Article Text

Download PDFPDF

Review: little evidence to support dose escalation of selective serotonin reuptake inhibitors in non-responders
  1. Stephan Heres, MD,
  2. Werner Kissling, MD,
  3. Stefan Leucht, MD
  1. Klinik und Poliklinik fuer Psychiatrie und Psychotherapie am Klinikum,
    rechts der Isar der Technischen Universitaet Muenchen, Germany

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

OpenUrlAbstract/FREE Full Text

Q Is dose escalation of selective serotonin reuptake inhibitors effective for treating major depressive disorders in people who do not respond to standard doses?


Embedded ImageDesign:

Systematic review.

Embedded ImageData sources:

MEDLINE, EMBASE, CINAHL, and PsycINFO (search date February 2005).

Embedded ImageStudy selection and analysis:

Included studies were randomised controlled trials (RCTs) or meta-analyses based on systematic reviews assessing the effects of dose escalation of selective serotonin reuptake inhibitors (SSRIs) for major depressive disorders not responding to standard SSRI doses. Studies were screened independently by two reviewers. Included studies were critically appraised, assigned a level of evidence, and data extracted using standardised forms and criteria. No meta-analysis was conducted due to heterogeneity of the timing of dose escalation.

Embedded ImageOutcomes:

Response (defined as ⩾50% reduction in Hamilton Rating Scale for Depression (HRSD) score or Clinical Global Impression (CGI) improvement or severity score ⩽2); remission (defined in studies as HRSD score ⩽7 or ⩽8).


Eleven studies met inclusion criteria (8 primary studies and 3 systematic reviews). The 3 systematic reviews had methodological weaknesses, including meta-analysis of heterogeneous studies. Results from 4 lower quality RCTs suggested that dose escalation before 4 weeks of treatment with the standard dose had no effect on response or remission. One good quality RCT found that dose-escalation after 6 weeks of treatment was less effective than continuing with the standard dose alone. Two lower quality RCTs provided limited evidence that dose escalation of fluoxetine after 8 weeks’ treatment may increase response compared with adding lithium or desipramine, although it was not clear if this increase was significant in both studies. Increasing doses were associated with higher study withdrawals in all included studies; this was related to increased side effects in some studies.


There is little evidence to support dose-escalation of SSRIs for improving depression in people who do not respond to standard doses. The available evidence is mostly of poor quality.


One of the eight primary studies identified appeared to be an uncontrolled study. It is not clear why this study was included, and its results are not described above. The weaknesses of some of the RCTs were the inclusion of people with minor depression, absence of placebo controls, lack of power to detect clinically relevant differences, inadequate reporting of data and inadequate reporting of blinding.


In their systematic review, Ruhé and colleagues report on the current literature on dose escalation for insufficient response to standard-dose selective serotonin reuptake inhibitors (SSRIs) in major depressive disorder. Eight dose escalation studies and three meta-analyses were identified. The authors found limited evidence for superior efficacy of dose escalation over dose maintenance only after 8 weeks of preceding non-response had elapsed before randomisation. Earlier dose increases after 4 or 6 weeks did not benefit the patient. Adequate methods were applied for the systematic review and the authors’ decision not to pool the results of the different studies in a meta-analysis appears appropriate due to significant heterogeneity of study designs.

Limitations of the identified studies addressed in detail by the authors were varying methodological quality, under-reporting of dropouts as well as dose increment schedules and different timing in dose escalation. Another potential source of bias is that patients randomised to dose escalation might discontinue the study due to side effects earlier than patients in the placebo group; their higher severity scores are reflected in the LOCF analysis, the most commonly applied statistical approach. The named limitations all sum up the methodological problems of three meta-analyses1,3 (which did also not find a clear efficacy of dose increase) identified by the authors.

In the face of the review, the lack of evidence becomes apparent regarding what psychiatrists should recommend to patients who do not, or only partially, respond to treatment with SSRIs for 4–8 weeks. Current treatment guidelines4 do name dose increase as an option but at the same time stress the low evidence level for this recommendation. Nevertheless an early dose increase after 4 weeks without clinical amelioration may be overhasty in the context of the current findings.

Based on the outcome of the current review, future studies should conduct randomisation after 8 weeks of insufficient treatment, a time point corroborated by data from Quitkin and colleagues5 finding an earlier switch to be premature. Furthermore they should report on dose tolerant patients in addition to LOCF data and collect information on SSRI-plasma levels, which may help identify additional prognostic factors for response.



  • For correspondence: Dr H G Ruhé, c/o Mrs M Haages, Out-patient Department of Psychiatry A3-255.1, Academic Medical Centre, PO Box 22660, 1100 DD Amsterdam, the Netherlands; h.g.ruhe{at}

  • Sources of funding: Not stated.