Content uploaded by Jd Pommier
Author content
All content in this area was uploaded by Jd Pommier on Apr 29, 2019
Content may be subject to copyright.
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
Metabolic syndrome and endocrine status
in HIV-infected transwomen
Jean-David Pommier
a
, Cedric Laou
enan
b,c,d
, Florence Michard
a
,
Emmanuelle Papot
a
, Paul Urios
e
, Anne Boutten
e
, Gilles Peytavin
f
,
Cecile Ghander
g
, Sylvie Lariven
a
, Gerald Castanedo
a
, David Moho
a
,
Rolland Landman
a,b
, Bao Phung
a
, Estela Perez
a
, Zelie Julia
a
,
Diane Descamps
h
, Pascale Roland-Nicaise
i
, Sylvie Le Gac
a
,
Yazdan Yazdanpanah
a,b,d
, Jean Guibourdenche
j,k
and Patrick Yeni
a,b,d
Background: HIV-infected transwomen face multiple specific issues. Economic and social
marginalization, sex work, substance abuse, hormonal consumption and silicone injection
may affect the course of HIV infection and lead to metabolic and endocrine complications.
Methods: A matched case–control study was performed between 2013 and 2015 in a
University Hospital and compared metabolic syndrome (MetS), thyroid and adrenal
functions in HIV-infected transwomen (i.e. cases) and cisgender HIV-infected men (i.e.
controls) matched for age and antiretroviral therapy. The interaction between hormonal
consumption, the course of HIV infection and antiretroviral therapy was also studied.
Clinical and biological data (CD4
þ
cell count, HIV RNA load, antiretroviral plasma
drug concentration, HDL, triglycerides, glucose, cortisol, thyroid stimulating hormone,
free thyroxine, prolactine) were measured.
Results: A total of 292 HIV-infected patients (100 cases and 192 controls) were prospec-
tively included. There was no difference between the two populations in terms of frequency
of MetS, but subclinical hypothyroidism and adrenal insufficiency were more frequent in
cases than in controls with, respectively, 12 vs. 3% (P<0.002) for hypothyroidism and 20
vs. 8% (P<0.001) for adrenal insufficiency. Prolactinemia, only performed in transwomen,
was often elevated (21%) but rarely confirmed as true active hyperprolactinemia (mono-
meric form) (3%). Although hormonal intake was frequent among transwomen (31%), no
impact on antiretroviral bioavailability and efficacy was detected.
Conclusion: In this study, no increase in the prevalence of MetS was detected in
HIV-infected transwomen patients. In contrast, adrenal and thyroid functions
abnormalities were frequent and should be systematically assessed in this popula-
tion. No impact of hormonal intake on antiretroviral bioavailability and efficacy was
detected. Copyright ß2019 Wolters Kluwer Health, Inc. All rights reserved.
AIDS 2019, 33:855–865
Keywords: adrenal insufficiency, HIV, hypothyroidism, metabolic syndrome,
prolactine, transgender, transwomen
a
Infectious and Tropical Diseases Department, Ho
ˆpitaux Universitaires Paris Nord Val de Seine (HUPNVS), AP-HP,
b
IAME, UMR
1137, INSERM,
c
Biostatistics Department, HUPNVS, AP-HP,
d
Paris Diderot University, Sorbonne Paris Cit
e,
e
Biochemistry
Department,
f
Pharmacology Department, HUPNVS, AP-HP,
g
Endocrinology and Vascular Diseases Department, Groupe
Hospitaliser Universitaire, La Piti
e Salep
etrie
`re,
h
Virology Department,
i
Immunology Department, HUPNVS,
j
Hormonal Biology
Department, Ho
ˆpitaux Universitaires Paris Centre (HUPC), AP-HP, and
k
Paris Descartes University, Paris, France.
Correspondence to Jean-David Pommier, Infectious and Tropical Diseases Department, Ho
ˆpitaux Universitaires Paris Nord Val de
Seine (HUPNVS), AP-HP, 21 rue Gutenberg, 75015 Paris, France.
E-mail: jdpommier@yahoo.fr
Received: 7 April 2018; revised: 15 November 2018; accepted: 22 November 2018.
See related paper on page 919
DOI:10.1097/QAD.0000000000002152
ISSN 0269-9370 Copyright Q2019 Wolters Kluwer Health, Inc. All rights reserved. 855
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
Introduction
Transgender is used to refer to people whose sex identity
or expression differs from their sex at birth [1].
Transwomen, defined as individuals who were assigned
male at birth but who identified as women, are known to
be at high risk of HIV infection. General HIV prevalence
in this population has been reported as ranging from 19.1
[2] to 27.7% [3]. Some authors have reported that HIV-
infected transwomen have a lower virological suppression
rate [4], a higher mortality rate [5] and a lower level of
treatment adherence [6] than HIV-infected cisgender
patients. In addition, transwomen face multiple specific
issues, for example economic and social marginalization,
sex work engagement, substance abuse and sexual abuses
[3,6–10]. Furthermore, their physical transition involves
a complex blend of surgical procedures, silicone
injections [11,12] and hormonal intake [13,14], leading
to frequent complications.
HIV-infected transwomen care is thus challenging and
there is a gap in HIV research regarding this particular
population. Various hormonal dysfunctions have been
described in transwomen populations. In particular,
estradiol oral intake has been associated with increased
levels of thyroid hormone-binding globulin in plasma
[15]. In addition, the occurrence of adrenal dysfunction
in this population who often treat siliconoma (silicone
injection complication) with self-administered cortico-
steroids is likely but has not been reported. In HIV-
infected cisgender patients, metabolic syndrome (MetS)
[16], hypothyroidism [17–19] and both adrenal insuffi-
ciency [20] and secondary Cushing syndrome have been
reported [21,22]. In a previous noncomparative study,
performed in HIV-infected transwomen patients, we
observed various endocrine issues and a possible high
frequency of MetS [23].
In the current study, we therefore compared the frequency
of MetS as primary objective and endocrine (adrenal and
thyroid) dysfunctions as secondary objective in HIV-
infected transwomen and HIV-infected cisgender men
matched control patients attending the infectious disease
department of a University Hospital. In addition, the
occurrence of pituitary dysfunction and the consequences
of interactions between estrogen–progestin with antire-
troviral therapy used to treat HIV infection were assessed in
HIV-infected transwomen as secondary objectives.
Methods
Study design and population selection
We performed a prospective-matched case–control study
(cases to controls ratio 1 : 2). Cases were HIV-infected
transwomen patients at least 18 years of age, regularly
attending the Infectious Diseases Department at Bichat
Claude-Bernard University Hospital (Paris, France). Con-
trols were HIV-infected cisgender male patients at least 18
years of age, attending the same Department and were
matched to cases on age (5 years) and on antiretroviral
treatment (currently treated with a protease inhibitor
containing regimen or not, or not treated), and were
randomly selected in the clinical cohort. The control group
was cisgender HIV-infected men patients, to avoid a sex bias,
as the HIV-infected transwomen were male at birth.
Data collection and laboratory testing
Patients were included from March 2013 to March 2015.
