Treatment can benefit all severities of major depressive disorder

Whether baseline severity of major depressive disorder (MDD) influences the efficacy of antidepressants is an important clinical question. We now have a clear answer: in the homogeneous population of Japanese patients, it does not. All patients benefit. How far this finding can be generalized to other groups with major depression is another question. But there is consensus on the need for personalized treatment.

The baseline severity of MDD does not influence the efficacy of acute treatment with antidepressant treatment. Once diagnosed as having MDD, patients who are at the mild end of the dimension benefit as much from treatment as those with severe disease.

That is the clear take-home message from the meta-analysis of individual patient data presented at EPA 2019 by Professor Andrea Cipriani, Warneford Hospital, University of Oxford, UK.

The model that best fitted the data showed that the interaction between baseline severity and treatment effect was far from statistically significant, with a p value of 0.49.1  

Patients at the mild end of MDD gain as much from treatment as those with severe disease

The meta-analysis included data from around 2,300 patients, with a mean age of 40 and a mean score on the Hamilton scale of 22-23, who were treated for 6-8 weeks. The proportion of female patients – 52% -- was lower than expected in routine practice.

Earlier studies of severity had flaws

Professor Cipriani and colleagues believe their work is an advance on earlier studies which had less statistical power and accounted for missing data by carrying forward values from the last observation – an approach which they say is no longer adequate.

Also, one of the four studies included patients with minor depression, ie people who did not meet the criteria for MDD. Minor depression, he emphasized, is different from meeting the diagnostic criteria for MDD but being less severely affected.

In the recent meta-analysis, appropriate sensitivity analyses were carried out. The initial model was based on four randomised controlled trials (RCTs), and tested using raw data and data adjusted for potential confounders, notably age and gender. The model was then applied to a further two RCTs; and the findings were replicated.

Clinical features and patient preferences – both should guide choice of treatment

Robust but limited findings

Including only time points at which all studies reported outcomes made no difference to the conclusions of the meta-analysis. Neither did excluding studies in which MADRS scores had to be converted into scores on the Hamilton scale so that there was only one outcome measure across all trials. Hence the conclusions of the meta-analysis can be regarded as robust.

However, there are important caveats. While the study was technically robust, the analysis involved data from only six RTCs; all patients were on monotherapy; and – importantly -- they were all conducted in Japan.

Japan was chosen because of its homogeneous population and because of a generally favorable approach to comprehensive release of data at the level of individual patients.

Matching treatments to individuals is the essence of a modern approach: we treat patients, not averages

The potential lack of generalizability of the new analysis suggests caution in urging that current guidelines such as those of NICE and APA should be updated in the light of its findings.

But there is a kind of corroboration from a meta-analysis of cognitive behavioral therapy (CBT) for depression. This again used data from individual patients, and it again found that MDD patients can expect benefit from CBT across the wide range of baseline severity.2

Why is depression different?

An interesting question is why baseline severity seems not to influence the efficacy of antidepressants or CBT for depression while severity does modulate the efficacy of antipsychotics and agents used in autism and mania.3

Patient preferences need to be part of the picture

One factor may be the heterogeneity of patients with an MDD diagnosis; and another may be that the effect size in treating depression is less than, for example, in schizophrenia.

Professor Cipriani’s passion is to advance personalized treatment in psychiatry. Ideally, choice of treatment would reflect factors such as gender, co-existing anxiety, duration of symptoms, family history, and cardiac risk.

It should also take into account patient preferences. In ongoing work in Oxford, patients are being asked to use a sliding scale to rate how concerned they are to avoid specific adverse events such as insomnia, nausea, sedation, tremor and weight gain. All are relevant to selection of the most appropriate antidepressant therapy.

Our correspondent’s highlights from the symposium are meant as a fair representation of the scientific content presented. The views and opinions expressed on this page do not necessarily reflect those of Lundbeck.


  1. Furukawa TA et al. Initial severity of major depression and efficacy of new generation antidepressants: individual participant data meta-analysis. Acta Psychiatr Scand. 2018 Jun;137(6):450-458.
  2. Furukawa TA et al. Initial severity of depression and efficacy of cognitive-behavioural therapy: individual-participant data meta-analysis of pill-placebo-controlled trials. Br J Psychiatry. 2017 Mar;210(3):190-196.
  3. Leucht S et al. Possibly no baseline severity effect for antidepressants versus placebo but for antipsychotics. Why? Eur Arch Psychiatry Clin Neurosci. 2018 Oct;268(7):621-623.