Systematic Review of the Long-Term Effects of Transgender Hormone Therapy on Bone Markers and Bone Mineral Density and Their Potential Effects in Implant Therapy

This study seeks to evaluate the long-term effects of pharmacologic therapy on the bone markers and bone mineral density of transgender patients and to provide a basis for understanding its potential implications on therapies involving implant procedures. Following the referred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and well-defined PICOT (Problem/Patient/Population, Intervention, Comparison, Outcome, Time) questionnaires, a literature search was completed for articles in English language, with more than a 3 year follow-up reporting the long-term effects of the cross-sex pharmacotherapy on the bones of adult transgender patients. Transgender demographics, time under treatment, and treatment received were recorded. In addition, bone marker levels (calcium, phosphate, alkaline phosphatase, and osteocalcin), bone mineral density (BMD), and bone turnover markers (Serum Procollagen type I N-Terminal pro-peptide (PINP), and Serum Collagen type I crosslinked C-telopeptide (CTX)) before and after the treatment were also recorded. The considerable variability between studies did not allow a meta-analysis. All the studies were completed in European countries. Transwomen (921 men to female) were more frequent than transmen (719 female to male). Transwomen’s treatments were based in antiandrogens, estrogens, new drugs, and sex reassignment surgery, meanwhile transmen’s surgeries were based in the administration of several forms of testosterone and sex reassignment. Calcium, phosphate, alkaline phosphatase, and osteocalcin levels remained stable. PINP increased in transwomen and transmen meanwhile, CTX showed contradictory values in transwomen and transmen. Finally, reduced BMD was observed in transwomen patients receiving long-term cross-sex pharmacotherapy. Considering the limitations of this systematic review, it was concluded that long-term cross-sex pharmacotherapy for transwomen and transmen transgender patients does not alter the calcium, phosphate, alkaline phosphatase, and osteocalcin levels, and will slightly increase the bone formation in both transwomen and transmen patients. Furthermore, long-term pharmacotherapy reduces the BMD in transwomen patients.


Introduction
The term "transgender" describes a population experiencing incongruence between their physical sex characteristics (assigned gender) and their gender identity (the extent to which people experience themselves to be like others of one gender) [1]. In some instances, as a result of the incongruence Meanwhile, the BTMs for bone resorption can be obtained from serum (collagen type I cross-linked C-telopeptide (sCTX), carboxyterminal telopeptide of type I collagen (ICTP), tartrate-resistant acid phosphatase (TRACP), tartrate-resistant acid phosphatase 5b (s-TRACP 5b), and urine (collagen type I crosslinked N-telopeptide (uNTX)), collagen type I cross-linked C-telopeptide (uCTX), total pyridoline (uPYD), and total deoxypyridoline (uDPD) [23].
Nearly 0.6% of U.S. adults identify themselves as transgender [24], and global transgender prevalence has been reported in the United Kingdom (0.5%) [25], Belgium (0.6%) [26], and the Netherlands (0.9%) [27]. It has been reported that the transgender population encounters difficulties regarding access to health care [28] and that healthcare providers are not suitably prepared to adequately serve the transgender community [28].
Focusing on the oral care for transgender patients, the literature in the field is very scarce and mainly centered around dental fear [29] and dental education related to the topic [30][31][32], with only one clinical case published that related to full mouth rehabilitation of a transgender patient [33]. Unfortunately, there are no references to the bone parameters of the transgender population linked to oral and maxillofacial surgery or implant procedures, and, as Ludwig et al. (2018) stated, "We cannot provide evidence-based dental care to a subset of the population, if that population has yet to be studied." [34].
Pharmacologic therapy may influence the bone structure of transgender patients receiving or rehabilitated with titanium implants or having oral and maxillofacial surgical procedures. Furthermore, the long-term effects of these therapies on bone mineral density, bone metabolism, and bone resorption are unknown.
Therefore, the goals of this systematic review were twofold: • First, to answer the following PICOT (Problem/Patient/Population, Intervention, Comparison, Outcome, Time) question: In adult transgender patients (transwoman and transmen), receiving long-term pharmacologic therapy, are the bone markers and bone mineral density affected differently? • Second, to provide a theoretical basis for a better understanding of the implications of the long-term pharmacologic therapy in the adult transgender patient on therapies involving orthopedic or dental implants.

