Immediate dual energy X-ray absorptiometry reveals a high incidence of osteoporosis in patients with advanced prostate cancer before hormonal manipulation


Mr N.J. Parr, Department of Urology, Ward 14, Arrowe Park Hospital, Wirral CH49 5PE, UK.



To examine the incidence of osteoporosis in patients with advanced prostate cancer (using forearm densitometry) before commencing androgen deprivation therapy (ADT), as osteoporotic fractures are more frequent in patients with prostate cancer who have undergone either medical or surgical castration, because of rapid loss of bone mass.


In all, 174 patients (mean age 74.6 years, range 46–90) with advanced prostate cancer presented over 2 years. Their forearm bone densitometry values were compared with those from 106 age-matched controls (mean age 74.3 years, range 66–90).


Of the 174 patients, 73 (42%) were osteoporotic (t score ≤− 2.5) and 65 (37%) were osteopenic (t score − 1 to − 2.4). This compares with a 27% incidence of osteoporosis in the control group (P = 0.022). There were also no significant correlations between prostate specific antigen levels, Gleason score, tumour stage, biochemical markers and the presence or absence of osteoporosis risk factors.


Patients with advanced prostate cancer requiring ADT have a high incidence of osteoporosis before treatment. In addition, osteoporosis in these men cannot be predicted from clinical or biochemical values. Therefore, bone densitometry should be used in all patients with advanced cancer requiring ADT, as the results have implications for the choice of cancer therapy and the prophylaxis for osteoporosis.


androgen deprivation therapy


alkaline phosphatase


bone mineral density


body mass index


dual-energy X-ray absorptiometry.


Carcinoma of the prostate is common and one of the leading causes of death in men [1]. A large proportion of men still present with advanced disease and in this situation androgen deprivation therapy (ADT) is the mainstay of treatment [2,3]. Castration may be achieved either surgically [4] or by administering LHRH analogues, with antiandrogen monotherapy being used less frequently because of a possible survival disadvantage, although this may be clinically insignificant [5].

Recently there has been increasing concern about the effect of castration on bone metabolism [6,7], as testosterone is essential for maintaining bone mass in men [8]. However, there is also evidence that prostate cancer itself is a significant risk factor for osteoporosis, and hence fracture, by causing disturbances in bone turnover and mineralization even before ADT [9,10]. This naturally has implications for the choice of cancer treatment and the need to assess suitable prophylactic agents to prevent further accelerated bone loss.

In this study we determined the incidence of osteoporosis in men with locally advanced (T3–T4) and/or metastatic prostate cancer (as determined by isotope bone scan or a PSA of > 50 ng/mL) requiring hormone manipulation therapy. Densitometry was used immediately before ADT, with an assessment of other concurrent risk factors for osteoporosis that could influence future fracture risk.


In all, 174 men with newly diagnosed advanced prostate cancer and requiring hormone manipulation were recruited from urology clinics from October 1999 to November 2002. The clinical stage (T) of the tumour was determined by a DRE. All patients had prostate biopsies taken to confirm the diagnosis of cancer and the Gleason sum score was used to assess the histological grade of the tumour. Biochemical variables included measurements of serum PSA, calcium, alkaline phosphatase (ALP) and creatinine.

Bone mineral density (BMD) of the nondominant forearm was measured by dual-energy X-ray absorptiometry (DEXA, DTX-200, Osteometer MediTech, Inc., Hawthorne, CA) within 1 week of the diagnosis and before ADT. The DTX-200 forearm system has a unique protocol for evaluating BMD at the ultradistal radius and ulna. The BMD result is expressed as the sd about the mean for a healthy age-matched population (z score). Osteoporosis was defined using the WHO criteria as a BMD of ≤− 2.5 sd below the mean for a healthy population group aged 20–40 years, termed the t score, of ≤− 2.5, with or without pre-existing fragility fractures. Osteopenia was defined by a t score of − 1.0 to −2.4 and a normal BMD as a t score of > − 1.0.

