Frequency, Timing, and Management of Erythrocytosis in Adolescents Receiving Gender-Affirming Testosterone Therapy—A Retrospective Study

Abstract

Erythrocytosis is a well-described effect of testosterone therapy in adults, but less is known about this phenomenon in transgender and gender-diverse adolescents. To explore this further, we performed a retrospective study in 158 adolescents who received gender-affirming testosterone therapy from January 2007 to April 2020 at a pediatric gender service. Erythrocytosis (hematocrit [Hct] > 0.50 L/L) developed in 9.5% (15/158) of adolescents. Frequency increased with time; in the majority (67%, 10/15), erythrocytosis occurred in the second year of treatment. Overall, our findings indicate erythrocytosis occurs commonly in testosterone-treated transgender adolescents and support the need for regular ongoing monitoring of Hct.

Introduction

While gender-affirming hormone therapy for transgender and gender-diverse adolescents is supported by numerous key professional associations and has been shown to improve mental health and quality of life, current treatment and monitoring recommendations are based primarily on clinician consensus.^1–3 Many adolescents seeking to masculinize use testosterone to affirm their gender. In testosterone-treated transgender adults, erythrocytosis (typically defined as hematocrit [Hct] >0.50 L/L) occurs commonly and increases with treatment duration.^4,5 Given the relationship between erythrocytosis and thromboembolic cardiovascular disease and the likelihood that testosterone treatment duration will be long term, identification of erythrocytosis may be of particular importance in transgender adolescents.⁶

To date, one prospective cohort study has reported Hct and erythrocytosis rates in testosterone-treated transgender adolescents. Specifically, Millington et al. found that Hct >0.50 L/L developed in 6.5% of adolescents while receiving testosterone treatment (n = 13/200).⁷ Two earlier retrospective studies reported slightly lower rates of erythrocytosis of 4.2% and 5.0% in transgender young people treated with testosterone, albeit in a smaller number of participants (n = 72 and 119).^8,9

In none of these previous studies were any subjects observed to have a Hct >0.54 L/L, which is the level above which clinical guidelines recommend cessation of testosterone therapy, dose reduction or therapeutic phlebotomy to reduce the risk of adverse events in hypogonadal cisgender men.¹⁰ However, subjects in two of the studies did nonetheless receive a reduction in their testosterone dose due to erythrocytosis.^7,9 The frequency at which such interventions were required was not readily apparent.

Based on these findings, there remains variation in opinion regarding optimal Hct monitoring for testosterone-treated adolescents.^7,8,11 To examine this further, we performed a retrospective chart review to determine the rates of erythrocytosis among an independent cohort of testosterone-treated transgender adolescents. In addition, we aimed to describe how the risk of erythrocytosis changes over time and the interventions used to manage erythrocytosis.

Methods

Participants and procedures

Adolescents who commenced testosterone between January 2007 and April 2020 at the Royal Children’s Hospital (RCH) Gender Service in Melbourne, Australia, were identified from departmental databases and our electronic medical record. The RCH Gender Service is a statewide pediatric referral clinic for transgender children and adolescents. Management is typically provided up until a young person’s 18th birthday, after which care is transferred to local adult services.

Gender affirming care provided at RCH follows national standards of care.² Decisions regarding testosterone formulation, initial dose, and rate of dose titration are determined by the treating clinician on an individualized basis. Typically, initial testosterone doses are lower with slower titration for adolescents who have not progressed beyond Tanner 3 and/or have received gonadotropin-releasing hormone analogue (GnRHa), whereas adolescents who are postmenarchal and not receiving GnRHa commence at higher doses.

Measures

Gender identity, as described by the adolescent at their initial appointment, was documented. Prior GnRHa use, testosterone treatment, body mass index (BMI), total testosterone, and Hct were extracted at four time points: in the 12 months prior to commencing testosterone (baseline), 0–6 months, 7–12 months, and 13–24 months of testosterone use. Adolescents were eligible for inclusion if Hct was performed at least once following commencement of testosterone and they had a Hct <0.50 L/L documented either (i) at baseline or (ii) in the 6 months prior to developing erythrocytosis. The RCH Human Research Ethics Committee approved the study (#36323). As this study retrospectively analyzed de-identified data collected during routine clinical care, individual participant consent was not obtained.

