Abstract
Osteoporosis is a major complication of long-term corticosteroid administration, but the magnitude of the effect in patients with chronic obstructive pulmonary disease (COPD) is not well defined. In a cross-sectional study, we evaluated the association between steroid use and vertebral fractures in 312 men, 50 yr of age or older, with COPD. Subjects were evaluated according to their corticosteroid use: Never Steroid Users (NSU) (n = 117), Inhaled Steroid Users (ISU) (n = 70), and Systemic Steroid Users (SSU) (n = 125). The prevalence of one or more vertebral fractures was 48.7% in the NSU group, 57.1% in the ISU group, and 63.3% in the SSU group. Compared with NSU, SSU were two times as likely to have one or more vertebral fractures: age-adjusted odds ratio (OR) = 1.80; 95% CI, 1.08 to 3.07. This relationship was primarily due to a strong association between continuous systemic steroid use and vertebral fractures: age-adjusted OR = 2.36; 95% CI, 1.26 to 4.38. In addition, fractures in SSU were more likely to be multiple and more severe. A weaker relationship existed between inhaled steroid use and vertebral fractures: age-adjusted OR = 1.35; 95% CI, 0.77 to 2.56 compared with NSU. These data indicate that vertebral fractures are common in older men with COPD; the likelihood of these fractures is greatest in those men using continuous systemic steroids. McEvoy CE, Ensrud KE, Bender E, Genant HK, Yu W, Griffith JM, Niewoehner DE. Association between corticosteroid use and vertebral fractures in older men with chronic obstructive pulmonary disease.
It is estimated that 14.6 million people in the United States have chronic obstructive pulmonary disease (COPD). Inhaled and oral steroids are frequently prescribed as part of the therapy for this condition, but little is known about the long-term side effects of this treatment. Previous small cross-sectional studies have demonstrated an association between chronic systemic corticosteroid use and lower bone density, primarily in areas with high trabecular bone content such as ribs and vertebrae, in COPD, and in other diseases (1-7). However, few studies have compared the prevalence of vertebral fractures, an anticipated result of low bone density, in patients who received corticosteroids with patients with the same disease process who did not receive steroids. Prior studies have also examined the prevalence of spinal osteoporosis in older men, but none has been of sufficient size to determine the prevalence of vertebral fractures in men with a specific disease such as COPD (3, 4, 7-10).
To determine the association between corticosteroid use and prevalent vertebral fractures in older men with COPD, we obtained thoracic and lumbar radiographs in 312 men older than 50 yr of age with documented COPD. The prevalence of vertebral fractures was compared among three groups: Never Steroid Users, Inhaled Steroid Users, and Systemic Steroid Users.
Subjects were identified through computer-generated pharmacy records at the Minneapolis Veterans Affairs Medical Center by their use of inhaled beta-agonists, inhaled corticosteroids, and oral prednisone during the preceding 2 yr. A random sample of each of the three groups was generated. Participants were volunteers who responded to a mailed invitation. These respondents were prescreened by telephone interview to determine whether they were qualified for further evaluation by study investigators. Subjects were considered eligible for the study if they met the following criteria: (1) male patients > 50 yr of age, (2) primary diagnosis of COPD, (3) FEV1/FVC ratio of < 70%, (4) smoking history of at least 20 pack-years, and (5) minimum of at least six refills of a beta-agonist inhaler within the past year as identified through pharmacy records.
Subjects were excluded for one or more of the following criteria: (1) a history of chronic renal failure (Cr > 2.5), cirrhosis, hyperthyroidism, hypogonadism, metabolic bone disease, malabsorption syndromes, Cushing's syndrome, thalassemia, lymphoma/leukemia, myeloma, rheumatoid arthritis, Paget's disease, and hyperparathyroidism; (2) use of phenytoin, heparin, androgens, or fluoride; (3) immobility or inability to ambulate without assistance. Written informed consent was given by all participants as approved by the Minneapolis VAMC Human Studies Committee and the University of Minnesota, Institutional Review Board, Human Subjects Committee.
