Stopping Is Not Neutral: Why Withdrawing Hormone Therapy Asks the Same Rigour as Starting It — Anatomy of Good Care (1/7)

The question we forget to ask

When a patient starts hormone therapy, we are thorough. We take a history, run baselines, talk through risks and benefits, document consent, and set a monitoring schedule. We treat initiation as what it is — a clinical decision with consequences that deserves care.

When the same therapy stops, that scrutiny often evaporates. A prescription lapses, a service is disrupted, a supply runs out, and the stopping passes without anything like the examination we gave the starting. The unspoken assumption is that withdrawal is simply the undoing of a treatment — a step back to a neutral baseline.

For a patient established on estradiol or testosterone, there is often no neutral baseline to step back to — and after gonadectomy, none at all. Starting and stopping are both interventions. We just only treat one of them that way.

What the body does when estradiol is withdrawn

The most familiar consequence of estrogen withdrawal is vasomotor: hot flushes and night sweats. The leading mechanistic account locates this in the hypothalamus, where a population of neurons (the KNDy neurons) helps regulate the body’s thermoneutral zone. Estrogen normally restrains them; when estrogen falls, they are disinhibited and the thermoneutral zone narrows, so ordinary shifts in temperature start triggering a flush response [1]. Notably, the better evidence suggests it is the downward swing in estradiol, rather than a low level in itself, that precipitates symptoms — which is precisely why withdrawal, the act of stopping, is the trigger [2]. (Mechanism characterised chiefly in cis menopausal cohorts and the animal/neuroendocrine literature; used here as the prevailing model for estrogen-withdrawal physiology, not as trans-specific outcome evidence.)

This is not uniform. Some people withdraw from estrogen with little vasomotor disturbance — in one controlled study, postmenopausal women with no prior history of flushes did not reliably develop them after abruptly stopping short-term estradiol [3]. Individual response varies, and we should say so. The point is not that withdrawal always produces a crisis; it is that withdrawal predictably does something, and that something is unpredictable enough to warrant attention rather than assumption.

The higher-stakes consequence is skeletal, and here the asymmetry is starkest. Estrogen is a brake on bone resorption. The loss of estrogen at menopause increases osteoclast activity and lowers bone mineral density [4]. (General/cis bone physiology.) And when menopausal hormone therapy is discontinued, bone loss can be clinically meaningful and relatively rapid: concentrated in the first two years, at around 1.5 to 2% per year, similar to the loss in the first two years after menopause, according to an earlier study cited by the review [5]. The same review notes that fracture risk does not seem to rise in the short term after stopping, and that an analysis of the Women’s Health Initiative trial found no increase in fractures after hormone therapy was stopped [5]. (Again, cis menopausal women — but the physiology of estrogen withdrawal on bone is the relevant read-across.)

For a trans woman who has had an orchidectomy, this stops being an analogy. After gonadectomy there is no endogenous gonadal estrogen to fall back on. Exogenous estradiol is the supply. Stopping it does not return her to some prior hormonal state — it leaves her in an unprotected, low-sex-steroid state. This is why specialist guidance names underutilisation of hormones after gonadectomy as a recognised osteoporosis risk factor, and advises that anyone who has had a gonadectomy and at least five years without hormone replacement should be considered for bone-density testing, regardless of age, while noting that there is insufficient evidence to guide bone-density testing in trans people generally [6]. The trans-specific data are consistent with this: in one cross-sectional study, trans women at least two years after gender-confirming surgery showed a high prevalence of low bone mass, statistically tied to low estradiol levels and poor adherence to replacement [7]. In a nationwide Dutch cohort, fractures in trans women aged 50 and over were more common than in age-matched reference men, even among women on long-term hormone therapy [8]. The authors attribute that excess to low bone density before treatment began rather than to the hormones; their measure was fractures seen in emergency departments over three years, not risk from the start of treatment; and in trans women they found no association between estradiol levels and fracture [8]. Our reading, which is an inference and not the authors’ finding, is that hormone therapy does not erase a pre-existing skeletal vulnerability; the data do not show what withdrawal would add to it. It is therefore not safe to assume that withdrawal would be benign.

