CRP vs ESR

What the research found.

• For spotting giant cell arteritis, CRP caught 74 percent of the cases and cleared 49 percent of the people who didn't have it. Its area under the curve was 0.675. The sedimentation rate caught 48 percent, cleared 72 percent, and came out at 0.644.

  Currier and colleagues, Rheumatology International, 2025

• In that series, 23 of 130 giant cell arteritis patients had a normal pair, meaning their sedimentation rate was normal and so was their CRP. The ones measured later after symptoms began were more likely to read normal. The two timings were a mean of 35 days against 20, at a P value of 0.03.

  Currier and colleagues, Rheumatology International, 2025

• Two databases recorded consecutive rheumatoid arthritis patients, where a normal CRP means under 10 milligrams per liter. That turned up in 44 percent of one group and 58 percent of the other, and both markers read normal in 33 and 42 percent. All three read abnormal in only 28 and 23 percent, those three being ESR, CRP, and rheumatoid factor.

  Sokka and colleagues, Journal of Rheumatology, 2009

• The usual age rule for the sedimentation rate takes two forms. For men you take your age and halve it, and for women you add ten to your age first and then halve it. Applying that rule removed an apparent age difference in a group of 263 patients with early rheumatoid arthritis.

  Ranganath and colleagues, Journal of Rheumatology, 2005

• One review pooled trials of following a protocol toward a combined target, and that target included an inflammation marker. More people reached remission than under ordinary care. The biggest trial reported 65 percent against 16 percent.

  Hock and colleagues, SN Comprehensive Clinical Medicine, 2021

The short answer.

Neither test wins, and where accuracy figures exist they split cleanly. For giant cell arteritis, CRP caught more cases, at 74 percent against 48 percent, so it misses fewer. The sedimentation rate cleared more healthy people, at 72 percent against 49 percent, so it raises fewer false alarms. The areas under the curve were 0.675 and 0.644, and both describe tests that contribute rather than decide.

The more useful comparison is speed. CRP gets measured directly and it moves within a day or two of a change in inflammation, so it's the better answer to whether something is happening now. The sedimentation rate moves slowly in both directions and can stay raised after the inflammation behind it has settled, which makes it the worse choice for tracking a fast change and a reasonable picture over a longer stretch.

What both do badly is rule disease out. About half the people turning up with rheumatoid arthritis had a normal CRP, across two ordinary-care databases, and 23 of 130 patients with giant cell arteritis had both markers normal. So sending both together helps, and it doesn't turn two imperfect tests into a reliable all-clear.

Side by side.

Two tests, one question, different routes

CRP and ESR are the two blood tests used most often to measure inflammation, and a rheumatologist often orders both. They answer a similar question by different routes, and understanding the difference between them explains why your two results sometimes disagree with each other. Neither is better than the other, and they get ordered together because each covers something the other misses.

CRP stands for C-reactive protein, and your liver makes it when it gets inflammation signals, so the test measures how much of that protein is there. ESR stands for erythrocyte sedimentation rate, and it measures how fast red blood cells settle in a tube over an hour. So one test looks at a substance and the other looks at a physical result.

Inflammation changes the proteins in blood plasma, and that change makes red cells settle faster, which is how the sedimentation rate measures inflammation at one remove. That difference explains everything else on this page. Hold on to it, because the rest follows from it.

Where the accuracy figures come from

Only one condition in our sources has a study comparing how well these two spot a disease, and that condition is giant cell arteritis. It ran in a fast-track clinic series, where CRP caught 74 percent of the cases and cleared 49 percent of the people who didn't have it, at an area under the curve of 0.675. The sedimentation rate caught 48 percent, cleared 72 percent, and came out at 0.644.

The split is clean, and it's the most useful thing to know about the pair. CRP misses fewer cases and raises more false alarms, while the sedimentation rate raises fewer false alarms and misses more cases. A test that catches 48 percent misses more than half of what it's asked about.

Those figures come from a group already suspected of the disease, which lifts the catch rate compared with an unselected group and lowers the clearing rate too. So read them as a rough guide to the size of the difference rather than as fixed properties of the tests. This page rests on one disease's numbers, and that's a limit worth stating rather than hiding.

Why the difference in speed is the practical one

CRP answers quickly, rising when inflammation starts and falling once the cause settles, and it moves within a day or two in both directions. So it's a reasonable measure of what's happening now, and a rheumatologist who wants to know whether you're inflamed today gets a better answer from it. That's the question this test was built to answer.

