INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Asthma is a chronic inflammatory disorder and inhaled corticosteroids (ICS) are considered the mainstay of its long-term treatment, as suggested by national and international treatment consensuses (1, 2). Low doses of ICS seem to minimally influence bone metabolism, but the long-term effects of higher doses are to be further documented (3, 4). Although reports showed variable results, daily doses equivalent to > 800 µg of beclomethasone were generally found to exert some influence on markers of bone metabolism (5). In this regard, we previously reported a lower serum osteocalcin and higher urinary phosphorus level in subjects using ICS at a mean daily dose of 1,140 µg, determined as the proportion of the prescribed dose taken when estimating compliance to the treatment, for a mean of 3.2 yr, compared with a matched group on low dose or not using ICS (8).

On the other hand, bone density seems unaffected by low doses of ICS whereas the influence of high doses is uncertain (8). Although high doses of ICS may have the potential to reduce bone density, the magnitude of this effect may be influenced by individual susceptibility to steroid-induced osteopenia, dietary and lifestyle-related factors such as sedentarity, a genetic predisposition, and estrogen replacement therapy (12- 15). Up to now, no reliable predictor of bone loss has been found. If we could find a marker that will predict who is at higher risk of bone loss after long-term therapy with corticosteroids, preventive measures could be suggested.

The present study aimed at determining if there were any differences in bone density over a period of 3 yr, between asthmatic subjects to whom high doses of ICS were prescribed compared with no or low doses for a period of at least 5 yr and to verify if the initial bone density level or biological markers of bone metabolism could predict who is at greater risk of bone loss.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Of 74 asthmatic subjects who took part in a previous evaluation of bone density and metabolism 3 yr ago, 51 agreed to come back for a reevaluation (8). These subjects had been recruited from the Laval Hospital asthma clinic and all had a diagnosis of asthma in keeping with the American Thoracic Society criteria (16). Twenty-eight were from the high-dose ICS group (Group HD), initially defined in the first study as taking a daily dose of beclomethasone or budesonide  800 µg for at least 18 mo and 23 from the control group (Group C) using no (n: 8) or < 500 µg of ICS daily (Table 1). All used an inhaled ₂-agonist on demand and none used oral corticosteroids regularly.

Excluded from the initial study were subjects with any other respiratory disease than asthma, endocrine disorders, or a systemic, metabolic, renal, or bone disease such as Paget's disease, osteoporosis, or malabsorption, that could have interfered with bone metabolism. Other exclusion criteria were conditions such as pregnancy or breastfeeding, alcoholism or drug abuse, poor compliance with medication use, and use of oral steroids for more than 2 wk or within the last 3 mo. All subjects were active although most of them were not doing exercise regularly.

Evaluation

Subjects' characteristics. Recordings of subjects' characteristics, including duration of asthma and medication (type of drug, prescribed doses, estimated compliance from questionnaire, duration, number of short-courses [< 15 d of systemic corticosteroid treatment]), were updated. Any factor that could have influenced bone metabolism such as changes in physical activity, calcium intake, and medication use was searched.

Dietary evaluation. A dietary evaluation was done for each subject and all subjects were asked to fill a 3-d diary chart. These last were compared with those filled 3 yr previously. Main nutritional parameters analyzed included proteins, phosphorus, magnesium, calcium, zinc, and vitamins A, C, and D.

Physical activity. A global evaluation of physical activity for the last 3 yr was done using a questionnaire on exercises performed. Intensity was graded semiquantitatively on a scale from 1 to 3 (weak, moderate, intense). The score was determined by the types of exercises practiced regularly and their intensity.

Pulmonary function tests. Expiratory flows and bronchodilator response were measured according to the recommendations of the American Thoracic Society on a wedge spirometer (17). Forced expiratory volume in one second (FEV₁) and forced vital capacity (FVC) were measured before and after the inhalation of 200 µg of albuterol. Predicted values were obtained from a study by Knudson and coworkers (18).

