Effects of testosterone therapy on BMI, blood pressure, and laboratory profile of transgender men: a systematic review
Summary
Testosterone is the main hormonal agent used for cross-sex hormone therapy in female-to-male transgender persons. Our aim was to systematically review the literature concerning the effects of testosterone on body mass index (BMI), blood pressure, hematocrit, hemoglobin, lipid profile, and liver enzymes in transgender men. PUBMED and EMBASE were searched for studies published until March 2017. Studies were included if they reported interventions with any dose of testosterone and comparison of variables before and during treatment. Of 455 potentially eligible articles, 13 were reviewed. Study duration ranged from 6 to 60 months, sample size ranged from 12 to 97 patients, and the most common treatment was parenteral testosterone undecanoate 1000 mg/12 weeks. Slight but significant increases in BMI were reported (from 1.3 to 11.4%). Three out of seven studies assessing the impact of different testosterone formulations on blood pressure detected modest increases or clinically irrelevant changes in this variable. In another study, however, two patients developed hypertension, which was resolved after cessation of testosterone therapy. Decreases in HDL-cholesterol and increases in LDL-cholesterol were consistently observed. Eight studies observed a relationship between testosterone and increased hemoglobin (range: 4.9–12.5%) and hematocrit (range: 4.4–17.6%), but discontinuation of androgen therapy was not necessary. In one study, two patients developed erythrocytosis (hematocrit >52%) after 9 and 12 months of treatment. One study analyzing testosterone formulations observed smaller increases in hemoglobin and hematocrit with testosterone gel. Six studies assessing liver function showed slight or no changes. Overall, the quality of evidence was low, given the lack of randomized clinical/controlled trials and the small sample sizes. In conclusion, exogenous testosterone administration to transgender men was associated with modest increases in BMI, hemoglobin/hematocrit, and LDL-cholesterol, and with decreases in HDL-cholesterol. Long-term studies are needed to assess the long-term risks of testosterone therapy, particularly as they relate to cardiometabolic risks such as diabetes, dyslipidemia and the metabolic syndrome.
Introduction
Gender dysphoria is characterized by a persistent desire to live and be accepted as a member of the opposite sex (World Health Organization, 1992; American Psychiatric Association, 2000, 2013). People who experience gender dysphoria require cross-sex hormones to induce secondary sexual characteristics of the desired sex. In this context, female to male transgender persons (transgender men) may desire treatment with androgen hormones to induce virilization (Moore et al., 2003; Gooren, 2005).
Testosterone is the main hormonal agent used for cross-sex hormone therapy in transgender men. Injectable testosterone esters have been traditionally used, administered in doses of 100–250 mg every 7–21 days (Pelusi et al., 2014). A long-acting testosterone undecanoate formulation which maintains more stable levels of testosterone is also available (Pelusi et al., 2014). In addition, transdermal testosterone formulations, administered as gel or patches, are another option for transgender men, which is available in many countries.
Few studies so far have evaluated the effects of testosterone therapy in transgender men. Some suggest that testosterone administration is safe and associated with low risk of adverse effects (Hembree et al., 2009; Coleman et al., 2012). Other studies, however, have reported hypertension, increased erythropoiesis, decreased high density lipoprotein (HDL), increased low-density lipoprotein (LDL), elevation of liver enzymes, obesity, and acne, among others, associated with this treatment in transgender individuals (Van Kesteren et al., 1997; Michel et al., 2001; Levy et al., 2003; Moore et al., 2003).
Therefore, the aim of the present systematic review was to summarize the available information regarding the effects of testosterone administration on clinical and metabolic variables of transgender men, especially body mass index (BMI), blood pressure, and hematologic and metabolic profiles.
Materials and Methods
Search strategy and study selection
This systematic review was performed in accordance with PRISMA guidelines (Stroup et al., 2000; Liberati et al., 2009). Medline (PUBMED) and EMBASE databases were searched for articles published until March/2017. There were no other limits except for the end date.
The following search strategy was developed for PUBMED and modified as needed for EMBASE: ‘transsexualism’ or ‘transgender person’ or ‘person, transgender’ or ‘persons, transgender’ or ‘transgenders’ or ‘transgender’ or ‘transgendered persons’ or ‘person, transgendered’ or ‘persons, transgendered’ or ‘transgendered person’ or ‘transsexual persons’ or ‘person, transsexual’ or ‘persons, transsexual’ or ‘transsexual person’ and ‘testosterone’ or ‘cross sex hormone therapy.’
The selection criteria for the studies were as follows: focus on transgender men, intervention with any dose of testosterone, and comparison of clinical and metabolic variables before and during/after treatment. The main outcomes of interest were BMI, blood pressure, hematocrit, hemoglobin, lipid profile, and liver enzymes.