The same guidelines are also plain about where the evidence runs out. Their chapter on feminising hormone therapy states that there is no evidence to support either continuing or stopping hormones in older trans women, and that, since the average age of menopause is around 49, it is reasonable for trans women who have had a gonadectomy to consider stopping hormone therapy around the age of 50 [16]. That is a finding against any claim that stopping must always be resisted, and we give it its full weight. It describes a considered decision, taken with the patient at a point the guidance names, which is the kind of stopping this piece argues for. It does not describe a lapse that no one decided on.

The same principle applies to testosterone

The same logic runs through testosterone, and it is sometimes missed because we picture testosterone as the “active” hormone and forget what it is protecting against. A trans man who has had bilateral oophorectomy has no ovarian source of sex steroids; in transmasculine people on testosterone, oophorectomy has been associated with significantly lower serum estradiol than pre-surgical levels [9]. The working clinical assumption is that he will continue testosterone indefinitely — and continuation is what guards against the long-term morbidity of an early surgical menopause. That assumption does not always hold: older studies reviewed by Giacomelli and Meriggiola found that trans men who stopped testosterone, or took it irregularly, after surgery had lower bone density, which rose again when treatment resumed [11]. Gonadectomy leaves an iatrogenic hypogonadal state that can harm bone unless gender-affirming hormones are consistently continued [11]; consistent with this, longitudinal data show measurable falls in bone-density Z-scores in the bilateral-oophorectomy subgroup — supporting ongoing monitoring and adequate replacement rather than a presumption of harm in every case [12]. Specialist guidance lists oophorectomy before age 45 without optimal hormone replacement as a recognised osteoporosis risk factor [6].

So the principle holds across both directions of care: after gonadectomy, hormone therapy is not an enhancement layered on top of a self-sustaining gonadal system. For practical clinical purposes, it is the main source of sex-steroid support. Stopping it is a withdrawal of physiological support, with a predictable direction of harm.

A caveat that sharpens the point rather than softening it: before gonadectomy the picture is more variable. A trans woman who stops estradiol and stops her anti-androgen will usually see endogenous testosterone production resume, which lends the skeleton some protection; the same holds in reverse for a trans man with intact ovaries. But if estradiol lapses while androgen suppression continues — for example, when one prescription is interrupted but the other continues — the patient may be left in a very low-sex-steroid state despite having gonads. The post-gonadectomy case is simply the version where that state is guaranteed rather than possible.

The wellbeing question, answered honestly

It would be easy, and rhetorically tempting, to claim that stopping hormones reliably harms mental health. The evidence doesn’t license a claim that strong, and overstating it would be its own kind of error. What the literature reasonably supports is that starting gender-affirming hormone therapy is associated with improvements in depression for many adults, although in the cohort cited here anxiety did not change significantly and only a fifth of participants completed follow-up [13] — and one randomised comparison found reductions in depressive symptoms within three months of initiation relative to delayed access [14] — though the picture is not perfectly uniform across cohorts.

The direct, high-quality evidence on the psychological effects of withdrawing established gender-affirming hormone therapy is limited. That limitation should make clinicians more careful, not less. We do not have evidence that abruptly stopping established hormone therapy is psychologically neutral; in the absence of that evidence, treating withdrawal as a non-event is not a cautious position — it is an unexamined one. The burden of proof sits with the act that changes the status quo, and stopping is that act.

