Cystatin C Versus Creatinine: Why Two eGFR Estimates Can Disagree

The short answer#

Creatinine and cystatin C are two different ways to estimate the same thing, the glomerular filtration rate (GFR), and they disagree because neither molecule is a clean readout of filtration alone. Creatinine is nudged by muscle, diet, and certain drugs; cystatin C is nudged by body fat, inflammation, thyroid status, and steroids. The 2024 KDIGO guideline keeps creatinine first-line but adds cystatin C when a decision hinges on an accurate GFR, and recent evidence suggests that when the cystatin C number sits well below the creatinine number, the difference itself points to higher risk.

Key points#

• Both markers estimate GFR, but each is distorted by non-kidney factors, so a laboratory can return two eGFR values that differ by 15 mL/min/1.73 m2 or more from a single draw.
• Creatinine reflects muscle mass and meat intake, so it can flatter frail patients and penalize muscular ones.
• Cystatin C sidesteps muscle but is raised by adiposity, corticosteroids, thyroid disease, and inflammation.
• KDIGO 2024 keeps creatinine as the default and recommends the combined equation (eGFRcr-cys) when precision changes management.
• A lower cystatin C estimate relative to creatinine is associated with higher mortality, but it more likely marks a sicker patient than causes the harm.

One filter, two speedometers#

Think of the kidney as an engine and the two markers as two speedometers wired to it. Both are trying to report the same speed, yet each has its own calibration quirk, so the needles rarely land on the exact same number. The size of the disagreement is not random. In the studies that examine it, a difference of 15 mL/min/1.73 m2 between the two estimates is the threshold used to call the pair meaningfully discordant. Understanding why the needles differ means looking at what each marker is really made of.

What creatinine measures, and what it misses#

Creatinine is a waste product of muscle metabolism. The kidney filters it freely at the glomerulus and also secretes a little of it through the tubules, so the blood level depends on far more than filtration. The most important confounder is muscle: a person with low muscle mass simply makes less creatinine, so a creatinine-based estimate (eGFRcr) can look reassuring while true filtration is poor.

The distortion runs in both directions. A muscular athlete or someone eating a high-meat diet can appear to have worse kidney function than they do, because they generate more of the marker. In the opposite corner, a person with advanced cirrhosis, sarcopenia, or a spinal cord injury can look deceptively healthy on creatinine because there is little muscle to supply it. Everyday medicines add more noise: trimethoprim and cimetidine block the tubular secretion of creatinine and raise the blood level without any real change in filtration.

What cystatin C adds, and its own blind spots#

Cystatin C is a small protein made at a fairly constant rate by nearly every nucleated cell in the body. It is freely filtered and, unlike creatinine, is not meaningfully secreted by the tubules. Because production is largely uncoupled from muscle, cystatin C usually gives you a truer picture in people at the extremes of muscle mass, where creatinine is least reliable.

It is not a perfect marker either. As the NIDDK notes, body fat, corticosteroid treatment, thyroid dysfunction, and inflammation can each raise cystatin C without any change in filtration. The useful point is that the two markers fail in different situations: each one covers a blind spot the other has, which is exactly why using them together tends to beat either one alone.

What KDIGO 2024 actually recommends#

The 2024 KDIGO guideline still opens with creatinine, for practical reasons: the assay is cheap, fast, and available everywhere. Its notable move is a graded recommendation (strength 1C) to use the combined creatinine and cystatin C equation, eGFRcr-cys, when creatinine alone is likely to be inaccurate and when the GFR value will actually drive a decision.

The support for combining the markers comes largely from race-free CKD-EPI equations published in the New England Journal of Medicine in 2021, which found the combined equation more accurate than either marker alone and narrowed the differences seen between Black and non-Black participants. KDIGO also asks laboratories to use enzymatic creatinine assays for consistency, and it flags three groups who benefit most from a confirmatory cystatin C: frail or multimorbid patients, anyone who needs a precise GFR for a specific decision such as chemotherapy dosing or transplant assessment, and those at high risk of kidney complications.

When the two disagree, the gap carries information#

You may be tempted to treat a discordant pair as clutter and simply average the two numbers. A 2025 systematic review and meta-analysis in Clinical Kidney Journal makes the opposite case. Pooling 18 studies that ranged from 373 to more than 363,000 participants, the authors set the reference direction as the cystatin C estimate running lower than the creatinine estimate. They defined a large discordance as the cystatin C value falling to 60 percent or less of the creatinine value, or a gap of at least 15 mL/min/1.73 m2 between them.

When the cystatin C estimate sat well below the creatinine estimate, all-cause mortality was substantially higher (hazard ratio 1.58, 95% CI 1.42 to 1.76), and cardiovascular events rose too (hazard ratio 1.32, 95% CI 1.25 to 1.39). The reverse pattern was reassuring: a higher cystatin C estimate relative to creatinine tracked with lower risk. One proposed mechanism is selective glomerular hypofiltration, in which the filtration of mid-sized molecules is impaired out of proportion to how creatinine is handled.

Reading the evidence with care#

Direction is the first thing to check. The prognostic weight sits on the side where cystatin C is the lower number, so before you draw any conclusion, ask which marker is lower.

The second point is that this is an association, not a clean causal arrow. The discordant pattern travels with muscle wasting, chronic inflammation, and multiple coexisting illnesses, and each of those raises mortality on its own. So the gap may be as much a barometer of overall illness burden as a specific kidney lesion; it flags a sicker patient rather than showing that cystatin C itself is harmful. Meaningful variation between the pooled studies and residual confounding reinforce that caution, though the consistent direction across 18 datasets is what makes the signal credible.

The practical payoff is concrete. A single confirmatory cystatin C can reclassify a CKD stage, change the dose of a narrow-margin drug, and correct a risk estimate that creatinine alone would have given you.