Why are there no randomised trials or prospective cohort studies of antiperspirant use and breast cancer, only retrospective case-control studies?

The short version: nobody can randomize people to use or avoid antiperspirant for decades, so the honest evidence is observational, and the one good case-control study found nothing.

This is a question about study design, not about whether antiperspirant causes cancer. A randomized trial would mean assigning thousands of healthy women to use or avoid antiperspirant and following them for decades until enough breast cancers accrued to compare. That is expensive, slow, and ethically awkward, because you would be asking people to change a daily hygiene habit for years on a hypothesis with no strong prior. Prospective cohorts have the same problem in slower form: you need huge numbers and long follow-up to catch a disease that is uncommon per year. Case-control studies are cheap and fast by comparison, which is why they dominate.

The one solid study here is a population-based case-control from Western Washington, 813 cases and 793 controls, run from 1992 to 1995 [1]. It found no increased risk from ever using antiperspirant (odds ratio 0.9), no signal from applying it within an hour of shaving (0.9), and no interaction with blade razors [1]. The deodorant figures were similarly null (1.2) [1]. A weaker study of 437 survivors reported that heavier underarm product use tracked with earlier diagnosis, but it had no control group and cannot separate cause from coincidence [2].

The retrieval surfaced nothing on funding, and nothing on a prospective cohort or trial. That absence is a fact about this search, not proof none exist. What would settle it: a large prospective cohort that records antiperspirant use before diagnosis and follows women for decades.

My call: the case-control evidence points away from a link, but the design cannot prove absence of a small long-term risk. Confidence: moderate.