Demographics and clinical data such as age, geographic
origin, antiretroviral therapy, corticosteroids, hormones,
other treatment use, alcohol consumption, weight and
height were collected. Waist circumference was measured
to the nearest 0.1cm at the uppermost lateral border of the
right ilium using a measuring tape. After 5 min in the seated
position, blood pressure (BP) was measured and controlled
three times with 1-min interval if elevated (130/
85 mmHg). Steroid use was defined as use of corticosteroids
within the last year, regardless of the routeof administration
and dosage. Current or past hormone use, types and route
of administrations were assessed in the cases. A morning
fasting blood test was performed, including plasmatic
creatinin, aspartate amino transferase, alanine amino
transferase, albumin, triglycerides, total cholesterol (TC),
HDL, LDL calculated according to the Friedwald formula,
glucose (Vista 1500; Siemens, Saint-Denis, France) and
serum hormone concentrations: insulin, cortisol, thyroid
stimulating hormone (TSH), free thyroxine (FT4) (Advia
Centaur; Siemens), adrenocorticotropic hormone
(ACTH) (ELSA-ACTH; IBA, Paris, France). In addition,
HIV-infection parameters were measured: HIV plasma
viral load and CD4
þ
cell count. Chronic Kidney Disease
Epidemiology Collaboration (CKD-EPI) equation was
performed to estimate glomerular filtration rate (GFR).
When morning cortisol levels (before 0930 h) were
measured below the normal range, a Synacthen test was
performed whenever possible, or alternatively, morning
cortisol level was measured again. When TSH levels were
above 4 UI/l, they were measured again at least 1 month
later, in combination with antithyroperoxidase antibodies
(TPO-Abs, Cobas; Roche Diagnostics, Meylan, France).
To measure the hormonal consumption impact, addi-
tional assessments were only performed in the trans-
women population: a serum hormonal investigation
including oestradiol, total testosterone, sex hormone
binding protein (SHBG), luteinizing hormone (LH),
follicle stimulating hormone (FSH) and prolactin
(Cobas). A prolactin concentration above the normal
range (>30 mg/l) was controlled using another assay more
specific for the monomeric bioactive prolactin (Delfia
Prolactin; Perkinelmer, Villebon-sur-Yvette, France).
For values remaining above the normal ranges, prolactin
was measured after polyethylene glycol (PEG) precipita-
tion to exclude macroprolactin and to confirm the
856 AIDS 2019, Vol 33 No 5
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
increase in the monomeric bioactive prolactin; in such
cases, a prolactin chromatography was performed. Free
testosterone was estimated by a formula based on total
testosterone, albumin and SHBG [24]. Finally, plasma
antiretroviral drug concentrations were measured.
Syndrome definitions
MetS was defined according to the International Diabetes
Federation (IDF) in 2005 [25]. True hypothyroidism was
defined by the combination of an increased serum TSH
above 4 mUI/l with FT4 under the nor mal range.
Subclinical hypothyroidism was defined by the combi-
nation of a serum TSH above 4 mUI/l and a normal
serum FT4 [26,27] with or without the presence of TPO-
Ab. Adrenal insufficiency was established by a low cortisol
response (serum cortisol below 500 nmol/l 60min after
250 mg synacthen injection) [28] or, when synacthen was
not available, by a 0830 h cortisol value confirmed at least
twice below 140 nmol/l [29,30]. An analysis of sensitivity
was also performed to assess the frequency of adrenal
insufficiency using a lower cut-off value of 82nmol/l for
basal cortisol. Adrenal insufficiency was classified as
probable in case of a cortisol/ACTH ratio below 20 [31].
Hyperprolactinemia was defined as a serum prolactin
above 30 mg/l confirmed by an increase in the
monomeric bioactive prolactine combining a mono-
meric-specific assay and PEG precipitation and finally
prolactin chromatography [32]. Low concentrations of
antiretroviral drugs were defined according to expected
laboratory values [33], considering medication dosage
and information on schedule intake.
Ethics consideration
All participants gave written informed consent and the
research was approved by the local Ethics Committee
(Comite de Protection des Personnes, Saint-Germain-
en-Laye, France. No. IDRCB: 2012-004382-40).
Statistical analysis
Quantitative variables were expressed as median (inter-
quartile range 25 – 75), and qualitative variables were
expressed as percentages. Cases and controls were
compared using a conditional logistic regression model.
Furthermore, a univariate and multivariate nonmatched
analysis using a logistic regression model were performed
in the overall population to study determinants associated
with MetS. Variables achieving a Pvalue less than 0.20 in
univariate analysis were entered into a multivariate
logistic regression analysis. Using a backward selection
method, a final model in which all determinants had a P
value less than 0.05 was obtained. A univariate and
multivariate analysis using a logistic regression model
were also performed, within the transwomen population
only, to study specific determinants related to this
population associated with hypothyroidism, adrenal
insufficiency and hyperprolactinemia. Analyses were
performed with SAS v9.3 (SAS Institute Inc., Cary,
North Carolina, USA). All tests were two-sided with a
type-I error fixed to 0.05.
Results
A total of 100 transwomen and 192 control patients were
included in this study. 92 cases had two controls and eight
cases had one control. In cases and controls, the median
age was 39 [34– 44] and 41 [36 – 47] years old,
respectively. Ninety-seven percent of cases (n¼97) and
controls (n¼187) were on antiretroviral treatment, and
protease inhibitors were used in 49 (n¼49) and 47%
(n¼91) in each group, respectively.
General clinical and biological features
Cases and controls were statistically different in terms of
geographical origin (P<10
4
). Most cases originated
from South America (97%), whereas controls were mostly
from Europe (57%) and sub-Saharan African countries
(27%). There was no difference between the two groups
regarding the number of patients treated for hypertension,
dyslipidaemia and diabetes, nor in terms of corticoste-
roids use. Regarding HIV risk factors, the number of
patients who have sex with men was higher (80 vs. 49%),
whereas the number of people who inject drugs (0 vs. 4%)
was smaller among cases rather than among controls,
respectively, Pless than 10
4
. A normal GFR was more
frequent among cases than among controls (86 vs. 67%,
respectively, P¼0.02). There were more hepatitis C virus
coinfected patients among controls than cases (12 vs. 2%,
respectively, P¼0.01) (Table 1).
HIV data
There was no statistical difference between cases and
controls in terms of the Centers for Disease Control and
Prevention (CDC) classification for HIV stage, prether-
apeutic HIV viral load, nadir CD4
þ
cell count, actual
CD4
þ
cell count and frequency of undetectable HIV viral
load (<50 copies/ml). The median duration of HIV
infection and length of treatment were higher among cases
than among controls, with 11.3 vs. 7.9 years, P¼0.02, and
8.6 vs. 4.1 years, P¼0.004, respectively (Table 1).
All patients but three cases and five controls were
receiving antiretroviral treatment. No statistical difference
was found in terms of the types of nucleoside reverse
transcriptase inhibitors (NRTIs) and protease inhibitors
used. Non-NRTI (NNRTI) use was more predominant
in cases than in controls with 53 vs. 44% (P¼0.02),
whereas antiintegrase use was insignificantly (P¼0.06)
less frequent with, respectively, 13 vs. 4%.
Metabolic syndrome
Elevated BP and elevated fasting blood glucose were more
frequent among controls than cases (40 vs. 16%, P<10
3
and 19 vs. 6%, P<10
3
, respectively). MetS was assessed
Metabolic syndrome and endocrine status in HIV-infected transwomen Pommier et al. 857
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
858 AIDS 2019, Vol 33 No 5
Table 1. Clinical and biological characteristics of HIV-infected transwomen cases (nU100) and matched cisgender men controls (nU192).
