J Bone Miner Metab (2017) 35:99–107
DOI 10.1007/s00774-015-0732-9
ORIGINAL ARTICLE
Bone mineral density and its determinants in men with opioid
dependence
Frank Gotthardt1 · Christine Huber1 · Clara Thierfelder1 · Leticia Grize2,3 ·
Marius Kraenzlin4 · Claude Scheidegger1 · Christian Meier4 
Received: 24 September 2015 / Accepted: 7 December 2015 / Published online: 8 January 2016
© The Japanese Society for Bone and Mineral Research and Springer Japan 2016
Abstract  Data on the influence of opioid substitution
therapy (OST) on skeletal health in men is limited. This
cross-sectional study aimed to determine the prevalence of
low bone mass in male drug users and to evaluate the rela-
tionship between endogenous testosterone and bone mass.
We recruited 144 men on long-term opioid maintenance
therapy followed in the Center of Addiction Medicine in
Basel, Switzerland. Data on medical and drug history, frac-
ture risk and history of falls were collected. Bone mineral
density (BMD) was evaluated by densitometry and serum
was collected for measurements of gonadal hormones
and bone markers. 35 healthy age- and BMI-matched
men served as the control group. The study participants
received OST with methadone (69 %), morphine (25 %)
or buprenorphine (6 %). Overall, 74.3 % of men had low
bone mass, with comparable bone mass irrespective of
OST type. In older men (≥40 years, n = 106), 29.2 % of
individuals were osteoporotic (mean T-score −3.0  ± 0.4
SD) and 48.1 % were diagnosed with osteopenia (mean
T-score −1.7 ± 0.4 SD). In younger men (n = 38), 65.8 %
of men had low bone mass. In all age groups, BMD was
significantly lower than in age-and BMI-matched controls.
In multivariate analyses, serum free testosterone (fT) was
significantly associated with low BMD at the lumbar spine
* Christian Meier
christian.meier@unibas.ch
1
Basel Center for Addiction Medicine, Basel, Switzerland
2
Biostatistics Unit, Swiss Tropical and Public Health Institute
Basel, Basel, Switzerland
3
University of Basel, Basel, Switzerland
4 Opioids are associated with gonadal insufficiency, which
Division of Endocrinology, Diabetology and Metabolism,
University Hospital, Missionsstrasse 24, 4055 Basel,
Switzerland
(p = 0.02), but not at the hip. When analysed by quartiles
of fT, lumbar spine BMD decreased progressively with
decreasing testosterone levels. We conclude that low bone
mass is highly prevalent in middle-aged men on long-term
opioid dependency, a finding which may partly be deter-
mined by partial androgen deficiency.
Keywords  Osteoporosis · Opioid dependence ·
Testosterone · Bone mineral density
Introduction
Osteoporosis is characterised by a progressive loss of bone
mass and bone microarchitecture resulting in decreased
bone strength. An estimated 30–50 % of women and
15–30 % of men will experience a fracture related to oste-
oporosis in their lifetime [1]. Because bone loss occurs
insidiously and is initially asymptomatic, osteoporosis
is often only diagnosed after the first clinical fracture has
occurred [2]. Several risk factors, such as advanced age,
low body weight, previous fragility fractures, comorbidities
and life-style factors as well as decreased mobility, have to
be taken into account when evaluating individual fracture
risk in the general population [3].
Low bone mass has been reported among illicit drug
users and opioid users on long-term methadone mainte-
nance treatment [4–6]. This might be due to life-style fac-
tors and comorbidities associated with long-term opioid
consumption such as smoking [4, 7], alcohol intake [8–10],
advanced liver disease and cirrhosis [11–14], and low
body-mass index or malnutrition [15, 16].
may further aggravate bone loss and fracture risk [17–22].
Long-term opioid therapy for either addiction or chronic
13

