Anti-inflammatory strategies in the treatment of schizophrenia

Abstract

Schizophrenia is a major mental illness with a lifetime prevalence of about 1%. Antipsychotic drugs, with a primary mechanism of action that involves dopamine receptor blockade, are the mainstay in the treatment of the disorder. However, despite optimum antipsychotic treatment, few patients return to pre-morbid levels; the treatment deficit includes refractory positive symptoms, negative symptoms, mood impairments, cognitive impairments, social impairments, and/or a variety of medication-related adverse effects, including extrapyramidal symptoms, metabolic disturbances, hyperprolactinemia, and others. To address these, antipsychotic treatment has been augmented with psychosocial interventions, cognitive rehabilitation, different kinds of electrical and magnetic brain stimulation, and a large range of drugs from the neuropsychiatric as well as, surprise, the general medical pharmacopeia. The pleomorphic pathophysiology of schizophrenia includes abnormalities in immunological and inflammatory pathways, and so it is not surprising that anti-inflammatory drugs have also been trialed as augmentation agents in schizophrenia. This article critically examines the outcomes after augmentation with conventional anti-inflammatory interventions; results from randomized controlled trials do not encourage the use of either aspirin (1000 mg/day) or celecoxib (400 mg/day), both of which have been studied for this indication during the past decade and a half.

Keywords:

• anti-inflammatory treatment
• antipsychotic augmentation
• aspirin
• celecoxib
• NSAIDs
• schizophrenia

Data from animal models of schizophrenia as well as patients with the illness suggest that aberrant inflammatory mechanisms and inflammatory markers are present in plasma as well as in the brain before, during, and after the onset of psychosis; these have been well reviewed, recently [1–4]. Clues about possible inflammatory origins include association with history of infection and presence of antibodies against specific organisms, presence of possible autoantibodies in serum and cerebrospinal fluid, changes in the levels of inflammatory cytokines, glial cell activation in the brain, and others (1–4). The inflammatory processes may be genetically driven [4] or resultant from prenatal or postnatal exposure to infection [3]. Inflammatory hypotheses converge with rather than supplant existing neurochemical hypotheses, such as the dopaminergic and glutamatergic hypotheses of schizophrenia [3]. If inflammation contributes even partly to the pathophysiology of schizophrenia, then treatments as ordinary as conventional anti-inflammatory drugs may strike at the root of the disease process [3]. What is the evidence, so far?

Among the conventional anti-inflammatory drugs, only two have been systematically studied as antipsychotic augmentation agents in double-blind, randomized controlled trials (RCTs). These are aspirin (two RCTs; pooled n = 270) and celecoxib (six RCTs; pooled n = 504). A recent meta-analysis [5] that examined these two drugs separately found that aspirin but not celecoxib was associated with significant attenuation of total psychopathology ratings; the benefit associated with aspirin was, however, small (Hedges’ g, 0.30). Another recent meta-analysis [6] that pooled the data from the two drugs and examined different domains of benefit found no significant benefit with anti-inflammatory drug augmentation for attenuation of Positive and Negative Syndrome Scale (PANSS) total scores or PANNS-negative scores; whereas there was significant decrease in PANSS-positive scores, the benefit was almost negligibly small (Hedges’ g, 0.19). Because of high heterogeneity, secondary analyses of PANSS total scores were conducted; these found a small advantage for aspirin but not celecoxib; for anti-inflammatory augmentation in RCTs of in-patients but not those that included out-patients; and for anti-inflammatory augmentation in RCTs that included first-episode patients but not in RCTs of only chronic patients (significant Hedges’ g range, 0.29–0.44). The findings of these analyses must be viewed with caution because all were exploratory.

Aspirin resulted in significantly improved outcomes in meta-analyses [5,6]. However, one [7] of the two aspirin RCTs has not yet been published although the abstract appeared in print in 2012. This large (n = 200), 16-week study [7] did not support aspirin augmentation. The smaller (n = 70), shorter (12-week), published RCT [8] obtained only a small effect size for PANSS total (Cohen’s d, 0.47) and PANSS-positive (Cohen’s d, 0.39) ratings. Absolute differences in scores between aspirin and placebo groups were small and of uncertain clinical significance – only 3.8 and 0.9 for PANSS total and PANSS-positive ratings, respectively. Both these RCTs dosed aspirin at 1000 mg/day, which could raise the risk of bleeding events (in the upper gastrointestinal tract and elsewhere) to unacceptably high levels during chronic therapy [9]. Both studies administered pantoprazole 40 mg/day to reduce the specific risk of gastrointestinal adverse events; the resultant increase in gastric pH could impair the absorption of some drugs [10].

