Studies
Why three Depo-Provera meningioma risk numbers differ
BMJ says 5.55 times. JAMA Neurology says 2.43 times. A Danish study says 4.55 times. Here's why, and what the numbers actually mean for you.
In short: three peer-reviewed studies put the depo shot’s meningioma risk between 2.43 and 5.55 times that of non-users. The numbers differ because the studies measured different populations and used different designs, not because any of them is wrong.
Why do BMJ, JAMA Neurology, and a Danish study report different numbers?
They’re answering slightly different questions, in different countries, using different study designs. BMJ reported women who used injectable medroxyprogesterone, the depo shot, had about 5.55 times the odds of a meningioma diagnosis. A JAMA Neurology study on the depo shot specifically found 2.43 times the risk. A newer Danish government study, part of a 12-formulation review of birth control methods, found an odds ratio of 4.55.
None of these numbers is wrong. Here’s what each one actually measured, side by side.
Three primary studies report meningioma risk figures in the same 2x to 6x range, not three conflicting results.
Point estimate for each study's headline figure. Whiskers show the reported 95% confidence interval where one was published.
| Study | Design | Figure | Source |
|---|---|---|---|
| BMJ 2024 (France) | National case-control | OR 5.55 (2.27-13.56) | BMJ 2024;384:e078078 |
| Danish Medicines Agency 2026 (Denmark) | Nested case-control, national registry | OR 4.55 (2.19-9.45) | JAMA Network Open 2026;9(7):e2622603 |
| JAMA Neurology 2025 (United States) | Matched cohort | RR 2.43 (1.77-3.33) | JAMA Neurology 2025, DOI 10.1001/jamaneurol.2025.3011 |
Why the numbers move: design, population, and what counted as “meningioma”
Study design. BMJ and the Danish study are both case-control studies. Researchers started with women already diagnosed with meningioma, then looked backward at their contraceptive history.
The JAMA Neurology study is a matched cohort study instead. Researchers started with a large group of medroxyprogesterone users, matched them to non-users with similar health profiles, and followed both groups forward in time. Case-control and cohort designs tend to produce different-sized numbers even when they’re picking up the same real signal. That alone explains a meaningful share of the gap here.
Population. Each study drew from a different country and a different group of women. BMJ’s case-control study drew on French national health records.
The Danish study matched 1,473 women diagnosed with meningioma to 14,717 women who weren’t, all from Denmark’s national health registers. JAMA Neurology followed a cohort of 10.4 million US women in the TriNetX database. It measured the 2.43 figure specifically inside a matched subgroup of 88,667 women who used injectable medroxyprogesterone.
What counted as meningioma. The Danish study’s case definition includes spinal meningioma, not only meningioma in the skull. That’s a slightly broader outcome than some other studies measure, and it can move the number too.
None of the three figures is “the real one.” All three are real, peer reviewed findings, measuring a related but not identical question. Reading them side by side, without knowing that, is what makes the numbers look like they disagree.
The Danish study, in full
The Danish figure deserves its own detail, because it’s the newest of the three and carries specific caveats that belong in the same breath as the number itself.
The Danish Medicines Agency ran this as a nested case-control study inside Denmark’s national health registers, not a cohort study. For injectable medroxyprogesterone, the odds ratio was 4.55, with a 95% confidence interval of 2.19 to 9.45.
That wide interval is a signal on its own. It means the number of women who were both exposed and diagnosed is small, so the point estimate carries more uncertainty than a tighter interval would. The agency tested 12 progestogen formulations in total and found a statistically significant association for 8 of them. Injectable medroxyprogesterone had the strongest signal of the 12, well above the next-highest formulation.
A few things this study does not show. The risk isn’t a flat, lifetime multiplier. It concentrates in current or recent use and, for most formulations tested, faded within about five years of stopping.
The study also hasn’t led to any regulatory action yet. The Danish Medicines Agency has referred the finding to the European Medicines Agency’s safety committee for review, and that committee hasn’t ruled. That’s a separate, earlier step than the EMA’s own 2024 conclusion on this drug class, and separate again from the FDA’s already-completed December 12, 2025 US label update. All three are real, but they’re three different events on three different timelines, not one. See the full regulatory timeline for how those events actually line up.
The baseline number that keeps the risk in proportion
A relative-risk figure only means something next to how common meningioma is in the first place. In the United States, the CBTRUS Statistical Report puts overall meningioma incidence at 10.15 per 100,000 people a year. Meningioma is diagnosed in women at roughly 13.90 per 100,000, against 6.02 per 100,000 in men.
Denmark’s own baseline, cited in the Danish study itself, runs a bit higher: 12.6 per 100,000 overall, 17.9 per 100,000 in women. Those two national baselines aren’t the same number, and we aren’t blending them into one. They simply confirm that meningioma is an uncommon diagnosis in both countries, well before any drug exposure is considered.
The JAMA Neurology study also reports a number called number needed to harm, or NNH, for the medroxyprogesterone subgroup: 1,152. In plain terms, researchers would need to follow roughly 1,152 women on this exposure to see one additional meningioma diagnosis beyond what would be expected anyway. That’s a real number, and it doesn’t describe a common event. We’re reporting it because it belongs next to the multiplier, not because it’s a reason to raise an alarm.
What none of this changes
Meningioma is a usually non-cancerous brain tumor, and none of these three studies changes that. Non-cancerous doesn’t mean minor. A meningioma diagnosis can still mean surgery, monitoring, and lasting neurological change.
None of these studies has been tested in court. The Depo-Provera litigation’s own hearing on this science, a Daubert hearing under Federal Rule of Evidence 702, hasn’t happened yet. It’s currently scheduled for September 18, 2026.
Until that hearing and any ruling that follows it, no court has decided whether this evidence establishes that Depo-Provera caused any individual woman’s meningioma. These three studies describe an association measured across populations. They aren’t a verdict on any one person’s case.
Find out where you stand
You don’t need to read a study to know if this applies to you. Two questions tell you where you stand, no name, phone, or email required yet.