Osteoporosis Before Lung Transplantation: Association with Low Body Mass Index, but Not with Underlying Disease



Due to progress in lung transplantation, post-transplantation osteoporosis becomes an important problem. We determined bone mineral density (BMD) in 74 lung transplantation candidates, among them 24 patients with cystic fibrosis, 16 with chronic obstructive pulmonary disease, 14 with pulmonary fibrosis, and 11 with pulmonary hypertension. The mean T score (± SD) was − 2.6 ± 1.3 at femoral neck (FN), − 2.2 ± 1.6 at Ward's triangle (WT) and −2.3 ± 1.5 at lumbar spine (LS). Osteoporosis was found in 61% of the patients at FN, 45% at WT and 50% at LS. Patients with different underlying lung diseases were similarly affected, not only those with cystic fibrosis but also others, including patients with pulmonary hypertension. No association was found between BMD and age, gender, menstrual condition in women and testosterone level in men. A negative correlation was found between chronic glucocorticoid use and T scores. Body mass index correlated positively (p < 0.01) with T scores at any site and the correlation was also significant for the 2 largest subgroups. Loss of lung function (FEV1) also was associated with lower T scores. No correlation was found between BMD and biochemical indices of bone turnover. Multivariate analysis revealed BMI and glucocorticoid use as independent risk factors. We conclude that osteoporosis is a very common condition in patients with end-stage pulmonary disease, independent of the underlying diagnosis. In view of additional bone loss under immunosuppressive treatment after lung transplantation, early diagnosis and prevention of osteoporosis in the pretransplant period should receive high priority.


Bone loss and fractures are common complications in patients after lung transplantation. Since survival after transplantation has markedly improved and the number of procedures has increased over the last few years, post-transplantation osteoporosis has become an emerging problem, which severely affects quality of life after surgery (1–3). Post-transplantation bone loss is an important side-effect of the immunosuppressive treatment (glucocorticoids and cyclosporine). Many patients with advanced lung disease already have multiple risk factors for osteoporosis before transplantation, e.g. long-term therapy with glucocorticoids, low body weight, low vitamin D levels, reduced physical activity, or a history of smoking.

Although bone loss has been recognized as an important problem, so far the prevalence of osteoporosis in patients awaiting lung transplantation has been reported only from a few centers (4–7).

Moreover, most studies dealing with osteoporosis before lung transplantation concentrated on the classical risk groups of patients with chronic obstructive pulmonary disease (COPD) or cystic fibrosis.

The purpose of this study was: (i) to determine the prevalence of osteoporosis before lung transplantation including patients with underlying diseases other than COPD or cystic fibrosis (e.g. patients with pulmonary hypertension); (ii), to investigate whether generally accepted risk factors for osteoporosis (age, gender, body weight, glucocorticoid therapy) would play a role in patients awaiting lung transplantation; and (iii) to determine whether there is a correlation between functional parameters (12-min walking test and forced expiratory volume in 1 s (FEV1)) and pretransplantation bone mass.

Materials and Methods

Study population

Among 221 patients evaluated for lung transplantation between November 1992 and May 1998 at the University Hospital of Zurich, 86 were accepted for lung transplantation (8). Osteoporosis data were collected prospectively. In 74 of these 86 patients osteoporosis work-up [including bone mineral density (BMD) measurement] was complete. These patients consisted of 41 women and 33 men with a mean age of 39 ± 12 years (range: 16–59 years). Patients showed end-stage respiratory failure due to cystic fibrosis (n = 24), COPD (n = 16), pulmonary fibrosis (n = 14), pulmonary hypertension (n = 11), bronchiectasis (n = 4), lymphangioleiomyomatosis (n = 3), sarcoidosis (n = 1), and silicosis (n = 1).

