Predictors of low bone mineral density in elderly males with chronic obstructive pulmonary disease: the role of body mass index

Abstract

Objective. The aim of this study was to investigate the relationships between nutritional indices (Body mass index (BMI), serum albumin), sarcopenia, bone mineral density (BMD) and the severity of their pulmonary obstruction in elderly patients with chronic obstructive pulmonary disease (COPD).

Methods. The method involved was a prospective transversal study; 82 males >65 years old, 41 stable patients with COPD and 41 healthy elderly individuals (controls). All subjects underwent spirometry, biochemical analyses and dual energy X-ray absorptiometry. The significance of the differences between mean values and prevalence rates was tested. The relationships between BMD and independent predictors were analysed by multiple linear regressions. Logistic regression models were applied on dichotomised variables.

Results. In patients with COPD, the prevalence of osteoporosis was higher in subjects with sarcopenia (46% vs. 0%; p < 0.05) and with BMI < 25.1 kg/m² (58% vs. 15%; p < 0.02). Multiple regression analysis indicated that BMI, appendicular skeletal muscle mass (ASMM), albumin, and forced expiration volume after 1 s (FEV1) explained the 70% of BMD variability at the hip and 56% at the spine. Logistic regression showed that a BMI < 25.1 kg/m² was independently associated with osteoporosis risk (OR = 10.0; 95%CI 1.3–76); no independent effect emerged for FEV1% (<and ⩾50%).

Conclusion. In elderly males with COPD, the BMI values ≤25 kg/m² are more strongly related to low BMD levels than albumin values. Among those patients, BMI values within the normal range for younger adults might point out a higher risk of osteoporosis.

Keywords:

• Body mass index
• albumin
• sarcopenia
• airflow obstruction
• osteoporosis

Introduction

Bone loss has been recognised as one of the systemic effects of chronic obstructive pulmonary disease (COPD) [1]. Many factors are involved in this process, including airflow obstruction, but nutritional disorders may have a key role [2–4]. It is known, in fact, that malnutrition and underweight are common among individuals suffering from COPD, supported by a negative energy balance due to a low dietary intake and a higher than normal resting energy expenditure [5,6]. Also muscle mass depletion (sarcopenia), which occurs physiologically with ageing and is exacerbated by COPD, may have an additional role -in combination with malnutrition- in bone frailty [7].

Body mass index (BMI) is widely used for nutritional assessment and several studies have shown that a low BMI is associated with bone loss in elderly men and women [8,9].

In our hypothesis, in patients with COPD the presence of a chronic inflammatory condition and the airflow obstruction per sè might cause bone loss even when nutritional indices suggest a normal nutritional status and muscle mass is not much depleted. It could be reasonable to assume that COPD might lead to bone loss and the onset of osteoporosis even at normal BMI values. It could be also likely that hypoalbuminemia and sarcopenia (measured by the appendicular skeletal muscle mass: ASMM) might represent useful indices for the early identification of patients with COPD at risk of bone mass loss, irrespective of their BMI.

The aims of this study were to investigate the relationship between some nutritional indices (BMI, serum albumin), sarcopenia (ASMM) and bone mineral density (BMD) at different sites in elderly patients with COPD, also in relation to the severity of their pulmonary obstruction.

Methods

Study subjects

This survey was conducted at the Geriatrics Department of Padua University on 82 males over 65 years of age, 41 of whom had COPD, diagnosed according to the European Respiratory Society criteria [10], (age range 66–85 years). The patients with COPD were recruited consecutively from among the outpatients attending the Respiratory Pathophysiology Department in Padua for routine check-ups. Only patients with COPD in stable clinical conditions were considered. Patients with fever, worsening respiratory symptoms, leukocytosis (WC > 11.00 × 10⁹/l), an increased erythrocyte sedimentation rate (ESR) (>34 mm/h), changes to their medication in the previous 30 days, or hospital admissions in the previous 6 weeks were considered clinically unstable and ruled out. A history of oral or systemic corticosteroid use for more than 3 months consecutively was also reason for exclusion from the study. The allowable drugs were beta 2 agonists (taken by 64% of the patients recruited), inhaled corticosteroids (56%) and anticholinergic agents (41%).

Forty-one healthy elderly people were considered as a control group, recruited on a voluntary basis from among the elderly people attending the Geriatrics Department for a check-up (age range 66–87 years). Their healthy condition was established from their clinical history, clinical examination, normal biochemical tests and normal lung function (forced expiration volume after 1 s: FEV1% > 80).