Materials and Methods
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed [35]. Electronic and manual searches were completed in Medline, EMBASE, and PubMed from January 2018 to December 2018 for the following search terms: "transgender AND bone health" OR "transgender AND cortical bone" OR " transgender AND trabecular bone" OR "transgender AND bone structure" OR "transgender AND bone metabolism" OR "transgender AND bone mineral density" OR "transgender AND bone loss" OR "cross-sex hormone AND bone health" OR "pharmacotherapy AND transgender" OR "cross-sex AND cortical bone" OR "cross-sex AND trabecular bone" OR "cross-sex AND bone structure" OR "cross-sex AND bone metabolism" OR "cross-sex AND bone mineral density" OR "cross-sex AND bone loss" OR "cross-sex AND bone health".
An additional manual search was completed within the references provided in the included manuscripts to identify other reports not returned by the electronic searches.

Study Inclusion Criteria
Long-term clinical studies completed in adult transgender populations (transwomen and transmen) that received pharmacological treatment were included as per the following inclusion criteria: • Papers published in the English language from January 1990 through to December 2018; • Publications reporting clinical studies with three or more years follow-up; • Publications reporting the biological effects of cross-sex pharmacologic therapy in the bone markers, bone metabolism, and bone mineral density of transgender patients; • Papers reporting the pharmacologic therapy used for transgender adults (including both retrospective and prospective studies).
If a study appeared duplicated, just the study that featured the most extensive follow-up, or the most recent study, was included.

Study Exclusion Criteria
• Articles published in languages other than English; • Publications reporting less than three years of follow-up; • Publications detailing the effects of cross-sex pharmacologic therapy in teenagers or young transgender patients; • Publications reporting effects of cross-sex pharmacologic therapy not including the effects on bone • Animal studies and in-vitro studies; • Other systematic reviews and meta-analyses; • Duplicated studies; • Case reports.
Two reviewers (R.D. and G.R.) completed initial independent searches using the search terms. After the initial search, the titles and abstracts of the returned articles were read and articles that satisfied the inclusion criteria were selected. Afterward, the included articles were read in full and evaluated for final eligibility. The discrepancies between the reviewers were solved with the participation of a third, blinded reviewer (P.S.).

Data Extraction
The following information was extracted from all the included manuscripts:

•
Transgender demographics: Number of males to female (transwomen) or female to male (transmen) patients included in each study; • The duration of hormone-therapy treatment expressed in years or months; • Type of hormone received by the patient and dosage.

Primary Outcomes
• Changes in bone metabolism marker levels calcium (mmol/L), phosphate (mmol/L), alkaline phosphatase (U/L), and osteocalcin (µg/L). Changes were measured at baseline and after treatment or at different time points for each of the bone metabolism markers.

•
Changes in the BTMs including Serum Procollagen type I N-Terminal pro-peptide (PINP) (ng/mL) for evaluation of the bone formation, and Serum Collagen type I cross-linked C-telopeptide (CTX) (ng/mL) for evaluation of bone resorption. Changes were recorded at baseline and at different time points.

•
Changes in the BMD. BMD values were registered at baseline and after the completion of treatment or at baseline and at different time points.
The obtained data were organized in tables ordered by year of publication from the oldest to the newest publication.

Risk of Bias
To determine the risk of bias in the included studies, the risk of bias assessment tool (RoB 2 tool) [36] was used. Five elements (domains) were assessed for each included manuscript including the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. Each domain was graded, and the risk of bias was scored as low risk of bias, some concerns, or high risk of bias, following the descriptors of the RoB 2 tool [36].

Statistical Analysis
Descriptive statistics, percentages, mean and standard deviations, and forest plot graphics were used for the presentation of the data. Meta-analysis was completed, if applicable, using a random effects model. Meta-analysis software (Comprehensive-Meta-analysis 3.0, Biostat, NJ, USA) was used for the statistical comparisons.

