CRP vs ESR for Inflammation: Which Test You Need, and When Doctors Order Both

Side by side

In 1991, researchers followed 140 people through the week after surgery for a broken hip, drawing blood each day. CRP had a clear shape: a sharp rise on day 2, then a fall by day 7. ESR had no shape at all. It was raised to varying degrees all week, and complications made no difference to it.5

That one study holds the main difference. CRP, or C-reactive protein, is a protein that rises in your blood when inflammation starts and falls as it settles. ESR, the erythrocyte sedimentation rate or "sed rate," measures how far red blood cells sink in a tube of blood in one hour. Both get called inflammation markers. They read different things.

Where they differ

Three things separate them: speed, ingredients, and what nudges them when you are well.

A fresh problem shows up in CRP first. A slow one shows up in the proteins ESR reads, and sometimes only there.From a 1991 study in Int Orthop

Speed first. CRP climbs and falls within days.5 ESR drifts. A 2010 study of 5,777 hospital patients found that its slow variation, plus frequent interference from other factors, made ESR "frequently misleading" in patients not already suspected of a specific disease.4

Then ingredients. A 2022 analysis of hospital lab records worked out what moves the ESR. Antibody proteins, the clotting protein fibrinogen and the platelet count pushed it up. Albumin, the main protein in blood, pulled it down.6 CRP tracked the ESR closely but had no direct effect on it.6

The third difference shows up in people with no disease at all. A 2021 study tested 1,472 adults in one Spanish town, none with a known inflammatory condition. Most, 1,123 of them, were normal on both. But 72 had a high ESR with a normal CRP, a pattern tied to older age, and 208 had the opposite pattern, tied to higher BMI.3 Age leans on one test, weight on the other.

Which is more sensitive, ESR or CRP?

CRP, by a modest margin, in the 2020 meta-analysis of 29 studies of acute inflammation. For bone and joint infections the two were nearly even. For other inflammatory conditions CRP picked up 86% of cases against 77% for ESR, and it cleared more of the people who had nothing wrong, though the studies varied widely.1

A 2019 paper put it more bluntly. ESR has higher false positive and false negative rates than CRP. And the Choosing Wisely campaign of the American Board of Internal Medicine has recommended against ESR, in favor of CRP, when no diagnosis has been made.10

Sensitivity is not the whole story, though. In a 2025 study at a fast-track clinic for giant cell arteritis, CRP caught 74% of cases but cleared only 49% of people without it, while ESR caught just 48% and cleared 72%.9 One test was better at finding the disease, the other at ruling it out.

What if ESR is high but CRP is normal?

It happens often, and it points somewhere. In 2,150 paired tests from one Texas hospital, results that landed far apart turned up in 212 patients, or 12%, split almost evenly between the two directions. About 1 in 8.2

The two groups told different stories. High CRP with low ESR went with infections of the urinary tract, gut, lungs and bloodstream, and with heart attacks and blood clots. High ESR with low CRP went with bone and joint infections, connective tissue diseases such as lupus, and strokes.2

The 2010 hospital study found a mismatch in 33% of its 5,777 patients, almost all of it high ESR with normal CRP.4 In a chart review of 74 of them, 32% had an inflammatory disorder that was fading, 28% had something interfering with the ESR, 32% had no explanation, and 8% had an active disease the CRP had missed.4 So this pattern most often means inflammation on its way out, or a nudged test. Now and then it means something real that CRP missed.

Why do doctors order both?

Habit explains part of it. In Korea in 2022, 46.64% of claims for either test included both, and the share passed 60% in orthopedic, brain and plastic surgery departments.11 In a US community health system, a pop-up rule in the ordering software cut the pairing by 42%.10

But there is a case for both in the right setting. The 2020 meta-analysis found that ESR and CRP together were more accurate than either alone.1 And a 2016 review stressed that neither test is sensitive or specific enough to stand on its own. Both work best when the doctor already suspects a particular disease, and mismatches are most common in long-term inflammatory diseases.7

The tiebreaker

The deciding question is what disease is on the table, and how long it has been going on.

Giant cell arteritis (GCA), an inflammation of the arteries in older adults that can threaten eyesight, is where ESR still earns its reputation. A 2020 meta-analysis of 68 studies measured how much each finding shifted the odds. An ESR over 100 mm/h made GCA about 3 times more likely. The absence of an ESR over 40 mm/h was the strongest reassurance in the set: it cut the odds to about a fifth. The absence of a CRP of at least 2.5 mg/dL (25 mg/L on most US lab reports) cut them less, to about a third.8

Even here, neither test is a verdict. In the 2025 clinic study, 23 people with confirmed GCA had normal results on both tests, against 107 with at least one high. Those with normal markers had waited longer between first symptoms and the blood draw: 35 days on average, against 20. The authors concluded that normal markers should not stop a GCA work-up.9

So the hip surgery patients and the arteritis patients answer the same question from opposite ends. A fresh problem, days old, shows up in CRP first and clearest. A slow disease that has smoldered for weeks shows up in the proteins ESR reads, and sometimes only there. Pick the test that matches the age of the problem. When nobody knows, start with CRP.