Novel Augmentation Strategies in Major Depression

Dan Med J. 2017 Apr;64(4):B5338.


Hypothesis The hypotheses of all the four included studies share the common idea that it is possible to augment the effect of antidepressant drug treatment by applying different interventions and with each intervention attain a clinically meaningful better effect compared to a control condition, and with minor side effects, thus improving the short- and medium-term outcome in major depression. Procedures Study design The basic study design has been the double blind randomised controlled trial (RCT). In the light therapy study, all patients were treated with sertraline for the whole of the study duration. In the first five weeks of the study, patients were randomised to treatment with either 60 minutes of bright white or 30 minutes of dim red light (sham condition). In the four weeks follow-up period, patients were treated with sertraline alone. In the Pindolol study, all patients were treated with venlafaxine and randomised to augmentation with either active or placebo matching pindolol tablets. In the PEMF study patients were continued on ongoing medication and randomised to augmentation with active or inactive (sham) 30 minutes daily PEMF treatment on weekdays. In the Chronos study all patients were treated with duloxetine and randomized to either a combination of three wake therapies with daily bright light treatment and sleep time stabilisation (wake group) or to daily exercise of minimum 30 minutes as an active control intervention (exercise group). The Chronos study was divided into: (1) a one-week run-in phase where duloxetine were started (and continued for the whole 29 week study period), (2) a one-week inpatient intervention phase where patient in the wake group did three wake therapies (sleep abstinence for the whole night and the following day until evening) in combination with daily light therapy and guidance on sleep time stabilisation and patients in the exercise group started a daily exercise program, (3) a seven week continuation phase where patient in the wake group continued light therapy and sleep time stabilisation and patients in the exercise group continued an individual exercise program, and (4) a 20 week follow-up phase with the same treatment elements but where duloxetine dosage could be adjusted or changed to other antidepressants. Recruitment Patients recruited for these studies were allocated from general practitioners, psychiatric specialist practices and for the lesser part from open psychiatric wards. Only a few patients were re-cruited through advertisements (in the PEMF and Chronos studies). Inclusion criteria Inclusion criteria were major depression according to the DSM-IV, including a depressive episode as part of a bipolar disorder. For the PEMF study, treatment resistance was a specific inclusion criterion. Duration of studies Study duration was nine weeks for the light therapy study, 19 days for the Pindolol study, five weeks for the PEMF study, and 29 weeks for the Chronos study. Assessments In all studies, assessments were done with clinician rated scales, patient self-assessment scales, including quality of life scales and a side effect scale. As clinician rated scales we used the Hamilton depression rating scale: the HAM-D17 and its 6 item subscale: the HAM-D6, the Bech Rafaelsen Melancholia scale (MES), and the Bech Rafaelsen Mania scale (MAS). As self-assessment scales we used the Major Depression Inventory (MDI), the Symptom Check-list (SCL-92), and the Preskorn scale. For side effects we used the UKU scale. Further scales used are mentioned in the specific study sections. Assessments in the light therapy study were done weekly for the first six weeks and finally after nine weeks; at four time points in the Pindolol study (baseline, days 6, 11 and 19), weekly for five weeks in the PEMF study and weekly for the first nine weeks of the Chronos study and thereafter every four weeks. The clinical setting for evaluation has been the Psychiatric Research Unit at Mental Health Centre North Zealand. For the Bright Light study, Pindolol and PEMF study patients were also seen at a psychiatric specialist practice in Copenhagen. Biochemical measures In the Light therapy study saliva cortisol was collected at baseline before start of light therapy and sertraline and blood was drawn for thyroid analysis. In the Chronos study saliva and 24 hour urine cortisol was collected in the patients randomised to the exercise group. Main results The main results from the Bright Light study covering the first five weeks of the study are given in the PhD thesis "Adjunctive bright light in nonseasonal major depression" defended and awarded on the 18 November 2004 at the University of Copenhagen. Results from the cortisol measurement and for the four weeks extension period were published in separate papers after the PhD thesis and are included in this thesis. Results from the Bright Light study Analysis of the saliva cortisol measurements taken at baseline of the study as cortisol awakening profiles (CAR) showed that patients responded differentially to light treatment according to their CAR levels (dichotomized to high or low about the mean). Thus, in the bright light group HAM-D17 scores were reduced by 15.7 (4.2) points for patients with a low CAR (below mean), and 11.4 (4.8) points for patients with a high CAR (above mean). In the dim light group the corresponding values were 11.1 (5.2) for patients with a low CAR and 11.3 (5.3) for patients with a high CAR. This interaction between CAR and treatment group was statistically significant (p = 0.006). Survival analysis, for the first five weeks of the study period, showed a statistically significant higher response rate (χ2= 9.6, p =0.002) and higher remission rate (χ2 = 12.5, p = 0.0004) for the bright light treated group versus the dim light treated group. At end of the