A new study protocol proposes to study the effect of progesterone in feminising hormone replacement therapy, a question subject to persistent community interest for decades. This protocol suffers from many of the same limitations as past work on the topic and will still not answer the questions that have interested the community for decades. Once again, solving these problems requires the inclusion of trans community science perspectives in HRT research.
Last year, we published a commentary (Ari et al., 2025) on the first clinical trial on progesterone use and breast growth in transfeminine individuals (Dijkman et al., 2023; Dreijerink et al., 2025). This trial, which was presented at an EPATH conference, suffered from methodological issues that prevented us from drawing any conclusions on the subject it purported to examine. We would’ve liked to comment on the results further but, sadly, the study remains unpublished a year later.
Very recently, a new protocol for a randomised controlled trial (RCT) on the broader topic of progesterone and feminising hormone replacement therapy (fHRT) has been published (Cromer et al., 2026). While it is exciting to see academia finally studying progesterone use in transfeminine people, these developments are too little too late, as the fHRT community has been interested in answers to how progesterone affects breast growth, mental state, and other outcomes for over two decades outside the realm of academic research (Aly, 2020; danabevan, 2014; Prometrium, 2005).
We find this new protocol disappointing, as its methodology – particularly in its capacity to assess the effect of progesterone on breast development – isn’t great, and will probably remain so as long as researchers refuse to engage with the community they’re trying to research.
Below are a few methodological issues we have identified in the protocol, including both those that repeat the mistakes of past studies such as Dreijerink et al. (Dreijerink et al., 2025) and some new ones.
The protocol mentions three past studies of progesterone’s effect on breast growth in fHRT, citations [7,17,19], referring to a past review from the same group (Patel et al., 2022).
[7] is “Effects of low-dose oral micronised progesterone on sleep, psychological distress, and breast development in transgender individuals undergoing feminising hormone therapy: a prospective controlled study” (Nolan et al., 2022), which found no significant effect of progesterone on breast development or psychological measures, but suffered from some limitations in study design.
[17] is “Medroxyprogesterone Acetate in Gender-Affirming Therapy for Transwomen: Results From a Retrospective Study” (Jain et al., 2019), about which it is said:
However, these findings are limited by the absence of randomization and potential selection bias.
While this is true, that article does not even even study progesterone, but only the progestin medroxyprogesterone acetate.
[19] is the aforementioned study (Dijkman et al., 2023; Dreijerink et al., 2025) and is described as:
although it lacked blinding and placebo control and has not yet reported results[…]
The authors make the reasonable conclusion:
Collectively, existing studies highlight the paucity of rigorous evidence and underscore the need for controlled trials to clarify the clinical role of progesterone in gender-affirming hormone therapy.
While we share the critiques of the earlier breast growth studies (Ari et al., 2025), the proposed protocol does not escape many of the same limitations that past work had, while introducing some new ones, as we argue in the following sections.
Progesterone in feminising HRT is typically administered either orally or rectally. A recent survey showed that 80% of those using progesterone did so orally, while 15% did so rectally (Chang et al., 2025). Rectal use is common, as are questions about its relevance and usefulness compared to oral use, in part because of the higher bioavailability and different metabolism of non-oral progesterone. Indeed, the absorption profile of micronised progesterone, and the distribution of metabolites, has dramatic differences between routes of administration (rectal/vaginal, intramuscular, or oral) (Choavaratana & Manoch, 2004; Norman et al., 1991; Simon et al., 1993), and the total exposure with oral administration is highly dependent on whether a dose is taken with food (Qin et al., 2022; Simon et al., 1993).
Yet, the protocol ignores this and does not even mention rectal use: there is no rationale for why only oral use was chosen, why the dose is to be taken before bedtime (and not, for instance, with the last meal of the day), and no considerations about adherence to oral use. This is not unlike the aforementioned progesterone trial (Dreijerink et al., 2025), in which we can only assume the authors didn’t bother engaging with the community. Had they done so, they probably would have known some individuals enrolled in the trial would take progesterone rectally – and that is what some individuals elected to do sua sponte, at least in some cases because enrolling in the study was their only means of accessing progesterone by official routes (see footnote 2 in our earlier article (Ari et al., 2025)).