Other exclusion criteria common to all participants were: BMI < 22 or ⩾30 and cognitive or physical impairments severe enough to prevent them from taking the tests. The choice of the common exclusion criteria was finalised to the creation of two groups comparable for all variables potentially associated with BMD.

The study design complied with the provisions of the Helsinki Declaration of 1995. Participants were fully informed as to the nature and purpose of the study and gave their informed consent.

All subjects took several tests on the same morning, as follows:

Physical function and smoking habits

The Rosow and Breslau [11] and NAGI [12] scales were applied to patients and controls. The NAGI scale investigates nine items (able in bending, stooping or crouching, pushing or pulling a great load like a chair, lifting loads inferior or superior to 4.5 kg, reaching shoulders, writing or using small objects, standing during 15 min, sitting during 30 min) and scores from 0 (totally able in the investigated physical functions) and 9 (totally unable). The Rosow and Breslau scale assesses three gross mobility items (doing heavy works at home or outdoor, climbing stairs, walking 800 m) and scores from 0 (totally able) or 3 (totally unable). Information on smoking habits was obtained from the patients, who were classified as non-smokers, ex-smokers and current smokers. For smokers, the number of cigarettes/day and of years of smoking habit was also recorded.

Anthropometry

Body weight was measured to the nearest 0.1 kg and height was measured to the nearest 0.1 cm with subjects wearing light clothing and no shoes. BMI was calculated as the weight in kilograms divided by the height in meters squared.

Pulmonary function tests

Lung function was assessed using a computerised water-sealed Stead-Wells spirometer (BAIRES System, Biomedin, Padua, Italy), which meets the 1994 ATS recommendations for diagnostic spirometry. An arterial blood sample was drawn from the radial artery for blood gas analysis of PaO₂ and PaCO₂.

Biochemical analyses

Blood samples were obtained in the morning after a 12-h overnight fast. Besides normal routine biochemical parameters, analyses were performed for principal nutritional indices as well as albumin, pre-albumin and retinol-binding protein (RBP). Among skeletal markers 25-OH D3, total calcium and phosphate were collected, while it was not possible to investigate the parathyroid hormone, the alkaline phosphatase, osteocalcin or other products of bone turnover.

Bone and body composition

Bone parameters, fat-free mass (FFM) and fat mass (FM) were measured using a dual-energy X-ray fan beam densitometer (Hologic QDR 4500 W) and dedicated software, rel 8.2. BMD, T-score and Z-score of the lumbar spine and the hip (neck, trochanter and total hip) were measured using standard protocols. The definition of osteoporosis and osteopenia was conventionally based on the number of standard deviations (T-score) below the mean BMD of a gender-specific, young adult reference population [13]: a T-score between −1.5 and −2.5 indicates osteopenia, a T-score < −2.5 indicates osteoporosis.

ASMM was also measured as the sum of the fat-free soft tissue masses of the arms and legs [14]. The ASMM index (ASMMI) was calculated as the ASMM divided by height squared. As a cut-off for a definition of sarcopenia, we used 7.26 kg/m², as proposed by Baugmartner for males, which represents the sex-specific ASMMI value two standard deviations below the ASMMI mean value for young adults [15]. We also identified a pre-sarcopenic condition in subjects whose ASMMI was between two and one standard deviations below the reference group's mean, i.e. with an ASMMI in the range of 7.26–8.40 kg/m².

Statistical analysis

Data were analysed using the Systat statistical software, rel. 9.1 for Windows (SPSS, Chicago). Results were expressed as means ± standard deviations. Differences between means in COPD and healthy individuals were evaluated using Student's unpaired two-sided t-test. The prevalence of specific conditions was expressed as a percentage, and differences between groups were evaluated using the χ-squared test, exact Fisher's test or the Cochran-Armitage test for trend, as appropriate. Analysis of variance or the Kruskal–Wallis test were applied, as appropriate, to compare two or more mean values among patients with COPD. Pearson's product-moment correlation coefficient (r) was used to measure simple linear associations between total hip, trochanter, neck of hip and lumbar BMD, and age, ASMM, BMI, serum albumin, FEV1%, FM, FFM, in each group. Stepwise multiple linear regression was used to assess the independent association between BMD at different sites and the significantly associated factors emerging from the previous simple linear analysis (i.e. BMI, serum albumin, ASMM, FEV1%), adjusted for confounding factors (age, smoking habit and use of medication), where necessary. The interaction between covariates was assessed using first- and second-order interaction terms in the regression models. Partial and multiple R² coefficients were estimated and R² was used as a criterion of fit. To test for multi-collinearity, a variance inflation factor was calculated for each independent variable in the model. Values above 2 were used to indicate a multi-collinearity problem in the model.

