Kidney Function Testing: Creatinine, eGFR and Cystatin C

The number that alarms people on a routine blood panel is usually the estimated glomerular filtration rate, printed with a flag and sometimes with a stage number attached. It looks like a direct measurement of how well the kidneys are filtering. It is not measured at all.

Almost every eGFR reported anywhere in the world is the output of an equation. The equation takes a blood creatinine concentration, adds the patient’s age and sex, and returns a number in units of millilitres per minute per standard body surface area. Nothing was filtered, timed or collected. The kidneys were never directly observed.

This does not make eGFR useless. The equations were built by comparing creatinine values against carefully measured filtration in large groups of people, and they work well across most of the population. But knowing that the number is an estimate derived from a muscle waste product explains nearly every situation where it misleads, and it explains why a second, quite different marker exists for exactly those situations.

What Creatinine Is and Where It Comes From

Creatinine is the breakdown product of creatine phosphate, an energy storage molecule found principally in muscle. Muscle converts a small, fairly constant fraction of its creatine pool into creatinine every day, and that creatinine enters the blood as waste with no useful function.

The kidneys clear it. Blood is filtered through the glomeruli, tiny tufts of capillaries whose walls act as a sieve, and creatinine passes into the filtrate freely because it is small and not bound to proteins. Most of it is then excreted in urine rather than being reabsorbed. If production is steady and clearance falls, blood concentration rises. That relationship is the entire basis of creatinine as a kidney marker.

Two complications sit inside this simple picture. First, the kidneys do not only filter creatinine; the tubules also secrete a modest amount directly into the urine. This secretion becomes proportionally more important as filtration declines, which means creatinine-based estimates tend to flatter kidney function somewhat in more advanced disease. Certain medications block that secretion, raising creatinine and lowering the calculated eGFR without any change in actual filtration, a distinction worth knowing when a new drug coincides with a new abnormal result.

Second, the relationship between creatinine concentration and filtration is not a straight line but a curve. At normal filtration, a substantial fall in kidney function produces only a small rise in creatinine, because the remaining nephrons compensate and because the starting concentration is low. Only after considerable function has been lost does creatinine begin to climb steeply. This is why creatinine is an insensitive early marker and why a value still inside the reference range does not exclude meaningful loss of filtration.

Diet contributes as well. Cooked meat contains creatinine directly, so a large meat meal raises blood creatinine transiently, and creatine supplementation can raise it modestly without any effect on the kidneys.

Why Muscle Mass Distorts Creatinine Readings

If creatinine comes from muscle, then two people with identical kidney function and different muscle mass will have different creatinine concentrations. This is not a subtle effect and it is the single largest source of error in kidney testing.

A heavily muscled person produces more creatinine daily. Their kidneys clear it perfectly well, but the steady-state concentration in blood settles higher, and an equation reading that value will report a lower eGFR than reality. Conversely, someone with very little muscle, whether from age, prolonged illness, limb loss, malnutrition or a neuromuscular condition, produces little creatinine. Their concentration sits low, and the equation reports an eGFR that overstates their filtration, sometimes substantially. In frail, older patients this direction of error is the more clinically dangerous, because it can lead to drug doses being calculated as though the kidneys were healthier than they are.

The equations attempt to compensate using age and sex, which correlate with muscle mass across a population. That correlation is real but loose, and it fails precisely for the individuals whose muscle mass departs from what their age and sex would suggest.

Several other situations disturb the marker. Acute muscle injury releases creatine and creatinine in quantity, raising values without kidney involvement. Pregnancy increases filtration substantially, lowering creatinine, so standard interpretation does not apply. Liver disease reduces creatine synthesis and is often accompanied by muscle wasting, so creatinine can be misleadingly low in exactly the patients where kidney function matters most.

How eGFR Equations Convert One Number Into Another

An eGFR equation is a statistical model. It was built by measuring true filtration rate in a large group of people using a marker cleared purely by filtration, then finding the mathematical relationship that best predicted those measured values from creatinine, age and sex.

Because it is a best fit to a population, it carries a stated accuracy rather than a precise answer. Performance is usually described in terms of how often an individual’s estimate falls within a given percentage of their measured value, and even for well-validated equations a meaningful minority of individuals fall outside that band. An eGFR is therefore a range wearing the costume of a single number.

The result is also indexed to a standard body surface area, which allows comparison between people of different sizes. That indexing is helpful for classifying chronic kidney disease and unhelpful for drug dosing, where what matters is the individual’s actual clearance rather than a size-normalised one. This is why dosing guidance sometimes uses a de-indexed value or an older creatinine clearance estimate that incorporates weight directly.

A further requirement is that the creatinine assay itself be standardised. Older methods measured creatinine with a chemistry that also responded to other substances in blood, giving results that ran systematically high. Because an equation calibrated to one assay produces wrong answers when fed values from another, laboratories now calibrate creatinine against an internationally recognised reference material. Modern equations assume that standardisation, which is one reason results from many years ago are not directly comparable with current ones.

The Move Away From Race-Based Coefficients

Widely used equations previously included a coefficient that raised the estimated GFR for patients recorded as Black. It was introduced empirically, because in the development datasets that adjustment improved the fit against measured filtration.

The reasoning behind it did not hold up. Race is a social category rather than a biological variable, it is recorded inconsistently, it has no defined value for people of mixed ancestry, and the differences in average creatinine that motivated the coefficient are better explained by average differences in muscle mass and diet within particular study populations than by anything intrinsic.