Osteoporosis risk factors were assessed using a questionnaire at the time of DEXA scanning for both patients and controls (Table 1). Body habitus was assessed by height, weight and body mass index (BMI, kg/m2). The control group consisted of age-matched volunteers recruited both from the local community and from those attending urology clinics with conditions other than prostate cancer. These volunteers had forearm BMD assessed by DEXA and their BMI measured. The protocol was reviewed and approved by the Local Ethics Research Committee.

Table 1.  The distribution of patients with osteoporotic risk factors in the groups with cancer and in the controls.
  • *

    Previous or current use of steroids including inhaled preparations for respiratory conditions;

  • † Exercise, judged by three arbitrary distances, i.e. mobility within the home, to shops, any distance;

  • ‡ rheumatoid arthritis, liver disease, thyroid disease, malabsorption.

Alcohol (> 24 units/week)  9  913  9
Steroids*  0  2  3  3
Exercise10  9  3  3
Osteoporosis-related systemic disease  0  2  3  0
Family history1318  9  8
Previous fragility (low trauma) fractures16  4  6  3
Low BMI  3  0  0  1

Group statistics of continuous data were assessed using independent-sample t-tests. Equal variance was assumed as the Levine's test for equality of variances was not significant. Fisher's exact test was used to compare binary variables between groups. Multiple linear regression anova was used to identify relationships between tumour or biochemical variables and bone mineral density in those with prostate cancer. Multiple comparisons were assessed using a one-way anova and posthoc Tukey tests. A statistically significance result was assumed at P ≤ 0.05.


The clinical characteristics of the 174 men with prostate cancer (mean age 74.6 years, range 46–90) and 106 age-matched controls (mean age 74.3 years, range 66–90) are shown in Table 2. There was no statistical difference between the control and prostate cancer group for height, weight or BMI. Subgroup analysis of the patients showed a significantly lower weight (P = 0.002) and BMI (P = 0.046) in the osteoporotic group.

Table 2.  Clinical characteristics of men with prostate cancer, the controls and the tumour characteristics
Mean (sd)NormalOsteopeniaOsteoporosis
N  36  65  73
Age, years  71.5 (8.34)  73.8 (6.70)  76.8 (6.30)
Weight, kg  79.7 (12.40)  79.0 (14.56)  70.4 (9.72)
Height, cm    1.73 (0.088)    1.72 (0.063)    1.69 (0.078)
BMI, kg/m2  26.5 (3.45)  26.6 (4.60)  24.5 (3.02)
N  24  51  31
Age, years  70 (3.48)  73 (5.31)  75 (6.27)
Weight, kg  83.6 (11.32)  75.7 (10.46)  71.8 (7.41)
Height, cm    1.75 (0.07)    1.73 (0.07)    1.71 (0.07)
BMI, kg/m2  27.4 (3.08)  24.1 (3.06)  24.5 (2.26)
Clinical staging
T1  16    8  10
T2    8  50  33
T3  64  12  39
T4  12  30  18
Gleason Sum Score
2–4  15  19    9
5–7  47  37  62
8–10  38  44  29
M0  17    9  12
M1  14    9  14
Mx  69  82  74
Biochemical results, median (range)
PSA, ng/mL  56.5 (0.5–678)  53 (0.4–980)  62 (1.7–2148)
Calcium, mmol/L    2.34 (2.0–2.49)    2.35 (1.94–2.54)    2.35 (1.91–2.6)
ALP, IU/L  86 (45–135)  82 (47–1750)  87 (26–3014)
Urea, mmol/L    6.5 (0.7–13.9)    6.2 (3.4–13.6)    6.2 (3.4–13.8)
Creatinine, µmol/L109 (88–144)106.5 (62–165) 111 (77–202)

Osteoporosis was diagnosed in 42% of men with newly diagnosed advanced prostate cancer before commencing hormone-manipulation therapy; 37% were osteopenic and 21% had normal BMD. This compares with 27%, 46% and 37%, respectively, for the age-matched controls, indicating a higher incidence of osteoporosis in the patients (P = 0.022).

