Death Risk Spikes With New Onset-Depression After RA Diagnosis

Sara Freeman

June 15, 2022

The development of depression after a rheumatoid arthritis diagnosis increased the risk for death "more than sixfold" when compared with having no depression at diagnosis, according to Danish researchers.

Cumulative mortality at 10 years was approximately 37% in patients with comorbid RA and depression versus around 13.5% of RA patients with no depression, Jens Kristian Pedersen, MD, PhD, of Odense (Denmark) University Hospital–Svendborg Hospital and the department of clinical research at the University of Southern Denmark, also in Odense, reported at the annual European Congress of Rheumatology.

"According to [antidepressant] exposure status, the cumulative mortality followed two clearly different paths," Pedersen said. "The mortality curves separated early and already within the first and second year of follow-up."

RA, Depression, and Mortality

Rates of depression in patients with RA are high, Pedersen said, and while it's previously been reported that their coexistence can increase mortality, this is the first time that the link has been investigated in a population newly diagnosed with RA.

In this study, Pedersen and collaborators wanted to look at the association in incident RA and defined depression as the first filling of an antidepressant prescription.

"Although antidepressants are used for different indications, we have recently described that in RA the most frequent indication for filling antidepressants is depression," he explained. Moreover, that research found that "the frequency of filling coincides with the occurrence of depressive disorder previously reported in the scientific literature."

Data Sourced From Multiple Danish Registers

To examine the mortality risk associated with newly diagnosed RA and new-onset depression, Pedersen described how five different Danish registers were used.

First, data from the DANBIO register were used to identify patients with incident RA living in Denmark over a 10-year period ending in December 2018. Although perhaps widely known as a biologics register, DANBIO is required by the Danish National Board of Health to collect information on all patients with RA, regardless of their treatment.

Next, the Danish National Prescription Register and Danish National Patient Register were consulted to obtain data on patients who had a first prescription for antidepressant treatment and information on those who developed a diagnosis of depression. Demographic, vital status, and socioeconomic data were collated from the Danish Civil Registration System and Statistics Denmark databases.

To be sure they were looking at incident cases of RA and new cases of depression, the researchers excluded anyone with an existing prescription of antidepressants or methotrexate, or who had a confirmed diagnosis of either disorder 3 years prior to the index date of Jan. 1, 2008.

This meant that, from a total population of 18,000 patients in the DANBIO database, there were just over 11,000 who could be included in the analyses.

Overall, the median age at RA diagnosis was 61 years, two-thirds were female, and two-thirds had seropositive disease.

New-Onset Depression in Incident RA

"During follow-up, about 10% filled a prescription of antidepressants," said Pedersen, adding that there were 671 deaths, representing around 57,000 person-years at risk.

"The majority died from natural causes," he said, although the cause of death was unknown in 30% of cases.

Comparing those who did and did not have a prescription for antidepressants, there were some differences in the age at which death occurred, the percentage of females to males, the presence of other comorbidities, and levels of higher education and income. These were all adjusted for in the analyses.

Adjusted hazard rate ratios were calculated to look at the mortality risk in patients who had antidepressant exposure. The highest HRR for mortality with antidepressant use was seen in patients aged 55 years or younger at 6.66, with the next highest HRRs being for male gender (3.70) and seropositive RA (3.45).

But HRRs for seronegative RA, female gender, and age 55-70 years or older than 75 years were all still around 3.0.

Depression Definition Questioned

"My only concern is about the definition of depression in your analysis," said a member of the audience at the congress.

"You used antidepressant use as a proxy of depression diagnosis, but it might be that most or many patients have taken [medication] like duloxetine for pain control, and you are just seeing higher disease activity and more aggressive RA."

Pedersen responded: "After the EULAR 2022 submission deadline, we reanalyzed our data using two other measures of depression.

"First, we use treatment with antidepressants with a positive indication of depression, according to the prescribing physician, and secondly, we used first diagnosis with depression according to ICD-10 Code F32 – 'depressive episode after discharge from hospital as an outpatient,' " he said.

"All definitions end up with a hazard rate ratio of about three. So, in my opinion, it doesn't matter whether you focus on one measure of depression or the other."

David Isenberg, MD, FRCP, professor of rheumatology at University College London, wanted to know more about the antecedent history of depression and whether people who had been depressed maybe a decade or 2 decades before, were more likely to get RA.

That calculation has not been done, Pedersen said, adding that the study also can't account for people who may have had recurrent depression. Depression treatment guidelines often recommend nonpharmacologic intervention in the first instance, "so we do not necessarily get the right picture of recurrent depression if we look further back."

Pointing out that the sixfold increase in mortality was impressive, another delegate asked about whether it was because of a higher disease activity or joint damage and if the mortality risk might be lower in patients who are in remission.

"We don't know that yet," Pedersen answered. "We haven't done the calculations, but there is the issue of residual confounding if we don't take all relevant covariates into account. So, we need to do that calculation as well."

The study was supported by the Danish Rheumatism Association. Pedersen had no conflicts of interest to disclose.

This article originally appeared on, part of the Medscape Professional Network.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.