100
pain induces hypogonadism due to central suppression of
hypothalamic secretion of gonadotropin-releasing hor-
mone, decreased secretion of gonadotropins in the pituitary
gland and impaired testicular function [22, 23]. In part,
this might be due to opioid-induced immune processes in
both the brain (hypothalamic inflammation with reduced
GnRH secretion) and the peripheral tissue [24]. In con-
trast to accelerated bone loss during androgen deprivation
therapy, less dramatic changes in BMD and bone turnover
are observed in men with hypogonadism due to organic tes-
ticular or hypothalamic–pituitary dysfunction, largely due
to less profound and more variable degrees of androgen
deficiency. However, most men with hypogonadism have
significantly lower bone density than age-matched controls
[25]. The relative contribution of opioid-induced hypog-
onadism to the development of osteoporosis in male illicit
drug users and opioid users on long-term methadone main-
tenance treatment remains uncertain.
Data on the effect of illicit drug use and substitution
therapy on bone mass in men are limited due to the small
number of men included in previous studies [4–6]. Within
a large cohort of men on long-term opioid substitution,
the aim of our study was (a) to evaluate the effect of opi-
oid substitution on bone mineral density (i.e. to explore the
proportion of men with osteoporosis) and bone turnover
markers, and (b) to determine whether endogenous testos-
terone levels (among other risk factors) are associated with
low bone mass and whether any such effect is independent
of other parameters related to fracture risk.
Materials and methods
Study population
This was a cross-sectional study of long-term opioid-
dependent men living in the area of Basel, Switzerland.
Subjects were recruited at the Outpatient Clinic of the
Basel Centre for Addiction Medicine between Novem-
ber 2011 and March 2013. Inclusion criteria were opioid
dependency (ICD-10 F11.2) for more than 10 years and a
written informed consent to participate in the study. Men
not willing to participate or who have withdrawn consent
were excluded. The local Ethics Committee for Human
Studies approved the protocol.
Data on medical history, history of drug use, psychiat-
ric diagnoses, physical activity and daily calcium intake
were collected. Fracture risk was assessed using the WHO
fracture risk assessment tool (FRAX®); history of falls and
prevalent fractures were recorded. All subjects had anthro-
pometric measurements taken.
Median duration of opioid consumption in study partici-
pants was 21 years (interquartile range IQR, 15–25). Active
13
J Bone Miner Metab (2017) 35:99–107
long-term consumption was reported for cocaine in 70.1 %,
alcohol in 46.5 % and tobacco in 95.1 % of the participants.
All study participants received opioid substitution ther-
apy (OST), comprising methadone in 69 % (daily dose,
83 ± 41 mg), morphine in 25 % (daily dose, 609 ± 211 mg)
or buprenorphine in 6 % (daily dose, 13 ± 7 mg). Addi-
tional active use during the last 30 days before study inclu-
sion was reported for heroin in 56 % [median drug use,
7 days (IQR, 3–27)] and for cocaine in 44 % [median drug
use, 2.5 days (IQR, 1–7)]. The daily alcohol consumption
was ≤39 g (equivalent to 4 standard drinks, i.e. 1 L beer
or 0.4 L wine per day) in 40 % of participants and >39 g in
23 % of study subjects. 54 (37 %) participants reported no
alcohol use at the time of the survey.
Psychiatric comorbidities were common, including per-
sonality disorders (47 %), mood disorders (46 %), anxiety
disorders (9 %), hyperkinetic disorders (8 %), post-trau-
matic stress disorders (6 %) and schizophrenia (6 %). Psy-
chiatric co-medications included antidepressants (22 %),
benzodiazepines (19 %), neuroleptics (11 %), methylpheni-
date (7 %) or a combination of them.
Three study participants were HIV-positive and 78
(54 %) were positive for hepatitis C virus (HCV) antibod-
ies. Of these, 36 (46 %) had spontaneously cleared the
virus, 32 (41 %) had on-going chronic hepatitis C (includ-
ing relapses/non-responders to previous interferon/ribavi-
rin therapy) and 10 (13 %) had successfully been treated.
During the observation period 72 % of men were diagnosed
as being vitamin D deficient and received supplementation
therapy.
Age- and BMI-matched healthy men were used as con-
trols to compare densitometric and fracture risk data. All
controls were recruited and assessed in the Division of
Endocrinology, University Hospital Basel.
Biochemistry
After an overnight fast, peripheral venous blood was col-
lected for laboratory assessments including routine chemis-
try, and measurements of gonadal hormones and biochemi-
cal markers of bone turnover.
Calcium and creatinine were analyzed by standard meth-
ods on an autoanalyzer (Hitachi System 704 analyzer;
Roche Diagnostics, Basel, Switzerland). Serum bone-
specific alkaline phosphatase (BAP) was determined by
an ELISA (MicroVueBAP, Quidel, San Diego/USA). The
intra- and interassay variations were <5.8 and 7.6 %, respec-
tively [26]. The parameters beta-CrossLaps (CTX), N-termi-
nal propeptide of Type I collagen (PINP) and 25OH-vitamin
D (25OHD) were measured in serum with electrochemilu-
minescence immunoassays (ECLIA) on the automated ana-
lyzer Elecsys 2010 (Roche Diagnostics, Rotkreuz, Switzer-
land) [27, 28]. The intra-assay variations were 2.1–3.5 %
J Bone Miner Metab (2017) 35:99–107 101
for CTX, 1.3–3.0 % for PINP and 1.7–7.8 % for 25OHD
[29]. Serum concentrations of total testosterone, luteinizing
hormone (LH) and sex hormone-binding globulin (SHBG)
were measured in serum with electrochemiluminescence
immunoassays (ECLIA) on the automated analyzer Elec-
sys 2010. Intra-assay coefficients of variation for all assays
were <5 %, and interassay coefficients of variation were
<10 %. Free testosterone (FT) was calculated using the for-
mula proposed by Vermeulen et al. [30].
Bone mineral density measurement
Bone mineral density at the lumbar spine and hip were eval-
uated by dual X-ray absorptiometry (DXA) using a Hologic
Discovery A densitometer (Hologic, Bedford, MA, USA).
According to the WHO criteria, osteoporosis is defined
as a BMD that is 2.5 SD or more below the average value
for young healthy women (T-score of ≤−2.5 SD) either at
the lumbar spine (L2–4), the femoral neck or hip. A single
densitometer was used throughout the study. To assess the
short-term precision of the system, single repeat measure-
ments in 20 patients were performed [31]. The coefficient
of variation of individual measurements was 1.1 % for the
spine, 1.4 % for the femoral neck, 1.9 % for the trochan-
teric region and 1.1 % for the total hip [29].
Statistical analysis
Subject characteristics, measurements and outcomes of
continuous character were summarized as means and
standard deviations or as medians and interquartile ranges,
depending on their distribution. Categorical variables were
summarized as frequencies and proportions. To examine
potential correlations between the outcomes and poten-
tial determinants, Pearson or Spearman correlation coef-
ficients were calculated depending on the variable types.
The association of the outcome with each of the relevant
factors was explored using linear regression. In a step fur-
ther, relationships between the outcomes and factors of
interest and potential determinants (including age, BMI,
opiates, cocaine and alcohol consumption, smoking habits,
HCV status, HIV status, physical activity, vitamin D defi-
ciency, nutritional calcium intake, serum albumin and free
testosterone) were investigated using generalized multivari-
able linear regression. Backward selection was performed
to select the covariates associated with the outcomes. The
factors were selected by excluding the most non-significant
factor each time until the adjusted R2 for the model ceased
to increase. The normality and heteroskedasticity of the
results were checked. The unit of analysis was the subject
and the level of significance was set at p < 0.05. The statis-
tical analysis was done using SAS 9.3 (SAS Institute Inc.
2010, Cary, NC, USA).
Results
Patient characteristics
Out of a total of 264 eligible men, 144 (54.5 %) agreed
to participate in the study and gave their written informed
consent. Compared to non-participants (n = 120), recruited
individuals were older, had higher body mass index and
smoked less tobacco. Duration of OST and consumption
period of opioids, cocaine and alcohol were comparable
between participants and non-participants (Table 1).
Bone mineral density and fracture risk
Data on bone mineral density (BMD) and fracture risk
in both 144 men with long-term opioid dependency
and 35 age- and BMI-matched healthy controls (age
44.7 ± 10.1 years, BMI 26.5 ± 4.4 kg/m2) are summarized
in Table 2.
In men with opioid dependency, BMD expressed as
mean Z-scores was −1.2  ± 1.1 SD at the lumbar spine,
−0.6 ± 0.9 SD at the femoral neck and −0.7 ± 0.8 SD at
the total hip. For all men, but also for subgroups of men
according to their age (below and above 40 years), BMD at
all measured sites was significantly lower compared to age-
and BMI-matched healthy controls.
Based on WHO DXA criteria and limited to the total
of 106 men older than 40 years, 31 (29 %) individu-
als were diagnosed as having osteoporosis (mean T-score
−3.0  ± 0.4 SD) and 51 (48 %) were diagnosed with
osteopenia (mean T-score −1.7  ± 0.4 SD). In younger
men (n  = 38), low bone mass (defined as Z-score <−1.0
SD) was observed in 25 (66 %), including 6 (16 %) with
Z-score <−2.0 SD (Table 3). Overall, 74.3 % of men had
low bone mass. Bone mineral density did not differ accord-
ing to the type of OST used (methadone, morphine or
buprenorphine).
Prevalent low-trauma fractures (occurring after the age of
20 years) were rare and only reported from men older than
40 years (n = 3). In contrast, fractures related to traumatic
events (mostly extremity fractures) were reported in 42
(40 %) men aged ≥40 years and in 10 (26 %) younger men.
According to femoral neck BMD and clinical risk factors,
fracture risk (expressed as 10-year probability of fracture;
WHO FRAX® algorithm) in men older than 40 years was
6.2 % for major fractures and 1.9 % for hip fractures. Falls
in the previous 12 months were reported in 50 (35 %) study
participants, with most (n = 45) reporting only one fall.
Based on correlation analyses, lumbar spine BMD was
associated with age (r = −0.22, p = 0.008), BMI (r = 0.29,
p < 0.001), duration of cocaine consumption (r  =  −0.17,
p  = 0.04), serum levels of free testosterone (r  = 0.16,
p = 0.06) and SHBG (r = −0.15, p = 0.07). Femoral neck
13