Data analyses

Mean Hct over time was calculated using generalized estimating equations. To assess change in Hct with time on testosterone, mean Hct was plotted relative to cohort-specific reference ranges. Fifth, 50th, and 95th centiles for Hct at baseline were estimated as a function of the factorial power of age using quantile regressions. Associations between erythrocytosis and potential risk factors such as obesity (BMI ≥95th centile) and testosterone formulation are described as odds ratios (OR) with a 95% CI. Testosterone levels in adolescents who did and did not develop erythrocytosis were compared using a two-sample t-test.

Results

During the study period, 179 adolescents commenced testosterone, of whom 158 were eligible for inclusion (see Table 1 for participant characteristics). Of the 21 adolescents excluded, 3 did not have a suitable baseline Hct recorded and 18 did not have Hct performed after commencing testosterone. Intramuscular formulations were commenced in 142/158 (89.9%) adolescents, of whom 83/142 (58.5%) started a long-acting preparation (testosterone undecanoate 1000 mg/4 mL 12 weekly) (see Table 1 for treatment details). Of the eligible participants, many subsequently transitioned to adult care (or were occasionally lost to follow up), and thus follow-up data were unavailable for 17 adolescents at 7–12 months, and 52 adolescents at 13–24 months.

Table 1.

Baseline Characteristics, Testosterone Treatment, and Monitoring Data for 158 Adolescents Eligible for This Study

Median age commenced testosterone (years) [IQR]	16.7 [16.0–17.6]
Gender identity, n (%)
Male	176 (98)
Nonbinary	3 (2)
Treatment with GnRHa, n (%)
Yesa	13 (8)
No	145 (92)
Obesity at baselineb (BMI ≥95th centile), n (%)	38/144 (26)
Adolescents with ongoing gender service engagement, n
0–6 months	158
7–12 months	141
13–24 months	106
Hct documented, n (%)
Baseline	133/158 (84)
0–6 months	128/158 (81)
7–12 months	113/141 (80)
13–24 months	81/106 (76)
Testosterone formulation commenced, n (%)
Intramuscular	142 (90)
Testosterone undecanoate	83 (53)
Testosterone esters/propionate	59 (37)
Transdermal 1% testosterone gel	15 (9)
Oral testosterone undecanoate	1 (1)
Testosterone dose commenced, median [IQR]
Intramuscular (mg/week), n = 142	83 [42–83]
Transdermal (mg/day), n = 15	25 [25–25]
Oral (mg/day), n = 1	40 mg
Testosterone dose at 12 months of testosterone treatment, median [IQR]
Intramuscular (mg/week), n = 103	83 [83–83]
Transdermal (mg/day), n = 13	37.5 [25–50]
Testosterone dose at 24 months of testosterone treatment, median [IQR]
Intramuscular (mg/week), n = 36	83 [83–83]
Transdermal (mg/day), n = 4	31 [22–41]

^a At the Royal Children’s Hospital gender service adolescents Tanner stage 4–5 are not typically offered GnRHa.

^b Baseline BMI was not documented in 14 adolescents.

BMI, body mass index; GnRHa, gonadotropin-releasing hormone analogue; Hct, hematocrit; IQR, interquartile range.

Mean Hct increased with time on testosterone and ultimately rose higher than the cohort’s baseline 95th centile (Fig. 1). Overall, erythrocytosis developed in 9.5% (15/158) of adolescents, with Hct levels ranging from 0.51 L/L to 0.59 L/L. Three individuals recorded an Hct >0.54 L/L. Risk of erythrocytosis increased over time, from 1.6% (2/128) at 0–6 months to 2.7% (3/113) at 7–12 months to 13.6% (11/81) at 13–24 months. Two-thirds (10/15) of adolescents who developed erythrocytosis did so in the second year of testosterone treatment—all 10 reached a stable testosterone dose by 12 months of treatment, with no further dose increase documented. Among adolescents for whom BMI data were available, odds of erythrocytosis were higher for adolescents with baseline obesity (BMI ≥ 95th centile) (13.2%, 5/38) compared to nonobese adolescents (6.6%, 7/106) (OR: 2.14; 95% confidence interval [CI] 0.64–7.21). One of the three adolescents with Hct >0.54 L/L was obese. Of the remaining two, one was overweight, and one had a BMI within the normal range.

FIG. 1.