five weeks of light treatment response rates were 66.7% versus 40.7 % and remission rates were 41.7 % versus 14.8 % for the bright versus dim light treated group. In the subsequent publication that covered the four weeks extension period where light treatment was discontinued, data showed that the attained differences in response and remission rates between groups were not sustained. The offset of effect was nearly complete after four weeks of continued treatment on sertraline only. Thus, at end-point, response rates were 79.2 % versus 75.9 % and remission rates were 60 .4 % versus 55.6% in the bright versus dim light groups. The conclusion reached was that bright light in non-seasonal depression should be used to achieve an earlier antidepressant response and that light therapy probably should be of longer duration. Results from the Pindolol study The results from the Pindolol study showed that pindolol did not augment the effect of venlafaxine for the whole sample. However, for those patients classified as slow metabolizers, based on their O-desmethylvenlafaxine/venlafaxine ratio (ODV/V), pindolol did augment the antidepressant effect. For patients classified as fast metabolizers, pindolol worsened the outcome. This interaction between ODV/V ratio and treatment group was statistically significant (p = 0.01). Results from the PEMF study The results from the PEMF Study showed that treatment with active versus sham PEMF augmented the effect of the ongoing anti-depressant medication treatment. Thus, patients in the active PEMF group attained a statistically significant greater score reduction from week one and at all subsequent assessments compared to the sham treated group (p < 0.01). Response and remission rates in the active PEMF group were also larger than in the sham treated group with response rates at endpoint of 61.0 % versus 12.9 % (p < 0.01) and remission rates of 33.9 % versus 4.1 % (p < 0.05). Results from the Chronos study The Chronos study, published in three papers, covers a one-week intervention phase, a seven weeks continuation phase, and a 20 weeks follow-up phase. Results from the intervention week showed that patient treated in the wake group, from the day after the first wake therapy, had en clinically and statistically significant better antidepressant effect compared to the exercise group. On the HAM-D6 scale (which does not contains sleep items), patients in the wake group had a response rate after the first wake therapy of 58.7% versus 13.7% i the exercise group (p <0.0001) and a remission rate of 38.6% versus 2.9% (p <0.0001). After the second recovery sleep (the night after the second wake therapy = dag 5) patients in the wake group had a response rate of 75.0% versus 25.1% in the exercise group (p <0.0001) and remission rates of 58.6% versus 6.0% (p <0.0001). Results from the continuation phase showed, on the HAM-D17 scale which was used at all the following assessments, at week two response rates of 41.4% in the wake group and 12.8% in the exercise group (p = 0.003) and remission rates of 23.9% versus 5.4% (p = 0,004). This clinically relevant and statistically significant difference between the wake and exercise groups was maintained at all the subsequent assessments with response rates of 71.4% versus 47.3% (p = 0.04) and remission rates of 45.6% versus 23.1% (p = 0.04), at week nine. Results from the 20 weeks follow-up phase showed a continued better effect in the wake group at all visits with HAM-D17 depression scored at week 29 of 7.5 (SE = 0.9) in the wake group versus 10.1 (SE = 0.9), (p = 0.02) in the exercise group. Remission rates were higher in the wake group with endpoint rates of 61.9% versus 37.9% (p = 0.01) in the exercise group. Response rates was only numerically, but not statistically, higher in the wake group with 74.6% versus 64.4% in the exercise group (p = 0.22). The sleep diary data showed a statistically smaller day-to-day variation in sleep onset, sleep midpoint, sleep offset and sleep duration in the wake group compared to the exercise group as a sign of better day-to-day sleep-wake cycle control in the wake group (p < 0.01). In the first nine weeks of the study patients in the wake group had a moderate sleep phase advance that diminished during the follow-up period. The hypothesised predictors for response to wake therapy were confirmed. Thus, in the wake group, a positive diurnal variation (morning worst, evening best) was associated with a better out-come, after the wake therapies, compared to a negative diurnal variation (morning best, evening worst). In the exercise group, the reverse was found, as a positive diurnal variation was associated with worse outcome, compared to a negative diurnal variation. This interaction between group and diurnal variation was statistically significant (p = 0.0004). The positive predictive value of response to the first wake therapy (i.e. maintaining response also at week two) was 56.3 % and the negative predictive value of non-response to the first wake therapy (i.e. maintaining no response also at week two) was 75.0 %. The impact of naps on depression severity was examined. In the wake group, patients who napped on the days after wake therapy compared to those patients not napping, had a more severe deterioration at the following assessment at week two (p = 0.02). Patients in the exercise group were able to perform exercise with a mean of 63.0 minutes/day (55.3) for the first eight weeks.

Publication types

  • Review

MeSH terms

  • Antidepressive Agents / therapeutic use*
  • Depressive Disorder, Major / therapy*
  • Double-Blind Method
  • Exercise Therapy / methods
  • Female
  • Humans
  • Male
  • Phototherapy / methods
  • Psychiatric Status Rating Scales
  • Randomized Controlled Trials as Topic
  • Severity of Illness Index
  • Transcranial Magnetic Stimulation / methods
  • Treatment Outcome


  • Antidepressive Agents