For each patient, history was taken with regard to the use of glucocorticoids and smoking. To investigate the influence of glucocorticoids on BMD, patients were divided into two groups: individuals receiving glucocorticoids for more than 6 months before transplantation were assigned to the first group ‘chronic use of glucocorticoids’; patients receiving no glucocorticoids or for less than 6 months (e.g. a glucocorticoid trial in COPD some time ago) were assigned to the second group. Use of inhaled steroids was not taken into consideration. (Patients were also asked for additional drug treatment including osteoporosis prevention or therapy. However, data analysis turned out to be unrevealing.)


BMD was determined by dual-energy X-ray absorptiometry (DEXA) with quantitative digital radiography (QDR 2000, Hologic Inc., Waltham, Massachusetts, USA). Examination was performed at three skeletal locations: (a) femoral neck (FN), (b) Ward's triangle (WT), and (c) lumbar spine (LS). The results of the measurements were expressed as grams per square centimeter (g/cm2), as T scores and as Z scores. The Z score utilizes age-matched reference ranges, the T score is defined as numbers of standard deviations below peak bone mass. Osteoporosis, as defined by the World Health Organization, is present when the T score is below −2.5. Osteopenia or low bone mass is defined by a T score between −1.0 and − 2.5 (9).

Laboratory measurements

Serum calcium, phosphate, creatinine and alkaline phosphatase were determined by Hitachi Autoanalyzer 747. Intact parathyroid hormone (PTH) was measured by an immunoassay (Allegro Intact PTH IRMA, Nichols Institute, San Juan Capistrano, CA, USA). 25-hydroxyvitamin D and osteocalcin were measured by radioimmunoassay (Incstar Corporation, Minnesota, and Nichols Institute, San Juan Capistrano, CA, USA). Serum total testosterone was determined by RIA using a commercial kit (CIS Biointernational, Oris Industries, Gif-Sur-Yvette, France). Urinary deoxypyridinoline was measured by an ELISA (Metra Biosystems Inc., Mountain View, CA, USA) and expressed relative to creatinine excretion. Serum and urine samples were taken in the morning (between 7 am and 8 am).

Statistical analysis

Results are expressed as mean ± SD. The unpaired t-test with Bonferroni correction was used to compare BMD between groups with regard to diagnosis, gender, menstrual condition, use of glucocorticoids and tobacco. Simple regression analysis was used to establish correlations between BMD and age, body mass index, parameters of lung function tests and laboratory parameters, and to analyze correlations between laboratory measurements, lung function tests and clinical data. Simple analysis of variance was used to analyze differences between diagnostic groups. Multivariate analysis of variance with covariates was used to detect independent risk factors for osteoporosis. The level of significance was defined as a p-value below 0.05.



Osteoporosis and osteopenia were very common findings in the study population, and only 9% of all patients had normal values for BMD at FN.

The mean pretransplantation BMD (expressed in g/cm2) and T scores were low at FN (0.64 ± 0.14), WT (0.55 ± 0.18), and LS (0.81 ± 0.13) (Table 1). Osteoporosis was found in 45 (61%) of the 74 patients at FN, in 33 (45%) at WT, and in 36 (50%) at LS. Osteopenia was found in 23 (31%) patients at FN, in 29 (39%) at WT, and in 24 (33%) at LS (Figure 1).

Table 1. : Age, weight, BMI, T scores, and Z scores of patients evaluated for lung transplantation grouped according to underlying disease
  1. T scores and Z scores were determined at femoral neck (FN), Ward's triangle (WT) and lumbar spine (LS) and showed no statistically significant differences between the subgroups. Z scores were significantly lower in CF patients than in the other patients at FN.