Calcium metabolismserum samples. Both 3 yr ago and at the time of this last evaluation, measurements of plasma cortisol, serum osteocalcin, phosphorus, calcium, creatinine, alkaline phosphatase, bone isozyme (Alkphase-B; Metra Biosystems, Inc., Mountain View, CA), and -glutamyl transpeptidase (GT) were obtained between 8:00 and 10:00 A.M., before the morning dose of ICS. Procollagen C (Prolagen-C; Metra Biosystems, Inc.) was only measured at the last evaluation.

Calcium metabolismurinary samples. A 2-h urinary sample was obtained after the first morning void was discarded for measurement of calcium, phosphorus, and, only at the last evaluation, N-telopeptides levels (Osteomask; NTX, Ostex, Seattle, WA). A 24-h urinary sample was obtained for measurement of creatinine and cortisol. All specimen were kept at 70° C before analysis. Osteocalcin was measured by standard radioimmunoassay (19). Calcium, phosphorus, creatinine, alkaline phosphatase, GT, and cortisol were measured by standard methods.

Bone density. Bone density was determined with the same apparatus both 3 yr before and at the last evaluation, a Hologic-QDR-2000-DX osteodensitometer (Hologic, Boston, MA). Both evaluations were made in a single plane and a second reading was obtained if the standard anteroposterior (AP) evaluation of lumbar spine was not adequate. A single AP and lateral view radiograph of the lumbar spine and AP view of the left hip completed the evaluation. Trabecular bone loss was assessed at the level of lumbar spine and Ward's triangle; cortical bone loss was measured at the level of femoral neck. T score and Z score were evaluated as well as the density. T score is the standard deviation from the mean normal for young normals (subjects < 35 yr old) and Z score is the standard deviation from the mean predicted value for age- and sex-matched control subjects. Osteoporosis was defined by a T score < 2.5 standard deviation according to the criteria of the World Health Organization. In regard to variations in bone mineral bone density with the technique used, we have a reported variation of + or  2.8% (SD) with lumbar spine and 3.5% with hip measurements.

Analysis

We looked at the relationships between the total cumulative dose or mean daily dose of ICS (prescribed dose × estimated intake) and bone density and metabolism parameters, as well as changes in these parameters in the two groups in the 3-yr period separating the two evaluations, in relation to changes in bone density. Means and standard deviations were determined for continuous variables and percentages for dichotomous variables. All comparisons were two-tailed for all outcome measures. The results were considered statistically significant if the p value was < 0.05 or if the 95% confidence interval did not include zero. The two groups were compared using a Student's t test or a Wilcoxon rank sum test, depending on whether the normality and variance homoscedasticity assumptions were encountered or not. Statistical analyses were made using the statistical program SAS (SAS Institute Inc, Cary, NC).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects' Characteristics

The subjects' characteristics are shown in Table 1. Over the last 3 yr, four subjects went from using a low (< 500 µg) dose of ICS to > 1,000 µg and six did the opposite. The results are, however, expressed in keeping with the original categorization of groups although analyses were made both for the whole group of subjects and by excluding those who changed groups compared with 3 yr ago. Only three patients took an inhaled nasal steroid regularly in addition to their asthma medication: fluticasone 100 µg/d (Group HD) and budesonide 100 (Group C) and 200 µg/d (Group HD).

The mean duration of asthma was 22.6 ± 2.7 yr in Group HD, and 18.4 ± 3.0 yr in Group C. The mean daily dose of inhaled corticosteroids used within the last 3 yr was estimated to be 983 ± 80 µg in Group HD and 308 ± 86 µg in Group C. The mean number of oral steroid treatments over the last 3 yr was 1.8 ± 0.44 in Group HD and 0 in Group C. Two subjects of Group HD used oral steroids for a 2-mo period and another intermittently for a total period < 2 mo.

No statistically significant difference was found for the different dietary parameters measured in the two study groups nor for the level of physical activity. Twenty-one patients in Group HD and 22 in Group C were doing regular exercise although generally of mild intensity (mean score ± SEM: Group HD, 5.0 ± 0.6 and Group C, 5.4 ± 0.2). No significant changes in weight were noted over the 3-yr follow-up period in both groups. There were little changes in the menopausal status of the patients during this time period, one woman of Group HD and two of Group C becoming menopausal during the course of the study.