102 J Bone Miner Metab (2017) 35:99–107

Table 1  Comparison of study Study participants Non-participants p value

participants (n = 144) with (n = 144) (n = 120)
non-participants (n = 120)
followed at the Basel Center for Age (years) 44.2 ± 8.0 40.8 ± 8.7 0.001
Substance Abuse
Body mass index (kg/m2) 25.7 ± 5.0 24.2 ± 4.8 0.035
Opioid substitution therapy
 Duration (years) 14 (7–18) 14 (5–17) 0.252
Opioid consumption
 N 144 (100 %) 120 (100 %) 0.096
 Consumption period (years)* 21 (15–25) 20 (13–23)
Cocaine consumption
 N 101 (70.1 %) 84 (70 %) 0.980
a
 Consumption period (years) 15 (5–23) 15 (4–21) 0.481
Alcohol consumption
 N 67 (46.5 %) 48 (40 %) 0.287
 Consumption period (years)a 20 (0–29) 17 (2–27) 0.780
Tobacco consumption
 N 137 (95.1 %) 120 (100 %) 0.017
 Pack-years 26 (17–35) 23 (14–32) 0.075
a
  Consumption period: from the first regular consumption of ≥3 consecutive days/week or >3 days/week.
Normally distributed data are shown as mean ± SD, non-normally distributed data are shown as median
(IQR). International Classification of Diseases (ICD-10): 1ICD-10 code: F11.22; 2ICD-10 code: F14.1/
F14.2; 3ICD-10 code: F10.1/F10.2; 4ICD-10 code: F17.1/F17.2

Table 2  Bone mineral density (BMD) and fracture risk in men with long-term opioid dependency (n = 144) and age- and BMI-matched male
controls (n = 35)
Control group Men with long-term opioid dependency
Age ≥ 40 years Age < 40 years All men Age ≥ 40 years Age < 40 years All men