Age (years)39.2 ± 11.726.8 ± 5.448.3 ± 5.350.3 ± 7.339.5± 11.841.4 ± 7.0
Weight (kg)56.7 ± 14.149.3 ± 5.959.3 ± 15.565.6 ± 16.362.5 ± 16.650.6 ± 6.5
BMI (kg/m2)20.2 ± 4.717.6 ± 1.620.7 ± 4.824.6 ± 5.621.5± 4.719.0 ± 2.5
T score FN− 2.6 ± 1.3− 2.8 ± 0.9− 2.9 ± 1.8− 2.4 ± 0.9− 2.2 ± 1.5− 2.6± 1.3
Z score FN− 2.0 ± 1.4− 2.7 ± 0.8− 1.9 ± 1.7− 1.3 ± 1.2− 1.6 ± 1.6− 2.0 ± 1.3
T score WT− 2.2 ± 1.6− 1.8 ± 1.3− 2.8 ± 2.0− 2.2 ± 1.0− 1.7 ± 1.7− 2.5± 1.4
Z score WT− 1.2 ± 1.5− 1.5 ± 1.3− 1.5 ± 1.8− 0.6 ± 1.2− 0.9 ± 1.9− 1.4± 1.4
T score LS− 2.3 ± 1.5− 2.7 ± 1.3− 2.8 ± 1.6− 1.7 ± 1.4− 1.4 ± 1.7− 2.4± 0.9
Z score LS− 2.0 ± 1.6− 2.6 ± 1.4− 2.4 ± 1.6− 1.2 ± 1.6− 1.1 ± 1.7− 1.9 ± 1.1
Figure 1.

Frequency distribution of T scores. The figure shows the frequency distribution of T scores in the whole group at femoral neck (A), Ward's triangle (B) and lumbar spine (C). Black columns mark T scores below − 2.5 (osteoporosis), hatched columns mark T scores between − 2.5 and − 1.0 (osteopenia) and crosshatched columns mark normal T scores.

T scores were low in all diagnostic groups, interestingly, also in patients with pulmonary hypertension. Osteoporosis was present in 16/24 (67%) with cystic fibrosis, in 11/16 (69%) with COPD, in 6/14 (43%) with pulmonary fibrosis, and in 6/11 (55%) with pulmonary hypertension. BMD and T scores did not significantly differ between the different groups (Table 1).


Female and male patients had a comparable age (39 ± 11 and 40 ± 13 years, respectively). In contrast to findings in healthy individuals, only a weak relationship between age and T score was detected. T scores correlated with age in a significant manner only at WT (r = 0.24; p < 0.05). However, patients with CF were considerably younger than patients with other respiratory diseases (Table 1) therefore these patients had particularly low Z scores (Table 1).


T scores did not differ significantly between female and male patients, but there was a trend to lower T scores in males at any site.

Of the 41 women, 12 were amenorrhoic and 27 were menstruating, being on average 15 years younger. For two patients after hysterectomy the estrogen status could not be estimated by clinical criteria. T scores were similar in women with and without menstruation (Table 2).

Table 2. : T scores and Z scores grouped according to gender, testosterone levels in men, and menstrual condition in women
 MenLow testo-
(< 9.4 nmol/L)
Normal testo-
(9.4–37 nmol/L)
  1. T scores and Z scores were determined and are shown in the whole group and separated according to sex and testosterone levels in men, and menstrual condition in women, respectively. No statistically significant differences were found. When expressed as Z scores, values were significantly lower in the premenopausal than in the postmenopausal women at femoral neck (FN), but not at Ward's triangle (WT) and lumbar spine (LS).