Steroid use. Subjects who met the study enrollment criteria were evaluated according to their exposure to steroids. Never Steroid Users were defined as those men using bronchodilators only with no exposure to systemic or inhaled corticosteroids. Systemic Steroid Users were defined as those study subjects who have used prednisone for more than 2 wk continuously in their lifetime. The minimum dose of prednisone required was 5 mg every day or 10 mg as needed. Systemic Steroid Users were further classified as Intermittent or Continuous Users. Intermittent Systemic Steroid Users were defined as those subjects using prednisone periodically for more than 2 wk but never continuously for more than 6 mo in their lifetime. Continuous Systemic Steroid Users were defined as subjects using prednisone on a daily or as-needed basis for at least six consecutive months in their lifetime. Systemic Steroid Users may also have used an inhaled steroid preparation. Inhaled Steroid Users are defined as those study subjects using an inhaled steroid preparation a minimum of four puffs a day for at least 6 mo of the previous year and a history of no more than two brief courses of oral corticosteroids (⩽ 1 wk) or maximum systemic steroid exposure of 2 wk in their lifetime. Medical and pharmacy records were used to verify use group in 100% of subjects and to determine dose and duration of systemic steroids. In addition to both medical and pharmacy record reviews, exposure status was determined by the subjects' answers to a series of questions in the self-administered questionnaire. Immediately after completion of the questionnaire the subjects' answers were reviewed with an investigator in the event additional clarification was necessary. Furthermore, subjects were requested to bring all current medications to the clinical evaluation.
Pulmonary function. We measured the FEV1 and FVC by an automated testing module (MG System 1070/1085; Medical Graphics, Saint Paul, MN) according to the guidelines of the American Thoracic Society (11). We evaluated subjective measurement of dyspnea with the Baseline Dyspnea Index (BDI), a validated multidimensional instrument based on three components: functional impairment, magnitude of task, and magnitude of effort (12). The score of each component is added to give the total BDI score. A score of 12 indicates no impairment in functional status, task performance, or effort caused by breathlessness. A score of zero indicates complete impairment in all activities because of breathlessness. Subjects who were limited in function, task, or effort because of medical problems other than breathlessness received no score (n = 27) and were excluded from the multivariate analysis.
Demographics, general health status, and health habits. A detailed self-administered questionnaire was used to elicit information regarding age, race, past medical history, and use of alcohol and tobacco. General health status was assessed with a modified Medical Outcomes Study (MOS) short-form survey. MOS questions relating to physical functioning, health perceptions, and pain were included in our questionnaire (13). General Health Index is scored on a scale of 1 to 5; a score of 1 indicates excellent health and a score of 5 indicates poor health. Activity Limitation is the sum of 10 questions pertaining to physical functioning; a score of zero indicates no limitation in activities, whereas a score of 20 indicates complete limitation in all activities of daily living. Although individual components have been previously tested (13), the aggregate questionnaire was not validated.
Other medications. Participants' current and previous use of theophylline, thiazide diuretics, calcium, and 1,25-dihydroxyvitamin D supplements was obtained by a self-administered questionnaire. Pharmacy and medical records were reviewed to verify medication use in all subjects.
Vertebral fracture evaluation. All participants obtained lateral lumbar and thoracic radiographs that were evaluated for the presence of vertebral deformities. All films were taken using a standardized 40″ tube-to-film distance. the films were focused on T 7-8 and L 2-3 with the patient on the table in the lateral position. Care was taken to align the person's spine as close as possible to the horizontal plane.