23 12 35 106 38 144
DXA lumbar spine
 BMD (g/cm2) 1.217 ± 0.118 1.251 ± 0.164 1.229 ± 0.134 0.923 ± 0.134# 0.947 ± 0.115§ 0.929 ± 0.129*
#
 Z-score (SD) 0.4 ± 1.4 0.5 ± 1.3 0.4 ± 1.3 −1.1 ± 1.2 −1.2 ± 1.1§ −1.2 ± 1.1*
 T-score (SD) −0.2 ± 0.9 na na −1.5 ± 1.2# na na
DXA femoral neck
  BMD (g/cm2) 1.049 ± 0.164 1.127 ± 0.169 1.077 ± 0.167 0.747 ± 0.117# 0.806 ± 0.126§ 0.763 ± 0.122*
#
 Z-score (SD) 0.3 ± 1.6 0.5 ± 0.8 0.4 ± 1.3 −0.6 ± 0.93 −0.6 ± 0.9§ −0.6 ± 0.9*
 T-score (SD) −0.5 ± 1.5 na na −1.3 ± 0.9# na na
DXA total hip
 BMD (g/cm2) 1.084 ± 0.179 1.192 ± 0.121 1.132 ± 0.161 0.893 ± 0.127# 0.921 ± 0.113§ 0.900 ± 0.124*
#
 Z-score (SD) 0.3 ± 1.4 0.9 ± 0.8 0.4 ± 1.3 −0.7 ± 0.9 −0.6 ± 0.7§ −0.7 ± 0.8*
 T-score (SD) 0.1 ± 1.3 na na −1.0 ± 0.9# na na
WHO FRAX®
 Major fracture risk (%) 2.4 ± 0.3 na na 6.2 ± 3.7# na na
 Hip fracture risk (%) 0.1 ± 0.1 na na 1.9 ± 2.6# na na
Fracture history
 Trauma-related (%) 1 0 1 42 (40.0) 10 (26.3) 52 (36.4)
 Low-trauma fractures (%) 0 0 0 3 (8.8) 0 3 (7.3)

Data are given as mean ± SD or number (percentage). Comparison with age- and BMI-matched healthy men: p < 0.01 compared with all men*,
men <40 years§ and men ≥40 years#

13
J Bone Miner Metab (2017) 35:99–107 103
Table 3  Percentage of men
with low bone mass according
to WHO criteria
n 35
Osteoporosis, n (%)
Osteopenia, n (%)
Low bone mass, n (%)
Rate of osteoporosis (defined as T-score ≤−2.5 SD) and osteopenia (defined as T-score between −1 and
−2.5 SD) in men aged ≥40 years and rate of low bone mass (defined as Z-score <−1 SD) in men aged
<40 years
BMD was related to age (r  =  −0.37, p < 0.001), BMI
(r = 0.30, p < 0.001), duration of opioid and cocaine con-
sumption (r = −0.20, p = 0.02 and r = −0.18, p = 0.03)
and serum SHBG levels (r = −0.32, p < 0.001).
One third of our patients were on psychiatric co-medica-
tion with intake of at least one drug such as antidepressants,
antipsychotics or anti-epileptics. There were no differences
in the prevalence of low bone mass or in the rate of oste-
oporosis in the group aged ≥40 years related to intake of
psychiatric drugs.
Biochemical characteristics
In men with long-term opioid dependency, bone turnover as
assessed by serum CTX was increased. Serum CTX levels
were inversely related to lumbar spine BMD (r  =  −0.16,
p = 0.06).
In the entire cohort, mean serum levels of total and free
testosterone were within the lower normal range. Total and
free testosterone (fT) levels were decreased in 50 % and
62.5 % of men, respectively. Serum free testosterone cor-
related with age (r = −0.30, p < 0.001), BMI (r = −0.29,
p < 0.001) and lumbar spine BMD (r = 0.16, p = 0.06), but
not femoral neck BMD. Mean serum levels for LH, pro-
lactin and SHBG were within the reference range. Serum
SHBG levels were elevated in 40 men (27.5 %) and higher
in the subgroup with chronic hepatitis C (p < 0.001). Serum
SHBG levels were related to increasing age (r  = 0.30,
p < 0.001), longer duration of cocaine and tobacco con-
sumption (r = 0.18, p = 0.03 and r = 0.23, p = 0.005), and
BMD at the femoral neck (r = −0.33, p < 0.001).
Multivariate regression models
In a multivariate analysis all determinants of bone mass
were considered simultaneously (including age, BMI,
cocaine and alcohol consumption, physical activity, vitamin
D deficiency, nutritional calcium intake, serum albumin
and free testosterone; Table 4). The level of free testoster-
one was significantly and positively associated with BMD
at the lumbar spine (β = 0.00024, p = 0.02). An increase in
100 pmol/L free testosterone would therefore translate into
Control group Men with long-term p Value
opioid dependency
144
0 31 (29.2) 0.006
8 (34.8) 51 (48.1) 0.021
5 (41.7) 25 (65.8) <0.001
an increase of lumbar spine BMD of 0.024 mg/cm2. Body
mass index was significantly associated with BMD at the
lumbar spine (β  =  −0.120, p  = 0.02 for the comparison
of underweight to normal weight; β  = 0.06, p < 0.01 for
the comparison of overweight to normal weight) (Table 4).
Tests for residuals normality (p = 0.06, Shapiro–Wilk test)
and heteroskedasticity (p  = 0.68, White test) were not
significant.
The level of free testosterone was positively but not sig-
nificantly associated with BMD at the femoral neck and
the hip, respectively (β = 0.00013, p = 0.17; β = 0.00009,
p = 0.34).
When the entire sample was analysed by quartiles of free
testosterone (fT, adjusted for age and BMI) BMD at the
lumbar spine decreased with decreasing concentrations of
circulating testosterone (Fig. 1). Lumbar spine BMD was
significantly lower only in men with fT values in the lower
two quartiles (fT < 135 pmol/L). When stratified by median
free testosterone levels, BMD was significantly lower in
men with fT values <135 pmol/L than in men with higher
fT levels (r = −0.06, p = 0.01). Serum SHBG levels were
comparable across the quartiles of free testosterone.
Discussion
This study comprising a large cohort of male illicit drug
users on long-term opioid maintenance therapy shows that
a large proportion of opioid users have impaired bone min-
eral density (BMD) with lower age-adjusted BMD values,
predominantly at the lumbar spine. Furthermore, serum
free testosterone was shown to be an independent predictor
of bone mass. Lumbar spine BMD decreased progressively
with decreasing concentrations of circulating free testos-
terone, with lowest values in men with advanced hypog-
onadism, i.e. serum free testosterone levels <135 pmol/L.
Kim et al. report decreased bone mass in 83 % of
patients enrolled in a methadone maintenance treatment
program. In this population of women and men (median
age 42 years, 36 % men) significant predictors of low
BMD were male gender, lower weight and heavy alcohol
use [4]. Comparably, a recent Swiss study of 19 younger
13

