News blog

Biomarker predicts recovery from a type of depression

Jiri Hodan

People who live with clinical depression must also suffer the ‘trial and error’ approach that psychiatrists take when prescribing antidepressants. Now, a study published this week signifies the beginning of the end of guesswork. In it, a blood test predicts who will respond well to a novel treatment for depression, and who might even fare worse.

“We haven’t had a test like this in psychiatry before,” says Andy Miller, a professor of psychiatry at Emory University and an author on the study in Archives of General Psychiatry. “There is no brain scan, no physiological measure that tells you whether a patient will respond to one drug more than another.”

The test identifies an inflammatory protein in blood, C-reactive protein or CRP, that indicates internal inflammation. Whereas 62% of depressed participants with high CRP levels improved after receiving the new treatment, only 33% of participants with high CRP levels in the placebo arm did.

The correlation was not entirely unexpected, because the drug suppresses inflammation, and Miller thinks that inflammation underlies depression in some people. To test whether a potent anti-inflammatory could soothe the malady, his team recruited 60 people who had lived with major depression for more than a decade and had received no relief from antidepressants.

Half of the participants received monthly treatments of the rheumatoid arthritis drug, Janssen’s Infliximab, and half received a placebo. Overall, Infliximab did not seem to work. However, when Miller’s team analysed how the subset of participants with high CRP faired, it turns out they responded well to the drug, with a relief from sadness, suicidal thoughts, anxiety and other symptoms.

Since the late 1980s, researchers have sporadically hypothesized that inflammation can lead to depression. The theory is that depressed behaviour might be beneficial in the short term because it reserves an injured animal’s energy for healing rather than romping around in the sunlight. Although the hypothesis has never received widespread support, researchers have found that some depressed patients indeed bear elevated levels of inflammatory proteins.

On the basis of the results from this relatively small study, a biologic drug such as Infliximab might be a better option in the anti-inflammatory realm than Cox-2 inhibitors such as aspirin, which come with unwanted side effects, says Miller. Although he knows of no Infliximab-like drug in development for depression, he says that companies might be encouraged by his team’s results. What’s more, with a biomarker to predict a response, companies will have a better chance of success.

Robert Dantzer, a neuroimmunologist at MD Anderson Cancer Center in Houston, Texas, notes that some of the participants in the low-CRP group fared worse on Infliximab than on placebo. Thus, the CRP test could be as important a tool for excluding depressed patients from taking anti-inflammatory therapies as for predicting responders.

Correction: An older version of this article mistakenly stated that 62% of depressed participants with high CRP levels responded well to the new treatment, compared with 33% of participants with low CRP levels.



  1. Report this comment

    Anna Babicz said:

    Very good news for people with depression. I’m waiting for next steps.
    By the way – after reading I found information about study “association between depression and elevated C-reactive protein” ( Is there any relation with those two studies?

  2. Report this comment

    Bernard Carroll said:

    This report is much ado about nothing. It is a classic of hyperanalyzing weak data to find the pony. Archives of General Psychiatry was taken in by the hyperbole – no surprise there. So was your reporter Amy Maxmen. Her statement that “Whereas 62% of depressed participants with high CRP levels responded well to the new treatment, only 33% of participants with low CRP levels did” is flat out wrong. The response rate among those with low CRP levels was 41%, and there was no significant difference in response rate between those with high or low CRP levels.

    It gets worse… for the more pertinent outcome of remission, results trended in favor of placebo (26%) over infliximab (9%).

    An especially outrageous claim made in the article by Dr. Miller is that “the number needed to treat for participants with a hs-CRP concentration greater than 5 was 3.45 (as opposed to 8-10 for conventional antidepressants).” As Dr. Miller should have realized, an NNT calculation for such a tiny sample (22 cases total!) is preposterous, especially as there was no significant difference anyway between response rates with placebo and active drug in this subgroup.

    Isn’t it time to restore quality standards in clinical psychiatric research, and for NIMH to stop funding this nonsense?

Comments are closed.