Cases, n¼100 Controls, n¼192 P
Sociodemographic data
Age (years) 39 (34– 44) 41 (36– 47)
Geographical origin
Sub-Saharan Africa 0 51 (27%)
Latin America 97 (97%) 12 (6%)
Asia 2 (2%) 5 (3%) <10
4
Europe 1 (1%) 109 (57%)
Middle East 0 15 (8%)
Consumptions
Alcohol 13 (14%) 6 (3%) 0.002
Hepatitis coinfections
HBV coinfection 11 (11%) 13 (7%) 0.24
HCV coinfection 2 (2%) 23 (12%) 0.01
Risk factors
Sex with women 0 58 (30%)
Sex with men 89 (89%) 104 (54%)
PWID 0 8 (4%) <10
4
Transfusion 1 (1) 2 (1%)
Unknown 10 (10) 20 (11%)
Biochemistry and hormonal data
Serum biochemistry
GFR (ml/min) 85 (86%) 127 (67%)
90 13 (13%) 56 (30%) 0.02
60– 90 1 (1%) 7 (3%)
<60 26 (26%) 26 (14%) 0.01
Elevated AST 14 (22%) 22 (12%) 0.51
Elevated ALT 30 (30%) 58 (30%) 0.96
Elevated GGT
Non-HIV treatment
Antihypertensive treatment 7 (7%) 22 (12%) 0.23
Lipid-lowering treatment 3 (3%) 19 (10%) 0.05
Diabetes treatment 2 (2%) 9 (5%) 0.25
Corticosteroids 13 (13%) 24 (13%) 0.95
Oral 6/13 (46%) 11/24 (46%) –
Topical 2/13 (15%) 12/24 (50%) –
Inhaled 3/13 (23%) 1/24 (4%) –
Injectable 1/13 (8%) 0 –
Antidepressant 5 (5%) 8 (4%) –
HIV infection
CDC classification
A 66 (66%) 136 (71%)
B 5 (5%) 17 (9%) 0.15
C 29 (29%) 37 (20%)
Duration of infection (years) 11.3 (5.3– 10.1) 7.9 (3.1– 12.4) 0.02
Immunovirological data
Nadir CD4
þ
cell count (cells/ml) 218 (151– 249) 249 (120– 371) 0.36
Zenith CD4
þ
cell count (cells/ml) 900 (668– 1159) 820 (623– 1078) 0.28
Current CD4
þ
cell count (cells/ml) 655 (510– 840) 570 (460– 790) 0.12
Pretherapeutic HIV plasma viral load (copies/ml) (log
10
) 5.0 (4.5–4.9) 5.0 (4.4–5.5) 0.82
Undetectable viral load (<50 copies/ml) 79 (81%) 169 (90%) 0.08
HIV therapy
Number of patients treated 97 (97%) 187 (97%) –
Duration of HIV therapy (years) 8.6 (3.9– 8.0) 4.1 (2.0 – 10.5) 0.004
HIV drugs used in regimen
Tenofovir 86/97 (89%) 151/187 (81%) 0.06
Protease inhibitors 49/97 (50%) 91/187 (49%) –
Darunavir 30/97 (31%) 65/187 (35%) 0.30
Atazanavir 13/97 (13%) 20/187 (11%) 0.66
Lopinavir 5/97 (5%) 5/187 (3%) 0.27
Nonnucleoside reverse transcriptase inhibitor 51/97 (53%) 83/187 (44%) 0.02
Rilpivirine 8/97 (8%) 29/187 (16%) 0.06
Efavirenz 32/97 (33%) 37/187 (20%) 0.01
Etravirine 9/97 (13%) 12/187 (6%) 0.39
Integrase inhibitors 4/97 (4%) 25/187 (13%) 0.06
Results are median and [IQR 25– 75] for continuous variables and number n(%) for categorical variables; Pvalues were obtained from conditional
logistic regressions. Alcohol, daily consumption more than 30 g/day; GFR, glomerular filtration rate; GGT, gamma-glutamyl transferase; HBV,
hepatitis B virus; HCV, hepatitis C virus; IQR, interquartile range; LH, luteinizing hormone; PRL, prolactin; PWID, persons who inject drugs.
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
in 259 patients (96 cases and 163 controls) in whom all
components were available. There was no difference in
terms of MetS frequency between cases and controls,
with a respective prevalence of 15% [95% confidence
interval (CI) ¼0.09 –0.24] and 26% (95% CI ¼0.19 –
0.33). Age and history of stavudine use were indepen-
dently associated with MetS [odd ratio (OR) ¼1.09/
year; 95% CI: 1.05– 1.14, P<10
4
and OR ¼3.97; 95%
CI: 1.40–11.74, P¼0.01, respectively] (Tables 2 and 3).
When restricting the analysis to cases, only SHBG median
Metabolic syndrome and endocrine status in HIV-infected transwomen Pommier et al. 859
Table 2. Metabolic syndrome, adrenal and thyroid functions in HIV-infected transwomen cases and matched cisgender men controls
(cases U100, controls U192).
Cases, n¼100 Controls, n¼192 P
Metabolic syndrome and components
Waist circumference >94 cm 38 (38%) 73/184 (40%) 0.78
BMI >30 kg/m
2
16 (16%) 23/191 (12%) 0.35
SBP 130 mmHg or DBP 85 mmHg 19 (16%) 76/190 (40%) 0.0008
HDL <1.03 mmol/l 49/98 (50%) 81/189 (43%) 0.28
FPG >5.6 mmol/l 6 (6%) 40/172 (23%) 0.0007
Triglycerides >1.7 mmol/l 35 (36%) 60/189 (32%) 0.53
Insulin (mUI/l) 8 (5.4 – 10.9) 10.1 (6.5–17.0) 0.003
Metabolic syndrome (ethnicity-specific norms) 15/96 (16%) 42/163 (26%) 0.06
Endocrine functions
Thyroid
TSH (mUI/l) 2.9 (1.8– 3.8) 1.8 (1.3– 2.4) <10
4
FT4 (pmol/l) 12.6 (10.9– 13.4) 11.9 (10.0–14.1) 0.50
TSH >4 mUI/l 17 (17%) 9 (5%) 0.002
T4 <12 pmol/l 42 (42%) 101 (53%) 0.06
Positive TPO-Ab 0/17 2/9
Adrenal
Cortisol (nmol/l) 298 (185– 392) 407 (315– 518) <10
4
ACTH (pmol/l) 3.0 (1– 6) 9.0 (9– 16) <10
4
Low ACTH level (2 pmol/l) 37 (37%) 12 (6%) <10
4
High ACTH level (>12 pmol/l) 2 (2%) 69 (36%) <10
4
Cortisol <140 nmol/l 12 (12%) 4 (2%) 0.002
Cortisol <82 nmol/l 7 (7%) 3 (2%) 0.04
Ratio cortisol : ACTH <20 3 (3%) 11 (6%) 0.26
Low cortisol response to Synacthen 8/31 2/15 0.23
Results are medians and [IQR 25–75] for continuous variables and numbers n(%) for categorical variables, Pvalues were obtained from
conditional logistic regressions. ACTH, adrenocorticotropic hormone; FPG, fasting glucose plasma concentration; FT4, free thyroxine; IQR,
interquartile range; TPO-Ab, thyroperoxydase antibody; TSH, thyroid stimulating hormone.
Table 3. Determinants for metabolic syndrome among HIV-infected transwomen cases (nU96) and cisgender men controls (nU163):
univariate and multivariate analysis.
