Tamoxifen and Cancer Risk

Tamoxifen is an unusual entry. The National Toxicology Program lists it as known to be a human carcinogen. It has been on that list since 2000, in the Ninth Report on Carcinogens. Its CAS number is 10540-29-1.

The same NTP profile says something else in the same breath. Conclusive evidence shows that tamoxifen lowers the risk of breast cancer in the other breast. That applies to women already diagnosed. It may also prevent or delay breast cancer in women at higher risk.

Both statements are true. This page is about holding them together.

Tamoxifen is a selective estrogen receptor modulator, often shortened to SERM. That name explains the whole problem.

It blocks estrogen in breast tissue. That is where the benefit comes from. The lining of the uterus is called the endometrium. There it does not block estrogen. It stimulates.

So one drug, two tissues, two opposite effects.

NCI's clinical summary on breast cancer prevention gives the size of the effect.

In high-risk women, tamoxifen cut two things by about 30% to 50%. Those were estrogen receptor-positive breast cancer and ductal carcinoma in situ. The course was 5 years of treatment.

The effect lasts well past the treatment. The drop in estrogen receptor-positive invasive breast cancer held for at least 16 years after starting. That is 11 years after the drug was stopped.

One line of that summary is often skipped. Breast cancer mortality was not affected. Fewer breast cancers occurred. Deaths from breast cancer did not fall in these trials.

NCI's summary reports a meta-analysis relative risk of 2.4 for endometrial cancer, with a 95% confidence interval of 1.5 to 4.0.

A second meta-analysis gives a useful comparison. The hazard ratio for endometrial cancer was 2.18 for tamoxifen. For raloxifene, another SERM, it was 1.09. Raloxifene does not raise endometrial cancer risk. It does raise clot risk.

Age changes the picture. NCI states that harms were significantly higher in women over 50 than in younger women.

The evidence behind the listing is deep. By the mid-1990s the effect had shown up in several places. That included one adequate cohort study, four adequate case-control studies, and fourteen randomized trials. The cohort study covered 87,323 women with breast cancer in NCI's SEER program. It found a significant excess of endometrial cancer among those given tamoxifen. The two largest randomized trials found a strong, significant link. Twelve smaller trials each fell short of significance. Pooled together, they showed 29 endometrial cancers in treated women against 14 in controls.

Endometrial cancer is not the only one, and the others are easier to forget.

NCI lists three categories, based on solid evidence:

• Endometrial cancer.
• Blood clots in vessels. These include pulmonary embolism, stroke, and deep vein thrombosis. The meta-analysis relative risk was 1.9, with a confidence interval of 1.4 to 2.6. A separate hazard ratio came to 1.73.
• Cataracts.

Now the detail that matters for planning. The endometrial cancer risk lasts 5 years after the drug is stopped. The clot and cataract risks do not.

NCI's clinical summary on endometrial cancer is direct about monitoring. Patients on tamoxifen who have abnormal uterine bleeding need a follow-up exam. They also need a biopsy of the endometrial lining.

The FDA has released a boxed warning on this. It includes data on the rise in uterine cancers linked to tamoxifen. A boxed warning is the strongest label warning the FDA issues.

Neither NCI nor the FDA publishes a routine screening interval here. There is no set ultrasound or biopsy schedule for women on tamoxifen with no symptoms. In the published guidance, the trigger is bleeding, not the calendar.

NTP also records that reviewers looked hard for other explanations. One review concluded the published results suggested a causal link but were not conclusive, pointing to confounders such as prior hysterectomy or hormone replacement therapy. IARC examined the same confounders and judged them unlikely to have had a major effect on the reported relative risks.

Two other options appear in NCI's prevention summary, and each shifts the balance differently.

Raloxifene raises clot risk but not endometrial cancer risk. Take a woman with a uterus who worries most about endometrial cancer. For her, that difference is the whole point.

Aromatase inhibitors work differently. They apply to women after menopause. Three years of exemestane cut breast cancer incidence by 65% against controls. Five years of anastrozole cut it by 53%, and that effect was still there at 11 years. One exemestane trial gives the absolute figure. It avoided 21 cancers among 2,280 participants over 35 months. That is a number needed to treat of about 100. Side effects were mainly hot flashes, up 8% in absolute terms, and fatigue, up 2%.

That number needed to treat is the kind of figure worth asking for on any preventive drug.

A listing on the Report on Carcinogens answers one question: can this substance cause cancer in people? For tamoxifen the answer is yes, for cancer of the uterine lining.

The listing does not weigh that against the breast cancers prevented. That weighing happens in the clinic. It uses one person's breast cancer risk, her age, whether she still has a uterus, and her clot history.

For more on how these categories work, see our explainers on what a carcinogen is and hazard versus risk. Our page on uterine cancer covers the cancer at the other end of this trade-off.

• National Toxicology Program, Report on Carcinogens, Fifteenth Edition: Tamoxifen, accessed August 6, 2026
• National Cancer Institute, Breast Cancer Prevention (PDQ) - Health Professional Version, accessed August 6, 2026
• National Cancer Institute, Endometrial Cancer Treatment (PDQ) - Health Professional Version, accessed August 6, 2026