Data extraction and quality control
Two investigators independently screened titles and abstracts of all articles selected in order to evaluate if the studies were eligible for inclusion in the systematic review. The selected articles were read in full for confirmation of eligibility and data extraction. Disagreements were resolved by discussion among all investigators. The following information was extracted from each individual study: name of the first author, publication year, country, number of subjects, duration of the follow-up, intervention, and evaluation time.
Results
Study selection
Figure 1 shows the flowchart of study selection. Initially, 455 potentially eligible articles were identified; 438 were excluded after reading the abstracts and/or titles; and 17 articles were read in full. After this step, three articles were excluded because the variables of interest were not reported, and one because a progestin was added to androgen treatment. Thirteen articles were thus included in the systematic review (Giltay et al., 2004; Berra et al., 2006; Jacobeit et al., 2007, 2009; Mueller et al., 2007, 2010; Chandra et al., 2010; Cupisti et al., 2010; Pelusi et al., 2014; Wierckx et al., 2014; Deutsch et al., 2015; Quirós et al., 2015; Fisher et al., 2016).
Characteristics of included studies
As shown in Table 1, study duration ranged from 6 to 60 months, sample size ranged from 12 to 97 patients, and the most common treatment was parenteral testosterone undecanoate 1000 mg/12 weeks, which was reported in eight of 13 articles (Jacobeit et al., 2007, 2009; Mueller et al., 2007, 2010; Cupisti et al., 2010; Pelusi et al., 2014; Wierckx et al., 2014; Quirós et al., 2015; Fisher et al., 2016). The design of most studies was observational (non-controlled).
| Study | Country | n | Treatment duration (months) | Mean age (years) | Treatment | Variables assessed |
|---|---|---|---|---|---|---|
| Giltay et al. (2004) | Netherlands | 81 | 7 | 36.7 (21–61)aa
Median and interquartile interval.
|
-Testosterone esters (250 mg/2 weeks) | BMI, blood pressure, lipid profile |
| -Oral testosterone undecanoate (160–240 mg/day) | ||||||
| Berra et al. (2006) | Italy | 16 | 6 | 30.4 ± 3.4 | Testosterone enanthate (100 mg) + testosterone propionate (25 mg/10 days) | BMI, blood pressure, lipid profile |
| Jacobeit et al. (2007) | Netherlands | 12 | 12 | 33.0 ± 6.0 | Testosterone undecanoate (1000 mg/12 weeks) | Lipid profile, hematocrit/hemoglobin |
| Mueller et al. (2007) | Germany | 35 | 12 | 29.6 ± 8.9 | Testosterone undecanoate (1000 mg/12 weeks) | BMI, blood pressure, lipid profile, liver function, hematocrit and hemoglobin |
| Jacobeit et al. (2009) | Netherlands | 17 | 36 | 34.0 ± 7.0 | Testosterone undecanoate (1000 mg/12 weeks) | Lipid profile, liver function, hematocrit and hemoglobin |
| Mueller et al. (2010) | Germany | 45 | 24 | 30.4 ± 9.1 | Testosterone undecanoate (1000 mg/12 weeks) | BMI, blood pressure, lipid profile, liver function, hematocrit and hemoglobin |
| Cupisti et al. (2010) | Netherlands | 29 | 12 | 29.9 (18–40)aa
Median and interquartile interval.
|
Testosterone undecanoate (1000 mg/12 weeks) | BMI |
| Chandra et al. (2010) | United States | 12 | 12 | 29.0 ± 9.0 | Testosterone cypionate or testosterone enanthate (50–125 mg/2 weeks) | BMI, blood pressure, lipid profile, liver function, hematocrit |
| Pelusi et al. (2014) | Italy | 45 | 12 | 28.2 (25.6–30.9)aa
Median and interquartile interval.
|
-Testosterone undecanoate (1000 mg/12 weeks) | BMI, lipid profile, liver function, hematocrit and hemoglobin |
| −29.4 (26.6–32.1)aa
Median and interquartile interval.
|
-Testosterone gel (5 mg/day) | |||||
| −30.9 (27.9–33.9)aa
Median and interquartile interval.
