The research team, led by Enrico Altiero Giusto and Simone Patergnani of the University of Ferrara together with colleagues from the Italian National Institute of Health, the University of Catania, the San Raffaele Scientific Institute in Milan, Kansai Medical University in Osaka, and AULSS 9 Scaligera in Verona, conducted the analysis according to the PRISMA reporting guidelines and registered the protocol in PROSPERO. The primary outcome was the incidence of primary breast cancer confirmed through cancer registries, a choice that anchors the findings to objectively verified diagnoses rather than self-reported disease. The investigators searched PubMed, Embase, and SCOPUS from database inception through September 2025, capturing observational cohort and case-control studies as well as randomised controlled trials. Two reviewers independently screened the records, with disagreements resolved by a third author, and only studies meeting rigorous criteria for exposure definition and data quality were included. Eligible studies had to draw on validated sources such as health registries, electronic prescription databases, national cancer registries, structured interviews conducted by qualified health professionals, or medical records, while investigations with unobjective exposure measures or inadequate definitions were excluded.

The scale of the pooled dataset is what sets this work apart. Across the twenty-three included studies, the researchers analysed 184 distinct antidepressant-related variables, spanning drug classes, exposure windows, cumulative durations, dosages, and prescription counts. This granular approach allowed the team to move beyond the blunt question of whether antidepressants are associated with breast cancer and instead ask a more clinically meaningful one: when, and in whom, might any association emerge? The statistical framework employed restricted maximum likelihood estimation, and study quality was assessed with the Newcastle-Ottawa Scale, providing a structured appraisal of the observational evidence base that dominates this field.

The headline findings are striking precisely because they are not uniform. Women who had used selective serotonin reuptake inhibitors, or SSRIs, prior to study baseline for at least one year showed a reduced incidence of breast cancer, with an odds ratio of 0.79 and a 95 percent confidence interval of 0.67 to 0.94. In contrast, cumulative SSRI exposure of zero to one year was associated with an increased risk, with an odds ratio of 1.08 and a confidence interval of 1.04 to 1.12. In other words, the direction of the association appears to flip depending on the exposure window: short-term SSRI use correlated with a modest elevation in risk, whereas prior, sustained use appeared protective. This temporal heterogeneity may explain much of the confusion that has characterised the literature to date, since studies that lumped all exposure durations together would have mixed these opposing signals into a single, potentially misleading estimate.

The biological plausibility of an antidepressant-breast cancer link has long been debated. SSRIs act on the serotonin transporter, and serotonergic signalling has been implicated in pathways relevant to cell proliferation and survival. Several antidepressants are also metabolised by the cytochrome P450 2D6 enzyme, which participates in the activation of the anti-cancer drug tamoxifen, raising questions about pharmacological interactions in patients with existing disease. Beyond neurotransmitter systems, the authors’ broader research programme has explored how intracellular calcium signalling and mitochondrial function intersect with cancer biology, and the hypothalamic-pituitary-adrenal axis, which is dysregulated in chronic stress and depression, influences glucocorticoid signalling that can affect tumour microenvironments. The meta-analysis itself does not establish mechanism, but its authors explicitly call for further research to clarify the underlying biological pathways that could account for the time-dependent associations they observed.

It is essential to interpret the effect sizes carefully. An odds ratio of 1.08 for short-term SSRI use represents a relative increase of roughly eight percent, which translates into a very small absolute risk change for an individual woman, given that breast cancer is a common disease but its baseline incidence per year remains low. Likewise, the 21 percent relative reduction associated with prior SSRI use of at least one year should not be read as evidence that antidepressants prevent breast cancer. Observational studies, however well designed, are vulnerable to confounding by indication, reverse causation, and healthy-user effects. Women who discontinue antidepressants or who have long completed treatment may differ systematically from those initiating therapy, and depression itself is associated with lifestyle factors, healthcare utilisation patterns, and screening frequency that can bias cancer detection. The meta-analysis cannot fully eliminate these limitations, and the authors are careful to frame their conclusions as informing risk assessment rather than establishing causation.

Nevertheless, the clinical implications are significant. The authors argue that their findings could support a personalised approach to antidepressant prescribing, particularly for long-term therapy and for patients who already carry elevated breast cancer risk factors, such as family history, genetic predisposition, or prior proliferative breast disease. Risk-benefit assessment in clinical practice, they suggest, should weigh both the established efficacy of antidepressants in treating depression and anxiety, which are serious and sometimes life-threatening conditions in their own right, against any potential influence on breast carcinogenesis. For the vast majority of patients, the mental health benefits of appropriate antidepressant treatment will continue to outweigh any uncertain oncological considerations, but the new data give clinicians a more nuanced evidence base for shared decision-making, especially when choosing among drug classes or planning the duration of maintenance therapy.

The methodological rigour of the meta-analysis deserves attention as well. By requiring validated data sources and objective exposure definitions, the team filtered out weaker studies that might have injected noise into the pooled estimates. The inclusion of both cohort and case-control designs allowed cross-checking of results across methodological traditions, and the analysis of 184 exposure-related variables represents one of the most comprehensive variable-level syntheses attempted in this field. The supplementary material, which includes the full search strategy, statistical methods, Newcastle-Ottawa assessments, and an extensive set of figures, provides the transparency needed for independent scrutiny. The work was funded through the Italian Department of Excellence programme at the University of Ferrara, with additional support to individual authors from the European Research Council, the Italian Association for Cancer Research, the Italian Ministry of Health, and other sources, and the authors declare no competing interests.

What makes this study resonate beyond specialist circles is the sheer ubiquity of its subject. Depression affects hundreds of millions of people worldwide, and breast cancer is the most commonly diagnosed cancer in women globally. Any medication taken by such a large fraction of the population will inevitably be scrutinised for long-term safety signals, and antidepressants have faced recurring questions about associations with outcomes ranging from bleeding risk to bone density. This meta-analysis does not close the book on the antidepressant-breast cancer question, but it reframes it in a way that is far more useful: the relationship is not a simple yes or no, but a function of exposure timing, cumulative duration, and patient context. As antidepressant prescribing continues to rise, the study’s message to researchers is to pursue the biological mechanisms behind these time-dependent patterns, and its message to clinicians is that treatment planning for women requiring antidepressants can now be informed by one of the largest evidence syntheses ever assembled on the topic, supporting evidence-based and individualised decisions at the intersection of mental health and cancer prevention.

Subject of Research: The association between antidepressant use, particularly SSRIs, and breast cancer risk across different exposure windows and cumulative treatment durations

Article Title: Antidepressants and breast cancer risk: a meta-analysis of exposure windows and cumulative use in nearly two million women

Article References: Giusto, E. A., Patergnani, S., Cutillo, M., Oteri, V., Guido, G., Giorgi, C., Pinton, P., & Fiorica, F. (2026). Antidepressants and breast cancer risk: a meta-analysis of exposure windows and cumulative use in nearly two million women. BMC Medicine. https://doi.org/10.1186/s12916-026-05169-0

Image Credits: AI Generated

DOI: 10.1186/s12916-026-05169-0

Keywords: antidepressants, SSRIs, breast cancer, meta-analysis, cancer epidemiology, psychopharmacology, drug safety, exposure windows, risk factors, women's health, BMC Medicine, pharmacoepidemiology