The sedimentation rate moves slowly in both directions. That means it can stay raised after the inflammation behind it has settled, and that lag makes it less useful for tracking a fast change and more useful as a picture over a longer stretch. In some conditions it has a well-established role of its own, and polymyalgia rheumatica and giant cell arteritis are the two that come up most.

So here is what that means for your own results. A trend tells you more than one value, and that's more true of the sedimentation rate than of CRP. Two sedimentation rates a week apart may say very little, while two CRP values a week apart during a treatment change can say a good deal.

Why the two results disagree

A raised sedimentation rate with a normal CRP is common, and it doesn't mean the lab made a mistake. The sedimentation rate gets moved by things that have nothing to do with inflammation, including low blood counts, kidney disease, pregnancy, extra body weight, and rising age. Any of those can lift the number with no inflammatory disease present.

Age deserves its own note, because the correction is standard and most lab reports don't apply it. For men you take your age and halve it, and for women you add ten to your age first and then halve it, which gives the usual upper limit. Applying that rule removed an apparent age difference between older and younger onset, in a group of 263 patients with early rheumatoid arthritis.

The reverse disagreement happens too, because CRP can be raised while the sedimentation rate stays normal. That happens most early in an inflammatory process, since CRP moves first. So when your two results disagree, the rheumatologist reads them against your symptoms and your exam, because neither number is the answer on its own.

What neither test tells you

Neither test says anything about what's causing the inflammation. Both rise in infection, both rise in an inflammatory arthritis, both rise in some cancers, and both rise after surgery or injury. So a raised result says something inflammatory is happening without saying what, which is why these tests support a diagnosis rather than making one.

A normal result also doesn't rule out active disease, and this is where the numbers are most striking. Two databases recorded consecutive rheumatoid arthritis patients, one in Finland and one in the United States, where a normal CRP means under 10 milligrams per liter. That turned up in 44 percent of one group and 58 percent of the other, and both markers read normal in 33 and 42 percent. All three read abnormal in only 28 and 23 percent, those three being the sedimentation rate, CRP, and rheumatoid factor.

Giant cell arteritis raises the stakes on that point. In one series, 23 of 130 patients with that disease had a normal pair, meaning their sedimentation rate was normal and so was their CRP, and that disease can take your sight suddenly and for good. So say your markers are normal and your joints are swollen and stiff, or say your head hurts in a new way, and the clinical picture comes first rather than the numbers.

The timing finding that runs the wrong way

One result from the giant cell arteritis study deserves separating out, because most people would guess it backwards. Patients measured later after symptoms began were more likely to read normal, at a mean of 35 days against 20, with a P value of 0.03. The intuition says a longer illness should push markers higher.

The study found the opposite link, and this page reports it as found. It offers no explanation, and the practical result stands without one. Take a normal marker several weeks into an illness, which counts for less reassurance than the identical result at the start.

That's the moment somebody is most likely to relax about a reassuring number. A normal result after weeks of feeling unwell is easy to read as an all-clear. Look at the one disease where this was studied, though, where the late-measured group was the most likely to read normal and every one of them had the disease.

How they get used over time

Say you're on treatment for an inflammatory arthritis. Then these tests get followed alongside joint counts and how you say you're doing, where a falling CRP after starting a new drug supports the impression that it's working and a rising one prompts a closer look. The trend across several results tells you more than any single value.

What the evidence supports is following a combined score rather than a marker on its own. One review pooled trials of following a protocol toward such a target, where that target included an inflammation marker, and more people reached remission than under ordinary care, with the biggest trial reporting 65 percent against 16 percent. So the marker there is one input into a score rather than the thing being watched.

This site holds no source on how either test performs for monitoring on its own, which is a different question from diagnosis, and a test can be reasonable at one and poor at the other. What the diagnostic data does establish is worth taking into any conversation about a reassuring result. A normal marker doesn't mean the disease is absent.

When each one fits.

• CRP answers whether something is inflamed now, because it moves within a day or two and it gets measured directly.
• The sedimentation rate gives a slower picture, so one value says less than a trend, and it can stay raised after the disease has settled.
• Both get sent together in practice, because they measure different things and each covers part of what the other misses.
• Read the sedimentation rate against your age and your sex, since the usual upper limit is a rule rather than the range most labs print.
• Treat a normal pair as weak evidence, which holds most in giant cell arteritis, where 23 of 130 patients with the disease had one.
• Expect one of them to be followed rather than both, if your diagnosis is settled and a combined score is steering your treatment.

Questions patients ask.