Hematocrit changes during testosterone therapy. Hematocrit (Hct) values measured at baseline, 0–6 months, 7–12 months, and 13–24 months of testosterone therapy for each adolescent are plotted by age at time of measurement. Adolescents who ultimately developed erythrocytosis (Hct >0.50 L/L) are represented by dashed lines. ^aOne adolescent had a baseline hematocrit of 0.23 L/L and was found to have severe iron deficiency anemia. Subsequent Hct of 0.50 L/L was measured after 4 months of both testosterone treatment (25 mg daily transdermal) and oral iron supplementation. ^bAnother adolescent had a peak hematocrit of 0.59 L/L and testosterone was ceased for 3 months.

Odds of erythrocytosis were similar in those who received intramuscular testosterone (9.9%, 14/142) compared to transdermal formulation (6.7%, 1/15) (OR: 1.53; 95% CI: 0.19–12.54) and those who received a long-acting formulation (9.7%, 13/134) compared with short-acting intramuscular formulations (8.3%, 2/24) (OR: 1.18, 95% CI: 0.25–5.60). Testosterone levels documented at the time of Hct >0.50 L/L ranged from 4.9 to 24.3 nmoL/L (141–700 ng/dL). Mean testosterone level during treatment was similar for adolescents who developed erythrocytosis (10.70 nmol/L or 308.61 ng/dL) and those who did not (11.92 nmol/L or 343.80 ng/dL) (p = 0.37).

No adverse outcomes were reported in adolescents who developed erythrocytosis. Recorded Hct >0.50 L/L prompted a change in clinical management or a recommended intervention to reduce Hct in 7 of 15 (46.7%) adolescents (4.4% of the entire cohort). One adolescent ceased testosterone for 3 months (peak Hct 0.59 L/L) and five (peak Hct 0.51–0.54 L/L) increased the interval between testosterone injections. One further adolescent (Hct 0.54 L/L) had therapeutic phlebotomy recommended at the time of transfer to adult care, though it is unknown whether this subsequently occurred.

Of the 15 adolescents with documented erythrocytosis, 11 were transferred to adult-based services shortly afterward and one had no further Hct measured, therefore we could not reliably determine whether erythrocytosis improved. Of the remaining three adolescents, two had an improvement in Hct to <0.50 L/L without intervention. The adolescent whose testosterone was withheld for 3 months had an improvement in Hct from 0.59 L/L to 0.52 L/L.

Discussion

This study of testosterone-treated transgender adolescents indicates that erythrocytosis occurs relatively commonly, with increasing prevalence over the initial two years of treatment. Compared to previous studies in transgender adolescents the rate of erythrocytosis within our cohort (10%) was around twofold higher.^7–9 Jarin et al. reported Hct >0.50 L/L in 3 of 72 (4.2%) of their testosterone-treated adolescents.⁸ However, Hct results beyond one year, where the greatest increase in our cohort was found, were not included in Jarin’s study, which might in part explain their lower rate of erythrocytosis.

The cohort of Laurenzano et al.—in whom 6/119 (5.0%) developed erythrocytosis—had a longer duration of follow-up (median 1.9 years).⁹ It should be noted the majority of these adolescents received a relatively lower dose of subcutaneous testosterone (25–50 mg weekly) compared to subjects in our study, who received intramuscular testosterone at a higher median dose (83 mg weekly). Millington et al reported development of erythrocytosis over a two-year period in 6.5% (13/200) of their adolescent cohort—97% of whom received subcutaneous testosterone at a median dose of 40 mg per week.⁷ This rate was again lower than our cohort.

Group differences in testosterone dose, frequency, formulation and/or administration route may explain differences in the rate of erythrocytosis observed among these studies. In transgender adults the testosterone formulation associated with highest risk for erythrocytosis remains unclear. Some studies have reported higher risk with long-acting preparations and others with short-acting testosterone esters.^5,12,13