|
-Testosterone enanthate (100 mg/10 days) | |||||
| Wierckx et al. (2014) | Belgium (Ghent) | 27 | 12 | 27.3 ± 8.5 | Testosterone undecanoate (1000 mg/12 weeks) | BMI, blood pressure, lipid profile, liver function, hematocrit |
| Norway (Oslo) | 26 | 21.7 ± 5.1 | ||||
| Quirós et al. (2015) | Spain | 97 | 24 | 28.6 ± 8.6 | -Testosterone gel (5 g/day) | BMI, blood pressure, lipid profile, hemoglobin |
| -Testosterone undecanoate (1000 mg/12 weeks) | ||||||
| Deutsch et al. (2015) | United States | 31 | 6 | 27.0 ± 6.9 | -Testosterone gel (5 g/day) | BMI, blood pressure, lipid profile |
| -Testosterone cypionate (50 mg/week) | ||||||
| -Testosterone patch (4 mg/day) | ||||||
| Fisher et al. (2016) | Italy | 26 | 24 | 33.9 ± 9.2bb
Age of all sample (transgender men and women).
|
-Testosterone undecanoate (1000 mg/12 weeks) | BMI |
| -Testosterone enanthatecc
Dose not informed.
|
||||||
| -Transdermal testosteronecc
Dose not informed.
|
- BMI, body mass index.
- a Median and interquartile interval.
- b Age of all sample (transgender men and women).
- c Dose not informed.
Most of the studies followed a protocol in which a second injection of testosterone undecanoate was prescribed 6 weeks after the first administration, with subsequent 1000 mg injections recommended every 12 weeks. However, one study (Quirós et al., 2015) prescribed either intramuscular testosterone undecanoate 1000 mg or 1% transdermal gel, according to patient preference and depending on the patient's clinical condition. Other testosterone formulations were also prescribed: testosterone patch (4 mg/day) (Deutsch et al., 2015), testosterone gel (1%, 5 g/day) (Deutsch et al., 2015; Quirós et al., 2015), testosterone cypionate (Chandra et al., 2010; Deutsch et al., 2015); testosterone enanthate (100 mg) associated with testosterone propionate (25 mg/10 days) (Berra et al., 2006; Fisher et al., 2016); testosterone esters (250 mg/2 weeks); and oral testosterone undecanoate (160–240 mg/day) (Giltay et al., 2004) (Table 1). Except for one study (Quirós et al., 2015), all others included only patients without previous treatment. In the study by Quirós et al. (2015), 13.4% of the patients had already used some hormone therapy.
Qualitative data synthesis
Most participants in the studies were around 30 years of age (Table 1). Ten studies included participants living in Europe, and two included participants living in the United States. Baseline prevalence of obesity and hypertension varied among the studies, as can be inferred from baseline BMI and blood pressure values (Tables 2 and 3).
| Study | BMI | Lipid profile | Liver enzymes (U/I) |
|---|---|---|---|
| Giltay et al. (2004) | 22.89 ± 4.53 and 24.46 ± 3.85; ↑6.8%, p < 0.001 | TC (mmol/L): 4.57 ± 0.87 and 4.53 ± 1.12; ↓0.8%, p = NS | NA |
| HDL (mmol/L):1.41 ± 0.43 and 1.14 ± 0.29; ↓19.1%, p < 0.001 | |||
| LDL (mmol/L): 2.72 ± 0.86 and 3.03 ± 1.14; ↑11.3%, p = 0.002 | |||
| TGL(mmol/L): 0.70 (0.55–1.20) and 0.80 (0.60–1.05); ↑14.2%, p = NS | |||