Results
The initial search returned 564 articles. After reading the titles and abstracts, 471 articles were excluded. The remaining 93 articles were read in full and 84 articles were excluded (according to the exclusion criteria). Finally, nine manuscripts that fulfilled the inclusion criteria were included for this review ( Figure 1).

Statistical Analysis
Descriptive statistics, percentages, mean and standard deviations, and forest plot graphics were used for the presentation of the data. Meta-analysis was completed, if applicable, using a random effects model. Meta-analysis software (Comprehensive-Meta-analysis 3.0, Biostat, NJ, USA) was used for the statistical comparisons.

Results
The initial search returned 564 articles. After reading the titles and abstracts, 471 articles were excluded. The remaining 93 articles were read in full and 84 articles were excluded (according to the exclusion criteria). Finally, nine manuscripts that fulfilled the inclusion criteria were included for this review ( Figure 1).

Transgender Demographics and Time under Treatment
Nine long-term studies featuring subjects who had undergone more than three years of pharmacotherapy treatment were included. Three studies were featured only male (M) to female (F) transgender patients, two studies were only included F to M transgender patients, and four studies featured both groups (M to F and F to M) [37][38][39][40][41][42][43][44][45]. In addition to the pharmacotherapy, all the patients received sex reassignment surgery. The shortest period of evaluation was 3.5 years [40], and the longest was 18 years [45]. The total population studied was 1640 patients (921 M to F and 719 F to M transgender patients) ( Table 1).

Transgender Demographics and Time under Treatment
Nine long-term studies featuring subjects who had undergone more than three years of pharmacotherapy treatment were included. Three studies were featured only male (M) to female (F) transgender patients, two studies were only included F to M transgender patients, and four studies featured both groups (M to F and F to M) [37][38][39][40][41][42][43][44][45]. In addition to the pharmacotherapy, all the patients received sex reassignment surgery. The shortest period of evaluation was 3.5 years [40], and the longest was 18 years [45]. The total population studied was 1640 patients (921 M to F and 719 F to M transgender patients) ( Table 1).

Type of Hormone Received and Dosage
Different treatment modalities and different dosages were received by the patients. No comparison was possible between studies based on the used drugs. Variations between hormone type/dose/time as well as variations to the treatment due to adjustments by the provider and the individual patient's characteristics precluded the comparisons.
• F to M Several forms of testosterone are administered

Bone Metabolism Marker Levels Before and After the Treatment
Calcium, phosphate, alkaline phosphatase, and osteocalcin were used in three of the studies. Data could not be compared due to the differences in population and methods of detection. In addition, the measuring units varied between mg/dL and mmol/L ( Table 2).

The Two Main Bone Turnover Markers
Both markers (PINP and CTX) were evaluated in four studies [40][41][42][43]. Meanwhile, CTX alone was evaluated in one study [45] (Table 3).  [38]. No differences in calcium, phosphate, alkaline phosphatase, or osteocalcin in the patients receiving hormone therapy compared to controls by Ruetsche et al. 2005 [39]. F to M: No differences in calcium or phosphate were observed. Reduction in alkaline phosphatase from 36-120 (Ukat/L) in controls compared to 6.3 ± 1.5 (3.4-11.4) in tests. No changes in osteocalcin in the patients receiving hormone therapy compared to controls [39]. No changes were found in alkaline phosphatase and osteocalcin in patients receiving hormone therapy compared to male and female controls [45]. Table 3. Bone turnover markers were evaluated in five of the articles. Variable results were obtained for Serum Procollagen type I N-Terminal propeptide (P1NP) and for Serum collagen type I crosslinked C-telopeptide (CTX). P1NP increased in transwomen and transmen. CTX showed similar values before and after treatment. PINP in M to F: Bone formation marker PINP increased from 32-49 ng/mL [40], and from 102-125 ng/mL [43]. PINP in F to M: Bone formation marker PINP increased from 40 ± 12-50 ± 24 ng/mL [42]. CTX in M to F: Bone resorption marker, decreased in one study from 0.36 ± 0.16-0.24 ± 0.14 ng/mL [40], and increased in another study from <0.58 ng/dL to 0.62-1.24 ng/dL [43]. CTX in F to M: Bone resorption marker CTX increased from 0.20 ± 0.10-0.36 ± 0.15 ng/mL [42], and was maintained within the normal range [41,43], and decreased from 400 ± 124-302 ± 190 ng/mL [45].
The method for the evaluation of the BMD was dual-energy X-ray absorptiometry (DXA) (90%). In addition, peripheral quantitative computed tomography (pQCT) was used for the evaluation of the bone architecture (10%) [40,41] (Table 1).
When evaluating the BMD variations related to the time of pharmacotherapy in transwomen (M to F), the range was about 0.4 gr/cm 2 over a 14 year period. Meanwhile, the BMD variation in transmen (F to M) was in the range of 0.6 gr/cm 2 over a nine year period (Figure 2). The BMD was evaluated at different anatomical locations: lumbar spine (100%), femoral neck (60%), and less frequently at the total hip, distal forearm, or the whole body [37][38][39][40][41][42][43][44][45].
The method for the evaluation of the BMD was dual-energy X-ray absorptiometry (DXA) (90%). In addition, peripheral quantitative computed tomography (pQCT) was used for the evaluation of the bone architecture (10%) [40,41] (Table 1).
When evaluating the BMD variations related to the time of pharmacotherapy in transwomen (M to F), the range was about 0.4 gr/cm 2 over a 14 year period. Meanwhile, the BMD variation in transmen (F to M) was in the range of 0.6 gr/cm 2 over a nine year period (Figure 2). The BMD value for the longest evaluation period for transwomen (17 year follow-up) was 1.08 gr/cm 2 and for transmen was 1.19 gr/cm 2 (18 years follow-up).