A confirmatory analysis was performed by applying a multivariate logistic model entering dichotomised predictors. The outcome binary variable was the presence or absence of osteoporosis.

The explanatory variables were categorised as follows: BMI on the first tertile of distribution (BMI < 25.1 kg/m²) and FEV1% on the 50% cut-off, indicating severity of obstruction [16]. The model was adjusted for use of inhaled corticosteroids.

The odds ratio (OR) and 95% confidence interval (CI) for each independently-associated condition were assessed. The level of significance was set at 0.05 for all tests.

Results

The mean age was similar in patients with COPD and control subjects (75.7 ± 5.3 years vs. 76.0 ± 5.8 years; p = 0.38). The Rosow–Breslau and NAGI scales indicated no differences in physical function between the two groups (2.0 ± 1.1 vs. 1.1 ± 1.4; p = 0.13 and 2.2 ± 2.1 vs. 1.4 ± 2.1; p = 0.32). Also, smoking habits did not differ significantly between groups. Anthropometric data, pulmonary indices, nutritional, body composition and bone parameters for the patients with COPD and healthy subjects are shown in Table I. As expected, no significant differences emerged between the anthropometric parameters of the two groups. The spirometric values were significantly lower in the COPD cases than in controls, the former being more hypoxemic and hypercapnic. Mean albumin was significantly lower in patients with COPD, while no significant difference emerged between COPD cases and controls for pre-albumin, RBP, 25-OH D3, total calcium or phosphate. Patients with COPD had significantly lower FFM and ASMM, both in kg and after correction for height squared. BMD was significantly lower in the lumbar region of patients with COPD (0.95 ± 0.17 vs. 1.08 ± 0.16, p < 0.001) and in the neck of the hip (0.73 ± 0.14 vs. 0.80 ± 0.13, p < 0.05), while no significant differences were seen for the other sites considered (Table II). The prevalence of osteoporosis was 29.7% in the COPD group and 5.0% among the healthy elderly individuals (p < 0.01).

Table I.  Anthropometry, blood gas and spirometric parameters, nutritional, body composition and bone indices, in patients with COPD and controls.

	COPD (41), mean ± SD	Controls (41), mean ± SD
Age	75.7 ± 5.3	76.0 ± 5.8
Height (cm)	167.8 ± 6.8	169.3 ± 6.4
Weight (kg)	71.8 ± 12.1	73.9 ± 9.1
BMI (kg/m^2)	25.4 ± 3.5	25.7 ± 2.6
PaO2 (mmHg)	75.0 ± 13.1**	98.38 ± 15.8
PaCO2 (mmHg)	41.6 ± 5.2*	36.8 ± 3.9
FEV1 (l)	1.4 ± 0.5**	2.9 ± 0.6
FEV1 (% predicted)	53.1 ± 17.6**	110.9 ± 22.5
FEV1/VC (%)	48.9 ± 9.8**	73.3 ± 8.3
Albumin (g/l) (n.v. 35–52)	38.3 ± 5.4**	43.0 ± 4.7
Prealbumin (mg/l) (n.v. 200–400)	277.0 ± 63.0	301.8 ± 72.2
RBP (mg/l) (n.v. 30–60)	47.6 ± 12.8	53.6 ± 13.9
25-OH D3 (nmol/l) (n.v. > 80)	53.1 ± 41.0	42.6 ± 37.6
Total calcium (mg/dl) (n.v. 8.1–10.4)	9.5 ± 0.4	9.2 ± 0.5
Phosphate (mg/dl) (n.v. 2.5–4.8)	2.3 ± 1.0	2.7 ± 0.4
FM (kg)	18.0 ± 6.1	17.8 ± 4.8
FFM (kg)	50.7 ± 6.2**	55.3 ± 6.2
ASMM (kg)	20.9 ± 3.2*	22.6 ± 3.2
ASMMI (kg/m^2)	7.4 ± 0.9*	7.8 ± 0.8
BMD lumbar (g/cm^2)	0.95 ± 0.17**	1.08 ± 0.16
T-score lumbar	−1.47 ± 1.53**	−0.23 ± 1.46
Z-score lumbar	−0.26 ± 1.48**	0.89 ± 1.78
BMD total hip (g/cm^2)	0.85 ± 0.18	0.93 ± 0.14
T-score total hip	−1.18 ± 1.22	−0.67 ± 1.11
Z-score total hip	−0.11 ± 1.07	0.35 ± 1.26
BMD neck (g/cm^2)	0.73 ± 0.14*	0.80 ± 0.13
T-score neck	−1.54 ± 1.18*	−0.92 ± 1.17
Z-score neck	0.01 ± 1.07*	0.64 ± 1.31
BMD trochanteris (g/cm^2)	0.661 ± 0.146	0.715 ± 0.128
T-score trochanteris	−0.95 ± 1.23	−0.45 ± 1.16
Z-score trochanteris	−0.18 ± 1.09	0.34 ± 1.29