| Berra et al. (2006) | 21.8 ± 2.9 and 22.8 ± 25.6; ↑4.5%, p < 0.001 | TC (mmol/L): 4.5 ± 0.8 and 4.7 ± 0.7; ↑4.4%, p = NS | NA |
| HDL (mmol/L):1.7 ± 0.4 and 1.5 ± 0.4; ↓11.7%, p < 0.005 | |||
| LDL (mmol/L):2.5 ± 0.8 and 2.8 ± 0.8; ↑12%, p = NS | |||
| TGL (mmol/L): 0.6 ± 0.1 and 0.7 ± 0.1, ↑16.6%, p = NS | |||
| Jacobeit et al. (2007) | NA | TC (mg/dL); 215.8 ± 58.5 and 196.7 ± 39.6; ↓8.8%, p < 0.05 | NA |
| HDL (mg/dL): 51.7 ± 10.8 and 52.2 ± 11.6; ↑0.9%, p = NS | |||
| LDL (mg/dL): 140.5 ± 47.0 and 118.3 ± 30.7; ↓15.8%, <0.05 | |||
| Mueller et al. (2007) | 23.94 ± 4.86 and 24.29 ± 4.64; ↑1.4%, p = 0.03 | TC (mg/dL): 187.26 ± 45.65 and 191.00 ± 42.09; ↑1.9%, p = NS | AST: 18.9 ± 6.3 and 21.7 ± 6.8; ↑14.9%, p = 0.05 |
| HDL (mg/dL): 59.00 ± 10.88 and 48.29 ± 9.77; ↓18.1%, p < 0.0001 | ALT:20.3 ± 9.9 and 24.5 ± 9.4; ↑20.4%, p = 0.05 | ||
| LDL (mg/dL):126.60 ± 35.27 and 133.49 ± 36.87; ↑5.4%, p = NS | GGT:15.8 ± 11.0 and 20.7 ± 12.2; ↑30.8%, p = 0.05 | ||
| TGL (mg/dL): 122.14 ± 54.67 and 152.43 ± 51.24; ↑24.7%, p = 0.02 | |||
| Jacobeit et al. (2009) | NA | TC (mg/dL): 218 ± 47 and 188 ± 42; ↓13.7%, p < 0.05 | AST:24 ± 9 and 23 ± 9; ↓4.1%, p = NS |
| HDL (mg/dL): 50 ± 11 and 51 ± 10; ↑2%, p = NS | |||
| ALT:22 ± 7 and 22 ± 7; 0%, p = NS | |||
| LDL (mg/dL): 139 ± 48 and 119 ± 46; ↓14.3%, p < 0.05 | |||
| TGL (mg/dL): 88 ± 14 and 87 ± 15; ↓1.1%, p = NS | |||
| Mueller et al. (2010) | 24.1 ± 4.5 and 24.2 ± 3.8; ↑0.4%, p = NS | TC (mg/dL):185.8 ± 33.4 and 185.8 ± 34.0; 0%, p = NS | AST: 19.6 ± 6.4 and 21.3 ± 4.5; ↑8.6%, p = NS |
| HDL (mg/dL): 61.8 ± 16.3 and 47.3 ± 13, ↓23.4%; p < 0.0001 | |||
| ALT: 19.9 ± 9.6 and 22.9 ± 4.6; ↑15%, p < 0.01 | |||
| LDL (mg/dL): 131.2 ± 32.4 and 141.4 ± 29.0; ↑7.7%, p = 0.05 | |||
| GGT:16.4 ± 12.7 and 25.2 ± 11.0; ↑53.5%; p < 0.0001 | |||
| TGL (mg/dL): 120.5 ± 64.0 and 148.3 ± 43.2; ↑23%, p < 0.01 | |||
| Cupisti et al. (2010) | 23.5 (21.6–26.0) and 24.2 (22.0–26.5); ↑2.9%; p = 0.001 | NA | NA |
| Chandra et al. (2010) | 27.5 ± 5.2 and 27.6 ± 3.9; ↑0.3%, p = NS | TC(mg/dL):184 ± 26 and 181 ± 34; ↓1.6%, p = NS | AST: 21 ± 5 and 25 ± 7; ↑19%; p = NS |
| HDL (mg/dL): 52 ± 11 and 40 ± 7; ↓23%; p < 0.001 | |||
| ALT: 19 ± 7 and 24 ± 10; ↑26.3%; p < 0.01 | |||
| LDL (mg/dL):113 ± 22 and 121 ± 29; ↑7.0%, p = NS | |||
| TGL (mg/dL);92 ± 72 and 94 ± 41;↑2.1%, p = NS | |||
| Pelusi et al. (2014) | 22.1 (19.5–24.6) and 22.4 (20.0–24.8); ↑1.3% a | TC(mg/dL): 161.5 (145.8–177.1) and 172.6(152.5–192.7); ↑6.8%a 161.3 (145.6–176.9) and 155.7(135.6–175.8); ↓3.4% b 174.4 (158.7–190.1) and 178.6 (158.7–198.7); ↑2.4% c p = NS | AST: 16.9 (13.7–20.1) and 19.4 (16.5–22.3); ↑14.7% a 20.2 (16.9–23.4) and 18.8 (15.9–21.7); ↓6.9% b 16.6 (13.7–19.5) and 18.4 (15.7–21.1); ↑10.8% c p = NS |
| 23.9 (21.2–26.6) and 24.3 (21.8–26.9); ↑1.6% b | |||
| HDL (mg/dL): 62.9 (52.9–70.8) and 58.9(50.167.7);↓6.3%a 67.8 (60.2–75.4) and 58.0(49.6–66.4); ↓14.4%b 70.2 (62.6–77.8) and 58.3 (49.9–66.6);↓16.9% c p < 0.0005 | |||
| LDL (mg/dL): 83.3 (67.3–99.3) and 98.9(79.8–118.2);↑18.7%a 82.0 (66.8–97.3) and 84.9(66.6–103.1); ↑3.5%b 92.6 (77.4–107.8) and 107.0 (88.7–125.3); ↑15.5% c p = 0.001 | |||