Number33 719412712
Age (years)39.8 ± 12.741.1 ± 13.837.4 ± 11.938.8 ± 10.948.9 ± 7.333.4± 8.4
T score FN− 2.8 ± 1.4− 2.8 ± 1.0− 3.2 ± 1.2− 2.5 ± 1.3− 2.6 ± 1.2− 2.2± 1.5
Z score FN− 2.1 ± 1.5− 2.1 ± 1.4− 2.5 ± 1.3− 1.9 ± 1.3− 2.3 ± 1.2− 1.2 ± 1.3
T score WT− 2.5 ± 1.4− 2.6 ± 1.2− 2.7 ± 1.2− 2.0 ± 1.7− 1.9 ± 1.7− 2.5± 1.9
Z score WT− 1.5 ± 1.3− 1.5 ± 1.5− 1.8 ± 1.2− 1.0 ± 1.7− 1.3 ± 1.7− 0.5± 1.7
T score LS− 2.7 ± 1.3− 3.0 ± 1.6− 2.9 ± 1.1− 2.0 ± 1.6− 2.2 ± 1.4− 1.6± 1.9
Z score LS− 2.4 ± 1.4− 2.7 ± 1.9− 2.6 ± 1.3− 1.6 ± 1.6− 1.9 ± 1.5− 1.9 ± 1.8

Testosterone levels were determined in 29 men. The average testosterone level was 14.6 ± 6.3 nmol/L. Seven (24%) patients had a testosterone level below the lower limit of normal. T scores were similar in patients with low and normal testosterone levels (Table 2). A correlation of testosterone levels with T scores was not found.

Body weight and BMI

Patients with cystic fibrosis had a lower BMI than patients with COPD, pulmonary fibrosis, or pulmonary hypertension.

Body weight and BMI correlated significantly with the T scores at all sites (Figure 2 and Table 3). Positive correlations between BMI and T scores were also found within the subgroups of patients with cystic fibrosis and COPD, whereas for the smaller subgroups numbers did not reach statistical significance.

Figure 2.

Correlation between T scores and BMI. The figure shows the correlation of T scores with BMI at femoral neck (A), Ward's triangle (B) and lumbar spine (C). Positive correlations between T scores and BMI were found within the subgroups of patients with cystic fibrosis and COPD, whereas for the smaller subgroups numbers did not reach statistical significance.

Table 3. : Summary of correlations between body weight, BMI, and FEV1 (%) given as r-values
Body mass
index (BMI)
FEV1 (%)
  1. p < 0.05 = *

  2. p < 0.01 = **

  3. Femoral neck (FN), Ward's triangle (WT), lumbar spine (LS).

BMD FN (g/cm2)0.62**0.55**0.27*
T score FN0.56**0.52**0.23*
Z score FN0.66**0.64**0.29*
BMD WT (g/cm2)0.41**0.38**0.25*
T score WT0.39**0.35**0.19
Z score WT0.56**0.57**0.26*
BMD LS (g/cm2)0.50**0.44**0.38**
T score LS0.49**0.47**0.37**
Z score LS0.49**0.48**0.37*

Glucocorticoid use, smoking

T scores were lower in patients with ‘chronic use of glucocorticoids’ than in patients without glucocorticoids at all 3 skeletal sites, significantly at WT and LS. The difference at FN was nearly significant (p = 0.07). Data are shown in Table 4.

Table 4. : T scores, Z scores, and glucocorticoid therapy
 No glucocorticoid
Chronic use of
  1. T scores and Z scores are shown in patients without and with a history of chronic glucocorticoid use. Chronic use of glucocorticoids was associated with lower T scores at femoral neck (FN), Ward's triangle (WT) and lumbar spine (LS), significantly *(p < 0.05) at WT and LS. Z scores were not significantly different at any location.

T score FN− 2.3 ± 1.6− 2.8 ± 1.3
Z score FN− 2.0 ± 1.2− 2.0 ± 1.5
T score WT*− 1.5 ± 1.5− 2.6 ± 1.5
Z score WT− 1.0 ± 1.4− 1.4 ± 1.6
T score LS*− 1.8 ± 1.3− 2.6 ± 1.5
Z score LS− 1.7 ± 1.3− 2.2 ± 1.7

Smoking, although a risk factor for osteoporosis in the normal population, was not associated with lower T scores.

Functional parameters: 12-min walking test and FEV 1

During evaluation for lung transplantation, a 12-min walking test was performed to estimate the capacity for physical performance and mobility. The mean distance was 465 ± 203 meters. Women and men reached similar distances (480 ± 211 meters and 453 ± 196 meters, respectively). There was no difference between the diagnostic groups. No correlation between walking distance and T score was found.