In their large 20-year follow-up study of transgender adults receiving testosterone, Madsen et al. demonstrated an increasing probability of developing erythrocytosis (defined as two Hct levels >0.50 L/L) as duration of testosterone treatment extended (10% after one year, 38% after 10 years).⁵ To date, few adolescent cohorts have examined the timing of development of erythrocytosis. In their cohort, Millington et al. reported highest rates in the first year of testosterone treatment, but also noted emergence of three new recorded cases of erythrocytosis on testing at 24 months.⁷ In comparison, our data indicate that, as in adults, the risk increases with duration of testosterone therapy, which has implications for monitoring recommendations in this population.⁵ Optimal Hct reference range values and clinical decision limits for transgender adolescents and men remain to be clarified.¹⁴ As in other studies, mean Hct for our cohort over time could be interpreted as within the “normal” male range (often considered 0.40–0.50 L/L) although the clinical relevance of this is uncertain.^7–9 To avoid imposing population reference intervals that may be less appropriate (e.g., for adult or cisgender populations) we used baseline (pretestosterone) Hct data to estimate reference ranges specific to our cohort, against which to assess change over time. While numbers in our dataset are not large enough for these estimations to be applied more broadly, they afford valuable population-specific comparisons of change over time with testosterone.

Importantly, the Hct level at which the risk of neuro-occlusive or cardiovascular events increases is not well-defined. Current clinical guidelines for gender affirming testosterone treatment attribute a moderate-high risk of adverse outcome with a hematocrit level of >0.50 L/L, although the guideline in cisgender men on which this level was based has more recently been revised downwards to >0.48 L/L.^3,10 Our observation that 7 of 15 (46.7%) of adolescents with erythrocytosis had changes to their medical management recommended in response to Hct levels >0.50 L/L speaks to clinician discomfort with Hct above this threshold in our service. Nonetheless here and elsewhere, some young people had spontaneous resolution of these levels without intervention.

Limitations

In adults, obesity and testosterone formulation have been identified as risk factors for testosterone-induced erythrocytosis.⁵ Among our adolescents, ORs suggest an association with obesity, but supporting evidence was weak. Moreover, smoking and obstructive sleep apnea are well-established contributors to erythrocytosis, but the retrospective nature of our study precluded examination of these factors.

Missing data in the context of transition to adult care or missed Hct monitoring over the study period are an additional limitation. Related to this, we were unable to follow-up Hct trajectory and/or outcome of erythrocytosis in most subjects, which was further compounded by our follow-up period of 24 months. Looking ahead, prospective trials with larger numbers of transgender adolescents that encompass a longer duration of testosterone treatment and key cardiovascular health outcomes (e.g., thromboembolic events) will be required to better understand the risk and significance of erythrocytosis in transgender individuals who commence testosterone in adolescence.

Conclusion

In this retrospective analysis of adolescents attending a tertiary gender service, ∼10% of adolescents receiving testosterone developed erythrocytosis, with rates of erythrocytosis increasing with time on testosterone. There remains uncertainty regarding optimal monitoring for erythrocytosis, particularly given the variation in testosterone dose and formulation worldwide. The relatively high frequency of erythrocytosis in our cohort supports regular Hct monitoring as currently suggested in international guidelines.³ Moreover, given the increased rate of erythrocytosis that we observed over time among our adolescents—as has been observed in transgender adults previously—long-term Hct monitoring beyond two years is likely to be prudent.

Abbreviations Used

BMI

Body mass index

CI

Confidence interval

GnRHa

Gonadotropin-releasing hormone analogue

Hct

Hematocrit

IQR

Interquartile range

OR

Odds ratio

RCH

Royal Children’s Hospital

Authors’ Contributions

A.M.: Conceptualization (lead), methodology (lead), formal analysis (lead), writing—original draft (lead). M.H.: Resources (lead), visualization (lead), formal analysis (supporting). K.C.P.: Writing—review and editing (equal), supervision (equal). M.O’C.: Conceptualization (supporting), writing—review and editing (equal), supervision (equal).

Data Availability

The data that support the findings of this study are available from the corresponding author, M.O’C., upon reasonable request.

Author Disclosure Statement

The authors declare they have no conflict of interest.

Funding Information

K.C.P. wishes to acknowledge the support of the Royal Children’s Hospital Foundation, the Hugh Williamson Foundation Trust, and the Australian National Health and Medical Research Council (GNT 2006529 and 2027186).

 

Cite this article as: McLean A, Hoq M, Pang KC, O’Connell M (2025) Frequency, timing, and management of erythrocytosis in adolescents receiving gender-affirming testosterone therapy—a retrospective study, Transgender Health 10:4, 401–406, DOI: 10.1089/trgh.2024.0233.