| TGL (mg/dL):72.5 (54.0–90.9) and 68.7(41.3–96.2); ↓5.2% a 60.8 (40.4–81.2) and 71.1(40.8–101.5) b 57.4 (38.1–76.7) and 62.6 (33.8–91.4); ↑9.0% c p = NS | |||
| 22.3 (19.9–24.6) and 23.6 (21.4–25.8); ↑5.8% c; p < 0.0005 | |||
| ALT:12.8 (9.5–16.1) and 17.4 (11.9–22.8); ↑35.9% a 15.2 (11.9–18.5) and 15.0 (9.6–20.4); ↓1.3% b 14.5 (11.5–17.6) and 15.3 (10.3–20.3) c ↑5.5% c p = NS | |||
| Wierckx et al. (2014) | 24.8 ± 5.3 and 25.6 ± 4.4; ↑3.2%; p = 0.01 | TC (mg/dL): 171.9 ± 28.1 and 178.2 ± 30.6; ↑3.6%, p = 0.04 | AST: 20.0 (17–23) and 24 (18–29.5); ↑20%; p = 0.01 |
| HDL (mg/dL): 56.3 ± 12.7 and 47.8 ± 10.7; ↓15%; p < 0.001 | |||
| ALT: 16.0 (11.5–20) and 20 (15–25); ↑25%; p = 0.02 | |||
| LDL (mg/dL): 98.4 ± 26.3 and 116.1 ± 28.9; ↑17.9%; p = 0.006 | |||
| TGL(mg/dL): 69.0 (51.7–89.5) and 81.1 (65.3–124.6); ↑17.5%; p < 0.001 | |||
| Quirós et al. (2015) | 25.0 ± 4.7 and 26.0 ± 3.8; ↑4%; p = 003 | TC (mg/dL):166.0 ± 35.1 and 175.6 ± 38.2; ↑5.7%; p = 0.001 | NA |
| HDL (mg/dL): 52.2 ± 12.2 and 45.4 ± 13.8; ↓13%; p = 0.001 | |||
| LDL (mg/dL): 103.8 ± 28.7 and 112.8 ± 30.3; ↑8.6%; p = 0.013 | |||
| TGL (mg/dL): 70.6 ± 30.7 and 102.3 ± 68.5; ↑44.9%; p < 0.001 | |||
| Deutsch et al. (2015) | 29.1 ± 11.2 and 30.0 ± 11.4; ↑3%; p = 0.024 | TC (mg/dL): 177 ± 38 and 178 ± 42; ↑0.5%; p = NS | NA |
| HDL (mg/dL): 58 ± 22 and 56 ± 29; ↓3.4%; p = 0.006 | |||
| LDL (mg/dL): 93 ± 33 and 97 ± 38; ↑4.3%; p = NS | |||
| TGL (mg/dL): 75 ± 56 and 73 ± 69; ↓2.6%; p = NS | |||
| Fisher et al. (2016) | 24.9 ± 0.5 and 27.7 ± 4.0; ↑11.4% p < 0.01 | NA | NA |
- Data are expressed as before and during treatment and %change. BMI, body mass index; TC, total cholesterol; TGL, triglycerides; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NA, not available; NS, not significant.
| Study | Blood pressure (mmHg) | HB | HT |
|---|---|---|---|
| Giltay et al. (2004) | SBP: 126.62 ± 13.14 and 122.02 ± 10.75; ↓3.6%; p = 0.005 | NA | NA |
| DBP: 79.80 ± 8.00 and 77.72 ± 6.81; ↓2.6%; p = 0.032 | |||
| Berra et al. (2006) | NA | NA | NA |
| Jacobeit et al. (2007) | NA | 13.8 ± 1.22 and 15.1 ± 0.68, ↑9.4%, p < 0.05 | 41.0 ± 3.6 and 44.3 ± 1.4, ↑8%, p < 0.05 |
| Mueller et al. (2007) | SBP:129.43 ± 13.38 and 133.71 ± 11.33;↑3.3%; p = 0.04 | 13.17 ± 1.35 and 14.83 ± 1.15; ↑12.6%; p < 0.0001 | 41.50 ± 3.33 and 46.25 ± 3.35, ↑11.4%, p < 0.0001 |
| DBP: 81.14 ± 8.14 and 84.00 ± 5.25; ↑3.5%; p = 0.02 | |||
| Jacobeit et al. (2009) | NA | 13.6 ± 1.2 and 16.0 ± 1.5; ↑17.6%; p < 0.05 | 41.0 ± 4.0 and 46.0 ± 4.0; ↑12.1%; p < 0.05 |
| Mueller et al. (2010) | SBP: 129.3 ± 12.0 and 135.1 ± 10.3: ↑4.4%; p = 0.01 | 13.2 ± 1.5 and 14.6 ± 0.7; ↑10.6%; p < 0.0001 | 41.2 ± 3.2 and 44.7 ± 3.7; ↑8.4%; p < 0.0001 |
| DBP: 81.0 ± 8.1 and 83.8 ± 9.1; ↑3.4%; p = 0.21 | |||
| Cupisti et al. (2010) | NA | NA | NA |
| Chandra et al. (2010) | MBP: 87 ± 14 and 91 ± 16; ↑4.5%; p = NS | NA | 40 ± 2 and 45 ± 5; ↑12.5%; p < 0.001 |
| Pelusi et al. (2014) | NA | 13.3 (12.7–13.8) and 14.6 (13.8–15.4); ↑9.7% a | 39.1 (37.7–40.5) and 43.4 (41.5–45.3);↑10.9% a |
