Chronic kidney disease is defined less by a single number than by two of them read together — the filtration rate and the albuminuria — and by time. A GFR is a snapshot; CKD is a snapshot that has lasted, and that carries risk. The discipline of this opening chapter is to define the condition precisely, stage it on both axes, and — most importantly — translate the stage into the risk that should drive everything the rest of the volume does.
What counts as CKD
KDIGO defines CKD as an abnormality of kidney structure or function, present for more than three months, with implications for health. Two parts of that sentence do real work. The chronicity — more than three months — is what distinguishes CKD from the acute kidney injury and acute kidney disease of the previous volume; a single low GFR is not CKD until it has persisted. And the 'or function' means CKD is diagnosed in two ways: a GFR below 60, which is itself abnormal, or a marker of kidney damage even when the GFR is preserved. The commonest damage marker is albuminuria, but the list also includes an abnormal sediment, electrolyte or tubular disorders, structural abnormalities on imaging, abnormal histology, and a history of transplantation. A consequence trips up the unwary: at a preserved GFR — category G1 or G2 — there must be a damage marker to call it CKD. A healthy person with a GFR of 75 and no albuminuria does not have CKD; the same GFR with an ACR of 50 does.
Staging by GFR and albuminuria
Staging runs on two axes. The GFR categories descend from G1 (90 or above) and G2 (60 to 89) through G3a (45 to 59) and G3b (30 to 44) to G4 (15 to 29) and G5 (below 15). The split of the old stage 3 into G3a and G3b matters because risk rises steeply across it. The albuminuria categories are A1 (below 30 mg/g), A2 (30 to 300), and A3 (above 300), measured by the albumin-to-creatinine ratio on a spot urine, ideally a first-morning sample. The modern framework asks for all three descriptors together — Cause, GFR, and Albuminuria, the CGA classification — because the cause shapes the disease, and GFR and albuminuria each carry independent prognostic weight. Staging by GFR alone, the old habit, throws away half the information.
The heatmap: from stage to risk
The reason to stage on two axes is the KDIGO heatmap, which arrays the GFR categories down one side and the albuminuria categories across the other and colours each cell from green through yellow and orange to red. The colour is risk — of progression, of end-stage kidney disease, of AKI, of cardiovascular events, and of death. The map's lesson is that GFR and albuminuria combine, so two patients with an identical GFR of 50 can sit in very different places: one at A1 in a yellow, low-risk cell, the other at A3 in a red, high-risk one. Reading the heatmap turns a static label into a prognosis and, with it, a management intensity — green cells are monitored, red cells are referred and treated aggressively. This risk-based reframing is the organising idea of contemporary CKD care.
Estimating GFR
GFR is almost always estimated rather than measured, and the modern standard is the 2021 CKD-EPI creatinine equation, which removed the race coefficient that earlier equations carried — a change made on grounds of both science and equity, since race is a social rather than biological variable and its inclusion systematically misclassified risk. Creatinine-based estimates inherit all the limitations met in the AKI volume: they depend on muscle mass, diet, and drugs that block tubular secretion, and they assume a steady state. When the estimate will drive a major decision — a transplant work-up, a drug dose with a narrow margin, a referral threshold — cystatin C, which does not depend on muscle, is used to confirm it, and a combined creatinine-cystatin C equation is the most accurate of the readily available estimates. Measured GFR is reserved for the rare case where precision is essential.
Albuminuria: marker and target
Albuminuria deserves its own emphasis because it plays two roles. As a damage marker, it defines CKD at a preserved GFR and places the patient on the albuminuria axis of the heatmap. As a risk factor, it predicts progression, cardiovascular disease, and death independently of the GFR — which is why it is not a footnote to the filtration rate but a co-equal axis. And, uniquely among the staging variables, it is modifiable: the blood-pressure, RAAS, and SGLT2-inhibitor therapies of Part 2 lower albuminuria, and that reduction tracks with slower progression. So the ACR is measured not once to label the patient but repeatedly to follow them, a marker that is also a target.
Confirming chronicity
Because the definition turns on permanence, confirming chronicity is part of making the diagnosis. A single abnormal GFR could be AKI, AKD, or CKD; what distinguishes CKD is that the abnormality has lasted more than three months, established by repeated measurements over that interval or by prior records showing the abnormality was already present. Supportive findings help when the history is thin: small, echogenic kidneys on ultrasound, or the anaemia and mineral-bone changes of established disease, point to chronicity, whereas normal-sized kidneys and an abrupt change point to an acute process. Getting this right matters because it changes everything downstream — an AKI that will recover is managed quite differently from a CKD that will progress.
Predicting the future: risk equations
The heatmap gives a categorical risk; the kidney-failure risk equation gives a number. Validated across large international cohorts, its four-variable form predicts the two- and five-year risk of end-stage kidney disease from age, sex, eGFR, and albuminuria — the same variables the heatmap uses, combined quantitatively. Its value is practical: it informs the timing of nephrology referral, of access planning, and of modality education, replacing the crude single-threshold referral of the past with a risk-calibrated one. A young person with a moderately reduced GFR but no albuminuria may have a very low five-year risk and need only monitoring, while an older person with the same GFR and heavy albuminuria may have a high risk and need preparation now. The equation lets the heatmap's logic drive concrete decisions about who needs what, and when.
Where the evidence is firm, and where it argues
The firm parts are the prognostic ones. The heatmap's risk gradient — that lower GFR and higher albuminuria each predict worse outcomes, and combine — is among the most reproduced findings in nephrology, drawn from consortium meta-analyses of millions of people. The kidney-failure risk equation's accuracy is well validated. The race-free equation is now the standard, though debate continues about the best single marker and about whether to default to cystatin C more widely. What the chapter does not claim is that any estimate is precise at the individual level: eGFR is an estimate with real error bars, and the wise clinician treats the trajectory and the risk category, not the third decimal place. The proper use of all this machinery is to stratify risk honestly and let that risk, rather than a number in isolation, guide care.
