g and >300 mg/g)
3. Subdivide CKD Stage 3 into two stages (GFR 30-44 and 45-59 ml/min/1.73 m2)

A less strong consensus emerged that it would be premature to change the current definition of CKD based on levels of GFR or presence of kidney damage. The following recommendations were also adopted by those present:

1. Make no change to the definition based on GFR (<60 ml/min/1.73 m2), irrespective of age or sex
2. Make no change to the level of albuminuria defined as a marker of kidney damage (urine ACR >30 mg/g)

These recommendations need to be codified by an appropriately constituted guidelines development group, which would involve a broader array of disciplines. The immense and unique data-base provided by the meta-analyses described at the Controversies Conference will supply an environment rich with aggregate data upon which to base new guidelines for the diagnostic, classification and prognosis of CKD.

As a consequence, for the time being, CKD prevalence estimates will remain unchanged, including a large fraction of the elderly population. However, a modified classification that includes cause of disease (if known), and albuminuria stages in addition to GFR stages will relate better to prognosis than the staging based solely on GFR. This may be particularly helpful in the great majority of elderly individuals with reduced GFR where albuminuria staging may better define their true prognosis. Improved information on prognosis can be helpful for a large number of management decisions, including decisions on whom to refer to nephrologists. As a consequence of this landmark meeting, designed to assess the controversies but not to develop new guidelines, it is anticipated that revision of the 2002 KDOQI clinical practice guidelines on definition and classification of CKD will be undertaken by KDIGO in the near future.

After the meeting, Dr. Glassock commented “The Controversies Conference was truly a “historical” event that will propel this entire field to a new level. The openness of the debate, the rigor of the questions and answers, and the immensity of the data and its analysis was truly remarkable. While much work remains to be done on refining the classification, diagnosis and prognosis of CKD, there is no doubt that the end-product will have as its origins the findings and discussions that were in evidence at the London meeting.”

In summarizing the outcome of the conference, Dr. Levey said, “The debate reflects a tension in our field caused by the paradigm shift about the basic perspective on CKD – from kidney failure as a life-threatening illness to earlier stages of kidney disease as the target for prevention, detection, evaluation and management. While change is always difficult, especially for those in its midst, the debate has been healthy, and the discussions and consensus should enable us to move on and work across disciplines to improve outcomes for our patients.”