104 J Bone Miner Metab (2017) 35:99–107

Table 4  Multivariate model to examine the determinants of lumbar spine BMD (g/cm2) in men with long-term opioid dependency
Predictor/factor Value Beta estimate Standard error p value

Intercept 0.631624 0.181787 0.0007

Age (years) 1 unit −0.002580 0.001636 0.1173
BMI classes Underweight −0.119959 0.050113 0.0182
Overweight 0.064207 0.022088 0.0043
Normal (ref.) 0.000000
Opiate intake duration (years) 10 units 0.017595 0.016639 0.2924
Cocaine intake duration (years) 10 units −0.019704 0.011590 0.0916
Nutritional calcium intake 1 unit 0.000011 0.000009 0.2377
Physical activity (20 min brisk walking, Never −0.003862 0.026799 0.8856
fitness training or sport) 1–2 times/month −0.084030 0.052260 0.1104
3–4 times/month 0.045973 0.055559 0.4096
1–2 times/week −0.057808 0.034009 0.0917
3–4 times/week −0.001330 0.032319 0.9672
≥5 times/week (ref.) 0.000000
Vitamin D deficiency No 0.028495 0.022206 0.2018
Yes (ref.) 0.000000
Albumin (g/L) 1 unit 0.007014 0.003358 0.0388
Free testosterone (pmol/L) 1 unit 0.000241 0.000103 0.0201

Ref. Reference level for comparisons within a categorical factor

measured sites than age-and BMI-matched healthy con-

Fig. 1  Mean lumbar spine BMD in function of free testosterone
(FT) quartiles, adjusted for age and BMI. Q1 FT < 89 pmol/L, Q2
FT 89–134 pmol/L, Q3 FT 135–206 pmol/L, Q4 FT > 206 pmol/L.
Squares represent the adjusted means and whiskers their 95 % confi-
dence intervals. The reference group consisted of men with testoster-
one concentrations in the lowest quartile (Q1)
heroin-dependent patients on injectable heroin mainte-
nance observed reduced bone mass with osteopenic values
in 58 % and osteoporotic values in 16 % [6]. Our study
comprised 144 men of comparable age but longer median
duration of opioid consumption (21 vs. 14 years). Never-
theless, we confirm the high prevalence of low bone mass
in 74 % of study participants. Of note, men with long-term
opioid dependency had significantly lower bone mass at all
trols, resulting in a significantly higher proportion of men
with osteoporosis or osteopenia as defined by WHO cri-
teria. Accordingly, 48 % of older men were found to have
values in the osteopenic range and 29 % had osteoporotic
values. Hence, this large study confirms earlier reports of
high prevalence of decreased bone mass, a finding observed
not only in women but equally in men with long-term illicit
drug use.
Several population-based and case–control studies indi-
cate that opioid drugs used for medical reasons such as
pain relief are associated with an increased risk of fractures
[32, 33]. The moderate risk of fractures observed among
morphine and opioid users has been observed even at very
low doses which suggest that, at least in part, the increased
risk might have been caused by other factors such as drug-
induced falls rather than a deterioration of bone structure
[32]. In contrast, data on the association between illicit
long-term opioid dependency and fracture risk are limited.
Although men in our cohort presented with some clinical
risk factors of osteoporosis, prevalent low-trauma fractures
were rare and only reported from men older than 40 years.
Fractures related to traumatic events occurred more fre-
quently. Our findings are in line with published cases of
fractures among patients of comparable age on opioid
maintenance therapy [5], which is also confirmed by a
low calculated absolute fracture risk (expressed as 10-year
probability of fracture based on femoral neck BMD and
clinical risk factors).

