Response

The Feature article by De Giorgi and colleagues ^{Reference De Giorgi, Hahn, Cheliotis-James, Smith, Penninx and Pillinger1} brings an insightful and necessary critical view on the field of immunometabolic psychiatry. In their comprehensive review, using the specific cases of autoimmune encephalitis, immunometabolic depression and heart health in psychosis, the authors try to answer one of the key questions of this line of research: why have immunometabolic approaches to diagnosis, prognosis or therapeutics not been incorporated into clinical practice in psychiatry? Over the past decades, accumulating evidence has consistently implicated immunometabolic mechanisms in the pathophysiology of serious mental illnesses (SMI). One of the key pillars of the immunometabolic hypothesis of SMI – that inflammatory and metabolic dysregulation is intrinsic, or independent from lifestyle or iatrogenic factors – has gradually been strengthened over time by clinical studies showing an elevated prevalence of immunometabolic abnormalities in drug-naïve patients; ^{Reference Chen, Wang, Lang and Zhang2} by epidemiological data demonstrating a prospective association between biomarkers of immunometabolic dysfunction and risk of onset of SMI; ^{Reference Watson, Simard, Henderson, Nutkiewicz, Lamers and Nasca3} and by multiple genetic studies, documenting large positive genetic correlations between SMI and inflammatory and metabolic diseases/traits. ^{Reference Fanelli, Franke, De Witte, Ruisch, Haavik and van Gils4} Conversely, neurobiological studies have consistently implicated brain and whole-body immunometabolic mechanisms in transdiagnostic populations.

Nonetheless, these findings have yet to be translated into clinical practice. No biomarkers related to inflammatory or metabolic function have been validated for diagnostic or prognostic applications. Multiple clinical trials have tested the therapeutic potential of inflammatory and metabolic agents. Despite some promising results, results have generally been mixed and no candidate agent has demonstrated efficacy. De Giorgi and colleagues elegantly appraise some of the major challenges of translational efforts. A connecting thread across the different cases assessed is the definition and categorisation question: how broadly could/should we define autoimmune psychosis? The implications of applying this concept to a relatively small proportion of atypical presentations of SMI, or the much larger overall population, are tremendous. Ultimately, this question can be answered by the trials results, specifically the success or failure of potentially disease-modifying immunotherapies in more ‘classical’ cases of SMI. More challenging is how to operationalise the immunometabolic phenotype, or subtype, of depression. Whereas autoimmune encephalitis can be identified by specific antibodies, immunometabolic depression has been proposed as a clustering of inflammatory (e.g. elevated levels of C-reactive protein (CRP) and cytokines), metabolic (e.g. obesity, dyslipidaemia, insulin resistance) and behavioural features (e.g. anergia and anhedonia). Although it can be argued that definitions based on clustering of distinct but interrelated phenomena are closer to biological reality than those based on monocausal explanations, ^{Reference Fried5} they pose some practical problems. Pragmatic criteria are needed; most studies have hitherto relied on simplified measures, such as elevated CRP or body mass index, to define subpopulations. This is understandable because these are widely available and easy-to-collect indicators, but there are some obvious limitations as these are also fairly non-specific, imprecise markers. De Giorgi and colleagues relevantly propose that precision medicine approaches are promising. Specifically, there is a need to identify and validate more refined biomarkers of immunometabolic function that are appropriate for neuropsychiatric applications. For example, the measurement and definition of insulin resistance is based on peripheral measures (e.g. Homeostatic Model Assessment for Insulin Resistance, Matsuda Index), which primarily estimate hepatic insulin resistance. The brain response to insulin (as a whole and across different brain regions) is fundamentally distinct and is not accurately captured by these measures. It is plausible, but remains to be tested, that measures of brain-specific insulin resistance can be more useful for prognostic, diagnostic and therapeutic translations in populations with SMI.

Moreover, the need for more refined measures is not restricted to biomarkers. Phenotypical characterisations of immunometabolic depression generally highlight anhedonia as its core symptomatic presentation. Anhedonia is itself a highly heterogeneous construct which, in its most commonly used definitions, incorporates at least partially distinct behavioural domains, such as the experience of pleasure, motivated behaviour and reward learning. ^{Reference Cooper, Arulpragasam and Treadway6} More clarity on the emotional and behavioural manifestations of immunometabolic depression, preferably beyond surface-level symptomatic expression, will be required to advance this line of research. Conversely, there is also increasing recognition of the need for more precise outcome measures for the development of novel interventions. Previous clinical trials testing the efficacy of inflammatory or metabolic treatments have typically used global depression scales as primary outcome measures, as required by regulatory agencies. These scales, such as the Montgomery–Åsberg Depression Rating Scale, were developed and calibrated to capture the effects of antidepressants and are conceivably not sufficiently sensitive to detect the effects of interventions with clearly distinct mechanisms of action and dynamics. Results from studies using alternative outcome measures, such as objective behavioural tasks, have been promising, ^{Reference Gill, Badulescu, Shah, Brudner, Phan and Di Vincenzo7} but extensive work needs to be done to validate and implement this approach.

Finally, the call for action from De Giorgi and colleagues to implement existing best practices on managing physical health in individuals with SMI is very timely. Despite accumulating evidence of the prevalence and impact of medical comorbidities in populations with SMI, and the availability of evidence-based interventions for mental illness and cardiometabolic disease prevention and early intervention, uptake in clinical practice is still shockingly low. ^{Reference De Hert, Cohen, Bobes, Cetkovich-Bakmas, Leucht and Ndetei8} As a result, the disparities in health outcomes and life expectancy between individuals with and without a SMI have increased over the past several decades. ^{Reference Lindstrӧm, Siersma, Kriegbaum, Grauers Willadsen, Bakkedal and Brodersen9} Although the transformative potential of better understanding the role of immunometabolic mechanisms in SMI aetiopathology is exciting, De Giorgi and colleagues do a great service by reminding us that there are several ‘low-hanging fruits’ ready for implementation practices that can immediately address a critical unmet need.