Considering the difference between oral and non-oral progesterone, this has the potential of being a major source of error, which the study should account for. Future studies could do so, for instance, by allowing individuals enrolled in a trial to choose wether they’d rather take progesterone (or placebo) orally or rectally, thus allowing for subgroup analysis. Another option would be to provide the benefits of being enrolled in the trial (like accessing progesterone) to all eligible patients requiring it, but only performing analysis on those willing to use progesterone by the oral route.
The appearance of breast growth, as measured by breast hemicircumference in the protocol, may be due to fluid retention in breast tissue caused by the action of mineralocorticoid-like metabolites of progesterone (Dillard et al., 2025; Taraborrelli, 2015), rather than to fibrotic changes and glandular development. The study does not account for this effect, which would revert with cessation of progesterone. The final measurement is taken immediately after the sixth month of the study, not after a washout period during which progesterone is not administered. This makes the study unable to isolate the transient water retention effect from permanent changes to breast tissue.
The sample size of n=34 (17 per group) is justified by a past study (Ginger et al., 2024) on the basis of variation in psychological well-being measurements. However, this sample is likely to be too small to yield conclusive results on other metrics, such as breast growth – which is the main effect for which people on fHRT take or have taken progesterone (Chang et al., 2025). Note that in the aforementioned breast growth study, a sample size of 15 individuals in 6 groups yielded very wide confidence intervals (Dreijerink et al., 2025) and thus inconclusive (preliminarily reported) results.
This is the only limitation of the protocol among those we discuss here that is acknowledged by the authors (the others being the short study duration, problems with psychological assessments and sleep measures, findings not being applicable to individuals not meeting the inclusion criteria, and inter-individual differences in HRT regimens).
The study purports to be “designed to assess […] [the] potential clinical role of progesterone in transfeminine care” (our emphasis). An exploratory trial with a small sample size would be understandable if we were in the early 2000s and this were the first time someone had ever thought about studying the potential relevance of progesterone in fHRT. However, this is far from the case: we do not need data on whether progesterone should be studied further, we need the further studies and data on what progesterone actually does. Going through the trouble of setting up a RCT only for it to be about what the community has already known for years is yet another example of epistemic injustice in trans science (Eric Llaveria Caselles, 2021).
The study will use breast hemicircumference (taken by wrapping a tape measure around the chest from the sternum to the spine at the widest point of the chest) as the only outcome measure to assess breast growth. The study motivates this choice by its superiority to Tanner staging as done by the earlier Nolan et al. (Nolan et al., 2022):
This builds upon the study by Nolan et al. which utilized a lower dose of micronized progesterone (100 mg vs 200 mg in our study), had a shorter follow-up period (3 months vs. 6 months in our study). In addition, their study assessed breast development via Tanner staging, which can be subjective, whereas our study uses a more objective breast hemi-circumference measurement.
However, the study could do better: chest circumference is likely correlated with breast volume and the development of breast structures, but it is not acknowledged as an optimal measure of breast development. In fact, other studies prefer breast volume as a measure, with circumference used as a simpler proxy (Blok et al., 2021):
Furthermore, in this study, breast development measured with 3D images appears to be a better reflection of hormone treatment induced breast development compared to breast-chest differences measured with a tape measure […]
As shown in this study, the curve of breast development measured with a tape measure flattened prematurely compared to the breast volume increase measured with 3D imaging. This is in line with the study of Kovacs et al., which showed that breast volume measurements using 3D imaging agreed better to volume calculations using MRI than anthropomorphic measurements did.[28] […]
Although this observation may lead to a preference of using 3D imaging over breast-chest circumference measurement, a tape measure could be a good alternative for centers with fewer resources.