The mean BMD in men with prostate cancer was 6.6% lower than the control group (P = 0.006). The t-score and z-score was significantly lower in the patients (Table 3). There was no statistically significant difference among the three BMD groups with cancer for PSA, tumour stage, grade and other biochemical markers (Table 1), but age correlated significantly with BMD (P < 0.001). Table 1 also shows the clinical T stage, histological grading of the tumour and bone metastasis status, as determined by isotope bone scan; there was no statistical difference between each BMD group.

Table 3.  The difference in mean (sd) t score, z score and BMD between patients and controls
t-score −  2.19 (1.43) −  1.81 (1.21)0.025
z-score −  0.56 (1.38) −  0.16 (1.19)0.013
BMD, g/cm3      0.48 (0.08)      0.51 (0.07)0.006

Smoking appeared to be the most common risk factor for osteoporosis in each BMD group, with family history featuring more highly in the osteoporotic and osteopenic groups. The incidence of previous fragility fractures was higher in the osteoporotic group, which were essentially vertebral fractures manifesting as back pain (Table 1).


Osteoporosis is a progressive systemic skeletal disease characterized by low bone mass and micro-architectural deterioration of bone tissue, with a consequent increase in bone fragility and susceptibility to fracture (WHO 1994). The socio-economic and individual consequences of resulting osteoporotic fracture are of considerable public health concern [11]. Osteoporosis results in > 200 000 fractures each year in the UK, causing severe pain and disability to individual sufferers, at an annual cost to the National Health Service of over £940 million [12]. More than a third of adult women will sustain one or more osteoporotic fractures in their lifetime. Although the condition is regarded as predominantly affecting women, men still have half the female lifetime risk, also making it an important male health concern [13]. Furthermore, men are twice as likely as women to die within a year of a hip fracture and are more likely to become dependent on nursing homes after such fractures [14].

Somewhat surprisingly the incidence of osteoporotic fractures has been reported to be greater than pathological fractures in men with prostate cancer receiving LHRH-analogue therapy [15]. Recognising this specific at-risk group is therefore of paramount concern. The present results suggest that a significant number of men diagnosed with advanced prostate cancer have osteoporosis and osteopenia before ADT. PSA levels are thought to relate to overall tumour burden, except in some cases of anaplastic disease. The median PSA levels in each of the three groups with normal BMD, osteopenia and osteoporosis were > 50 ng/mL. Clinicians are usually reluctant to withhold ADT in patients with levels above this, even when asymptomatic, as most will develop symptoms and/or complications in the near future if untreated.

There were no statistically significant differences among each of the three BMD groups in histological grade and clinical stage, biochemical variables and the presence of osteoporotic risk factors. Thus, it is impossible to predict which patients are most likely to be osteoporotic from these presenting features, without formal measurements of BMD. There was a greater incidence of osteoporosis in men with prostate cancer than in the age-matched controls. Unlike some studies comprising only healthy community volunteers, the present controls were mainly patients with other urological and medical conditions. This emphasizes the significance in the current study; the results show that advanced prostate cancer is a risk factor for osteoporosis, even before ADT.

BMD is an important determinant of fracture risk [16]; prospective studies show that the risk of fracture increases progressively with decreasing BMD. The predictive value of BMD for fracture is at least as good as that of blood pressure for stroke [17]. Measuring BMD using DEXA is most accurate in skeletal regions with a substantial trabecular bone component, the more metabolically active region of bone than cortical bone. Axial osteometry, particularly at the hip, is usually the preferred choice for predicting risk of fracture. However, no one technique serves all the functions of skeletal assessment (diagnosis, prognosis and monitoring of treatment).

