Depression – let’s get personal! A report from an ECNP scientific symposium

Early signs of success

Trials of antidepressant drugs usually assess clinical outcomes weeks or months after trial commencement. However, André Tadić (Germany) suggested that clues to a therapy’s likely efficacy – if any - may be obtained earlier, possibly after just 2 weeks.

...after just 2 weeks differences between them were apparent – differences that were predictive of final outcome.

He described two prospective clinical studies that go some way to supporting his contention. In the first study, different anti-depressive therapies were compared and after just 2 weeks differences between them were apparent – differences that were predictive of final outcome. In the second study, the processing of a biological genetic marker correlated with response/non-response to anti-depressive therapy early in the course of a clinical study of patients with MDD. While much remains to be done, these prospective studies do indeed suggest that predictors of clinical response may be apparent much earlier in the course of treatment than is currently thought.

The future of depression research - spit and sunshine!

Professor Brenda Penninx (The Netherlands) was interested in predicting the course of a disease long-term rather than its short term outcome or remission rate. She reported the findings of the Netherlands Study of Depression and Anxiety – a study that followed patients for 2 years. What emerged was the identification of factors that were both easy to identify and correlated with poor outcome ( age, accompanying co-morbid anxiety disorder, longer symptom duration, greater symptom severity, older age at onset) and, perhaps more importantly, factors, such as gender, that did not correlate. However, as Professor Penninx commented, none of these factors help identify the underlying pathophysiology in depression.

Depression involves dysregulation of stress mechanisms and endocrine systems including the hypothalamus-pituitary-adrenal (HPA) axis. HPA axis function can be assessed by measuring cortisol in saliva, levels of which are known normally to peak upon awakening. Interestingly, in patients with depression - and in their children - this cortisol peak differs from that seen in controls and, therefore, may be indicative of the trait of depression. Furthermore, those with the lowest cortisol peak/trough had the poorest outcomes.

In addition to cortisol levels at awakening, other putative markers of depression are being studied. Plasma levels of brain derived neurotrophic factor (BDNF) are diminished in patients with depression compared to normal and may permit a measure of the state of major depressive disorder. Vitamin D, C-reactive protein, IL-6, IL-1 and TNF are also being investigated further and preliminary finding indicate that the levels of these markers may be predictive of poor treatment response and may be a useful area for future research.

Once may be enough

However, it may well be that predictors of successful antidepressant therapy outcomes are apparent very early in the course of managing depression. As Professor Harmer outlined, even exposure to a single dose of an antidepressant can initiate a change to a patient’s response to emotional cues including their response to the happy facial expression predictor. Importantly, in a caring environment, those patients showing the greatest response at such an early stage in the management of depression responded better to long-term antidepressant therapy than those showing a lesser response.

Interestingly, psychological interventions such as CBT could also produce beneficial effects similar to anti-depressants after just a single session. However, unlike responses to anti-depressants, changes following a single session of CBT were not immediate. Never the less they do occur – and delegates were keen to know which CBT regimen had been used to generate such an effect!

OK computer

Indeed, machine learning using the technique of functional MRI parameters showed 84% accuracy in identifying patients with treatment-resistant depression

Finally, Eric Ruhe (The Netherlands) explained that machines can be ‘trained’ to recognise the changes in brain activity predictive of short- and long-term outcomes in MDD. Indeed, machine learning using the technique of functional MRI parameters showed 84% accuracy in identifying patients with treatment-resistant depression. Although still early days, he suggested that in the prediction of treatment outcomes, it may be the early changes from a baseline parameter that will prove the most informative.