Metabolic syndrome Univariate analysis Multivariate analysis
Present, n¼57 Absent, n¼202 POR 95% CI POR 95% CI
Clinical data
Age (years) 45 (24–66) 40 (25– 67) <10
4
1.08 (1.04– 1.12) <10
4
1.09 (1.05– 1.14)
Alcohol consumption 7 (12%) 33 (16%) 0.45 0.71 (0.28–1.62)
Corticosteroids use 6 (10%) 30 (15%) 0.41 0.67 (0.24– 1.61)
HCV coinfection 8 (14%) 14 (7%) 0.10 2.19 (0.83–5.42)
HBV coinfection 5 (9%) 18 (9%) 0.97 0.98 (0.31–2.59)
HIV infection
Duration of infection (years) 12.6 (7.7–15.5) 8.4 (3.9 – 12.2) <10
4
1.08 (1.03– 1.13)
Current CD4
þ
cell count (cells/ml) 620 (480– 910) 590 (470–810) 0.31 1.00 (0.99–1.00)
Nadir CD4
þ
cell count (cells/ml) 248 (120– 340) 231 (144–331) 0.71 1.00 (0.99–1.00)
HIV viral load before treatment 5.0 (4.4 – 5.4) 5.0 (4.5– 5.4) 0.60 0.92 (0.78–1.08)
CDC stage: C vs. B/A 13 (23%) 41 (21%) 0.71 1.15 (0.55–2.28)
Undetectable viral load (<50 copies/ml) 47 (87%) 167 (84%) 0.57 0.78 (0.30–1.78)
HIV therapy
Duration of HIV therapy (years) 8.8 (2.1– 13.9) 5.2 (2.5– 10.4) 0.01 1.07 (1.01– 1.13)
HIV drugs used in regimen
Protease inhibitors 27 (47%) 94 (47%) 0.91 1.03 (0.57– 1.86)
Protease inhibitors for more than 1 year 20 (36%) 70 (35%) 0.83 1.07 (0.57–1.98)
NNRTI 27 (47%) 97 (48%) 0.93 0.97 (0.54 – 1.76)
NNRTI for more than one year 21 (37%) 76 (38%) 0.89 0.96 (0.52– 1.75)
History of stavudine use 17 (30%) 24 (12%) 0.002 3.15 (1.54– 6.4) 0.01 3.97 (1.40– 11.74)
History of didanosine use 13 (23%) 36 (18%) 0.39 1.37 (0.65 –2.75)
Results are median and [IQR 25–75] for continuous variables and number n(%) for categorical variables, OR 95% CI ¼odd ratio and 95%
confidence interval. Pvalues were obtained from univariate and multivariate logistic regressions. Alcohol consumption, daily alcohol
consumption; HBV, hepatitis B virus; HCV, hepatitis C virus; IQR, interquartile range; NNRTI, nonnucleoside reverse transcriptase inhibitor.
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
level was found 0.6-fold lower in cases affected by MetS
compared with cases not affected by MetS.
Hypothyroidism
The median TSH level was 1.6-fold higher and the
prevalence of subclinical hypothyroidism was signifi-
cantly higher in cases than in controls (12 vs. 3%,
respectively, P¼0.002). The prevalence of true hypo-
thyroidism was not statistically different although
remained higher in cases than in controls (5 vs. 2%,
P¼0.69, respectively). TC concentration was signifi-
cantly higher in both cases with clinical and subclinical
hypothyroidism (Tables 2 and 4).
Hypothyroidism was associated with an elevated BMI
(OR ¼1.23/kg; 95% CI: 1.04–1.46, P¼0.02) and the
use of steroids (OR ¼7.15; 95% CI: 1.57–35.05,
P<0.01) in the multivariate logistic regression. In
contrast, the duration of antiretroviral therapy was found
to be protective (OR¼0.84/year of ART; 95% CI:
0.70–0.97, P¼0.03).
Adrenal dysfunction
The prevalence of adrenal insufficiency was significantly
higher in cases than in controls (17 vs. 3%, respectively,
P<10
3
). This difference was confirmed (12 vs. 3%,
respectively, P<10
3
) in an analysis of sensitivity in
which a lower cut-off (82 nmol/l) was used for basal
cortisol. When combining adrenal insufficiency and
probable adrenal insufficiency, the frequency remained
higher in cases (20 vs. 8%, respectively, P<10
3
).
Neither clinical symptoms nor biological signs of adrenal
insufficiency such as hypotension, hypoglycaemia, was
found in patients from the adrenal insufficiency group.
The occurrence of adrenal insufficiency in the cases was
associated with a lower level of ACTH (OR ¼0.74/
pmol/l; 95% CI: 0.55–0.92, P¼0.02) in the multivariate
logistic regression (Tables 2 and 5).
Hyperprolactinemia (investigated among cases
only)
Hyperprolactinemia was frequent (21%) in the case
group. When controlled with a technique more specific
for active monomeric prolactin, this ratio decreased to
11% (n¼11). A true increase in the active monomeric
prolactin was confirmed in three patients (3%) only. Two
of them underwent pituitary MRIs; one was considered
as normal and one revealed a pituitary hyperplasia. Only
low free testosterone level (<0.225 nmol/ml)
(OR ¼3.79; 95% CI: 1.01–15.17, P¼0.02) and
tuberculosis (TB) history (OR ¼4.00; 95% CI: 1.27 –
13.17, P¼0.05) were associated with hyperprolactinemia
in multivariate logistic regression (Supplemental 3,
http://links.lww.com/QAD/B439).
860 AIDS 2019, Vol 33 No 5
Table 4. Determinants for hypothyroidism in 100 HIV-infected transwomen: univariate and multivariate analysis.
Hypothyroidism Univariate analysis Multivariate analysis
Present, n¼17 Absent, n¼83 POR 95% CI POR 95% CI
Clinical data and biochemestry
Age (years) 38 (32 –43) 39 (34– 45) 0.73 0.99 (0.92– 1.05)
BMI (kg/m
2
)28.7 (25.8– 29.3) 25.7 (24.3–28.5) 0.03 1.16 (1.012 –1.34) 0.02 1.23 (1.04– 1.46)
Corticosteroids use 5 (29%) 8 (9%) 0.04 3.90 (1.04– 13.87) 0.01 7.15 (1.57 – 35.05)
Silicone presence 14 (82%) 56 (67%) 0.23 2.25 (0.66 – 10.35)
Tuberculosis history 4 (23%) 21 (26%) 0.88 0.91 (0.24 – 2.90)
Cocaine use 6 (27%) 17 (22%) 0.49 1.50 (0.43– 4.66)
HBV coinfection 0 (0%) 11 (13%) 0.2 0.3 (-inf, 1.9)
Total cholesterol (mmol/l) 4.8 (4.6 –5.2) 4.3 (3.8–4.8) 0.05 0.73 (0.99– 3.07)
Hormonal dysfunction and treatment
Adrenal insufficiency 4 (23%) 16 (19%) 0.74 1.29 (0.33–4.22)
Hyperprolactinemia
a
(>30 mg/l) 5 (30%) 17 (20%) 0.42 1.62 (0.46 – 5.04)
Low free testosteron level (<0.225 nmol/ml) 4 (23%) 33 (39%) 0.21 2.14 (0.69 –8.13)
SHBG (nmol/l) 55.6 (40.5 –79.6) 58.5 (45.3–85.4) 0.21 0.99 (0.96– 1.01)
Hormonal
a
treatment 3 (18%) 28 (34%) 0.20 0.42 (0.09–1.42)
Past or current use of injectable hormone
a
6 (35%) 35 (42%) 0.6 0.75 (0.24 – 2.16)
HIV infection
CDC stage: C vs. B/A 7 (41%) 21 (25%) 0.19 0.97 (0.21 – 3.46)
Duration of infection (years) 11.3 (3–13.3) 11.3 (5.6 – 13.6) 0.27 0.94 (0.85–1.05)
Current CD4
þ
cell count (cells/ml) 760 (440– 920) 650 (510 – 820) 0.55 1.00 (0.999 – 1.002)
Nadir CD4
þ
cell count (cells/ml) 185 (82– 191) 247 (184 –318) 0.08 0.996 (0.991– 1.001) 0.05 0.994 (0.988–0.999)
HIV viral load before treatment (log
10
) 5.4 (4.4 – 5.9) 5.0 (4.5– 5.3) 0.28 1.62 (0.71 – 4.13)
Undetectable viral load (<50 copies/ml) 5 (29%) 16 (19%) 0.35 2.07 (0.68 – 6.09)
HIV therapy
Duration of therapy (years) 6 (1.8– 8.9) 8.8 (4.3–11.8) 0.11 0.91 (0.80–1.02) 0.03 0.84 (0.70– 0.97)
Drugs used in regimen
Protease inhibitors 7 (41%) 42 (51%) 0.16 0.42 (0.11– 1.31)
NNRTI 10 (59%) 41 (49%) 0.54 1.74 (0.50–5.48)
History of stavudine use 3 (17%) 15 (18%) 0.97 1.39 (0.48 – 4.00)
Results are median and [IQR 25–75] for continuous variables and number n(%) for categorical variables, OR 95% CI ¼odd ratio and 95%
confidence intervals. Pvalues were obtained from univariate and multivariate logistic regressions. IQR, interquartile range; NNRTI, nonnucleoside
reverse transcriptase inhibitor; SHBG, sex hormone binding protein.