All radiographs were evaluated independently at two institutions by experienced skeletal radiologists (H.K.G./W.Y. at University of California, San Francisco and E.B. at VA Medical Center, Minneapolis) according to semiquantitative methods previously published (14). Vertebrae T4-L4 were graded on visual inspection without direct vertebral measurement as normal (Grade 0), mild (Grade 1, 20–25% reduction in anterior, middle, and/or posterior height and a reduction of area 10 to 20%), moderate (Grade 2, 26 to 40% reduction in any height and a reduction in area 21 to 40%), or severe (Grade 3, > 40% reduction in any height and area) (14). Vertebral deformity resulting from epiphysitis was identified and excluded by conventional criteria: the presence of irregular vertebral end-plate outline, sclerotic deformities of the end plate, and wedging (14). All readings were done without knowledge of the subjects' corticosteroid usage. There was reasonable agreement in the radiographic interpretations between institutions (kappa = 0.68). No significant differences in the results of the age and multivariate adjusted analyses were observed using each institution's interpretation independently (H.K.G./W.Y. or E.B.). Because the results did not differ significantly by radiologist observer, we present the results based on analyses using one institution's (H.K.G./ W.Y.) interpretation.
Sample size requirements were based on a comparison of vertebral fracture prevalence between Systemic Steroid and Never User groups using a standard formula for a cross-sectional study. The prevalence of vertebral deformities in each of the three groups was estimated from a pilot study and literature review (8, 15-18). The prevalence values used were: 10% Never Users, 26% Inhaled Steroid Users, and 40% Systemic Steroid Users. Given 80 subjects in the Inhaled Steroid User group, 80 subjects in the Systemic Steroid User group, and 160 subjects in the Never User (control) group, our study had a 90% power (α = 0.05 [two-sided]) to detect 16% difference in prevalence of vertebral fractures between the Inhaled Steroid User and Never User groups and 95% power (α = 0.05 [two-sided]) to detect a 30% difference in prevalence of vertebral fractures between Systemic Steroid User and Never User groups. χ2 was used to perform crude comparisons, and the Bonferroni method was applied for multiple pair-wise comparisons. Multiple logistic regression techniques were used to perform age-adjusted and multivariate adjusted analyses. Biologically plausible confounders, variables that differed significantly between groups, and significant independent predictors were included in the multivariate model to both help explain the association and to adjust for potential bias. The a priori list of variables identified as potential confounders included: FEV1% of predicted, pack-years of tobacco use, current smoking status, Activity Limitation Score (ALS), General Health Status Score (GHS), and BDI. Independent predictors of vertebral fracture included in the model were: age, weight, and age of smoking initiation. Variables in the model were included as either continuous: age, weight, pack-years of tobacco use, age of smoking initiation, FEV1% of predicted, ALS, GHS, and BDI; or categorical: current smoking status and dependent variable, vertebral fracture. Odds ratios (ORs) and 95% CI were determined in all multivariate logistic regression analyses. An α of 0.05 was used for all statistical tests. Analyses were done using SPSS software (SPSS Inc., Chicago, IL).
Four hundred eleven men were seen for evaluation; 312 (75.9%) who met all eligibility criteria were enrolled. Forty-four patients were excluded based upon insufficient pack-years of tobacco use, five because of phenytoin use and three because of testosterone use. The remaining 47 patients were excluded based on spirometry results showing no evidence of airflow obstruction. The positive response rate to the mailed invitation was 28.5% in Never Users, 34.8% in Inhaled Steroid Users, and 35.8% in Systemic Users groups. A random 10% sample of nonresponders in each group were interviewed by telephone. Of the 51 patients sampled, 47.0% (n = 24) failed to respond because of death or illness. The remainder of the surveyed nonresponders had moved out of state (n = 9), felt the drive was too far (n = 7), were ineligible for the study (n = 4), could not remember why they chose not to volunteer (n = 6) or reported that the study would interfere with a work schedule (n = 1). No significant differences were observed among the three groups.
