Long-term characterisation of the relationship between change in depression severity and change in inflammatory markers following inflammation-stratified treatment with vortioxetine augmented with celecoxib or placebo

Abstract

Background

Major depressive disorder (MDD) is a highly prevalent condition with a substantial incidence of relapse or treatment resistance. A subset of patients show evidence of low-grade inflammation, with these patients having a higher likelihood of more severe or difficult to treat courses of illness. Anti-inflammatory treatment of MDD has been investigated with mixed results, and no known studies have included assessments beyond cessation of the anti-inflammatory agent, meaning it remains unknown if any benefit from treatment persists. The objective of the present study was to investigate treatment outcomes up to 29 weeks post-cessation of celecoxib or placebo augmentation of an antidepressant, and how concentrations of selected inflammatory markers change over the same period.

Methods

The PREDDICT parallel-group, randomised, double-blind, placebo-controlled trial (University of Adelaide, Australia) ran from December 2017 to April 2020. Participants with MDD were stratified into normal range or elevated inflammation strata according to screening concentrations of high sensitivity C-reactive protein (hsCRP). Participants were randomised to treatment with vortioxetine and celecoxib or vortioxetine and placebo for six weeks, and vortioxetine alone for an additional 29 weeks (35 total weeks). Following a previous publication of results from the six-week RCT phase, exploratory analyses were performed on Montgomery–Åsberg Depression Rating Scale (MADRS) scores, response and remission outcomes, and selected peripheral inflammatory markers across the entire study duration up to week 35.

Results

Participants retained at each observation were baseline N=119, week 2 N=115, week 4 N=103, week 6 N=104, week 8 N=98, week 22 N=81, and week 35 N=60. Those in the elevated hsCRP celecoxib-augmented group had a statistically significantly greater reduction in MADRS score from baseline to week 35 compared to all other groups, demonstrating the greatest clinical improvement long-term, despite no group or strata differences at preceding time points. Response and remission outcomes did not differ by treatment group or hsCRP strata at any time point. Changes in hsCRP between baseline and week 35 and Tumour Necrosis Factor-α (TNF-α) concentrations between baseline and week 6 and baseline and week 35 were statistically significantly associated with MADRS scores observed at week 6 and week 35 respectively, with reducing TNF-α concentrations associated with reducing MADRS scores and vice versa in each case. A post-hoc stratification of the participant cohort by baseline TNF-α concentrations led to significant prediction by the derived strata on clinical response at weeks 6, 8 and 35, with participants with elevated baseline TNF-α less likely to achieve clinical response.

Interpretation

The present analysis suggests for the first time a possible longer-term clinical benefit of celecoxib augmentation of vortioxetine in inflammation-associated MDD treatment. However, further research is needed to confirm the finding and to ascertain the reason for such a delayed effect. Furthermore, the trial suggests that TNF-α may have a stronger relationship with anti-inflammatory MDD treatment outcomes than hsCRP, and should be investigated further for potential predictive utility.

Clinical Trials Registration: Australian New Zealand Clinical Trials Registry (ANZCTR), ACTRN12617000527369p. Registered on 11 April 2017, http://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?ACTRN=12617000527369p.