a
Hormone, estrogen– progestin or antiandrogen. Hyperprolactinemia measured after first dosage.
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
Impact of hormonal consumption on the gonadal
axis (investigated among cases only)
Thirty-one percentage of cases reported hormonal intake
and 60% reported consumption in the past. Hormone
intake is detailed in Table 6 (Supplemental data, http://
links.lww.com/QAD/B439). Regarding serum sexual
hormones, 37, 11 and 10% of cases displayed, respectively,
a low free testosterone level (<0.225 nmol/ml), a low
FSH level (<1.5 UI/l) and a low LH level (<1.5 UI/l)
(Table 7, Supplemental data, http://links.lww.com/
QAD/B439). Patients reporting hormonal consumption,
regardless of type and administration route, more
frequently displayed low free testosterone (77 vs. 18%,
P<10
4
), LH (36 vs. 2%, P<10
4
) and FSH (32 vs. 2%,
P<10
4
) level, and increased SHBG (median: 68 vs.
56 nmol/l retrospectively, P¼0.03) than patients without
hormonal consumption (Supplemental 1, http://
links.lww.com/QAD/B439).
Impact of hormonal consumption on
antiretroviral therapy and HIV infection
The proportion of cases with undetectable viral load was
not different in cases and in controls (81 vs. 90%,
respectively, P¼0.08; Table 1). Furthermore, when we
compared cases treated with antiretroviral and consuming
hormones to cases treated with antiretroviral without
hormone consumption, we found no difference in terms
of the median of CD4
þ
levels (740 vs. 630 cells/ml,
P¼0.25, respectively) and proportion with undetectable
HIV viral load (84 vs. 80%, P¼0.78). Cases with
hormone consumption were more often treated with
NNRTI than cases without hormone consumption (68
vs. 45%, respectively, P¼0.05) (Supplemental 2, http://
links.lww.com/QAD/B439).
Antiretroviral drugs plasma concentrations were available
in 93 out of 97 treated cases. The proportion of patients
having a concentration within the expected range for all
molecules of the antiretroviral treatment, as well as within
different types of antiretroviral used, was not statistically
different among hormones users vs. nonhormones users
(74 vs. 51%, respectively, P¼0.12). Similarly, there was no
difference in the proportion of patients with a low
concentration in one antiretroviral molecule only (10 vs.
21%, P¼0.18) or in all molecules (13 vs. 9%, P¼0.77),
respectively, for hormone users vs. nonhormone users.
Metabolic syndrome and endocrine status in HIV-infected transwomen Pommier et al. 861
Table 5. Determinants for adrenal insufficiency in 100 HIV-infected transwomen: univariate and multivariate analysis (nU100).
Adrenal insufficiency Univariate analysis Multivariate analysis
Present, n¼20 Absent, n¼80 POR 95% CI POR 95% CI
Clinical data
Age (years) 34.5 (31–45) 40 (36–44) 0.10 0.95 (0.88 – 1.01)
BMI (kg/m
2
)26.1 (24.1– 28.2) 26.1 (24.3 – 28.8) 0.82 0.98 (0.85 – 1.12)
Corticosteroids use 3 (15%) 10 (13%) 0.72 1.23 (0.20 – 5.53)
Silicone presence 14 (70%) 56 (70%) 1 1.00 (0.35– 3.10)
Tuberculosis history 2 (10%) 23 (29%) 0.10 0.28 (0.04 –1.06)
Cocaine use 7 (35%) 16 (20%) 0.16 2.15 (0.71 – 6.20)
HBV coinfection 4 (20%) 7 (8%) 0.16 2.30 (0.75–6.68)
Hormonal
a
treatment 3 (15%) 28 (35%) 0.10 0.32 (0.07– 1.08)
Past or current use of
injectable hormone
a
12 (60%) 29 (36%) 0.06 2.64 (0.98– 7.46)
Hormonal data
Hypothyroidism 4 (20%) 13 (16%) 0.74 0.31 (0.01– 6.12)
Hyperprolactinemia (>30 ng/l) 3 (15%) 19 (24%) 0.55 0.57 (0.15– 2.14)
Low free testosteron level
(<0.225 nmol/ml)
5 (25%) 32 (40%) 0.22 2.00 (0.70– 6.64)
SHBG (nmol/l) 52 (40–63) 60 (45 – 85) 0.04 0.98 (0.95 –1.00)
ACTH (pmol/l) 1 (1– 9) 4 (1– 27) 0.01 0.71 (0.53 – 0.90) 0.02 0.73 (0.54 – 0.92)
HIV infection
CDC stage: C vs. B/A 4 (20%) 24 (30%) 0.38 0.58 (0.15– 1.79)
Duration of infection (years) 10.9 (4.3 – 13.6) 11.3 (6.6– 13.6) 0.21 0.94 (0.85– 1.03)
Current CD4
þ
cell count (cells/ml) 690 (535– 698) 655 (507– 847) 0.79 1.00 (0.99–1.00)
Nadir CD4
þ
cell count (cells/ml) 259 (193– 333) 221 (145– 306) 0.46 1.00 (0.99 –1.00)
HIV viral load before
treatment (log
10
)
5.0 (4.5– 5.3) 5.0 (4.5–5.4) 0.36 0.71 (0.33– 1.51)
Undetectable viral load
(<50 copies/ml)
5 (25%) 16 (20%) 0.62 1.33 (0.39– 4.04)
HIV therapy
Duration of HIV therapy (years) 6.0 (2.6–10.4) 8.7 (4.4– 12) 0.10 0.91 (0.81 – 1.01)
HIV drugs used in regimen
Protease inhibitors 10 (50%) 39 (49%) 0.54 1.36 (0.50– 3.68)
NNRTI 7 (35%) 44 (55%) 0.58 0.75 (0.26– 2.04)
History of stavudine use 5 (25%) 13 (16%) 0.35 1.72 (0.53–5.55)
Results are median and [IQR 25–75] for continuous variables and number n(%) for categorical variables, OR 95% CI ¼odd ratio and 95%
confidence interval. Pvalues were obtained from univariate and multivariate logistic regressions. HBV, hepatitis B virus; IQR, interquartile range;
NNRTI, nonnucleoside reverse transcriptase inhibitor; SHBG, sex hormone binding protein.
a
Hormone, estrogen– progestin or antiandrogen. Hyperprolactinemia measured after first dosage.