Risk of Bias Assessment
None of the nine studies were randomized (high risk of bias), there were no deviations from the intended interventions (low risk of bias), there were eight studies missing at least one of the outcomes' data (high risk of bias), when measured, the outcomes were properly assessed (low risk of bias), and there were no problems with the selection of the reported result (low risk of bias). Overall, there were some concerns related to missing information on outcomes and the lack of randomization (Figure 3). The BMD value for the longest evaluation period for transwomen (17 year follow-up) was 1.08 gr/cm 2 and for transmen was 1.19 gr/cm 2 (18 years follow-up).

Risk of Bias Assessment
None of the nine studies were randomized (high risk of bias), there were no deviations from the intended interventions (low risk of bias), there were eight studies missing at least one of the outcomes' data (high risk of bias), when measured, the outcomes were properly assessed (low risk of bias), and there were no problems with the selection of the reported result (low risk of bias). Overall, there were some concerns related to missing information on outcomes and the lack of randomization (Figure 3).
. Figure 3. Risk of bias assessment. From the nine included studies, eight had some concerns and only one study showed an overall high risk of bias. S (study identification), Weight (sample size/number of patients per study).

Discussion
The purpose of this systematic review was to evaluate the bone markers and BMD of transwomen (M to F) and transmen (F to M) patients after long-term pharmacotherapy treatment for feminization or virilization with or without sex reassignment surgery.
The obtained information might provide the clinicians with a reference for the bone characteristics of transgender patients receiving long-term hormone therapy and a baseline for studying the future implant site and the peri-implant bone characteristics in this patient population.

Transgender Demographics and Time under Treatment
These data were gathered to understand the trends of the transgender population under pharmacotherapy supplementation. The results showed that the demographics of adult transgender M to F and F to M populations were comparable and the long-term follow-up studies were 100% based on European populations. The longest reported follow-up period extended up to 18 years [45].

Discussion
The purpose of this systematic review was to evaluate the bone markers and BMD of transwomen (M to F) and transmen (F to M) patients after long-term pharmacotherapy treatment for feminization or virilization with or without sex reassignment surgery.
The obtained information might provide the clinicians with a reference for the bone characteristics of transgender patients receiving long-term hormone therapy and a baseline for studying the future implant site and the peri-implant bone characteristics in this patient population.

Transgender Demographics and Time under Treatment
These data were gathered to understand the trends of the transgender population under pharmacotherapy supplementation. The results showed that the demographics of adult transgender M to F and F to M populations were comparable and the long-term follow-up studies were 100% based on European populations. The longest reported follow-up period extended up to 18 years [45].