13
J Bone Miner Metab (2017) 35:99–107 105
Life-style factors, including smoking and alcohol con-
sumption, are known determinants of low bone mass, pre-
dominantly in men [7–9, 34]. Although most of the men
were long-term smokers and about half of them were regu-
lar alcohol consumers, both factors were not related to low
bone mass in our cohort. The negative association between
alcohol intake and bone mass in our study is likely to be
related to lower amounts of daily intake, supporting a dose-
dependent effect of alcohol on bone mass [8]. Furthermore,
chronic alcoholism is associated with other risk factors
which may contribute to the pathogenesis of bone dis-
ease. Such determinants, like poor nutrition, leanness and
chronic liver disease, were not prevalent in our cohort. Neg-
ative influences on bone mineral density and fracture risk
with the use of antidepressants, antipsychotics and anti-
epileptics are described [33]. One third of our patients were
on at least one of these medications. There were, however,
no differences in the prevalence of low bone mass or osteo-
porosis in our cohort related to psychiatric co-medications.
Hypogonadism has a detrimental effect on bone metabo-
lism. Men with overt hypogonadism either due to organic
testicular or hypothalamic–pituitary dysfunction or age-
associated androgen deficiency present with dose-depend-
ent decreases in bone mass [25, 35]. Long-term opioid
therapy for either addiction or chronic pain induces hypo-
gonadism [17–21] due to central suppression of hypo-
thalamic secretion of gonadotropin-releasing hormone,
decreased secretion of gonadotropin-releasing hormone in
the pituitary gland and impaired testicular function [23]. In
a large cross-sectional study including patients with opioid
dependence from methadone clinics, the average testoster-
one level in men receiving methadone treatment was signif-
icantly lower than in controls [18]. Thereby, the methadone
dose was inversely associated with testosterone levels indi-
cating that the relationship between methadone and testos-
terone is dose-dependent.
It can be assumed that opioids, specifically long-term
opioid dependency, may induce bone loss as result of par-
tial androgen deficiency and thereby, at least in part, con-
tribute to the observed low bone mass in illicit drug users.
Based on multivariate analyses, we observed that circu-
lating free testosterone levels were associated with low
BMD at the lumbar spine. Lumbar spine BMD was sig-
nificantly lower only in men with free testosterone levels
<135 pmol/L, with progressive decline in parallel to lower
testosterone values. This finding supports a detrimental
effect of opioid-induced androgen deficiency only in men
with pronounced androgen deficiency. Hypothetically this
could translate into an increased risk of fractures in opioid
users with severe androgen deficiency; however, fracture
prevalence in our cohort was too small to draw any conclu-
sion on the association between androgen status and frac-
ture risk.
Testosterone deficiency is a recognized comorbidity in
patients with decompensated cirrhosis and hepatocellular
carcinoma [36]. Dysregulation of gonadal hormones with
decreased free testosterone levels and increased serum lev-
els for sex hormone-binding protein (SHBG) has also been
shown in patients with compensated liver disease [37]. Sex
hormone-binding globulin is a plasma glycoprotein that
binds with high affinity to sex hormones thereby regulating
the free, biologically active form of testosterone. In liver
failure, the rise in SHBG levels is assumed to result from
induction by estradiol, increased liver SHBG synthesis or
reduced SHBG clearance intrinsic to the liver disease itself
[38–40]. Our data confirm the influence of liver disease on
serum SHBG levels, as its levels were higher in the sub-
group of men with chronic hepatitis C. However, lower
free testosterone levels were independently related to lower
bone mineral density in multivariate analyses with correc-
tion for hepatitis C status, suggesting an independent role
of testosterone on bone mass in men with long-term opioid
dependency.
This present study’s findings should be interpreted
within the context of its strengths and limitations. In con-
trast to previous published studies, our study is strength-
ened by a large number of men on long-term opioid main-
tenance therapy with complete sets of data on skeletal
health. Our cohort is representative of male illicit drug
consumers as demonstrated by the similar duration of OST
and consumption period of opioids and cocaine in study
participants compared to non-participants from our clinic.
As with any cross-sectional study, our findings should not
be interpreted as an indication of any causal relationship
between circulating testosterone and bone mineral den-
sity. Several potential determinants such as lifetime opioid
exposition and additional heroin use or nutritional factors
were not quantifiable. The potential impact of chronic hep-
atitis C on bone mass loss could be underestimated as most
of our patients with chronic hepatitis C had only mild liver
disease without cirrhosis. One limitation of the study is that
we did not measure serum estradiol (E2) levels, thereby
not excluding the possibility that the observed association
between testosterone and BMD is driven by E2 (after aro-
matization from testosterone) rather than by testosterone
itself. Finally, we used the Vermeulen formula for calculat-
ing free testosterone. This method takes into account circu-
lating SHBG and albumin levels and has been shown to be
correlated with free testosterone measured by equilibrium
dialysis in various patient groups [30]. As the equilibrium
dialysis method is not easily accessible, calculated free tes-
tosterone is regarded as an acceptable method.
We conclude that low bone mineral density is highly
prevalent in middle-aged men on long-term opioid mainte-
nance therapy and that, among others, androgen deficiency
is an independent determinant of bone mass. Nevertheless,
13