What stopping properly actually looks like

Here is the quiet tell in all of this. When clinicians write about how to come off sex-hormone therapy, they do not describe a non-event — they describe something to be anticipated, planned, and monitored. Symptom recurrence after stopping is common; and while trials have not actually established that a gradual taper outperforms abrupt cessation, that is rather the point — stopping is a clinical event that has to be managed however it is done, not a default that manages itself [15]. For short-acting oral estrogen regimens, intermittent missed doses produce fluctuating exposure rather than a smooth reduction; a planned dose reduction or formulation change is a different thing from an unmanaged lapse.

In other words, the considered way to stop is itself an active clinical process — assessed, planned, monitored, supported. Which is the whole argument in miniature: if stopping safely requires all of that, then stopping is not a neutral default. An abrupt, unmanaged cessation isn’t the absence of an intervention. It is the unmanaged version of one.

The principle

In a patient who is stable on treatment, and absent a patient-led decision to stop or a specific clinical contraindication, continuation is the clinically safer default. Withdrawal is the deviation, and when it is unplanned or clinician-driven it requires its own justification and its own evidence — held to the same standard we would demand before we ever began.

This is not an argument against stopping when the patient wants to stop, or when a specific medical reason makes stopping necessary. It is an argument against treating interruption, refusal, or unmanaged withdrawal as clinically neutral.

A patient stable on treatment is owed the same rigour on the way out as on the way in. Anything less — when withdrawal is imposed by a clinician, a system, or a lapse in access rather than chosen by the patient — is not caution. It is a second intervention, performed without the same consent and clinical care we required at initiation.

