Understanding Serum Creatinine

Serum creatinine is the most-ordered marker of kidney function and one of the most misread numbers in medicine. Understanding what it is — and what moves it — explains most of the errors in creatinine clearance calculation.

Where creatinine comes from

Creatine is stored in skeletal muscle as a phosphate energy reserve. Roughly 1–2% of it converts spontaneously and non-enzymatically to creatinine each day, at a rate proportional to total muscle mass.

That gives creatinine two properties that make it useful:

• Production is near-constant for a given individual, day to day.
• It is freely filtered at the glomerulus and not reabsorbed.

And one that limits it: production depends on muscle mass, which varies enormously between people.

Creatinine is cleared almost entirely by the kidneys — mostly by filtration, but with a meaningful contribution from active tubular secretion. That secretion is why creatinine clearance overestimates true GFR by 10–20%, and why the gap widens as filtration falls.

Why the reference range misleads

A typical adult reference range is roughly 0.7–1.3 mg/dL (60–115 µmol/L) for men and 0.6–1.1 mg/dL (50–95 µmol/L) for women.

It is a population range, not a personal one — and creatinine is a marker where the personal baseline matters far more.

Consider two patients, both with a creatinine of 1.1 mg/dL, both reported as normal:

• A 30-year-old, 90 kg man whose baseline is 1.1. Nothing has changed.
• An 82-year-old, 50 kg woman whose baseline is 0.6. Her creatinine has nearly doubled.

The second patient has lost a large fraction of her kidney function. Every laboratory report says normal.

This is why calculating matters more than reading. It is also why a trend across previous results is more informative than any single value. Our article on creatinine clearance in elderly patients works through the geriatric case in detail.

The curvilinear relationship

Creatinine sits in the denominator of every clearance equation, so the relationship to kidney function is a curve, not a line.

| Serum creatinine | CrCl (70 kg, 60-year-old man) | Change from previous row | | --- | --- | --- | | 0.8 mg/dL | 97 mL/min | — | | 1.0 mg/dL | 78 mL/min | −19 | | 1.5 mg/dL | 52 mL/min | −26 | | 2.0 mg/dL | 39 mL/min | −13 | | 3.0 mg/dL | 26 mL/min | −13 | | 4.0 mg/dL | 19 mL/min | −7 |

The steepest losses happen inside and just above the reference range. By the time creatinine reaches 2 mg/dL, further rises represent progressively smaller absolute losses — most of the function is already gone.

The clinical consequence: a small rise within the reference range is a bigger deal than a large rise outside it. A patient moving from 0.8 to 1.1 over two years, entirely "normal" throughout, has lost about a quarter of their estimated clearance.

What moves creatinine without touching the kidneys

This is where misinterpretation happens.

Raises creatinine, filtration unchanged

• Trimethoprim — blocks tubular secretion. Rises of 0.2–0.4 mg/dL are typical and reversible.
• Cimetidine — same mechanism. Occasionally used deliberately to make creatinine clearance approximate true GFR.
• Cobicistat, dolutegravir, ritonavir — inhibit tubular transporters. Expected on initiation, not a sign of nephrotoxicity.
• Creatine supplementation — more substrate, more creatinine.
• A large cooked-meat meal — cooking converts creatine to creatinine; absorption can raise serum levels transiently. A reason for a fasting sample where precision matters.
• Intense exercise — transient rise from muscle turnover.
• Dehydration — a genuine prerenal reduction in filtration, but reversible and not chronic disease.

Lowers creatinine, filtration unchanged

• Low muscle mass — the elderly, cachexia, malnutrition, paraplegia, amputation.
• Advanced liver disease — reduced hepatic creatine synthesis. Cirrhotic patients commonly have creatinine well below their true kidney function, a known reason for underestimating impairment in that group.
• Vegetarian and vegan diets — no dietary creatine intake.
• Pregnancy — increased plasma volume and a genuine rise in GFR.

Assay considerations

Older Jaffe methods are susceptible to interference from bilirubin, ketones and some cephalosporins. Modern enzymatic, IDMS-traceable assays are more specific and read lower. Equation coefficients — including the 0.413 in bedside Schwartz — assume IDMS-traceable values. Using an equation with values from a differently calibrated assay introduces systematic error.

Creatinine in acute kidney injury

The most important limitation. Serum creatinine is a lagging indicator.

If filtration stops completely, creatinine rises at a rate set by production and volume of distribution — typically 1–2 mg/dL per day. So a patient whose kidneys failed this morning has a near-normal creatinine this afternoon.

Every creatinine-based estimating equation assumes steady state. In evolving AKI that assumption is false, and the equation returns a number that is not merely imprecise but systematically too high, exactly when overdosing is most harmful.

In suspected AKI, look at the trend and at urine output, not at a single value, and do not dose confidently from a calculated clearance.

When to use something else

Cystatin C is produced by all nucleated cells, so it does not depend on muscle mass. Consider it for amputees, cachexia, paraplegia, very high muscle mass, or where creatinine-based and clinical impressions disagree — see CrCl in amputees, children and race for how each of those breaks the weight assumption. The CKD-EPI 2021 combined creatinine–cystatin C equation is the most accurate of the standard set — available on the eGFR calculator.

Cystatin C has its own confounders: thyroid disease, corticosteroids, and smoking all affect it.

A timed urine collection measures the patient directly, and is worth the effort where estimating equations are weakest. An incomplete collection is worse than an estimate, so run the adequacy check — how the test works and how to tell it went wrong.

Summary

• Creatinine reflects muscle mass as much as kidney function.
• The reference range is a population range. The personal baseline matters more.
• The relationship to clearance is curvilinear — small rises inside the range matter most.
• Several common drugs raise creatinine without changing filtration.
• Liver disease, low muscle mass and vegetarian diets lower it independently of the kidneys.
• In AKI, creatinine lags. Trend and urine output beat a single value.
• Do not read the number. Calculate with it.