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
Finally, a low concentration of antiretroviral drugs was not
related to the type of hormone intake (i.e. antiandrogen,
oral or injectable estrogen, estrogen and progestin).
Discussion
To our knowledge, this is the first matched case–control
study comparing HIV-infected transwomen with a
cisgender male population for metabolic and hormonal
dysfunctions and identifying an increased risk of subclinical
hypothyroidism and adrenal insufficiency in the trans-
women population. This study also addressed the issue of
pituitary dysfunction and possible interactions between
antiretroviral therapy and hormones used in HIV-infected
transwomen without finding any deleterious impact of
combining hormonal and antiretroviral therapies.
No difference in the prevalence of MetS was found
between the two populations regarding the IDF
definition although median duration of HIV infection
and length of treatment were higher in cases vs.
controls. As previously reported in HIV-infected
patients [16,34,35], age and stavudine use history were
identified as risk factors for MetS but duration of HIV
infection and length of treatment were not identified as
significant determinants in multivariate regression. No
transwomen specificities such as hormonal intake or
silicone injection were identified as risk factors for
MetS in cases.
Frequency of subclinical hypothyroidism was high among
HIV-infected transwomen patients and higher than the
control group. Prevalence of true hypothyroidism was in
the range of that previously reported in HIV populations
(3.6–12.6%) using the same definition [17,26,36 – 40]. It
was not related to MetS as there was no difference in
terms of MetS between cases and controls [41].
Hypothyroidism in the transwomen population was
mostly subclinical (normal FT4), moderate (TSH
between 4 and 11 mUI/l) and was not related to
thyroiditis, as TPO-Ab was always negative. It was
associated with a recent introduction of antiretroviral and
a low CD4
þ
nadir count, in line with data from a previous
study [40] pointing an association between hypothyroid-
ism and immune reconstitution. Furthermore, a chronic
increase in cortisol levels is known to modulate TSH
secretion [42]. Thus, interpretation of TSH level has to be
made with caution as a lower cortisol level with a higher
TSH level was observed in cases vs. controls. Higher BMI
and cholesterol levels were observed in the hypothyroid-
ism group and found in both groups of subclinical and
true hypothyroidism. No association was identified
between the occurrence of hypothyroidism and trans-
women specificities in terms of hormones use as
previously suspected [15], silicone presence, drug use
or serum sexual hormone concentrations.
In this study, HIV-infected transwomen patients demon-
strated evidence of a high rate of adrenal insufficiency
(probable or confirmed) (20%), with no association with a
late CDC stage of HIV disease. Although adrenal
insufficiency was previously described in HIV-infected
patients, it has been reported as infrequent, associated with
a wide fluctuation in both cortisol and ACTH [43–48]
plasma concentrations, often linked to advanced disease
[44] and opportunistic infections [49]. Adrenal insuffi-
ciency in HIV-infected transwomen patients was associated
with an inappropriately low ACTH level pointing out a
hypothalamic– pituitary – adrenal insufficiency in these
patients. We hypothesize that transwomen patients often
chronically consumed corticosteroids without notifying
their doctors, to treat some silicone inflammations for
instance, resulting in a chronic negative feedback on the
pituitary gland. The high frequency of self-medicated
noncommercialized injectable hormones use among the
cases with adrenal insufficiency (60 vs. 36% in the absence
of adrenal insufficiency, P¼0.06), as well as paucisympto-
matic adrenal insufficiency emphasizes the hypothesis of a
possible hidden self-medication with steroids. However,
adrenal insufficiency was not associated with protease
inhibitor use in our study [22]. Of note, given the low
prevalence of hypothyroidism (20%) or TB history (10%)
in the adrenal insufficiency group, there is no argument for
a Schmidt syndrome or Addison disease as a possible cause
for adrenal insufficiency.
Hyperprolactinemia is frequent in the transwomen popula-
tion and is known to be associated with estrogen intake [50].
The prevalence we observed is similar to that previously
reported in a transwomen population [51]. Because of
several reports of prolactinomas occurring after long-term
estrogen therapy [52,53], serum prolactin follow-up and
pituitary imaging in case of increased concentration are
recommended [50]. However, no prolactin adenoma was
observed among transwomen in this study as an increase in
the monomeric bioactive prolactin was observed in only
three out of 21 elevated prolactin with two MRI performed
without detectable adenoma. Multivariate analysis on
hyperprolactinemia points out that patients are taking
different hormone dosages; it could be that only those in
whom hormone intake resulted in low testosterone level are
at risk of high prolactin level.
Hormonal intake did not affect imunovirological results
as well as antiretroviral drug plasma concentrations in the
transwomen population. No association between the use
of hormones and a low concentration of antiretroviral
drugs was observed. Furthermore, 77% of patients taking
both antiretroviral drugs and hormones were in the
recommended range of serum testosterone levels
(<0.225 nmol/ml for free testosterone) for transwomen
patients under hormonal therapy [50].
Based on these results, we recommend thyroid and
adrenal function screening in HIV-infected transwomen
862 AIDS 2019, Vol 33 No 5
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
population to discuss a potential supplementation and
surveillance for metabolic and cardiovascular complica-
tions in those with subclinical hypothyroidism especially
at antiretroviral treatment initiation. Prolactin levels
should be cautiously interpreted in the absence of
clinical features. If prolactinemia is increased, we
recommend controlling the result by using another
technique or using a PEG precipitation, to restrict
pituitary MRI to true active monomeric prolactin
elevations only. In settings where none of these
techniques are available, all patients with increased
prolactin without a known cause should be offered
pituitary imaging to rule out a tumor. Although HIV-
infected transwomen patients should be closely moni-
tored regarding their hor monotherapy, no harmful
consequence of combining antiretroviral and hormonal
therapies was observed in the respective efficiency of
these two classes of drugs. This should reinforce HIV-
infected transwomen patients’ confidence in and
compliance to antiretroviral therapy. Smoking is
strongly discouraged in such patients, as it increases
the cardiovascular risk [50].
The current study has several caveats. Because trans-
women mostly originated from South America, the
results reported here cannot be generalized to other
ethnical and social groups without caution. To our
knowledge, no increased frequency of subclinical
hypothyroidism or adrenal dysfunction has been
reported in South American population, but MetS is
more frequent in US Hispanics and South Americans
than in other groups [54–56]. Considering endocrine
function, the adrenal insufficiency definition was
weakened in some patients by the absence of a
synacthen testing being performed. The median
duration of HIV infection and length of treatment
were higher in cases vs. controls; however, as none of
these variables was associated with endocrine dysfunc-
tion, such differences probably do not impact the
observed differences in endocrine functions between
both groups. Finally, the small number of patients
undergoing hormone therapy and the difference of
antiretroviral types used, precluded specific analyses on
interactions with antiretroviral therapy.
In conclusion, this study essentially focused on
metabolic and hormonal abnormalities in HIV-infected
transwomen, and showed an increased risk of subclinical
hypothyroidism and adrenal insufficiency as compared
with HIV-infected cisgender male patients. Therefore,
specific attention should be given to thyroid and adrenal
dysfunction screening in HIV-infected transwomen, and
further investigations are needed to better understand
the mechanism of adrenal insufficiency in the HIV-
infected transwomen population. Although hyperpro-
lactinemia was frequent in these patients, an increase in
monomeric bioactive prolactin was rare and no
prolactinoma was observed. In addition to metabolic
and hormonal data, this study also emphasizes the
absence of deleterious impact of combining hormonal
and antiretroviral therapies.