| 13.5 (12.9–14.0) and 14.1 (13.2–14.9); ↑4.4% b | 40.6 (39.1–42.2) and 42.6 (40.5–44.6); ↑4.9 b | ||
| 12.6 (12.0–13.2) and 14.3 (13.5–15.2); ↑13.4% c; p < 0.0005 | 38.4 (36.8–39.9) and 43.1 (40.9–45.2); ↑12.2% c; p < 0.0005 | ||
| Wierckx et al. (2014) | SBP: 111.5 ± 12.6 and 115.6 ± 11.7; ↑3.6%; p = 0.05 | NA | 40.8 ± 2.9 and 45.8 ± 3.0; ↑12.2%; p < 0.001 |
| DBP: 70.2 ± 10.5 and 72.5 ± 9.2; ↑3.2%; p = NS | |||
| Quirós et al. (2015) | SBP: 118.2 ± 9.1 and 120.4 ± 13.0; ↑1.8%; p = NS | 13.5 ± 12.1 and 15.3 ± 18.0; ↑13.3%; p < 0.001 | NA |
| DBP: 75.2 ± 8.9 and 76.7 ± 9.4; ↑1.9%; p = NS | |||
| Deutsch et al. (2015) | SBP:120 ± 23 and 123 ± 14; ↑2.5%; p = NS | NA | NA |
| DBP: 72 ± 16 and 70 ± 16; ↓2.7%; p = NS |
- HB, hemoglobin; HT, hematocrit; SBP, systolic blood pressure; DBP, diastolic blood pressure; MBP, mean arterial blood pressure; NA, not available; NS, not significant.
The prevalence of smoking was high in the five studies that reported this cardiovascular risk factor: 37% (Mueller et al., 2007), 30% (Chandra et al., 2010), 51% (Pelusi et al., 2014), 20% (Wierckx et al., 2014), and 60% (Quirós et al., 2015). Slight but significant increases in BMI were observed in most studies, ranging from 1.3 to 11.4%, during testosterone treatment (Table 2). In three studies using bioimpedance (Berra et al., 2006) or dual-energy X ray absorptiometry (DXA) (Pelusi et al., 2014; Wierckx et al., 2014) to assess body composition, an increase in lean mass and reduction in fat mass were also detected during androgen treatment. One study observed a lean mass gain of up to 14.6% after only 6 months of androgen therapy (Berra et al., 2006). Among the studies that evaluated body composition by DXA, lean mass gain ranged from 8.5% (Pelusi et al., 2014) to 12.3% (Wierckx et al., 2014) within 12 months of follow-up. Three other studies reported no changes in BMI during testosterone therapy (Chandra et al., 2010; Cupisti et al., 2010; Mueller et al., 2010), but one had a sample limited to only 12 patients (Chandra et al., 2010). While Mueller et al. (2010) found no changes in BMI, a significant increase in 4.7% in lean mass was detected.
Seven studies (Giltay et al., 2004; Mueller et al., 2007, 2010; Chandra et al., 2010; Wierckx et al., 2014; Deutsch et al., 2015; Quirós et al., 2015) assessed the impact of various testosterone formulations on blood pressure levels in transgender men. One of these studies (Mueller et al., 2007) observed a slight increase in systolic and diastolic blood pressure during testosterone treatment; two studies found a modest increase only in systolic blood pressure (Mueller et al., 2010 and Wierckx et al., 2014) while two did not observe significant changes (Deutsch et al., 2015 and Quirós et al., 2015) and one reported a decrease in systolic and diastolic blood pressure levels (Giltay et al., 2004). In one study, blood pressure was expressed as mean arterial pressure, and no changes were observed (Chandra et al., 2010) (Table 3).