The characteristics of participants are shown in Table 1. Systemic Steroid Users were more likely to be ex-smokers, have a greater pack-year of tobacco use, report a poorer general health status, be more limited in activities, have a lower FEV1%, complain of more dyspnea, and use of theophylline, 1,25-dihydroxyvitamin D, and calcium supplements than Never Users. Inhaled Steroid Users were more likely to be ex-smokers than Never Users and less likely to use theophylline, 1,25-dihydroxyvitamin D, or calcium supplements than Systemic Steroid Users.
| Characteristic | Never Steroid User (n = 117 ) | Inhaled Steroid User (n = 70) | Systemic Steroid User (n = 125) | |||
|---|---|---|---|---|---|---|
| Mean age, yr (SD) | 68.3 (7.5) | 69.7 (6.3) | 68.9 (7.2) | |||
| Current smoker, n (%) | 50 (42.7)*,† | 17 (24.3)† | 15 (12.0)* | |||
| Mean pack-years | ||||||
| of cigarette use (SD) | 67.0 (33.1) | 67.1 (32.4) | 57.4 (30.7) | |||
| Mean age regular | ||||||
| smoking initiation, yr (SD) | 16.8 (3.7) | 16.9 (3.3) | 17.1 (3.9) | |||
| Mean years since | ||||||
| last cigarette use (SD) | 7.7 (10.4)* | 8.8 (9.4) | 11.3 (9.9)* | |||
| Alcohol consumption | ||||||
| Any alcohol past year, n (%) | 65 (56) | 34 (49) | 59 (47) | |||
| Mean drinks/wk (SD) | 4.2 (9.2) | 2.9 (6.6) | 2.6 (6.1) | |||
| General Health Status Index, | ||||||
| mean (SD) | 3.6 (0.85)* | 3.7 (0.71) | 3.9 (0.85)* | |||
| Activity Limitation Score, | ||||||
| mean, (SD) | 10.8 (5.5)* | 12.5 (4.5) | 13.5 (5.3)* | |||
| Blocks walked per day (SD) | 9.7 (10.9)* | 7.9 (10.0) | 5.6 (8.3)* | |||
| Baseline Dyspnea Index, | ||||||
| mean (SD) | 6.0 (2.1)* | 5.4 (2.2) | 4.6 (4.1)* | |||
| History of | ||||||
| Diabetes, n (%) | 18 (15.4) | 7 (10.0) | 25 (20.0) | |||
| Hypertension, n (%) | 52 (44.4) | 36 (51.4) | 63 (50.4) | |||
| Osteoarthritis, n (%) | 48 (41.0) | 25 (35.7) | 47 (37.6) | |||
| Impotence, n (%) | 30 (25.6) | 18 (25.7) | 41 (32.8) | |||
| Medication use | ||||||
| Thiazide, n (%) | 25 (21.4) | 26 (37.1) | 37 (29.6) | |||
| Theophylline, n (%) | 21 (18.0)* | 27 (38.6)‡ | 101 (80.8)*,‡ | |||
| Calcium supplement, n (%) | 49 (41.8)* | 22 (31.4)‡ | 87 (69.6)*,‡ | |||
| Vitamin D supplement, n (%) | 29 (24.8)* | 18 (25.7)‡ | 65 (52.0)*,‡ | |||
| Mean weight, lbs (SD) | 191.1 (43.9) | 183.6 (38.1) | 180.2 (33.7) | |||
| Mean height, inches (SD) | 67.7 (9.2) | 68.0 (1.9) | 67.2 (8.5) | |||
| Mean FEV1, L (SD) | 1.65 (0.60)* | 1.48 (0.57) | 1.36 (0.58)* | |||
| Mean FEV1, % pred (SD) | 57.1 (19.3)* | 53.5 (19.2) | 48.3 (19.9)* |
The mean duration of prednisone use in Systemic Steroid Users was 163.4 ± 213.7 wk (range, 2.5 to 1,300.0 wk), and the mean cumulative lifetime systemic corticosteroid dose was 19.5 ± 24.8 g (range, 0.2 to 124.0 g). Among Intermittent Users of prednisone, the average cumulative corticosteroid dose was 2.8 ± 2.7 g (range, 0.2 to 10.0 g). Continuous Users averaged 18.5 ± 7.9 mg of prednisone per day with a mean cumulative lifetime dose of 31.2 ± 26.6 g (range, 1.8 to 124.0 g). Duration of prednisone use in Intermittent Users ranged from 2.5 to 72 wk, with a mean duration of 16.6 ± 15.0 wk; in Continuous Users mean duration of use was 5.2 ± 4.4 yr, with a range of 1.0 to 25 yr. Concomitant use of an inhaled steroid preparation was reported in 92.8% of the Systemic Steroid Users.