Acknowledgements
We thank our funder IMEA. We also want to thank
Dorothee Valois, MD, Stanislas Harent, MD and
Adrianna Pinto, MD for their input in the organization
of the study, as well as ACCEPTESS-T and ARCAT
patient community groups for their support during
the study.
Authors’ contributions: J.-D.P., C.L., P.Y., F.M., J.G.,
C.G. and P.U. have designed the study. G.C., E.P., S.L.,
F.M., D.M. and E.P. recruited and managed patients.
S.L.G., E.P., D.M., R.L., B.P. and J.-D.P. coordinated the
study. J.G., P.U., A.B., G.P., D.D., P.R.-N. were in charge
of all laboratory testing, analysis and interpretation. J.-D.P.
and C.L. were in charge of all statistics analysis. R.L., B.P.
and Z.J. were in charge of the screening and all the
datasets. J.-D.P. and C.L. drafted the article under the
supervision of P.Y., J.G. and Y.Y. All authors reviewed
the article.
The study was funded by Institut de Medecine et
d’Epidemiologie Appliquee, Hopital Bichat, Paris,
France (IMEA).
Conflicts of interest
There are no conflicts of interest.
References
1. Johnson CA. Off the map: how HIV/AIDS programming is failing
same-sex practicing people in Africa Open Society Foundations.
2007.
2. Baral SD, Poteat T, Str€
omdahl S, Wirtz AL, Guadamuz TE,
Beyrer C. Worldwide burden of HIV in transgender women:
a systematic review and meta-analysis. Lancet Infect Dis 2013;
13:214–222.
3. Herbst JH, Jacobs ED, Finlayson TJ, McKleroy VS. Estimating
HIV prevalence and risk behaviors of transgender persons
in the United States: a systematic review. AIDS Behav 2008;
12:1–17.
4. Santos G-M, Wilson EC, Rapues J, Macias O, Packer T, Ray-
mond HF. HIV treatment cascade among transgender women
in a San Francisco respondent driven sampling study. Sex
Transm Infect 2014; 90:430–433.
5. San Francisco Department of Public Health. HIV/AIDS epide-
miology annual report. San Francisco, CA: ; 2012, Available
from https://www.sfdph.org/dph/files/reports/RptsHIVAIDS/An-
nualReport2011.pdf. [Accessed 28 November 2015].
6. Bowers JR, Branson CM, Fletcher JB, Reback CJ. Predictors of
HIV sexual risk behavior among men who have sex with men,
men who have sex with men and women, and transgender
women. Int J Sex Health 2012; 24:290–302.
7. De Santis JP. HIV infection risk factors among male-to-female
transgender persons: a review of the literature. J Assoc Nurses
AIDS Care 2009; 20:362–372.
Metabolic syndrome and endocrine status in HIV-infected transwomen Pommier et al. 863
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
8. Operario D, Soma T, Underhill K. Sex work and HIV status
among transgender women: systematic review and meta-ana-
lysis. J Acquir Immune Defic Syndr 2008; 48:97–103.
9. Operario D, Nemoto T. HIV in transgender communities:
syndemic dynamics and a need for multicomponent interven-
tions. J Acquir Immune Defic Syndr 2010; 55 (Suppl 2):S91–
S93.
10. Bowers JR, Branson CM, Fletcher J, Reback CJ. Differences in
substance use and sexual partnering between men who have
sex with men, men who have sex with men and women and
transgender women. Cult Health Sex 2011; 13:629–642.
11. Wilson E, Rapues J, Jin H, Raymond HF. The use and correlates
of illicit silicone or ‘fillers’ in a population-based sample of
transwomen, San Francisco, 2013. J Sex Med 2014; 11:1717–
1724.
12. Wiessing LG, van Roosmalen MS, Koedijk P, Bieleman B,
Houweling H. Silicones, hormones and HIV in transgender
street prostitutes. AIDS 1999; 13:2315–2316.
13. Rapport Haute Autorit
e de Sant
e: Situation actuelle et perspec-
tives d’
evolution de la prise en charge m
edicale du transsex-
ualisme en France. 2009. Available from https://www.has-
sante.fr/portail/plugins/ModuleXitiKLEE/types/FileDocument/
doXiti.jsp?id=c_894407. [Accessed January 2015]
14. Revol M, Servant J-M, Banzet P. Surgical treatment of male-to-
female transsexuals: a ten-year experience assessment. Ann
Chir Plast Esthet 2006; 51:499–511.
15. Bisschop PH, Toorians AW, Endert E, Wiersinga WM, Gooren
LJ, Fliers E. The effects of sex-steroid administration on the
pituitary-thyroid axis in transsexuals. Eur J Endocrinol 2006;
155:11–16.
16. Nguyen KA, Peer N, Mills EJ, Kengne AP. A meta-analysis of the
metabolic syndrome prevalence in the global HIV-infected
population. PLoS One 2016; 11:e0150970.
17. Beltran S, Lescure FX, El Esper I, Schmit JL, Desailloud R.
Subclinical hypothyroidism in HIV-infected patients is not
an autoimmune disease. Horm Res 2006; 66:21–26.
18. Parsa AA, Bhangoo A. HIV and thyroid dysfunction. Rev Endocr
Metab Disord 2013; 14:127–131.
19. Hoffmann CJ, Brown TT. Thyroid function abnormalities in
HIV-infected patients. Clin Infect Dis 2007; 45:488–494.
20. Silva GA, Andrade MC, Sugui Dde A, Nunes RF, Pinto JF, Eyer
Silva WA, et al.Association between antiretrovirals and thyroid
diseases: a cross-sectional study. Arch Endocrinol Metab 2015;
59:116–122.
21. Yoganathan K, David L, Williams C, Jones K. Cushing’s syn-
drome with adrenal suppression induced by inhaled budeso-
nide due to a ritonavir drug interaction in a woman with HIV
infection. Int J STD AIDS 2012; 23:520–521.
22. Hyle EP, Wood BR, Backman ES, Noubary F, Hwang J, Lu Z,
et al.High frequency of hypothalamic– pituitary – adrenal axis
dysfunction after local corticosteroid injection in HIV-infected
patients on protease inhibitor therapy. J Acquir Immune Defic
Syndr 2013; 63:602–608.
23. Pommier JD, Laouenan C, Lariven S, Michard F, Castanedo G,
Hamet G, et al. Analysis of social and medical characteristics of
HIV-infected transgender patients. 13th European AIDS Con-
ference/EACSe; 2011 September.
24. Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of
simple methods for the estimation of free testosterone in
serum. J Clin Endocrinol Metab 1999; 84:3666–3672.
25. Alberti KG, Zimmet P, Shaw J, IDF Epidemiology Task Force
Consensus Group. The metabolic syndrome – a new worldwide
definition. Lancet 2005; 366:1059–1062.
26. Hollowell JG, Staehling NW, Flanders WD, Hannon WH,
Gunter EW, Spencer CA, et al.Serum TSH, T(4), and thyroid
antibodies in the United States population (1988 to 1994):
National Health and Nutrition Examination Survey (NHANES
III). J Clin Endocrinol Metab 2002; 87:489–499.
27. Almandoz JP, Gharib H. Hypothyroidism: etiology, diagnosis,
and management. Med Clin North Am 2012; 96:203–221.