What about the six celecoxib RCTs? As already stated, in meta-analysis, celecoxib did not outperform placebo [5,6]. Two of these RCTs were completed more than a decade ago; both failed to find an advantage for celecoxib and were never published in full [11,12]. One of these unpublished RCTs [11] was the largest (n = 270) and the longest (11 weeks) trial among the six. The other unpublished RCT [12] found PANSS total score benefits only in an exploratory analysis of patients with illness duration of <2 years.

Of the remaining four celecoxib RCTs, which were all published, one [13] found no benefits with celecoxib in continuously symptomatic out-patients receiving antipsychotic medication. Two more, both from the same team [14,15], found small benefits on some outcome measures, but the results at the study end point were statistically significant in only one of these [15]. In the other [14], acceleration of response was observed, but no statistically significant separation of groups was found at the end of the study. The last celecoxib trial [16] obtained strikingly positive results. This was a small (n = 60), short (8 weeks) trial that emerged from Iran. It obtained an enormous effect size (Hedges’ g, 0.93). This result must be viewed with caution; there is a large number of psychopharmacology RCTs from this part of the world, almost all of which have large, positive results.

All of the six celecoxib studies dosed the drug at 400 mg/day; patients receiving the drug are at risk of bleeding [17] and cardiac adverse events [18] with long-term treatment.

There are no data on the effects of aspirin or celecoxib on other important treatment targets in schizophrenia, such as social and neuropsychological deficits; given the absence of improvement in negative symptoms, improvement in these domains seems rather unlikely. So, the general conclusion that one may draw appears to be that anti-inflammatory drugs such as aspirin and celecoxib do not offer clinically worthwhile benefits in the augmentation of antipsychotic drugs in patients with schizophrenia [19]; furthermore the risk–benefit ratio is unfavorable. A note is made here that it is not easy for the null hypothesis to be rejected in add-on studies. So, if anti-inflammatory mechanisms truly mediate schizophrenia pathology, then anti-inflammatory treatments may yield statistically and clinically significant effect sizes when used in monotherapy. However, it is unlikely that such RCTs will ever be conducted given the pre-eminence of dopamine receptor antagonists in the treatment of the disorder.

What we do not know is whether anti-inflammatory agents are of value in subsets of patients, such as those preselected for antipsychotic refractoriness. Whereas this might be a worthwhile subject for future research, preliminary data suggest that patients who do not respond adequately to their ongoing antipsychotic may not benefit from celecoxib augmentation [13]. Another line of enquiry would be to examine the benefits of aspirin or celecoxib in patients in whom the baseline levels of markers of inflammation cross thresholds of interest. If benefits are detected, a further possibility would be to examine whether the benefits are enduring; that is, whether the course of the illness is improved.

Interestingly, other drugs with anti-inflammatory activity have also not shown consistent benefits in schizophrenia. For example, epigallocatechin gallate did not outperform placebo in a pilot study in patients with schizophrenia or bipolar disorder [20]. There are also many studies on omega-3 fatty acids, minocycline, and neurosteroids. They were recently reviewed [5,21] and are not considered further here.

On a parting note, immunological treatments have been approved for a variety of autoimmune disorders, ranging from rheumatoid arthritis to multiple sclerosis. If immunological mechanisms (which are related to anti-inflammatory mechanisms) are involved in schizophrenia, immunological approaches may merit study, whether or not conventional anti-inflammatory treatments work [3]. In this regard, infliximab, a TNF-alpha antibody, was found to reduce depression rating scores at 12 weeks in patients with treatment-resistant depression and elevated baseline TNF-α levels [22]. The finding encourages proof-of-concept schizophrenia studies. However, enthusiasm must be tempered by the knowledge that there is no evidence of active brain inflammation as seen, for example, in patients with multiple sclerosis who are experiencing an acute exacerbation of illness.

Financial & competing interests disclosure

C Andrade was recently a Principal Investigator and Medical Monitor for a multi-center, investigator-initiated, Indian investigation of the efficacy of Sensoril, a branded formulation of Ashwagandha. The study was funded by Natreon Inc. (USA). C Andrade has received funding from the Indian Council for Medical Research, Department of Biotechnology, and Corcept Therapeutics. The funding has supported the purchase of laboratory supplies and the salaries of research officers. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.