The mean duration of inhaled corticosteroid use in the Inhaled Steroid Use group was 3.5 ± 2.6 yr (range, 0.5 to 11 yr), with an average number of puffs used per day of 11.8 ± 6.4 (range, 4 to 32 puffs). The Systemic Steroid Users averaged 4.6 ± 3.2 yr (range, 0.3 to 20 yr) of inhaled corticosteroid use and a mean of 13.9 ± 6.4 puffs per day (range, 4 to 32 puffs). Beclomethasone dipropriate and triamcinolone acetonide were the only inhaled corticosteroid preparations used by our subjects.
The percentage of participants with at least one vertebral fracture observed on radiographs of the lumbar and thoracic spine was 48.7% in the Never Users group, 57.1% in the Inhaled Steroid Users group, and 63.3% in the Systemic Steroid Users group. Thoracic vertebral deformities were more common than lumbar deformities; 49% of the subjects had at least one thoracic fracture, whereas 16.5% of subjects were observed to have at least one lumbar deformity. This distribution of fracture location was consistent across all three groups. Multiple vertebral fractures (⩾ 2) were more often observed in the Systemic Steroid Users group (35.8%) than in either the Never Steroid User (23.0%) or the Inhaled Steroid User (17.1%) groups (p < 0.05 for comparison between Systemic Steroid Users and Inhaled Steroid Users). As illustrated in Figure 1, subjects with six or more vertebral fractures were observed exclusively in the Systemic Steroid Users Group (n = 10). Of those subjects with one or more vertebral fracture(s), severe fractures were significantly more common in the Systemic Steroid Users (22.1%) group than in either the Never Steroid Users (7.7%) or the Inhaled Steroid Users (3.5%) group (p < 0.05 for both comparisons) (Figure 2).
Fig. 1. Percent of subjects in Never Steroid Users, Inhaled Steroid Users, and Systemic Steroid Users groups with specified number of vertebral fractures.
[More] [Minimize]
Fig. 2. Percent of fractures that are graded mild, moderate, or severe within Never Steroid Users, Inhaled Steroid Users, and Systemic Steroid Users groups. The asterisk indicates a significant (p < 0.05) difference in percent of fractures for comparison between Systemic Steroid Users and both Never Steroid Users and Inhaled Steroid Users. The dagger indicates p < 0.05 for comparison between Systemic Steroid Users and Never Steroid Users.
[More] [Minimize]Significant independent predictors of one or more vertebral fracture(s) included in the multivariate analysis are increasing age (p < 0.02), earlier age of smoking initiation (p < 0.03), and greater weight (p < 0.02). Activity level, FEV1%, current smoking status, pack-years of tobacco use, BDI, and general health status were also included in the multivariate analysis based on the potential significant confounding effect of these variables. We performed regression analyses that included alcohol use, calcium and 1,25-dihydroxyvitamin D supplementation, impotence, diabetes, theophylline, osteoarthritis, and thiazide diuretic use in the multivariate model and found no significant effect on the reported ORs; therefore, these variables were not included in the multivariate model.
