Condition: PMS en PMDD

A 2024 Cochrane review pooled 34 randomized trials of fluoxetine, paroxetine, sertraline, escitalopram and citalopram against placebo. SSRIs probably reduce overall self-rated premenstrual symptoms (SMD -0.57, 95% CI -0.72 to -0.42; 12 studies, 1742 participants; moderate-certainty). Continuous dosing was more effective than luteal-phase dosing (continuous SMD -0.69 vs luteal -0.39; subgroup difference P = 0.03). Adverse effects were consistently more common than placebo, including nausea (OR 3.30), insomnia (OR 1.99), sexual dysfunction or decreased libido (OR 2.32) and fatigue (OR 1.52). SSRIs are the first-line drug treatment for PMDD and for severe PMS.

A 2024 systematic review and meta-analysis pooled 44 studies with 48 independent samples (50,659 participants). The point prevalence was 3.2% (95% CI 1.7% to 5.9%) for confirmed PMDD and 7.7% (5.3% to 11.0%) for provisional (recall-based) diagnosis, with very high heterogeneity (I2 = 99%). Restricting to community-based samples with a confirmed diagnosis gave 1.6% (1.0% to 2.5%) with low heterogeneity (I2 = 26%). The authors conclude that studies relying on provisional diagnosis produce artificially high rates. PMS in a milder, broader sense is far more common; PMDD is the severe, diagnostic end of the spectrum.

The Carolina Premenstrual Assessment Scoring System (C-PASS) study followed 200 women recruited for retrospectively reported premenstrual emotional symptoms and had them complete two to four months of daily ratings on the Daily Record of Severity of Problems (DRSP). Standardized C-PASS scoring agreed with expert clinician diagnosis in an estimated 98% of cases. Consistent with earlier work, retrospective reports of premenstrual symptom increases were a poor predictor of the prospective diagnosis. This is why DSM-5 requires symptoms to be confirmed by prospective daily ratings over at least two cycles; without it, a diagnosis is only provisional.

In a landmark NIH experiment, 20 women with PMS were given leuprolide (a GnRH agonist that suppresses ovarian function) or placebo. The 10 whose symptoms improved on leuprolide were then given estradiol and progesterone, each for four weeks, in a double-blind crossover while ovarian suppression continued. Symptoms returned significantly when the hormones were added back. Fifteen women without PMS given the identical regimen had no mood change. The conclusion is that in women with PMS the symptoms represent an abnormal behavioral response to normal hormonal changes, which is why treatments that quiet the response (SSRIs) or switch off the cycle (GnRH agonists, some contraceptives) both work.

A 2012 meta-analysis of 22 controlled trials of psychological and drug treatments for PMS and PMDD found cognitive behavioral interventions produced small-to-medium standardized effect sizes at post-treatment (range d+ 0.24 to 0.70) and maintained effects at follow-up (d+ 0.46 to 0.74), which was only assessed for the psychological treatments. Serotonergic antidepressants in the same analysis gave comparable post-treatment effects (d+ 0.29 to 0.58). The authors judged the overall efficacy of both approaches as not fully satisfactory and suggested combining them.

A 2023 Cochrane review pooled five randomized trials (858 women, most diagnosed with PMDD) of a combined oral contraceptive with drospirenone and low-dose ethinylestradiol against placebo. It may improve overall premenstrual symptoms (SMD -0.41, 95% CI -0.59 to -0.24; 2 RCTs, N=514; low-quality) and functional impairment in productivity, social activities and relationships. It also increased total adverse effects (OR 2.31) and withdrawals due to adverse effects (OR 3.41), with more breast pain, nausea and intermenstrual bleeding. The placebo also had a significant effect. Whether it beats other combined pills, or helps beyond three cycles or in milder cases, is not known.

A 2025 Cochrane review of 11 randomized trials (275 women) found GnRH agonists without add-back therapy improved global symptoms compared with placebo (SMD -1.23, 95% CI -1.76 to -0.71; 9 RCTs; high-certainty evidence for this outcome). The trade-off is that suppressing ovarian function induces a hypoestrogenic state with menopausal side effects such as hot flashes, and long-term use risks osteoporosis; women were more likely to withdraw due to adverse effects (RR 4.24). Add-back estrogen or progestogen is used to offset the menopausal state, but the review found insufficient evidence to confirm that add-back preserves efficacy, and one trial suggested estrogen-plus-progestogen add-back may worsen symptoms.

