Statistics from Altmetric.com
Studies were identified by searching Medline (1966–99), EMBASE/Excerpta Medica (1988–98), PsycLIT (1974–97), CINAHL (1982–99), and the Cochrane Controlled Trials Register database; scanning the reference lists of identified articles; and contacting the manufacturers of SSRIs.
Studies were selected if they were randomised, double blind, placebo controlled trials investigating SSRIs in the management of PMS.
Data were independently extracted in duplicate on study design, patient characteristics, drug type and dose, outcome measures (primary outcome measure was reduction in overall PMS symptoms), and side effects.
15 studies met the selection criteria involving 904 women (570 allocated to active treatment and 435 allocated to placebo, including 101 in crossover trials). The 2 most studied SSRIs were fluoxetine (7 trials) and sertraline (5 trials). Most of the trials presented continuous data, thus an overall standardised mean difference was calculated using a random effects model. The overall standardised mean difference for reduction in PMS symptoms in favour of SSRIs was –1.1 (95% CI –1.4 to –0.8), which is equivalent to an odds ratio of 6.9 (CI 3.9 to 12.2). In 7 trials where data could be extracted for a comparison between physical and behavioural symptoms, SSRIs were found to be effective in treating both with no statistically significant variance in the overall standardised mean differences. No difference in the effectiveness of SSRIs existed when comparing continuous and intermittent doses or between trials funded by pharmaceutical companies and those funded otherwise. Withdrawal from the studies due to side effects was 2.5 times greater (CI 1.6 to 3.7) in the SSRI group than in the placebo group.
In patients with severe premenstrual syndrome, selective serotonin reuptake inhibitors reduce symptoms.
This well done review by Dimmock et al shows that SSRIs improve symptoms in PMS. DSM-IV criteria for the severe variant of PMS, premenstrual dysphoric disorder, are stringent and require prospective confirmation of the symptoms and their unique occurrence during the luteal phase and disappearance during the follicular phase for 2 consecutive cycles. Depression may occur concomitantly. There is no accepted pathophysiological abnormality which means we have no diagnostic lab test for this difficult to define condition. It has, however, been shown that women with PMS, compared with those without PMS, exhibit an abnormal response to normal hormonal changes.1
As the authors have shown in a previous review,2 other pharmacological agents also work in PMS, suggesting that we try those with the lowest side effect profile after carefully identifying patients for treatment. But, importantly, it would not be appropriate to limit treatment to medication with these often severely distressed patients. Easily deployed non-drug interventions work well. Cognitive behavioural treatment is effective in patients with medically unexplained symptoms.3 Furthermore, careful attention to the provider patient relationship in a patient centred atmosphere has equally well documented benefits, as recent reviews note.4, 5 Thus, one can best help symptomatic patients with a multidimensional intervention, one small part of which is an SSRI.
Source of funding: no external funding.
For correspondence: Dr P W Dimmock, Academic Department of Obstetrics and Gynaecology, Keele University and North Staffordshire Hospital, Stoke-on-Trent ST4 6QG, UK. Fax +44 (0)1782 747319.
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.