Bone Metabolism Markers and Their Potential Effect on Implant Therapy
Calcium and phosphates are responsible for calcium homeostasis and participate in the acid-base balance and also facilitate the release of growth factors embedded in bone [38]. The action of the osteoclasts on the calcified bone matrix facilitates its dissolution and releases calcium ions into the blood to form blood calcium [46]. In a parallel action, the calcium contained in the blood flow can be deposited onto the bone to form bone calcium, mediated by the osteoblasts. These phenomena are regulated by enzymes and hormones (vitamin D, calcitonin, parathyroid hormone, and other metabolites) [47].
In the presence of metabolic imbalances produced by the cross-sex pharmacologic therapies, the blood calcium balance system can be altered, thereby playing an important role as an ethiological factor for pharmacologically induced osteoporosis [48].
These pharmacologic therapies can also alter the phosphorus/calcium ratio. The reduction of phosphorus can alter the calcium absorption, while increased phosphorus concentrations can increase the oxidative stress as well as the hormonal balance between phosphates, calcium, and vitamin D. This might be conducive to adverse effects on mineral metabolism and increased bone loss [49,50].
Alkaline phosphatase and osteocalcin are both bone formation markers [50,51]. Bone alkaline phosphatase (ALP) regulates bone mineralization [51]. There are also liver ALP isoforms that differ only by posttranslational modifications. However, the immunoassays with monoclonal antibodies better recognize bone isoforms [52]. Meanwhile, osteocalcin is a major non-collagenous protein synthesized by osteoblasts and odontoblasts, and its circulating levels are highly specific for bone formation. It is degraded and excreted by the kidneys [53].
Calcium and phosphate are an essential part of bone metabolism, and their depletion can result in reduced bone mineral density, changes in the bone structure (increased trabecular spacing and reduction of cortical bone thickness), osteoporosis, and delayed osseointegration [54]. Not all osteoporotic signs are the result of calcium, phosphate, or vitamin D deficiency, but these are important factors associated with optimal bone health [55,56]. Their depletion can also result from low intake, vitamin D deficiency, and changes in the metabolism induced by disease or medications [57,58].
The results of the present review showed that calcium and phosphate levels, as well as alkaline phosphatase and osteocalcin, remained within similar values after the long-term pharmacotherapy for transgender M to F and F to M patients, thereby demonstrating that the administered therapies had minimal effects on calcium/phosphate balance and alkaline phosphatase and osteocalcin levels [37,45].

Two Main Bone Turnover Markers
Following the recommendations by the International Osteoporosis Foundation (IOF) and the International Federation of Clinical Chemistry (IFCC) for the quantification of the bone turnover (bone resorption and bone formation processes), the N-terminal pro-peptide of type I procollagen (PINP) and C-telopeptide of type I collagen (CTX-I) were recorded [59]. Respectively, PINP measures bone formation and CTX-I measures bone resorption [59].
During transgender hormone therapy, the goal is similar to the goal of the antiresorptive treatment-to lower the PINP by at least 10 ng/mL and <35 ng/mL during bone resorption. Meanwhile, during bone formation, the goal is to raise the PINP by at least 10 ng/mL to achieve a level of >69 ng/mL [60].
Therefore, it seems that the long-term administered pharmacotherapy for M to F and F to M transgender patients can produce a slight increase in the bone formation rates evaluated with the PINP [40][41][42][43]. It should also be considered that these PINP values can change over time, induced by factors such as age, metabolism changes, and non-compliant treatment interruption [37][38][39][40][41][42][43][44][45].

Bone Mineral Density (BMD)
The BMD condition of M to F and F to M transgender patients receiving long-term cross-sex pharmacotherapy is contradictory. For Sosa et al. (2003) [38], Ruetsche et al. (2005) [39], Van Caenegem et al. 2012 [42], and Wiepjes et al. 2018 [44], both M to F and F to M patients will possess a stable or increased BMD compared to matched male or female controls.
The authors explain their findings based on the protective effects of estrogens in M to F patients against bone resorption, mediated by increased serum levels of estradiol [38], mediated by IGF1 (insulin-like growth factor) [39]. Meanwhile, in F to M patients, the preservation or increase of the BMD could be produced by the long-term effects of testosterone, reduced estrogen levels, and a muscle mass increase, which, all together, might result in reduced resorption rates [38,39,42,44].
On the other hand, according to Schlatterer et al. (1998) [43], the BMD of M to F patients was reduced, and signs of osteoporosis of the lumbar spine and distal arm were observed, but F to M patients did not show reduced BMD. Apparently, the effects of muscle mass reduction, non-compliance to the treatment (which can result in androgen deficiency), inadequate estrogen dosage, and sedentary lifestyles produced the bone resorption experienced by the M to F transgender patients [37,[39][40][41]43] (Table 1).