BMI, body mass index (weight ÷ height squared); FEV1, forced expiration volume after 1 s; VC, vital capacity; FFM, fat-free mass; FM, fat mass; ASMM, appendicular skeletal muscle mass; FFMI, FFM index; FMI, FM index; ASMMI, ASMM index; RBP, retinol-binding protein; BMD, bone mineral density; n.v., normal values.

COPD vs. controls (*p < 0.05,**p < 0.01).

Table II.  Simple and multiple partial R² for linear relationships between hip and lumbar spine BMD and BMI, albumin, ASMM and FEV₁.

	Total hip BMD		Lumbar BMD^†
	Simple linear, R^2	Multiple partial, R^2	Simple linear, R^2	Multiple partial, R^2
BMI	0.60**	0.62***	0.17**	0.00
ASMM (kg)	0.45**	0.01	0.20**	0.00
Albumin(g/l)	0.26**	0.00	0.15**	0.12*
FEV1 (l)	0.14*	0.08**	0.18*	0.04
	Total multiple, R^2 = 0.70		Total multiple, R^2 = 0.56
	Trochanteris BMD		Neck BMD
	Simple linear, R^2	Multiple partial, R^2	Simple linear, R^2	Multiple partial, R^2
BMI	0.41**	0.42***	0.33**	0.32***
ASMM (kg)	0.39**	0.01	0.35**	0.01
Albumin (g/l)	0.28**	0.03	0.28**	0.05*
FEV1 (l)	0.11*	0.08*	0.28*	0.09*
	Total multiple, R^2 =0.54		Total multiple, R^2 = 0.47

^†Regression model adjusted for inhaled corticosteroids (partial R² = 0.40; p < 0.001).

BMD, bone mineral density; BMI, body mass index; ASMM, appendicular skeletal muscle mass; FEV1, forced expiration volume after 1 s.

*p < 0.05; **p < 0.01;***p < 0.001.

In patients with COPD, the prevalence of osteoporosis increased significantly from 0% in the non-sarcopenic group to 46% in the sarcopenic one (Figure 1). According to BMI (cut off: 25.1 kg/m²), the prevalence of osteoporosis in COPD subjects, rose significantly (to 58%) in the lower BMI group (Figure 1) (p < 0.02). In patients with COPD, osteopenia and osteoporosis were not related to PaO₂, PaCO₂, age or smoking habits (data not shown). Only the lumbar BMD value and T-score were statistically associated with inhaled corticosteroids (p < 0.01), their values being lower in steroid users (data not shown).

Figure 1.  Prevalence of normal BMD (open columns), osteopenia (gray columns), osteoporosis (black columns) in patients with COPD by ASMMI class (normal: ASMMI ⩾8.4; pre-sarcopenic: 7.26 < ASMMI < 8.4; sarcopenic: ASMMI < 7.26)*, and BMI dichotomised on first tertile of distribution (25.1 kg/m²)^§. *The statistical analysis was performed comparing the prevalence of osteoporosis between groups: Cochran-Armitage test for trend: p < 0.05; ^§Exact Fisher's test: p < 0.02.

Simple correlation analysis (Table II) on the patients with COPD indicated that BMI, ASMM, serum albumin and FEV1 were significantly associated with BMD in the hip (total, trochanter and neck) and lumbar spine.

In the multiple linear regression (Table II), the model explained 70% of total hip BMD variability (54% in the trochanter and 47% in the neck) and 56% of the spine BMD. At all hip sites, BMI and FEV1 retained their independent role in explaining BMD, while ASMM failed to reach significance. Among the femoral sites, albumin significantly entered the models only for the neck of the hip. At spinal level, only albumin had an independent role in the model, where the inhaled corticosteroids explained 40% of the total variability of the BMD. No collinearity effects emerged with a variance inflating factor greater than 1.7.

In the logistic regression model, a BMI < 25.1 kg/m² was independently associated with the risk of osteoporosis (OR = 10.0; 95%CI: 1.3–76.9). COPD severity resulted not an independent predictor of osteoporosis (p = 0.98).