A 1998 multicenter, double-blind, placebo-controlled trial randomized 466 women (of 497 enrolled) with moderate-to-severe, prospectively documented premenstrual symptoms to 1200 mg/day of elemental calcium as calcium carbonate or placebo for three cycles. By the third treatment cycle, calcium produced a 48% reduction in total luteal-phase symptom scores from baseline versus 30% on placebo, with all four symptom factors (negative affect, water retention, food cravings, pain) significantly reduced. The benefit emerged from the second cycle onward.

A 2025 systematic review and meta-analysis of 82 observational studies (101,413 women) examined links between physical activity and menstrual health. Lower activity was associated with higher odds of PMS (OR 1.22, 95% CI 1.03 to 1.45; 18 studies) and of menstrual pain (OR 1.67; 21 studies). Study quality was low (mean 3.3 out of 7), most designs were cross-sectional, and physical activity was self-reported. The authors describe activity as a promising, low-cost, accessible approach while calling for standardized, prospective research.

A 2017 systematic review and meta-analysis found 17 randomized trials of Vitex agnus-castus for PMS, 14 poolable. Thirteen of 14 reported positive effects on total symptoms. The pooled effect in placebo-controlled trials was large (Hedges g -1.21, 95% CI -1.53 to -0.88), but heterogeneity was extreme (I2 = 91%) and most trials carried a high risk of bias, with signs of publication bias. The authors state the pooled effect should be read as exploratory and, at best, an overestimate, and call for high-quality trials against placebo, SSRIs and the pill.

A 1999 systematic review of nine randomized placebo-controlled trials (940 women) found the odds of improvement in overall premenstrual symptoms with vitamin B6 were 2.32 (95% CI 1.95 to 2.54) relative to placebo, and 1.69 (1.39 to 2.06) for premenstrual depressive symptoms across four trials (541 women). The authors caution that conclusions are limited by the low quality of most trials and that doses up to 100 mg/day are likely to help. Above roughly 100 mg/day, long-term B6 carries a risk of sensory peripheral neuropathy, which caps how much is sensible to take.

A 1996 systematic review searched for controlled trials of evening primrose oil for PMS and found only seven placebo-controlled studies, with clear randomization in five. Inconsistent scoring made a formal meta-analysis inappropriate. The two most well-controlled studies showed no beneficial effect, although the reviewers noted the trials were small enough that modest effects could not be excluded. Their conclusion was that on the available evidence evening primrose oil is of little value in managing PMS.

A 2024 multicenter, randomized, placebo-controlled trial enrolled 144 regularly menstruating women with PMS and a Chinese-medicine pattern of Liver depression, Spleen deficiency and blood-heat, at eight sites in China. They received Jiawei Xiaoyao pill (a modified Xiao Yao San, 12 g/day) or matching placebo for three cycles. The formula reduced luteal-phase Daily Record of Severity of Problems (DRSP) scores more than placebo (P = 0.002 for the full analysis set), covering psychological and somatic symptoms, with no significant difference in adverse events. This aligns the trial evidence with the traditional Liver Qi stagnation reading behind the Xiao Yao San family.

A 2018 systematic review and meta-analysis of 19 studies found alcohol intake was associated with a moderate increase in PMS risk (OR 1.45, 95% CI 1.17 to 1.79), with heavy drinking carrying a larger increase (OR 1.79, 95% CI 1.39 to 2.32). The included studies were observational (cohort, case-control and cross-sectional), and the authors specifically call for future work to avoid cross-sectional designs and to test whether there is a threshold below which alcohol has no effect.

A 2018 Cochrane review of five trials (277 women) found acupuncture may give a greater reduction in mood and physical PMS symptoms than sham (for example a mean difference of about -9 points on the DRSP mood and physical scales in one trial, n=67, low-quality evidence), and acupressure may reduce the proportion of women with moderate-to-severe symptoms (RR 0.64). Most findings rested on single small trials rated low or very low quality, and no study compared acupuncture or acupressure against currently recommended treatments such as SSRIs. The reviewers called for larger trials with validated outcomes and proper blinding.