106
the prevalence of fractures related to osteoporosis was low,
indicating that fracture risk might be driven largely by age
and only in part by the observed opioid-induced decrease in
serum testosterone.
Acknowledgments  We would like to thank all patients for their
willingness to participate in the study. We thank the staff involved in
the Basel Center for Addiction Medicine and the Division of Endocri-
nology, University Hospital Basel for their patience in dealing with
our study participants, with special thanks to Mrs. Kradolfer. The
study was supported by unrestricted educational grants from “Frei-
willige Akademische Gesellschaft Basel”, “Verein für Drogenarbeit
Basel”, and Roche Pharma (Switzerland).
Compliance with ethical standards  and osteoporosis in patients referred for a bone mineral density
Conflict of interest  Nothing to declare.
References
1. Rachner TD, Khosla S, Hofbauer LC (2011) Osteoporosis: now
and the future. Lancet 377:1276–1287
2. Vestergaard P, Rejnmark L, Mosekilde L (2005) Osteoporosis is
markedly underdiagnosed: a nationwide study from Denmark.
Osteoporos Int 16:134–141
3. Kanis JA, McCloskey EV, Johansson H, Cooper C, Rizzoli R,
Reginster JY (2013) Scientific Advisory Board of the European
Society for Clinical and Economic Aspects of Osteoporosis and
Osteoarthritis (ESCEO) and the Committee of Scientific Advi-
sors of the International Osteoporosis Foundation (IOF). Euro-
pean guidance for the diagnosis and management of osteoporosis
in postmenopausal women. Osteoporos Int 24:23–57
4. Kim TW, Alford DP, Malabanan A, Holick MF, Samet JH (2006)
Low bone density in patients receiving methadone maintenance
treatment. Drug Alcohol Depend 85:258–262
5. Grey A, Rix-Trott K, Horne A, Gamble G, Bolland M, Reid IR
(2011) Decreased bone density in men on methadone mainte-
nance therapy. Addiction 106:349–354
6. Dursteler-MacFarland KM, Kowalewski R, Bloch N, Wiesbeck
GA, Kraenzlin ME, Stohler R (2011) Patients on injectable dia-
cetylmorphine maintenance have low bone mass. Drug Alcohol
Rev 30:577–582
7. Jutberger H, Lorentzon M, Barrett-Connor E, Johansson H,
Kanis JA, Ljunggren O, Karlsson MK, Rosengren BE, Redlund-
Johnell I, Orwoll E, Ohlsson C, Mellstrom D (2010) Smoking
predicts incident fractures in elderly men: Mr OS Sweden. J
Bone Miner Res 25:1010–1016
8. Kim MJ, Shim MS, Kim MK, Lee Y, Shin YG, Chung CH,
Kwon SO (2003) Effect of chronic alcohol ingestion on bone
mineral density in males without liver cirrhosis. Korean J Intern
Med 18:174–180
9. Berg KM, Kunins HV, Jackson JL, Nahvi S, Chaudhry A, Harris
KA Jr, Malik R, Arnsten JH (2008) Association between alcohol
consumption and both osteoporotic fracture and bone density.
Am J Med 121:406–418
10. Malik P, Gasser RW, Kemmler G, Moncayo R, Finkenstedt G,
Kurz M, Fleischhacker WW (2009) Low bone mineral density
and impaired bone metabolism in young alcoholic patients with-
out liver cirrhosis: a cross-sectional study. Alcohol Clin Exp Res
33:375–381
11. Nakchbandi IA (2014) Osteoporosis and fractures in liver dis-
ease: relevance, pathogenesis and therapeutic implications.
World J Gastroenterol 20:9427–9438
13
J Bone Miner Metab (2017) 35:99–107
12. Trautwein C, Possienke M, Schlitt HJ, Boker KH, Horn R, Raab
R, Manns MP, Brabant G (2000) Bone density and metabolism in
patients with viral hepatitis and cholestatic liver diseases before
and after liver transplantation. Am J Gastroenterol 95:2343–2351
13. Giouleme OI, Vyzantiadis TA, Nikolaidis NL, Vasiliadis TG,
Papageorgiou AA, Eugenidis NP, Harsoulis FI (2006) Pathogen-
esis of osteoporosis in liver cirrhosis. Hepatogastroenterology
53:938–943
14. Li M, Xu Y, Xu M, Ma L, Wang T, Liu Y, Dai M, Chen Y, Lu
J, Liu J, Bi Y, Ning G (2012) Association between nonalco-
holic fatty liver disease (NAFLD) and osteoporotic fracture
in middle-aged and elderly Chinese. J Clin Endocrinol Metab
97:2033–2038
15. Asomaning K, Bertone-Johnson ER, Nasca PC, Hooven F,
Pekow PS (2006) The association between body mass index
examination. J Womens Health (Larchmt) 15:1028–1034
16. Forrester JE (2006) Nutritional alterations in drug abusers with
and without HIV. Am J Infect Dis 2:173–179
17. Bliesener N, Albrecht S, Schwager A, Weckbecker K, Lichter-
mann D, Klingmuller D (2005) Plasma testosterone and sexual
function in men receiving buprenorphine maintenance for opioid
dependence. J Clin Endocrinol Metab 90:203–206
18. Bawor M, Dennis BB, Samaan MC, Plater C, Worster A, Varen-