Discussion

The groups of healthy elderly individuals and patients with COPD did not differ statistically in terms of age, weight, height and BMI, so any differences between their nutritional and bone variables would be attributable to the COPD. These patients, in fact, had significantly lower values for FFM, ASMM, ASMMI and a 14% greater prevalence of sarcopenia than controls. A reduced FFM is a known feature of patients with COPD [17], particularly involving the peripheral skeletal muscles. Age-related muscle loss is accelerated by COPD, especially in the acute phase, when higher levels of circulating pro-inflammatory cytokines cause a higher protein turnover, hypoxemia and acidosis, further restricting physical activity [18]. Similar factors may be responsible for the loss of bone mass observed in patients with COPD [19,20]: as in these previous studies, there was a significantly higher prevalence of osteopenia and osteoporosis in our sarcopenic group of COPD cases. The role of pro-inflammatory cytokines may be crucial to osteoporosis in association with inflammatory disease [21]. Unfortunately, more specific markers of inflammation, such as high-sensitivity CRP, IL1, IL6, etc., were not measured in our patients, who were selected for their normal ESR and leukocyte values as a clinical expression of stable COPD.

Mean BMD, T-scores and Z-scores were lower in our patients with COPD than in controls at all sites, though the differences were only significant for the neck of the hip and the spine. These two sites consist mainly of trabecular bone and have a higher metabolic turnover, making them more sensitive to the catabolic effect of pro-inflammatory cytokines.

BMI has an important role among the numerous factors affecting bone mass in healthy elderly people, as demonstrated in a number of large epidemiological studies [8]. This association has been attributed to biomechanical forces [22] and to a greater aromatisation of androgens at subcutaneous adipose tissue level [23]. In our patients with COPD, BMI explained 60% of the variability in total hip BMD and, in the logistic regression model, the risk of osteoporosis increased 10-fold in the BMI class < 25.1 kg/m². In the older generally, BMI values below 22 (considered normal in younger adults) are already associated with a greater risk of malnutrition, osteoporosis and mortality [9,24]. In men with COPD, BMI values within the normal range (22–25 kg/m²) may already suffice to increase the risk of osteoporosis by comparison with the overweight (BMI > 25.1 kg/m²), the former probably being more sensitive to the effect of weight load on bone, despite their normal nutritional status. A BMI < 25 kg/m² in a patient with COPD should alert the clinician to the possibility of osteoporosis.

In the present study the severity of bronchial obstruction expressed by FEV1 resulted linearly associated with BMD values, although the risk of osteoporosis proved independent of the dichotomised severity of bronchial obstruction. This apparent discrepancy may be due partly to the fact that continuous variables were treated as categorical variables, resulting in the loss of some information. Actually, the severity of lung disease was found inversely correlated with BMD in several previous studies on the general population. The clinical distinction between moderate and severe lung disease, anyway, appears to be useless in distinguishing between patients with COPD with and without a low bone density.

Unlike BMI and FEV1%, sarcopenia no longer correlated significantly with femoral BMD after adjusting for BMI and the other investigated variables. This is probably partly due to the strength of the BMI variable, which also includes muscle mass. In fact, BMI expresses nutritional status and both adipose and muscle mass deficiencies, whereas the parameter used for sarcopenia (ASMM) only considers limb muscle mass, not total muscle mass. Further studies are needed here, using different indicators of sarcopenia, such as strength and muscle function.

Albumin was significantly lower in patients with COPD than in controls and significantly associated with BMD, even in multiple regression for lumbar and neck of hip BMD. The lower mean albumin values may be due to malnutrition and/or chronic inflammation. However, conflicting results have been reported on the relationship between serum albumin and BMD. In our earlier study, underweight hypoalbuminemic subjects had a lower T-score than underweight cases with normal serum albumin levels [25]. Two large cross-sectional studies, nevertheless, failed to confirm this correlation between albumin and BMD [26,27].

The present study has some shortcomings. The first is the lack of specific markers of inflammation, as already discussed above. The second is the absence of women in the study: this is due to the limited number of elderly women among the outpatients attending the Respiratory Pathophysiology Department in Padua for routine check-ups during the time of the study. Another limit could be the restricted number of patients due to severe inclusion criteria. More general selection criteria would bring in a number of factors very likely to interfere with the relationships between BMD and the variables studied, and adjusting these factors in the analysis would demand a much larger sample size.

In conclusion, our findings suggest that BMI expresses the relationship between nutritional status and BMD in elderly COPD males better than other nutritional indices, and that age-related sarcopenia seems not involved in bone mass loss. Severity of bronchial obstruction seems associated to BMD even if not related to a higher risk of osteoporosis, while BMI values < 25 kg/m², though still within the normal range for younger adults, carry a higher risk of osteoporosis in elderly men with COPD.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.