Sources

1. Vasomotor Symptoms During Menopause: A Practical Guide on Current Treatments and Future Perspectives. International Journal of Women’s Health, 2023. — Hypothalamic KNDy-neuron mechanism and thermoneutral-zone narrowing on estrogen withdrawal. (Population: cis menopausal women.)
2. Paradigm shift in the pathophysiology of vasomotor symptoms: effects of estradiol withdrawal and progesterone therapy. ScienceDirect, 2020. — Evidence that the downward swing in estradiol, not low estradiol alone, drives vasomotor symptoms. (Population: peri/menopausal women; presented as a model.)
3. Hammar M, et al. Postmenopausal women without previous or current vasomotor symptoms do not flush after abruptly abandoning estrogen replacement therapy. Maturitas, 1999;31(2):117–122. PMID 10227004. — Individual variation; abrupt withdrawal did not reliably induce flushes in women with no prior history. (Population: postmenopausal women.)
4. Khosla S, Oursler MJ, Monroe DG. Estrogen and the skeleton. Trends in Endocrinology and Metabolism, 2012. PMCID PMC3424385. — Estrogen is the principal restraint on bone resorption; its loss increases osteoclast activity and lifespan and lowers bone mineral density. (Population: general/cis bone physiology; women and men.)
5. Anagnostis P, et al. Antiosteoporosis therapy after discontinuation of menopausal hormone therapy: a systematic review. Hormones (Athens), 2024;23:339–344. DOI 10.1007/s42000-024-00526-1. — Background, citing Tremollieres et al. 2001: after MHT is stopped, bone loss occurs mostly within the first 2 years, at 1.5–2% per year, similar to the loss in the first 2 years of menopause. Discussion: “Fracture risk does not seem to increase in the short term after MHT discontinuation”; a post-hoc analysis of the Women’s Health Initiative trial found no increase in fractures after cessation. (Population: cis menopausal women.)
6. Radix A, Deutsch MB. Bone health and osteoporosis. UCSF Center of Excellence for Transgender Health — Guidelines for the Primary and Gender-Affirming Care of Transgender and Gender Nonbinary People, 2nd ed., 2016. — Underutilisation of hormones after gonadectomy listed as a known osteoporosis risk factor; oophorectomy before age 45 without optimal replacement listed as a risk factor. People who have had a gonadectomy and at least 5 years without hormone replacement “should also be considered for bone density testing, regardless of age” (graded X C W); the guideline states that there is insufficient evidence to guide recommendations for bone-density testing in transgender women or men. (Trans-specific.)
7. Motta G, et al. Fracture risk assessment in an Italian group of transgender women after gender-confirming surgery. Journal of Bone and Mineral Metabolism, 2020;38(6):885–893. DOI 10.1007/s00774-020-01127-9. — Cross-sectional study of 57 trans women at least 2 years after surgery (lumbar spine DXA). Low bone mass (Z-score ≤ −2) in 40% by male reference ranges, 30% by female ranges; significantly associated with low estradiol and with low compliance, measured as a medication possession ratio below 80%. No fragility fractures; 93% had low vitamin D. (Trans-specific.)
8. Wiepjes CM, et al. Fracture Risk in Trans Women and Trans Men Using Long-Term Gender-Affirming Hormonal Treatment: A Nationwide Cohort Study. Journal of Bone and Mineral Research, 2020;35(1):64–70. DOI 10.1002/jbmr.3862. — Fractures seen in emergency departments nationwide, 2013–2015. In trans women ≥50 on long-term hormone therapy, 4.4% vs 2.4% in age-matched reference men (OR 1.90, 95% CI 1.32–2.74); similar to age-matched reference women (OR 1.05). The authors attribute the excess to low pre-treatment bone density “instead of the effects of hormonal treatment”, describe the three-year window as a point prevalence, and found no association between estradiol and fracture in trans women (OR 0.99 per 10 pmol/L). Our reading of it in the text is an inference, not the authors’ finding. (Trans-specific.)
9. Kumar S, Bertin E, O’Dwyer C, et al. Serum estradiol levels decrease after oophorectomy in transmasculine individuals on testosterone therapy. Asian Journal of Andrology, 2023;25:309–313 (published online 2022). PMID 36124534; DOI 10.4103/aja202262. Retrospective chart review, Vancouver, BC (12 cases vs 12 matched controls). — Serum estradiol fell significantly below pre-surgical levels after oophorectomy (P = 0.02). (Trans-specific.)
10. Barrera EP, Grimstad FW, Boskey ER. Young Adult Patients with Testosterone Management Concerns after Gender-Affirming Hysterectomy and Bilateral Oophorectomy: A Case Series. Journal of Pediatric and Adolescent Gynecology, 2023;36(1):89–91. DOI 10.1016/j.jpag.2022.07.010. — No longer cited (11 October 2026). We do not hold this paper and had not read it in full, so it cannot support a statement under our own standard. The sentence it supported now rests on [11]. The entry stays, struck through, so that the numbering of the sources below it does not change.
11. Giacomelli G, Meriggiola MC. Bone health in transgender people: a narrative review. Therapeutic Advances in Endocrinology and Metabolism, 2022. DOI 10.1177/20420188221099346. — Surgical removal of the gonads produces an iatrogenic hypogonadal state that can negatively affect bone homeostasis if gender-affirming hormone therapy is not consistently continued. (Trans-specific.)
12. Sanna E, et al. Bone health in transgender assigned female at birth people: effects of gender-affirming hormone therapy and gonadectomy. Frontiers in Endocrinology (Lausanne), 2024;15:1416121. PMID 39391880. — Longitudinal DXA: the bilateral-oophorectomy subgroup showed significant Z-score reductions at lumbar spine, femoral neck and total hip at 5 years (and lumbar spine at 10 years), alongside lower estradiol — supporting monitoring and adequate replacement. (Trans-specific.)
13. Aldridge Z, Patel S, Guo B, Nixon E, Pierre Bouman W, Witcomb GL, Arcelus J. Long-term effect of gender-affirming hormone treatment on depression and anxiety symptoms in transgender people: A prospective cohort study. Andrology, 2021;9(6):1808–1816. DOI 10.1111/andr.12884; PMID 32777129. — 18-month prospective cohort, no control group; 178 of 906 eligible (20%) completed follow-up. Depression (HADS-D) fell significantly (mean change −2.05, 95% CI −2.72 to −1.38); anxiety (HADS-A) did not change significantly (−0.31, 95% CI −0.97 to 0.36; P = .37). (Trans-specific.)
14. Nolan BJ, Zwickl S, Locke P, Zajac JD, Cheung AS. Early access to testosterone therapy in transgender and gender-diverse adults seeking masculinization: a randomized clinical trial. JAMA Network Open, 2023;6(9):e2331919. DOI 10.1001/jamanetworkopen.2023.31919; PMID 37676662. — Three-month open-label RCT (64 randomised, 62 analysed), immediate vs delayed testosterone. By the authors’ 5-point threshold, a clinically significant reduction in depressive symptoms (PHQ-9, mean difference −5.6, 95% CI −6.8 to −4.4), and a significant reduction in suicidality (SIDAS, −6.5, 95% CI −8.2 to −4.8). The primary outcome, gender dysphoria (GPSQ, −7.2), fell by less than the authors’ 11-point clinical threshold. Addresses initiation, not withdrawal. (Trans-specific.)
15. Menopause: diagnosis and management — evidence review on starting and stopping HRT. National Institute for Health and Care Excellence (NICE). NCBI Bookshelf NBK343480. — Discontinuation treated as a planned clinical decision with anticipated symptom recurrence; four RCTs comparing tapered vs abrupt cessation found tapering did not reliably prevent symptom recurrence (taper-vs-abrupt superiority not established). (Population: cis menopausal women.)
16. Deutsch MB. Overview of feminizing hormone therapy. UCSF Center of Excellence for Transgender Health — Guidelines for the Primary and Gender-Affirming Care of Transgender and Gender Nonbinary People, 2nd ed., 2016. — “There is no evidence to support continuation or cessation of hormones for older transgender women. Since the mean age of menopause in the U.S. is 49, it is reasonable in transgender women who have undergone gonadectomy to consider stopping hormone therapy around age 50.” (Trans-specific.)