Age-adjusted and multivariate-adjusted ORs for the association between inhaled and systemic steroid use and the presence of prevalent vertebral fractures using Never Users as the referent group are shown in Table 2. Systemic Users were nearly two times more likely to have one or more prevalent vertebral fracture(s) as Never Users (age-adjusted OR = 1.80; 95% CI, 1.08 to 3.07). This observed effect was primarily due to the association between continuous systemic use and vertebral fracture; compared with Never Users, Continuous Systemic Users were 2.4 times as likely to have one or more vertebral fracture(s). Multivariate adjustment for potential confounders did not explain the association between systemic steroid use and vertebral fracture (multivariate-adjusted OR = 2.16; 95% CI, 1.14 to 4.11). Although there appeared to be a relationship between intermittent systemic use and vertebral fracture, this association did not reach statistical significance (age-adjusted OR = 1.25; 95% CI, 0.66 to 2.54; multivariate-adjusted OR = 1.38; 95% CI, 0.71 to 2.69).
| Age-adjusted Odds Ratio (95% CI) | Multivariate-adjusted Odds Ratio†(95% CI) | |||
|---|---|---|---|---|
| Inhaled steroid users | 1.35 (0.77–2.56) | 1.38 (0.71–2.69) | ||
| Systemic steroid users | 1.80 (1.08–3.07) | 2.16 (1.14–4.11) | ||
| Intermittent, n = 52 | 1.25 (0.66–2.54) | 1.55 (0.72–3.32) | ||
| Continuous, n = 73 | 2.36 (1.26–4.38) | 2.99 (1.38–6.49) |
The association between dose and duration of steroid use and prevalent vertebral fracture is shown in Table 3. Both increasing dose and duration were associated with an increase in the likelihood of one or more vertebral fractures. For every 3 yr of systemic steroid use (1 standard deviation) an elderly white male with COPD is 41% more likely to have a vertebral fracture. In this same population for every 24.8 g of cumulative prednisone use (1 standard deviation), the likelihood of having a vertebral fracture is increased by 35%.
| Variable | Mean (1 SD) | Age-adjusted Odds Ratio for 1 SD Change (95% CI) | ||
|---|---|---|---|---|
| Lifetime duration of prednisone use, wk | 163 (214) | 1.41 (0.89–2.23) | ||
| Lifetime dose of prednisone, g | 19.5 (24.8) | 1.35 (0.88–2.08) |
Cushing (19) first reported skeletal decalcification as a characteristic feature of adrenal hyperplasia secondary to ACTH-producing pituitary tumors. Although Cushing's syndrome is a rare disease, the administration of supraphysiologic doses of systemic corticosteroids has become a standard treatment for some diseases, including COPD. Osteoporosis is perhaps the most feared complication of long-term corticosteroid administration, but the magnitude of the clinical consequences in patients with COPD is difficult to judge for a number of reasons. Most importantly, the majority of studies evaluating the severity and potential mechanisms of corticosteroid-induced bone loss in patients with COPD do not include appropriate disease-matched control subjects (3, 7, 9). That is, patients with COPD treated with corticosteroids are not compared with equivalent patients not treated with corticosteroids. This is particularly important because COPD and its principal identified cause, cigarette smoking, may by themselves be significant risk factors in the development of osteoporosis (20, 21). Patients using corticosteroids also tend to be severely limited by their disease, and inactivity may pose another osteoporosis risk factor (22).
Our data suggest that vertebral fractures, a manifestation of osteoporosis, are highly prevalent in white men with COPD, regardless of whether they have been exposed to corticosteroids. Although corticosteroids significantly increase this risk, the baseline prevalence is much higher than anticipated based upon previously published results (8, 10, 17, 18, 23-25), suggesting that factors related to the disease itself may promote the development of osteoporosis. In our analysis, we have controlled for several factors that may contribute significantly to bone loss: disease severity with FEV1%, activity level, general health, and BDI, tobacco exposure with age of smoking initiation, current smoking status, and pack-years of cigarette use, age, and weight. We also evaluated the effects of other potential confounders such as history of impotence, alcohol use, diabetes, and diuretic use; however, adjustment for these variables did not account for the relationship between systemic steroid use and an increased likelihood of vertebral fracture.