The practical consequences were significant. Because the coefficient raised the reported eGFR, it made kidney function look better than the alternative calculation for a group already experiencing higher rates of kidney failure. A higher reported eGFR delays referral to specialist care, delays consideration of transplant listing, and changes eligibility thresholds for some medications.

Revised equations that omit race have been developed and widely adopted, retaining creatinine, age and sex as inputs. They perform slightly less well than the older equations did in the specific datasets used to build them, which is the expected consequence of removing a variable that improved statistical fit, and this trade-off was accepted deliberately. The change means that a person’s reported eGFR may have shifted when their laboratory switched equations, without any change in their kidneys, which is worth knowing when comparing results across a transition period.

Cystatin C as an Independent Filtration Marker

Cystatin C is a small protein produced at a relatively steady rate by essentially all nucleated cells in the body, not by muscle specifically. It is filtered freely at the glomerulus and then almost entirely reabsorbed and broken down by the tubules rather than appearing in urine.

Its value lies in failing differently from creatinine. Where creatinine’s weakness is muscle mass, cystatin C is largely indifferent to it, making it far more reliable at either extreme of muscularity, in limb loss or neuromuscular disease, and in frail older adults.

It has its own confounders. Thyroid dysfunction shifts it in both directions, corticosteroids raise it, and inflammation, obesity and smoking all affect it. It is also more expensive and not on every routine panel, though assay standardisation has improved considerably.

The most useful application is combination. An equation using both markers performs better than either alone, and the agreement between them is itself informative. When creatinine-based and cystatin C-based estimates agree closely, confidence in the value is high. When they diverge substantially, that divergence is a signal that one of the markers is being distorted, and it prompts a look at muscle mass, thyroid status, medications and recent illness. Guidelines increasingly recommend cystatin C as a confirmatory test when a creatinine-based eGFR is borderline and the result would change management.

Albumin-to-Creatinine Ratio and Early Damage

Filtration rate and kidney damage are separate things, and a test measuring one can be entirely normal while the other is well advanced. The test for damage is albuminuria.

Healthy glomeruli allow almost no albumin through, because the filtration barrier discriminates by both size and electrical charge and albumin is excluded by each. When that barrier is damaged, albumin leaks into the urine, and this leak frequently appears years before filtration measurably declines. In diabetes and hypertension, the two leading causes of chronic kidney disease worldwide, albuminuria is often the first detectable abnormality.

Measuring it uses a ratio rather than a concentration, because urine concentration varies enormously with hydration. Dividing the albumin concentration by the creatinine concentration in the same sample corrects for this, since creatinine excretion is comparatively steady through the day. A single early-morning sample is generally sufficient, avoiding the collection errors that plague timed collections.

Several factors raise albumin excretion temporarily: vigorous exercise in the preceding day, fever, urinary infection, heart failure and marked hyperglycaemia. Because of this, an isolated raised ratio is confirmed by repeating it, ideally with two of three samples abnormal over a period of months before it is regarded as persistent.

The classification of chronic kidney disease uses both axes deliberately, grading filtration and albuminuria separately, because the combination predicts outcomes better than either alone. Someone with a moderately reduced eGFR and no albuminuria has a considerably different outlook from someone with the same eGFR and heavy protein loss.

Why a Single Low eGFR Is Not a Diagnosis

Chronic kidney disease is defined by abnormality persisting for at least three months. A single reduced eGFR does not meet that definition, and treating it as though it does causes a great deal of unnecessary distress.

The reasons a one-off value can be low are numerous and mostly reversible. Dehydration reduces blood flow to the kidneys and raises creatinine, and rehydration reverses it. Acute illness of almost any kind can do the same. A large meat meal or creatine supplementation before the test raises creatinine directly. Intense exercise in the preceding day raises it through muscle turnover. Several common medications, including some blood pressure agents, anti-inflammatories and certain antibiotics, alter creatinine or filtration, and some do so by blocking tubular secretion without changing filtration at all.

Analytical variation adds a little more, and biological variation within an individual adds more still, so two values from the same stable person weeks apart will differ.

The appropriate response to a first low value is therefore to repeat it after an interval, with attention to hydration, recent illness, exercise, diet and medication changes, and to check urine albumin at the same time. If the value has normalised, no chronic disease is present, though the episode itself may be worth understanding. If it persists, the classification and the search for a cause proceed properly, with the trajectory over successive measurements mattering far more than any single figure.

The most important number in kidney testing is not the current eGFR but the slope. A stable value slightly below the reference range in an older adult, with no albuminuria, often reflects the gradual decline that accompanies ageing and may never cause a problem. A value falling steadily across several measurements, particularly alongside albuminuria, is the pattern that warrants attention regardless of where it currently sits.

That is why kidney results are best read as a series rather than a snapshot. The equation converting creatinine into eGFR carries real uncertainty for any individual, but that uncertainty is largely consistent from test to test in the same person on the same assay. It therefore cancels out when you look at change over time, which is exactly where the clinically useful signal lives.

The practical way to hold all of this is to treat eGFR as a well-calibrated guess rather than a reading. It is derived from a muscle waste product, adjusted by variables that stand in loosely for body composition, and reported to a precision the underlying method does not support.

Read alongside a urine albumin-to-creatinine ratio, repeated over months rather than acted on once, and interpreted with attention to muscle mass and medications, it does its job well. Read as a single definitive measurement of an organ, it produces both false alarms and false reassurance, and it produces the second more often than most people realise.