Premenstrual dysphoric disorder is characterised by symptoms confined to the premenstrual phase of the menstrual cycle. Confirmed diagnosis requires prospective monitoring of symptoms over two cycles, otherwise the diagnosis is provisional. We aimed to measure the point prevalence of premenstrual dysphoric disorder.

We searched for studies of prevalence using MEDLINE, EMBASE, PsycINFO and PubMed. For each study, the total sample size and number of cases were extracted. The prevalence across studies was calculated using random effects meta-analysis with a generalised linear mixed model. Potential sources of heterogeneity were explored by meta-regression and subgroup analyses. Pre-registration was with PROSPERO (CRD42021249249).

44 studies with 48 independent samples met inclusion criteria, consisting of 50,659 participants. The pooled prevalence was 3.2 % (95 % confidence intervals: 1.7 %–5.9 %) for confirmed and 7.7 % (95 % confidence intervals: 5.3 %–11.0 %) for provisional diagnosis. There was high heterogeneity across all studies (I² = 99 %). Sources of heterogeneity identified by meta-regression were continent of sample (p < 0.0001), type of sample (community-based, university, high school) (p = 0.007), risk of bias (p = 0.009), and method of diagnosis (p = 0.017). Restricting the analysis to community-based samples using confirmed diagnosis resulted in a prevalence of 1.6 % (95 % confidence intervals: 1.0 %–2.5 %), with low heterogeneity (I² = 26 %).

A small number of included studies used full DSM criteria in community settings.

The point prevalence of premenstrual dysphoric disorder using confirmed diagnosis is lower compared with provisional diagnosis. Studies relying on provisional diagnosis are likely to produce artificially high prevalence rates.