The suggestion that factors associated with COPD itself may contribute to bone loss is supported by other reports (4, 26). In one recent cross-sectional survey, serum osteocalcin levels (a measure of bone formation) and bone density measurements were measured in 19 men with COPD requiring prednisone daily for at least 1 yr, 25 men with COPD using bronchodilator therapy only, and 90 healthy age-matched control subjects (4). Compared with the control subjects, the investigators reported significantly lower osteocalcin levels and spinal bone density in both groups with COPD. Other investigators have published similar results in patients with rheumatoid arthritis (26).
The use of systemic corticosteroids clearly increases the likelihood that an older man with COPD will have evidence of significant osteoporosis. This risk is primarily related to chronic daily prednisone use of fairly substantial doses (18 mg/d) and was not associated with intermittent use. It is important to recognize that our two groups of Systemic Steroid Users were significantly different in their exposure to prednisone. The Continuous Users on average were taking prednisone daily for at least 5 yr, whereas the Intermittent Users were exposed to prednisone for an average of 4 mo during their lifetime. The actual threshold of dosing or duration at which the risk of having observed vertebral fractures could not be determined from this study. A larger prospective study is needed to determine whether an intermittent schedule will minimize the deleterious effects of corticosteroids on bone. Inhaled steroid use also was associated with an increased likelihood of a vertebral fracture; however, this relationship did not reach statistical significance, and our study lacked the power to rule out a type II error.
Our study has several other limitations. It is unlikely that we were able to control for all confounders given the cross-sectional study design. This population of older men was not heterogeneous, being primarily white and northern European in descent. Therefore, these results may not be generalized to other populations. In addition, the semiquantitative method of vertebral fracture determination may have been “oversensitive” given that the majority of observed fractures were mild. Even though the semiquantitative method of detecting and grading vertebral deformities has been found to reliably correlate with the more objective morphometric methods (14), the excess of mild deformities may arise from difficulties in recognizing non-fracture-associated anomalies. Quantitative methods, however, are not without problems. Morphometry is time-consuming, and variations in radiographic quality create difficulties in the placement of points required for digitization. Furthermore, the placement of these points for digitization is subjective, and clear differences in quantitative approaches exist. Small subjective differences in point placement can result in significant variance with mild or borderline wedge deformities making fracture/nonfracture discrimination inconsistent (14). Although the potential for an inflated prevalence exists, our data differ significantly from results reported by investigators using the same, semiquantitative method. Kiel and colleagues (25), in their examination of the Framingham Cohort, reported a 36% prevalence of vertebral fractures in older men (⩾ 85 yr); this is substantially lower than the 48% prevalence observed in Never Users whose mean age was nearly 20 yr younger.
In summary, there is a high prevalence of vertebral fractures in older white men with COPD independent of their steroid use. Systemic Steroid Users are twice as likely to have at least one vertebral fracture than are Never or Inhaled Steroid Users. In addition, Systemic Steroid Users tend to have more multiple and severe vertebral fractures than either Never or Inhaled Steroid Users. The question of whether corticosteroids are efficacious in the treatment of COPD is currently being addressed (27). If these trials show that corticosteroid therapy is beneficial, then recognition of potential adverse effects is essential. It is apparent that larger prospective studies with disease-matched control subjects are needed to determine the actual extent of the corticosteroid's effect on bone loss and fracture risk. More importantly, these prospective studies are needed to establish whether prevalent vertebral fractures in this population predict risk of clinically significant outcomes such as back pain, disability, and nonvertebral fractures. If so, evaluation of preventive strategies should be implemented.
The writers wish to acknowledge Dr. Paul McGovern for biostatistical advice and Mary Markhert for assistance in performing pulmonary function tests.
