Around a half of depressed patients recommended drug therapy discontinue before a clinician would think that appropriate, and many patients stop within three months. Historical data suggest that the tricyclics suffered more from this problem than the more recent SSRIs, but compliance with therapy continues to be far from optimal. And adverse events are in large part responsible.

Of course, not all patients are at the same risk of experiencing particular side-effects. And Stefan Kloiber (Center for Addiction and Mental Health, University of Toronto, Canada) offered the hope – and some intriguing data in support – that combining our knowledge about clinical and biological risk factors may enable us to predict which patients are most at risk of adverse events with specific agents. Individualization of therapy should in turn improve adherence and hence, we expect, outcome.

We can’t change baseline weight or disease severity, but we can choose which drug we recommend

Pharmacogenomics and long QT

Given the importance of the cytochrome P450 system, genetic variants in CYP enzymes are potentially relevant to risk of toxicity. But there is a lack of randomised trial evidence that use of CYP pharmacogenomic tests improves outcome relative to treatment as usual.

Drug-induced prolongation of QT interval is an important phenomenon, and the risk of its occurrence is increased by being female, having existing heart disease, electrolyte disturbance, use of diuretics, longer QT interval at baseline, and mutations associated with congenital long QT syndrome.  Such factors might eventually be built into a predictive tool, but we don’t have it yet.

Something we might be able to predict and influence is risk of weight gain, much of which is drug-specific. In the Munich Antidepressant Response Signature (MARS) project, that studied pharmacologic treatment of MDD in more than seven hundred patients, weight change over five weeks ranged from a reduction of almost 10kg to an increase of 12kg. But the mean was a gain of 0.7kg.

Much weight gain drug specific and hence avoidable

Clinical factors associated with weight gain, which were validated in a University of Münster cohort of patients, were a low to normal baseline BMI, severity of depression, presence of psychotic symptoms, and use of medication known to be associated with increase in weight.

Based on the findings of these two observational studies, Dr Kloiber and colleagues developed a risk score of 0-4 made up the following factors: BMI of 25kg/m² or less, HAMD score of greater than 20, presence of psychotic symptoms and use of an antidepressant known for its potential to increase weight. Risk of weight gain rises markedly with a score of 3 and rises further with a score of 4.

A member of the audience asked whether any of the risk factors for weight gain could be reduced by the physician or the patient.

We can’t do much about baseline BMI or the severity or nature of depressive symptoms, said Dr Kloiber. But for a patient with two or three risk factors already, we can choose not to give a drug that is particularly associated with weight gain, since that would substantially increase their chances of experiencing this adverse event.

There are also genetic factors, of course. A genome-wide study (in preparation and needing replication) found a signal in an SNP associated with a gene previously linked to metabolic or weight disorders. This gene polymorphism may be a plausible player in the story.

Immune system biomarkers and depression

Evidence for the involvement of inflammatory cytokines was reviewed by Michelle Roche (National University of Ireland, Galway).

High IL-6 levels in children increase risk of depression onset by the age of 18 years. Ketamine-induced antidepressant effects are associated with reduction in IL-6 and TNF-alpha; and anti-TNF antibodies may be effective in patients with MDD and high baseline levels of C-reactive protein.

There is post-mortem evidence of activated microglial cells – which release pro-inflammatory mediators -- in the brains of patients with depression; and data from in vivo PET studies using radiolabelled tracer for translocator protein (expressed by microglia) links the extent of their activation to greater severity of depression.

Taken together, a good body of data suggests a connection between the immune system and the brain in depression that may well result in relevant effects on neurogenesis and synaptic plasticity. Inflammation is, of course, not specific to depression – being present, for example, in auto-immune diseases. But these, in turn, seem to increase the risks of depression.