Mean FEV1 was 1.0 ± 0.7 L. FEV1 correlated with T scores at LS. Mean FEV1 percent of predicted was 32 ± 20%. Similarly, FEV1 percent of predicted correlated with T scores at FN and LS (Table 3).

Laboratory values

The results of the laboratory measurements are summarized in Table 5. 25-hydroxyvitamin D levels were available in 63 patients. Although the mean levels were in the normal range, in 16 patients (25%), the concentration was below the lower limit of normal [cystic fibrosis: 4 of 22 (18%), COPD: 5 of 12 (42%), pulmonary fibrosis: 2 of 14 (14%), pulmonary hypertension: 2 of 11(18%)]. 25-hydroxyvitamin D levels correlated positively with albumin (r = 0.40; p < 0.01) and negatively with parathyroid hormone concentrations (r = − 0.26; p < 0.05). Other associations with 25-hydroxyvitamin D were not found, including BMD values. Osteocalcin, a marker of bone formation, was determined in 61 patients. Twenty-nine (48%) of the patients had an osteocalcin level below the normal range. Osteocalcin levels correlated negatively with age (r = − 0.31; p < 0.05) and positively with deoxypyridinoline excretion (r = 0.24; p = 0.06). Deoxypyridinoline excretion (in a morning spot urine) was measured in 63 patients. Forty-three (68%) patients had an increased value. A relation between osteocalcin or deoxypyridinoline to BMD was not found. Parathyroid hormone was determined in 62 patients. Five (8%) patients had an increased level. Parathyroid hormone correlated in a positive manner with serum creatinine (r = 0.48; p < 0.01) and alkaline phosphatase (r = 0.30; p < 0.05) but not with the urinary deoxypyridinoline excretion, and in a negative fashion with vitamin D (r = − 0.26; p < 0.05). A significant correlation between parathyroid hormone levels or any other laboratory parameter and BMD was not found.

Table 5. : Laboratory parameters of bone metabolism
Parameters Reference values
  1. Results are given in mean ± SD.

Calcium 2.3 ± 0.12.1–2.6 mmol/L
Phosphate 1.0 ± 0.20.6–1.3 mmol/L
Albumin  41 ± 736–50 g/l
Creatinine  81 ± 19< 105 µmol/L
25-OH-Vitamin D  18 ± 1110–42 µg/L
Testosterone14.6 ± 6.39.4–37.0 nmol/L
PTH  39 ± 3610–65 ng/L
Osteocalcin 4.3 ± 3.03.4–11.7 µg/L
Urinary deoxypyridinoline/
creatinine ratio
 9.6 ± 9.32.5–5.0 nmol/mmol

Independent risk factors

To identify independent risk factors for osteoporosis, stepwise analyses of variance with covariates for T scores at FN, WT and LS were performed. Independent variables were age, gender, diagnostic groups (5 levels), BMI, glucocorticoid use, and FEV1% of predicted. BMI (p < 0.0001) and glucocorticoid use (p < 0.02) turned out to be independent risk factors for all analyses. Age was an independent risk factor only at WT. No other factors were found. Although BMI and glucocorticoid use are highly significant risk factors for osteoporosis, their predictive value is relatively poor (adjusted R2 between 0.30 and 0.37).


In this study, osteoporosis was found to be very common in patients with various end-stage lung diseases awaiting lung transplantation. Overall, osteoporosis was present in 60% of patients, and osteopenia in 31%. Only 9% of our lung transplantation candidates showed normal bone mass. Other studies also found that decreased bone mass is a widespread condition in lung transplantation candidates (4–6, 10). However, the proportion of patients with osteoporosis was higher at our institution.

