While not all of our patients who were born female will identify as female, in this topic, the issue is specifically a product for women vs men based on physiology and how that's treated in the U.S., so this particular article will use "women" where appropriate, but we mean all people under our care in gynecology.
The reason for this difficulty is cited is a lack of long-term safety data, particularly regarding cardiovascular disease and breast cancer risks (Obstetrics and Gynecology. 2025. Kling JM.)
In 2004, the FDA advisory committee acknowledged that a testosterone patch was effective in treating hypoactive sexual desire disorder (HSDD) in women, but voted against approval, citing insufficient documentation of long-term safety data (The Lancet 2015).
The BLISS trial, a cardiovascular safety study, was designed to meet this FDA requirement, intended to show 5 years of safety data, however, as efficacy trials failed to show benefit over placebo, the program was ultimately discontinued (American Heart Journal 2012).
Most RCTs have been limited to <6 months of treatment, but women with CV risks were systematically excluded from study populations, making it impossible to generalize safety to the broader population (Clinical Endocrinology 2019).
In stark contrast, sildenafil (trade name Viagra) was approved for men with erectile dysfunction by the FDA without requiring long-term safety data (FDA 2024).
It is well documented that women have been excluded and underrepresented in clinical trials and this is a problem with direct consequences for the safety and efficacy of medical treatments provided to women. The basis of this exclusion was based on assumptions that male and female bodies functioned similarly apart from reproduction (JAMA 2025). The FDA excluded women of childbearing age as "protection-by-exclusion" (Journal of the American College of Cardiology 2024), thus excluding very large numbers of women.
Despite the absence of an approved product for women, an estimated >2 million testosterone prescriptions are written annually for women in the US, many for compounded preparations that lack pharmacokinetic profiling and quality assurance (NEJM 2024).
Notably, a transdermal 1% testosterone cream has been approved in Australia, demonstrating that regulatory approval is achievable in other areas of the world (NEJM 2024).
The cost and complexity of a multi-year CV/breast cancer outcomes trial is prohibitive for a product that would generate relatively modest revenue (NEJM 2024), which means no one has designed a new study (American Heart Journal 2012) and this issue has remained stagnant even now, in 2026.
The primary available testosterone in the US is Androgel, which is formulated for men but has been safety tested, so the preferred approach is a fractionated dose of an FDA-approved male transdermal testosterone gel, targeting approximately 300 mcg/day (about one-tenth of the standard male dose), with the goal of achieving serum total testosterone levels within the premenopausal physiologic range (NEJM 2024).
With Androgel, dosing for women is challenging. In practice, we have recommended that our patients get the pump or the gel packets and place the gel into a syringe to give themselves ½ cc per day (for example) to get the "right" dose, however, this is messy, challenging and pharmacies tend to balk at this kind of SIG on a prescription. Androgel is also somewhat concentrated, so it wouldn't be difficult to accidentally get too much and achieve supraphysiologic doses.
Well, why not just open the pump and draw out the amount into the syringe? Because this is a controlled substance, the pumps don't actually open. One pump per day is often the recommendation for men, but could you imagine trying to gauge a 10th of a pump? If we're worried about the harms of testosterone, the FDA should be more committed to getting a medication approved that is less risky for women to get too much.
Join our email list for early access to new courses, monthly clinical letters from our founders, and enrollment notifications. No noise, we only send when we have something to say.