Data on changes in lipid profile are presented in Table 3. While decreases in HDL-c and increases in LDL-c were consistently observed in most studies, unfavorable changes in total cholesterol and triglycerides were less uniform and varied between studies.
Eight studies assessed hemoglobin and/or hematocrit levels (Table 3), and all observed a relationship between testosterone administration and increased hemoglobin and hematocrit levels. Reference values for males were not exceeded in any of these studies, as inferred from the study results or, in three studies, explicitly reported (Jacobeit et al., 2007; Mueller et al., 2007, 2010); discontinuation of androgen therapy was not required. The increase in hemoglobin ranged from 4.9% (Pelusi et al., 2014) to 12.5% (Chandra et al., 2010), while the increase in hematocrit levels ranged from 4.4% (Pelusi et al., 2014) to 17.6% (Jacobeit et al., 2009). The only study performing a separate analysis of different testosterone formulations observed smaller increases in hemoglobin and hematocrit in patients treated with testosterone gel (Pelusi et al., 2014).
The six studies assessing liver function showed only slight or no changes during testosterone administration (Table 2). The increases did not exceed male reference values.
Adverse effects
Severe adverse effects were not reported by any of the 12 studies. However, in one study (Mueller et al., 2007), two patients developed hypertension. Arterial blood pressure returned to normal after cessation of testosterone therapy. In another study (Wierckx et al., 2014), two patients developed erythrocytosis (hematocrit levels above 52%, the upper limit of the normal adult male range) after 9 and 12 months of treatment. In the same study, two subjects were switched from testosterone undecanoate to short-acting testosterone esters after 9 and 12 months, mainly because of muscle and joint aches.
In one study, hair loss was associated with the duration of androgen administration, and was present in about 50% of subjects after 13 years (Giltay et al., 2004). Hair loss was not associated with BMI, blood pressure or lipid profile during the first 3–4 months of androgen therapy. Acne was observed in approximately 15% of patients in two studies (Mueller et al., 2007, 2010).
Discussion
The present systematic review summarizes the literature on the effects of testosterone therapy on BMI, blood pressure, metabolic profile, hematocrit, hemoglobin, and liver enzymes in transgender men. Although only a few studies are available, there is agreement regarding the relative safety of short term testosterone treatment in transgender men (Gooren et al., 2008; Irwig, 2016). The studies analyzed describe various periods of exposure to different testosterone formulations and show only modest unfavorable changes in BMI, lipid profile, and hematological variables.
Indeed, until now, only two meta-analyses of data on testosterone treatment of transgender men have been published (Elamin et al., 2010; Klaver et al., 2016): the first of these studies addressed cardiovascular risk events, blood pressure, and lipid profile (Elamin et al., 2010), and the second examined the effects of cross-sex hormone therapy on body composition (Klaver et al., 2016). Elamin et al. (2010) found mild changes in lipids and in blood pressure levels. However, because of methodological limitations, the data were inconclusive regarding relevant outcomes such as mortality, stroke, myocardial infarction, or venous thromboembolism.
Sex steroid hormones are important determinants of regional fat deposition (Elbers et al., 1999). In this study, only a very modest increase in BMI was observed in most, but not all, studies. This finding is in line with data from the European Network for Investigation of Gender Incongruence (ENIGI), which did not report changes in BMI after 1 year of parenteral testosterone treatment (Van Caenegem et al., 2015). In turn, a recent meta-analysis about the effects of testosterone on body composition in 354 transgender men found an increase in total body weight and lean body mass, and a decline in body fat (Klaver et al., 2016). These results may reflect the use of different androgen formulations and/or treatment duration, which ranged from 3 to 24 months. Interestingly, in our review, the studies analyzing body composition also reported a significant increase in lean mass, suggesting a link between increased BMI and lean mass gain.
Regarding the effects of testosterone treatment on blood pressure in transgender men, mild, clinically irrelevant, or no changes were found in most studies. However, one study reported two case of hypertension, with return to normal blood pressure values after cessation of testosterone therapy. Other studies beyond the ones included in the present review have reported similar increases (Emi et al., 2008) and decreases (Gooren et al., 2008) in blood pressure during/following testosterone treatment in transgender men. This discrepancy could be explained by differences in ethnicity or treatment duration. In addition, no data are yet available regarding older patients or hypertensive transgender men receiving cross-sex hormone therapy.
In the present review, the effect of the various testosterone formulations seemed to be more unfavorable for HDL- and LDL-cholesterol as compared to the other lipid profile-related variables. Indeed, discrepant results were observed for total cholesterol and triglycerides. A previous systematic review and meta-analysis addressing the effects of different androgen formulations on cardiovascular risk has reported an increase in triglycerides after testosterone treatment of transgender men (Elamin et al., 2010).