BMD Changes and their Potential Relationship to Dental Implants
The local properties of the future implant bed (bone mineral density and bone structure), as well as the primary implant fixation (primary stability), are essential factors that can reduce micromotion and will allow immediate loading protocols [61,62]. Moreover, bone quality, quantity, implant geometry, and surgical technique have been considered factors that can influence the presence of micromotion or the presence of implant stability [63]. When the bone mineral density is low, the primary implant stability cannot be achieved unless certain modifications are completed during the implant bed preparation, for example-the use of specific implant designs (tapered, self-tapping) [64], under-drilling, bone condensation, or osseo-densification techniques [65].
There is also a link between low bone density, osteoporosis, and implant failure [66] and low bone density and lower osseointegration [67]. However, recent systematic reviews showed that the implant survival rates and marginal bone loss of unloaded implants were similar in implants inserted in low-density bone compared to implants inserted in normal-density bone [68,69]. Furthermore, it was reported that low bone mineral density values found in a group of patients with osteoporosis and osteopenia did not influence the implant osseointegration after 24 months of follow-up [70]. However, when the implant requires higher primary stability (i.e., immediate loading protocols, single-body implants), the bone mineral density is a factor that should be considered before the load protocols are applied [71]. Marquezan et al. demonstrated that there is a direct association between BMD and primary implant stability (as the bone density increase, the primary stability increase). [71].

Hypothesis for the Effects of Hormone Therapy for Transgender Patients on BMD and Its Potential Relation to Dental Implants
It seems that there is some risk of reduced BMD in M to F transgender patients receiving long-term cross-sex pharmacotherapy. Therefore, when performing dental implant procedures in such patients, the precautions followed in osteoporotic patients should be considered [66][67][68][69]. In addition, based on the bone changes observed in the present study over the long-term follow-up, it seems reasonable to monitor bone parameters before procedures involving dental implants in transgender patients [72]. Finally, when considering the risk factors for long-term implant survival (diabetes mellitus, age, smoking, and immediate loading), the BMD was the most critical factor determining implant survival (lower BMD values resulted in lower implant survival rates) [73].

Strengths and Limitations of the Present Work
The strengths of the present paper are that this is the first systematic review that has compiled the effects of the cross-sex pharmacotherapy on bone metabolism markers, BTMs, and BMD. Additionally, strict inclusion and exclusion criteria and adequate calibration were followed by the investigators involved in the data collection and data analysis. Finally, a hypothesis for the potential effects of long-term hormone therapy on dental implant therapy was provided.
There are limitations to the present work-firstly, the low number of studies included and the preclusion of any statistical comparisons; secondly, the exclusion of the effects of short-term (less than 3 years) cross-sex pharmacotherapy and their effects in younger patients. Moreover, the multiple variables that, at a certain point, can affect the bone metabolism and structural characteristics of this population (sex-reassignment surgery and aging) further limit the comparison of treatments.

Recommended Future Steps
Given the lack of information regarding the effects of hormone therapy in transgender patients on bone healing, implant osseointegration, peri-implant health, and implant survival, clinical studies compiling such information are recommended.

Conclusions
Within the limitations of this systematic review, the following conclusions can be drawn: • Long-term pharmacotherapy for transgender patients does not alter the calcium, phosphate, alkaline phosphatase, and osteocalcin bone markers. • Long-term pharmacotherapy for transgender patients will slightly increase the bone formation, expressed with increased PINP turnover markers. • Long-term cross-sex pharmacotherapy for M to F transgender patients will produce a slight reduction in bone mineral density.