Immediate access to static DEXA scanners is variable, with many regions having a significant burden of referrals and hence prolonged waiting times. Early detection of osteoporosis in men with prostate cancer before ADT may influence the choice of cancer treatment, and more immediate access to BMD assessment is essential to avoid unnecessary treatment delay.

Peripheral DEXA devices, e.g. the forearm scanner used here, confer greater accessibility because of their portability and provide a credible alternative means of measuring BMD. Furthermore, the distal forearm and calcaneum were originally the two common skeletal sites for bone densitometry assessment [18]. Most modern systems are designed to make limb positioning easier, thereby optimizing reproducibility, with a 2.5% coefficient of variation in the ultra-distal radial site [19]. The high mobility of the forearm and low precision error from bone densitometry at this region are further advantages [20].

The ultra-distal radius region has comparable amounts of trabecular bone (65%) to that of the hip or spine, making it an attractive site for bone densitometry [21], and providing precise BMD measurements with a high predictive value for fracture risk. Data from a recent study exploring the prevalence of osteoporosis in men already being treated for prostate cancer with ADT suggested that the BMD of the radius might be preferable to axial osteometry [22]. Furthermore, BMD measurements at the spine may be falsely high because of vertebral osteophytes and calcification of paravertebral structures [23], a common finding in the age group studied, and of particular note in prostate cancer is the presence of spinal metastases.

Bone densitometry DEXA at the distal forearm gives a comparatively low dose of radiation and is neither physically demanding for the patient nor requires advanced technical expertise by the physician, and takes ≈ 5 min. This system enabled measurements of BMD within a week of the clinical diagnosis of cancer, showing its practical advantages. However, an awareness of all osteoporosis risk factors is equally important in the overall assessment of fracture risk.

The relationship between prostate cancer and its distal non-metastatic effects on bone metabolism may be influenced by several mechanisms. Prostate cancer has profound effects on skeletal metabolism and causes mild parathyroid overactivity [24], which occurs in the presence of active sclerotic metastasis, thereby affecting osteoblast activity [25]. Other possible mechanisms include interference with physiological bone remodelling by the abnormal release of the hormones and paracrine factors normally involved in the modulation of osteoblastic and osteoclastic activity [26], occurring predominantly in trabecular bone [27]. Studies of bone turnover markers are necessary to explore these mechanisms further. Factors that may influence collective bone turnover may depend on the patient's age, stage of the neoplastic process, presence of bony metastasis, and possibly other yet unknown tumour characteristics. A histological study reported osteoporosis both adjacent and distal to skeletal metastases, implying that the mechanisms are probably complex [28]. A better understanding of these processes should lead to improved treatment of prostate cancer.

There are no previous studies examining the incidence of osteoporosis in patients with advanced prostate cancer at the time of diagnosis. Greater public awareness has led to earlier presentation of the disease, particularly in North America, where there is now a tendency to commence ADT at a much earlier stage. Investigators found no increase in osteoporosis before ADT [6], but the mean (sd) PSA was only 4.33 (7.70) ng/mL, signifying a much smaller tumour bulk than in the present patients. Furthermore, previous studies on fracture outcome in men with prostate cancer receiving ADT may have underestimated the true incidence of bone fractures by excluding patients with recent osteoporotic related fractures [12]. However, this is exactly the group which is predicted to be at greatest risk after commencing treatment with LHRH analogues.

More recent evidence suggests that early ADT may improve survival in prostate cancer [29]; nevertheless, such treatment schedules will result in men with no distant metastases being given LHRH analogues over potentially longer periods, which is likely to increase skeletal fragility. Therefore, an alternative form of hormone manipulation, e.g. antiandrogens, which preserve testosterone levels [30], or prophylactic administration of bisphosphonate, may be preferable [31].

In summary, this study suggests that all men with advanced prostate cancer should have their BMD measured before treatment, to select appropriate treatment for those who are already osteoporotic. Furthermore, we advocate the long-term monitoring of BMD in patients receiving ADT. The use of portable peripheral DEXA scanners allows practical, accurate and rapid measurements of BMD.