but M, Daiter J, Marsh DC, Desai D, Steiner M, Anglin R, Coote
M, Pare G, Thabane L, Samaan Z (2014) Methadone induces tes-
tosterone suppression in patients with opioid addiction. Sci Rep
4:6189
19. Hallinan R, Byrne A, Agho K, McMahon CG, Tynan P, Attia
J (2009) Hypogonadism in men receiving methadone and
buprenorphine maintenance treatment. Int J Androl 32:131–139
20. Rajagopal A, Vassilopoulou-Sellin R, Palmer JL, Kaur G, Bruera
E (2004) Symptomatic hypogonadism in male survivors of can-
cer with chronic exposure to opioids. Cancer 100:851–858
21. Daniell HW (2002) Hypogonadism in men consuming sustained-
action oral opioids. J Pain 3:377–384
22. Bawor M, Bami H, Dennis BB, Plater C, Worster A, Varenbut
M, Daiter J, Marsh DC, Steiner M, Anglin R, Coote M, Pare G,
Thabane L, Samaan Z (2015) Testosterone suppression in opi-
oid users: a systematic review and meta-analysis. Drug Alcohol
Depend 149:1–9
23. Katz N, Mazer NA (2009) The impact of opioids on the endo-
crine system. Clin J Pain 25:170–175
24. Zhang G, Li J, Purkayastha S, Tang Y, Zhang H, Yin Y, Li B,
Liu G, Cai D (2013) Hypothalamic programming of systemic
ageing involving IKK-beta, NF-kappaB and GnRH. Nature
497:211–216
25. Meier C, Seibel MJ, Handelsman DJ (2009) Testicular dysfunc-
tion and bone. In: Orwoll E, Bilezikian J, Vanderschueren D
(eds) Osteoporosis in men, 2nd edn. Academic Press, San Diego,
pp 423–434
26. Gomez B Jr, Ardakani S, Ju J, Jenkins D, Cerelli MJ, Daniloff
GY, Kung VT (1995) Monoclonal antibody assay for measuring
bone-specific alkaline phosphatase activity in serum. Clin Chem
41:1560–1566
27. Garnero P, Borel O, Delmas PD (2001) Evaluation of a fully
automated serum assay for C-terminal cross- linking telopeptide
of type I collagen in osteoporosis. Clin Chem 47:694–702
28. Schmidt-Gayk H, Spanuth E, Kotting J, Bartl R, Felsenberg D,
Pfeilschifter J, Raue F, Roth HJ (2004) Performance evaluation
of automated assays for beta-CrossLaps, N-MID-Osteocalcin
and intact parathyroid hormone (BIOROSE Multicenter Study).
Clin Chem Lab Med 42:90–95
29. Keel C, Kraenzlin ME, Kraenzlin CA, Muller B, Meier C (2010)
Impact of bisphosphonate wash-out prior to teriparatide therapy
in clinical practice. J Bone Miner Metab 28:68–76
J Bone Miner Metab (2017) 35:99–107 107
30. Vermeulen A, Verdonck L, Kaufman JM (1999) A critical evalu-
ation of simple methods for the estimation of free testosterone in
serum. J Clin Endocrinol Metab 84:3666–3672
31. Gluer CC, Blake G, Lu Y, Blunt BA, Jergas M, Genant HK
(1995) Accurate assessment of precision errors: how to measure
the reproducibility of bone densitometry techniques. Osteoporos
Int 5:262–270
32. Vestergaard P, Rejnmark L, Mosekilde L (2006) Fracture risk
associated with the use of morphine and opiates. J Intern Med
260:76–87
33. Takkouche B, Montes-Martinez A, Gill SS, Etminan M (2007)
Psychotropic medications and the risk of fracture: a meta-analy-
sis. Drug Saf 30:171–184
34. Kanis JA, Johnell O, Oden A, Johansson H, De Laet C, Eisman
JA, Fujiwara S, Kroger H, McCloskey EV, Mellstrom D, Melton
LJ, Pols H, Reeve J, Silman A, Tenenhouse A (2005) Smoking
and fracture risk: a meta-analysis. Osteoporos Int 16:155–162
35. Snyder PJ (2001) Effects of age on testicular function and con-
sequences of testosterone treatment. J Clin Endocrinol Metab
86:2369–2372
36. Pignata S, Daniele B, Galati MG, Esposito G, Vallone P, Fiore
F, Ricchi P, Pergola M (1997) Oestradiol and testosterone blood
levels in patients with viral cirrhosis and hepatocellular carci-
noma. Eur J Gastroenterol Hepatol 9:283–286
37. Nguyen HV, Mollison LC, Taylor TW, Chubb SA, Yeap BB
(2006) Chronic hepatitis C infection and sex hormone levels:
effect of disease severity and recombinant interferon-alpha ther-
apy. Intern Med J 36:362–366
38. De Besi L, Zucchetta P, Zotti S, Mastrogiacomo I (1989) Sex
hormones and sex hormone binding globulin in males with com-
pensated and decompensated cirrhosis of the liver. Acta Endo-
crinol (Copenh) 120:271–276
39. Guechot J, Chazouilleres O, Loria A, Hannoun L, Balladur P,
Parc R, Giboudeau J, Poupon R (1994) Effect of liver transplan-
tation on sex-hormone disorders in male patients with alcohol-
induced or post-viral hepatitis advanced liver disease. J Hepatol
20:426–430
40. Iturriaga H, Lioi X, Valladares L (1999) Sex hormone-binding
globulin in non-cirrhotic alcoholic patients during early with-
drawal and after longer abstinence. Alcohol Alcohol 34:903–909
13