Pulmonary Function Tests

Mean FEV₁ and FVC for the last evaluation are found in Table 1. Group HD had more airflow obstruction than the control group, with a mean baseline FEV₁ and FVC of 68.4 ± 3.5% and 86.6 ± 2.8% compared with 79.6 ± 0.2% and 92.3 ± 3.3%, respectively. After administration of inhaled salbutamol, mean FEV₁ and FVC values were 79.8 ± 3.2% and 97.4 ± 2.7% for Group HD and 90.2 ± 3.3% and 99.8 ± 2.9 for Group C.

Bone Metabolism: Serum and Urinary Parameters

Measurements of serum and urinary parameters are found in Table 2. At the last evaluation, the serum levels of calcium, phosphorus, creatinine, alkaline phosphatase, GT, and plasma cortisol were not significantly different between the two groups (p > 0.05). Serum osteocalcin and procollagen C, and urinary N-telopeptides and urinary cortisol were also similar in the two groups and not correlated with daily dose of inhaled corticosteroids intake (Figure 1).

View larger version (25K): [in this window] [in a new window]   Figure 1.   There was no significant correlation between daily dose of inhaled corticosteroids and osteocalcin, procollagen C, N-telopeptides, and urinary cortisol (p > 0.05, n = 51).

If we compare these results to those obtained 3 yr ago, in the control group, we observed a statistically significant decrease of serum calcium and phosphorus and an increase in osteocalcin (p < 0.001) as well as an increase in urinary creatinine and phosphorus (p < 0.05). These last changes and an increase in alkaline phosphatase and GT were observed in Group HD for the serum parameters (p < 0.05). Urinary creatinine and phosphorus were also increased in the two groups and urinary calcium in Group HD (p < 0.05) (Tables 2 and 3). However, all those changes were small. Comparison of the 3-yr changes in the two groups also showed an increase in urinary calcium in Group HD and a decrease in Group C (p < 0.05) (Table 3). There was no statistically significant correlations between the mean ICS dosage over the 3 yr and the changes in serum and urinary parameters (p > 0.05). The reduction in serum calcium was correlated with the change in bone density (lumbar spine, r = 0.30, p = 0.03, n = 51), but there was no correlation between changes in bone density and the other parameters' changes over the 3-yr period.

Bone Density

Bone density results are shown in Table 2 and Figures 2A and 2B. There were no significant differences in bone mineral density, T score, or Z score at the lumbar spine and hip levels between the two groups both on baseline (3 yr ago) and on the last evaluation (p > 0.05). If we look at the changes in bone density over the last 3 yr, they are minimal (not significant [NS]) and similar in the two groups.

View larger version (18K): [in this window] [in a new window]   View larger version (27K): [in this window] [in a new window]   Figure 2.   (A) There was no significant correlation between inhaled corticosteroids daily dose and bone density parameters, for the whole group analysis (p > 0.05, n = 51). (B) Bone density parameters were not significantly different in the two groups on the last densitometry done at the end of the 3-yr follow-up period.

Furthermore, for the two groups, no significant correlation was found between the change in bone density, and either initial bone metabolism markers, initial bone density, or the level of physical activity (Figure 3). There was no significant correlation between the daily dose of inhaled corticosteroid taken and the change in bone density for the whole group (r = 0.058, p = 0.68) or for Group C alone (r = 0.039, p = 0.86). However, an inverse correlation was found for Group HD between the ICS daily dose and the change in bone density (r =  0.49, p = 0.008), but the magnitude of this change was small. When we only analyze subjects showing a reduction in bone density over the 3-yr period using ICS, mean change (reduction) in bone density per 1,000 µg of ICS was not increasing linearly, being quite less from 0 to 1,000 µg compared with 1,000 to 2,000 with respective values of 0.02 and 0.08 g/cm². Median change was 2.8% in bone density (lumbar spine) per 1,000 µg of ICS (n: 23).