To our knowledge, this is the first study to report that patients with different underlying lung diseases are similarly affected. Thus, not only patients with cystic fibrosis or COPD have significantly decreased bone mass, but also patients with pulmonary hypertension. In previous studies, specific risk factors for osteoporosis were attributed to specific patient groups. However, we found that risk factors considered typical for cystic fibrosis (11, 12), such as low body weight and malnutrition [low albumin, low vitamin D (in 25% of patients), and sex hormone deficiency] were equally present in patients with other underlying diseases. Remarkably, there was no correlation between vitamin D or sex hormone deficiency and low BMD. In contrast, low body weight significantly correlated with low BMD in all patient groups.

In the general population higher age, female gender, and low body weight are accepted risk factors for osteoporosis. In contrast, neither age nor gender nor sex hormone deficiency appeared to have a major impact on BMD in our study population (Table 2). However, a strong association of BMD with body weight and BMI was detected. Our results are in accordance with the study of Aris et al. (5), who found a similar correlation in lung transplantation candidates. Similarly, in the general population, a protective influence of high body weight on BMD has been found in the Framingham study (13). More recent studies suggest that weight loss is associated with loss of BMD (14).

Patients with COPD or pulmonary fibrosis often have a history of chronic glucocorticoid therapy. Osteoporosis is a well-known, dose-dependent side-effect of glucocorticoid therapy (15). Glucocorticoids exert adverse effects on BMD by several mechanisms, including decreased bone formation (16). In our study, BMD was significantly lower in patients with a history of chronic use of glucocorticoids.

Since physical activity has been shown to be important in the prevention of osteoporosis (17, 18), we investigated if the 12-min walking test, a global test of cardiopulmonary capacity, correlates with BMD. We did not find a significant correlation between BMD and the 12-min walking distance. One explanation is that the walking test at the time of evaluation does not sufficiently reflect the physical activity over past years, or that the impact of reduced physical activity on BMD might be overridden by other risk factors. Possibly, loss of body weight is a more stable indicator for loss of muscle mass than walking distance at evaluation for lung transplantation.

In contrast, lower FEV1 values were associated with lower BMD. This might be due to the fact that lower FEV1 reflects not only obstruction to airflow but also loss of muscle mass.

Specific risk factors for osteoporosis in patients with pulmonary hypertension are not known. Recently, mutations in the bone morphogenetic protein-signaling pathway (BMPR-II) have been described in familial and sporadic forms of primary pulmonary hypertension (19, 20). Increased susceptibility to osteoporosis could be associated with these mutations.

No associations of BMD and biochemical indices of bone turnover were found. A negative correlation between parathyroid hormone and vitamin D and a positive correlation between parathyroid hormone and creatinine were detected as expected. Nevertheless, secondary hyperparathyroidism is not a frequent problem in patients awaiting lung transplantation since neither elevated values for parathyroid hormone or creatinine nor low 25-hydroxyvitamin D levels were associated with low T scores. Overall, there was a trend for a positive correlation between serum osteocalcin levels and urinary deoxypyridinoline, but osteocalcin levels were often low despite increased deoxypyridinoline levels, suggesting impaired coupling of bone formation to previous resorption.

According to multivariate analysis, BMI and glucocorticoid use were strong and independent risk factors for osteoporosis.

In conclusion, we showed that osteoporosis is a very frequent condition in patients with advanced lung disease. Intriguingly, patients with different underlying diagnosis were similarly affected.

Osteoporosis goes along with considerable morbidity and decreased quality of life, as also shown for patients with bone disease after lung transplantation (21). Therefore, early diagnosis, prevention, and treatment of osteoporosis in any patient with advanced lung disease should receive high priority.


The authors thank R. Naef, lung transplantation coordinator, for collecting data, and Dr T. Bianda for helpful discussions.

This work was supported in part by the Silva Casa Foundation, Switzerland, the Swiss National Science Foundation (No. 32–46808.96), and the Zurich Lung League.