In addition, considering that the treatment of hypogonadal men involves testosterone replacement in similar doses to those administered to transgender men, data regarding that population may also be relevant to assess the effects of testosterone treatment on lipid profile in transgender men. In this sense, a recent review and meta-analysis including 16 studies and 1921 hypogonadal individuals to evaluate the efficacy and safety of testosterone therapy found that exogenous testosterone was associated with lower serum concentrations of total cholesterol only after long-term therapy (Guo et al., 2016). In turn, a study with 120 hypogonadal men showed an increase in LDL levels after 8 years of treatment (Permpongkosol et al., 2016). Therefore, further long-term clinical trials are needed to confirm the impact of androgen treatment on LDL-cholesterol in transgender men.
Erythropoiesis is one of the many physiological functions stimulated by androgens. Studies suggest that androgens act indirectly to stimulate erythropoietin and directly to stimulate erythropoiesis in bone marrow (Shahani et al., 2009; Bachman et al., 2014). In a recent meta-analysis about the adverse effects of testosterone treatment in hypogonadal men, Fernandez-Balsells et al., (2010) found a positive association between increased testosterone levels and higher hematocrit and hemoglobin, which is in agreement with our review. In fact, we found a mild increase in hematocrit and hemoglobin in some studies; the final values, however, were still within the physiological male range. While a study of long-term and side effects of cross-sex hormone therapy in transgender people did not observe cardiovascular events or deep venous thrombosis in transgender men (Wierckx et al., 2012), increases in hematocrit and hemoglobin related to sex-hormone therapy may represent a concern in older transgender men, in those who are obese, dyslipidemic and/or hypertensive, and therefore at higher risk of cardiovascular events such as thrombosis. Indeed, the goal of cross-sex hormone treatment for transgender men should be to achieve equal or slightly lower male physiological levels of serum testosterone to maintain adequate androgenization (Hembree et al., 2009; Traish & Gooren, 2010; Meriggiola & Gava, 2015).
Regarding liver enzymes, the studies included in our systematic review again showed inconsistent results; AST and ALT levels remained unchanged in some studies, while others reported increases in these variables after testosterone treatment. It should be noted, however, that the reported increases were not clinically relevant (Mueller et al., 2010). Our data are consistent with the results reported by other studies showing no significant alterations (Meriggiola et al., 2008; Traish & Gooren, 2010). Nevertheless, a retrospective study of morbidity aspects in 293 transgender men found a transient elevation (<6 months) in liver enzymes in 13 patients, and a persistent elevation (more than 6 months) in 20 patients during testosterone treatment (Van Kesteren et al., 1997). Therefore, long-term studies are needed to assess the impact of androgen therapy on liver function. Meanwhile, monitoring liver enzymes might be restricted to transgender men presenting other medical conditions related to risk of potential liver dysfunction, such as obesity, dyslipidemia, insulin resistance associated to type 2 diabetes.
Limitations of the present systematic review are the small number of published articles and the small sample sizes. However, similar analyses are not available in the literature, and since the impact of androgen treatment on transgender men is not fully established, the present findings may serve as basis for future studies.
In conclusion, the present review showed that exogenous testosterone administration to transgender men was associated with modest increases in BMI, hemoglobin/hematocrit, and LDL-cholesterol and decreases in HDL-cholesterol. Less consistent results were observed for blood pressure values, total cholesterol, triglycerides and liver enzymes. Overall, the studies included in this review had low quality of evidence, mainly because of the absence of clinical randomized trials or controlled studies and to low sample sizes, suggesting that caution is still necessary when prescribing hormone therapy to transgender men, especially those presenting risk factors. Further studies are needed in order to verify the impact of age and cardiometabolic comorbidities on safety, and to investigate if specific formulations are more adequate for transgender men at metabolic or cardiovascular risk.
Acknowledgements
We thank Cintia Tusset for her assistance with data collection. Support for this work was provided by Conselho Nacional de Desenvolvimento Científico e Tecnológico/Brazilian National Institute of Hormones and Women's Health (CNPq INCT 573747/2008-3), Brazil; the funding sources had no influence in the writing or decision to submit the article for publication.
Disclosures
The authors declare no competing interests.
Authors’ Contributions
IV was involved in data collection, analysis/interpretation, drafting of article. TF was involved in data collection, analysis/interpretation, critical revision of article. PZ was involved in data analysis/interpretation, critical revision of article. PMS was involved in concept/design, data analysis/interpretation, drafting of article, critical revision of article. All authors read and approved the final manuscript.