View larger version (21K): [in this window] [in a new window]   Figure 3.   There was no correlation between changes in bone density and initial serum osteocalcin levels, bone density, or physical activity (assessed on a 0 to 20 scale according to the duration and intensity of regular physical exertion). There was, however, a weak correlation between bone loss and mean daily inhaled corticosteroid dose in the high-dose ICS group.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

No significant reduction in bone density was observed over a period of 3 yr in a group of asthmatic patients using beclomethasone or budesonide at a dose of about 1,000 µg daily for at least 5 yr, in comparison with patients using no or < 500 µg/d. On the other hand, we found variable but generally small changes in biological markers of bone metabolism over this time period. Neither baseline bone density nor any of the baseline biological parameters measurements could help to predict changes in bone density over a 3-yr period. A correlation was found between bone loss after 3 yr and ICS daily dose in Group HD although changes were small.

In the last decade, increasing doses of ICS have been used to treat asthma. However, both patients and physicians are concerned with potential effects of high doses of these agents on bone density. Many studies have been conducted on the effects of ICS on bone metabolism but their long-term effects have not yet been documented and potential predictors of bone loss remain to be identified. Osteocalcin is a marker of bone formation released by the osteoblast; it is also released during bone resorption. This last can also be assessed by hydroxyproline changes. New markers of bone resorption are now available, including urinary N-telopeptides and serum procollagen C (20).

Variable changes in bone markers were found in patients using ICS at different doses although high doses seemed able to modify those last bone markers (such as serum osteocalcin and urinary hydroxyproline) in most subjects (5, 8, 10). Kerstjens and coworkers reported two studies on the effect of ICS on bone metabolism (7). These investigators found a decrease in osteocalcin concentrations and in another marker of bone turnover, procollagen type I carboxyterminal propeptide (PICP), with a small increase in serum alkaline phosphatase, after a 4-wk treatment with a dose of ICS > 800 µg/d. No change was observed in type I collagen carboxyterminal propeptide (ICTP), a collagen degradation product, nor in urinary hydroxyproline. After 2.5 yr of treatment in 70 patients using 800 µg/d of beclomethasone compared with 85 only on bronchodilators, no difference was found in PICP between the treatment groups whereas an increase in ICTP was observed in the ICS group. The researchers concluded that long-term changes in bone turnover during treatment with inhaled corticosteroids should not be deduced from short-term studies of single serum parameters of bone metabolism.

In another study, Bootsma and coworkers found a decrease in osteocalcin and in PICP in asthmatic subjects taking a dose of beclomethasone 1,500 µg/d whereas fluticasone 750 µg/d did not change those variables of bone formation (23). Fluticasone and beclomethasone did not increase ICTP or deoxypyridinoline crosslinks markers of bone metabolism over a period of 6 wk.

In regard to bone metabolism, the different reports on the effect of low doses of ICS for periods up to 2 yr suggest that they have minimal influence on bone loss even in groups considered at risk of osteopenia such as postmenopausal women and elderly patients (8, 15, 24). For high doses of ICS equivalent to 1,000 µg of beclomethasone or more, the studies have, however, reported variable results. In a study by Ip and coworkers in women, a correlation was found between bone density and daily inhaled dose of ICS (mean 932 µg for a duration of 40 ± 43 mo (11). Packe and coworkers also found a reduction in bone density with high doses of ICS for a period of 3 years (10). Hanania and coworkers studied a group of asthmatic patients on ICS at a mean dose of 1,323 µg/d for a median duration of 24 mo and found that mean Z scores at the femoral neck were below normal predicted values (9). It was not reduced at the level of lumbar spine and Ward's triangle. When the dose of ICS was corrected for body mass index, it correlated negatively with bone density and adrenal function.

In some of those previous studies, however, the confounding influence of concomitant or intercurrent use of oral steroids could not be ruled out and long-term compliance was not assessed or estimated. On the other hand, as compliance to inhaled corticosteroids can be low, these observations could have underestimated the influence of ICS on bone density. For those long-term studies, it is however difficult to determine what was the compliance to the inhaled drugs. In our study, we estimated what proportion of the dose prescribed was taken by patients. Although this remains an imperfect method of assessment, our results are quite reassuring, as for mean daily doses of about 1,000 µg daily of beclomethasone or budesonide, bone density was not significantly different between our two groups and no significant change in bone density had occurred in the last 3 yr. Although we need to look at the effects of ICS on bone density over many years, it is nevertheless generally considered that a 2- to 3-yr period is sufficient to evaluate significant changes in bone density.