28. Bornstein SR, Allolio B, Arlt W, Barthel A, Don-Wauchope A,
Hammer GD, et al.Diagnosis and treatment of primary adrenal
insufficiency: an endocrine society clinical practice guideline. J
Clin Endocrinol Metab 2016; 101:364–389.
29. Kazlauskaite R, Evans AT, Villabona CV, Abdu TAM, Ambrosi B,
Atkinson AB, et al.Corticotropin tests for hypothalamic– pi-
tuitary– adrenal insufficiency: a metaanalysis. J Clin Endocrinol
Metab 2008; 93:4245–4253.
30. Ospina NS, Al Nofal A, Bancos I, Javed A, Benkhadra K, Kapoor
E, et al.ACTH stimulation tests for the diagnosis of adrenal
insufficiency: systematic review and meta-analysis. J Clin En-
docrinol Metab 2016; 101:427–434.
31. Lee MK, Vasikaran S, Doery JC, Wijeratne N, Prentice D.
Cortisol: ACTH ratio to test for primary hypoadrenalism: a
pilot study. Postgrad Med J 2013; 89:617–620.
32. Fahie-Wilson M, Smith TP. Determination of prolactin: the
macroprolactin problem. Best Pract Res Clin Endocrinol Metab
2013; 27:725–742.
33. Pretorius E, Klinker H, Rosenkranz B. The role of therapeutic
drug monitoring in the management of patients with human
immunodeficiency virus infection. Ther Drug Monit 2011;
33:265–274.
34. Shlay JC, Visnegarwala F, Bartsch G, Wang J, Peng G, El-Sadr
WM, et al.Body composition and metabolic changes in anti-
retroviral-naive patients randomized to didanosine and stavu-
dine vs. abacavir and lamivudine. J Acquir Immune Defic Syndr
2005; 38:147–155.
35. Mbunkah HA, Meriki HD, Kukwah AT, Nfor O, Nkuo-Akenji T.
Prevalence of metabolic syndrome in human immunodefi-
ciency virus – infected patients from the South-West region
of Cameroon, using the adult treatment panel III criteria.
Diabetol Metab Syndr 2014; 6:92.
36. Quirino T, Bongiovanni M, Ricci E, Chebat E, Carradori S,
Martinelli C, et al.Hypothyroidism in HIV-infected patients
who have or have not received HAART. Clin Infect Dis 2004;
38:596–597.
37. Wiener M, Lo Y, Klein RS. Abnormal thyroid function in older
men with or at risk for HIV infection. HIV Med 2008; 9:544–
549.
38. Madeddu G, Spanu A, Chessa F, Calia GM, Lovigu C, Solinas P,
et al.Thyroid function in human immunodeficiency virus
patients treated with highly active antiretroviral therapy
(HAART): a longitudinal study. Clin Endocrinol (Oxf) 2006;
64:375–383.
39. Grappin M, Piroth L, Verges B, Sgro C, Mack G, Buisson M, et al.
Increased prevalence of subclinical hypothyroidism in HIV
patients treated with highly active antiretroviral therapy. AIDS
2000; 14:1070–1072.
40. Bongiovanni M, Adorni F, Casana M, Tordato F, Tincati C,
Cicconi P, et al.Subclinical hypothyroidism in HIV-infected
subjects. J Antimicrob Chemother 2006; 58:1086–1089.
41. Gutch M, Rungta S, Kumar S, Agarwal A, Bhattacharya A, Razi
SM. Thyroid functions and serum lipid profile in metabolic
syndrome. Biomed J 2017; 40:147–153.
42. Benker G, Raida M, Olbricht T, Wagner R, Reinhardt W,
Reinwein D. TSH secretion in Cushing’s syndrome: relation
to glucocorticoid excess, diabetes, goitre, and the ‘sick euthyr-
oid syndrome’. Clin Endocrinol (Oxf) 1990; 33:777–786.
43. Biglino A, Limone P, Forno B, Pollono A, Cariti G, Molinatti
GM, et al.Altered adrenocorticotropin and cortisol response to
corticotropin-releasing hormone in HIV-1 infection. Eur J En-
docrinol 1995; 133:173–179.
44. Raffi F, Brisseau JM, Planchon B, R
emi JP, Barrier JH, Grolleau
JY. Endocrine function in 98 HIV-infected patients: a prospec-
tive study. AIDS 1991; 5:729–733.
45. Dobs AS, Dempsey MA, Ladenson PW, Polk BF. Endocrine
disorders in men infected with human immunodeficiency virus.
Am J Med 1988; 84 (3 Pt 2):611–616.
46. Verges B, Chavanet P, Desgres J, Vaillant G, Waldner A, Brun
JM, et al.Adrenal function in HIV infected patients. Acta
Endocrinol (Copenh) 1989; 121:633–637.
47. Villette JM, Bourin P, Doinel C, Mansour I, Fiet J, Boudou P,
et al.Circadian variations in plasma levels of hypophyseal,
adrenocortical and testicular hormones in men infected with
human immunodeficiency virus. J Clin Endocrinol Metab 1990;
70:572–577.
48. Membreno L, Irony I, Dere W, Klein R, Biglieri EG, Cobb E.
Adrenocortical function in acquired immunodeficiency syn-
drome. J Clin Endocrinol Metab 1987; 65:482–487.
49. Guarda LA, Luna MA, Smith JL, Mansell PW, Gyorkey F, Roca
AN. Acquired immune deficiency syndrome: postmortem find-
ings. Am J Clin Pathol 1984; 81:549–557.
50. Hembree WC, Cohen-Kettenis P, Delemarre-van de Waal HA,
Gooren LJ, Meyer WJ, Spack NP, et al.Endocrine treatment of
transsexual persons: an Endocrine Society clinical practice
guideline. J Clin Endocrinol Metab 2009; 94:3132–3154.
864 AIDS 2019, Vol 33 No 5
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
51. Asscheman H, Gooren LJ, Assies J, Smits JP, de Slegte R.
Prolactin levels and pituitary enlargement in hormone-treated
male-to-female transsexuals. Clin Endocrinol (Oxf) 1988;
28:583–588.
52. Gooren LJ, Assies J, Asscheman H, de Slegte R, van Kessel H.
Estrogen-induced prolactinoma in a man. J Clin Endocrinol
Metab 1988; 66:444–446.
53. Kovacs K, Stefaneanu L, Ezzat S, Smyth HS. Prolactin-
producing pituitary adenoma in a male-to-female
transsexual patient with protracted estrogen administration.
A morphologic study. Arch Pathol Lab Med 1994; 118:
562–565.
54. Llabre MM, Arguelles W, Schneiderman N, Gallo LC, Daviglus
ML, Chambers EC, et al.Do all components of the metabolic
syndrome cluster together in U.S. Hispanics/Latinos? Results
from the Hispanic Community Health study/Study of Latinos.
Ann Epidemiol 2015; 25:480–485.
55. Benziger CP, Bernab
e-Ortiz A, Gilman RH, Checkley W,
Smeeth L, M
alaga G, et al.Metabolic abnormalities are com-
mon among South American Hispanics Subjects with normal
weight or excess body weight: the CRONICAS Cohort Study.
PLoS One 2015; 10:e0138968.
56. Ford ES, Li C, Zhao G. Prevalence and correlates of metabolic
syndrome based on a harmonious definition among adults in
the US. J Diabetes 2010; 2:180–193.
Metabolic syndrome and endocrine status in HIV-infected transwomen Pommier et al. 865