Version history. v1.0, 29 June 2026 — first published. v1.1, 11 October 2026 — corrected after every source was re-read against the article while preparing our evidence to the UN. All the changes below are corrections; none reflects new evidence. Bone-density testing: we said the UCSF guidelines “recommend” it after gonadectomy and five years without replacement; they say such people “should be considered” for testing, and that the evidence to guide testing is insufficient (text and source 6). Stopping around 50: we omitted that the same guidelines find no evidence either way on continuing or stopping hormones in older trans women, and call it reasonable to consider stopping around 50 after gonadectomy; this is now stated at full weight (new source 16). Menopausal hormone therapy: we said bone loss after stopping mirrors that in “early menopause”; the review compares it with the first two years after menopause, citing an earlier study, and also says fracture risk does not seem to rise in the short term after stopping, which we had left out (text and source 5). Fracture data: we did not give the Dutch authors’ own explanation, low bone density before treatment, or that their measure was a three-year count of emergency-department fractures, or that they found no link between estradiol and fracture in trans women; and we presented our own inference without labelling it (text and source 8). Low bone mass: the 40% figure uses male reference ranges (30% by female ranges), the study is cross-sectional, and adherence was a possession ratio (text and source 7). Depression and anxiety: the cohort we cited found improvement in depression but no significant change in anxiety, with 20% follow-up; its source note claimed a statement on heterogeneity that the paper does not make (text and source 13). The randomised trial: its depression result is clinically significant by the authors’ threshold; its suicidality result is significant, and its primary outcome fell short of the authors’ clinical threshold (source 14). Testosterone interruption after oophorectomy: the sentence rested on a case series we do not hold and had not read; it now rests on a review we hold (source 11), and source 10 is struck through. Citation: source 9 now names its authors and its volume year, 2023. Recorded in our Evidence Transmission Register as ET-068. Sources 1 to 3, on vasomotor symptoms, have not yet been re-read against this article; that check is outstanding.