Furthermore, predictive value of a treatment may be better assessed when we have baseline data before the onset of the treatment. This was not possible to do in the present study as the patients were already on this medication for a significant number of months when we started the follow-up. Corticosteroid-induced bone loss may be higher in the first 6 to 12 mo of initiating systemic corticosteroid therapy, but long-term corticosteroid use is also associated with significant bone loss (27, 28).

The lower T and Z scores observed in our groups of asthmatic patients may reflect an effect of asthma per se or the effects of other factors such as a reduction in physical exertion from fear of symptoms and/or insufficiently controlled asthma. In this regard, physical activity was low in the two groups studied. On the other hand, we found no influence of dietary factors and the study groups were similar in regard to the different nutritional parameters. In more severe asthma, another contributing factor to bone loss may be intercurrent intake of oral corticosteroid; the use of oral corticosteroids was however low in our patients.

Stead and coworkers reported that in a group of 11 women with asthma, most of them postmenopausal, using high doses of ICS for more than 1 yr was associated with a reduction of mineral density of the lumbar spine of 13.4% compared with a group of healthy age-matched control women (29). There was, however, no evidence of increased bone loss on follow-up and these changes may have been present before the introduction of ICS. Toogood and coworkers found that bone density was lower with the use of high doses of ICS and with the duration of past prednisone therapy (15). Bone density decreased in relation with the dose of ICS although this decline was more evident with doses equivalent to more than 1,000 µg/day of beclomethasone. Bone density increased however with estrogen replacement therapy.

In the present study, we found a relationship between daily ICS dose and bone density reduction over 3 yr in the high-dose group. The initial lumbar Z scores for both groups are quite similar to the values predicted by relating the mean daily dose of inhaled corticosteroid to the linear regression of the lumbar Z score on daily dosage of inhaled corticosteroid reported by Toogood and coworkers (15). However, although there seems to be an effect of these agents on bone mass, it is quite small at the dosing assessed and of uncertain clinical relevance. Recently, the relation between ICS dose and duration and bone density was assessed, taking into account potential confounders including weight, exercise, and oral steroid use, in 47 patients taking an ICS at a mean current dose of 620 µg/d for a mean duration of 7.8 yr (30). There was no significant difference in mean bone density values between those men or women who were and those who were not on ICS. However, on multivariate analysis, cumulative ICS dose was associated with a reduction in posterior-anterior (P-A) and lateral lumbar spine bone mineral density in women, equivalent to a 0.11 standard deviation reduction in bone density per 1,000 µg/d ICS per year after adjustment for potential confounding factors. This is in keeping with our findings.

The need to have higher doses of corticosteroid may also reflect asthma severity and a more intense underlying inflammatory process which can possibly affect bone metabolism. In this regard, the possibility that a local inflammatory process as the one found in asthma could influence bone density or metabolism has not been adequately studied. Other conditions such as rheumatoid arthritis can induce a quite significant systemic inflammatory process with the release of cytokines such as interleukin-1 but even here, a possible effect of inflammation on bone metabolism remains hypothetical.

Up to now, to our knowledge, increased prevalence of fractures in patients on ICS has not been shown, whatever the dose, although there is an hypothetical increase in this risk with high doses. Even if bone markers are often abnormal in those patients, our study suggests that they cannot predict bone loss or clinical outcome. In this case, even initial bone density could not predict the rate bone loss.

In conclusion, we found that although we could demonstrate a correlation between daily ICS dose and bone loss over time in the high-dose group, these changes were minimal over a period of 3 yr and not different between a group using high versus low doses of ICS for more than 5 yr. Bone markers and initial bone density could not predict the rate of bone loss. Although these data are reassuring in regard to the use of doses of ICS up to about 1,000 µg/d in equivalent of beclomethasone, in adult asthmatic subjects, further evaluations should be done over longer time periods and with higher doses of ICS. Means of identification of patients susceptible of bone loss after corticosteroid therapy should be searched.