Fortunately, 3D scanning methods have recently been studied (Dijkman et al., 2024) by the same Amsterdam group that studied progesterone and breast growth (Blok et al., 2021; Dijkman et al., 2023; Dreijerink et al., 2025), with some scanning methods being relatively inexpensive (able to run on a modern smartphone) at a modest accuracy loss relative to specialised 3D scanners. The protocol thus falls short of that study in the measurement of that covariate.
The choices of outcome measures in the protocol are puzzling to us:
The primary outcome is change in psychological distress at 6 months. Secondary outcomes include changes in sleep quality, gender congruence/life satisfaction, breast hemi-circumference, blood pressure, lipid measures, estradiol, creatinine, alanine aminotransferase, and aspartate aminotransferase.
We can expect psychological distress to be correlated with, or even causally affected by, the other covariates studied, including gender congruence, life satisfaction, and indeed breast development. Moreover, while some of the effects of progesterone and its metabolites in the brain are understood (Guennoun, 2020), psychological assessments are a far more derived measure of these effects than, for instance, breast measurements are of progesterone’s effects in breast tissue.
It is thus unclear why psychological distress is chosen as the primary measure when we have more pressing mysteries to solve, noting again that breast development is the main reason for which individuals on fHRT take progesterone (Chang et al., 2025). Even so, as we have written here, the protocol is not great as a study of breast development. Perhaps the focus on psychological tests is to excavate a niche for the new study distinct from others that focus on breast growth.
Furthermore, no justification is given for the inclusion of liver enzyme and lipid measures, nor for the inclusion of estradiol in the target measures while testosterone is excluded (even though it is well known that most progestogens – including progesterone (Aufrère & Benson, 1976; Sundsfjord et al., 1971; Tollan et al., 1993) – have some level of antigonadotropic effects and that testosterone affects breast development (Meyer et al., 2020), among other target measures). It seems strange to us that serum estradiol is measured, but blood levels of testosterone – and of progesterone itself! – are not.
Overall, some specific data that is the object of frequent speculation, like breast volume, gonadotropin suppression, or even libido or anxiety (to stay in the realm of what can be assessed by questionnaires), is absent from the outcome measures, while very vague and overarching ones (like “life satisfaction”) are present. Likewise, in the risk-related outcomes, irrelevant measures like liver enzymes and creatinine are present, but not actual areas of concern, like insulin resistance.
Serum estradiol levels in a range of 100-300 pg/mL are used as an inclusion criterion. This is not in line with past studies:
However, we welcome this change because filtering for a sufficiently high estradiol level is a reasonable prerequisite to studying fHRT effects in general.
As discussed above, progesterone has antigonadotropic effects and can thus suppress gonadal testosterone production; however, testosterone ranges do not figure in the inclusion criteria. Including only those with testerone suppressed before initiation of progesterone could be more informative.
Finally, the protocol excludes participants who have been on GAHT for under 6 months. This is in line with the criteria in Nolan et al. (2022) (although the mean duration of treatment for participants was over two years); Dijkman et al. (2023) protocol used 12 months. The reasons for this choice are unclear, as breast development is far from complete after 6 months of estradiol treatment. Including only those on HRT for several years could be more informative, especially in light of concerns related to early initiation of progesterone (Aly, 2020).
Knowledge on the effects of progesterone has been accumulated by the fHRT community – notably the DIY HRT community – over decades. This knowledge continues to be expanded and distributed outside the realm of academic publication outlets while remaining backed and informed by the latest work published in mainstream venues (alix, 2025; Aly, 2020; danabevan, 2014). Conversely, academic researchers, despite their increased interest in trans HRT, have scarcely engaged with us. Their refusal to consider our perspectives, experiences, and scientific knowledge (or ignorance of even the possibility of it) has a concrete negative impact on the quality of their research and on its usefulness to the groups they are studying. We can only recommend researchers in the future engage in community-driven research (CDR); some reading can be found elsewhere on this website.