Supported by NIH Training Grant HL-07741, Minnesota Medical Foundation and School of Public Health, University of Minnesota.
| 1. | Seeman E., Wagner H., Offord K., Kumar R., Johnson W., Riggs B.Differential effects of endocrine dysfunction on the axial and appendicular skeleton. J. Clin. Invest69198213021309 |
| 2. | Gallagher J., Aaron J., Horsman A., Wilkinson R., Nordin B.Corticosteroid osteoporosis. Clin. Endocrinol. Metab21973355368 |
| 3. | Brandli D., Golde G., Greenwald M., Silverman S.Glucocorticoid-induced osteoporosis: a cross-sectional study. Steroids561991518523 |
| 4. | Praet J., Peretz A., Rozenberg S., Famaey J., Bourdoux P.Risk of osteoporosis in men with chronic bronchitis. Osteoporos. Int21992257261 |
| 5. | Laan R. F. J. M., van Riel P. L. C. M., van de Putte L. B. A., van Erning L. J. T. O., van Hof M. A., Lemmens J. A. M.Low-dose prednisone induces rapid reversible axial bone loss in patients with rheumatoid arthritis. Ann. Intern. Med1191993963968 |
| 6. | Bikle D., Halloran B., Fong L., Steinbach L., Shellito J.Elevated 1,25-dihydroxyvitamin D levels in patients with chronic obstructive pulmonary disease treated with prednisone. J. Clin. Endocrinol. Metab761993456461 |
| 7. | Reid I., Heap S.Determinants of vertebral mineral density in patients receiving long-term glucocorticoid therapy. Arch. Intern. Med150199025452548 |
| 8. | Mann T., Oviate S., Wilson D., Nelson D., Orwoll E.Vertebral deformity in men. J. Bone Miner. Res7199212591265 |
| 9. | Adinoff A., Hollister J.Steroid-induced fractures and bone loss in patients with asthma. N. Engl. J. Med3091983265268 |
| 10. | Riancho J., Gonzalez-Macias J., Del-Arco C., Amado J., Freijanes J., Anton M.Vertebral compression fractures and mineral metabolism in chronic obstructive lung disease. Thorax421987962966 |
| 11. | American Thoracic SocietyStandardization of spirometry: 1987 update. ATS statement. Am. Rev. Respir. Dis136198712851298 |
| 12. | Mahler D., Weinberg D., Wells C., Feinstein A.The measurement of dyspnea contents, interobserver agreement, and physiologic correlations of two new clinical indexes. Chest851984751758 |
| 13. | Stewart A., Hays R., Ware J.The MOS short-form general health survey: reliability and validity in a patient population. Med. Care261988724735 |
| 14. | Genant H., Wu C., Kuijk C. V., Nevitt M.Vertebral fracture assessment using a semiquantitative technique. J. Bone Miner. Res8199311371148 |
| 15. | Francis R., Peocock M., Marshall D., Horsman A., Aaron J.Spinal osteoporosis in men. Bone Miner51989347357 |
| 16. | Cummings S., Kelsey J., Nevitt M., O'Dowd K.Epidemiology of osteoporosis and osteoporotic fractures. Epidemiol. Rev71985178208 |
| 17. | Johansson C., Mellstrom D., Rosengren K., Rundgren A.A community-based population study of vertebral fractures in 85-yr-old men and women. Age Ageing231994388392 |
| 18. | Santavirta S., Konttinen Y., Heliovaara M., Knekt P., Luthje P., Aromaa A.Determinates of osteoporotic thoracic vertebral fracture: screening of 57,000 Finnish women and men. Acta Orthop. Scand631992198202 |
| 19. | Cushing H.The basophil adenomas of the pituitary body and their clinical manifestations. Bull. Johns Hopkins Hosp501932137195 |
| 20. | Daniell H.Osteoporosis of the slender smoker: vertebral compression fractures and loss of metacarpal cortex in relation to postmenopausal cigarette smoking and back of obesity. Arch. Intern. Med1361976298 |
| 21. | Seeman E., Melton L. J., O'Fallon W. M., Riggs B. L.Risk factors for spinal osteoporosis in men. Am. J. Med751983977 |
| 22. | Sinaki, M. 1988. Osteoporosis: Etiology, Diagnosis. Raven Press, New York. 457–465. |
| 23. | Cooper C., Atkinson E. J., O'Fallon W. M., Melton L. J.Incidence of clinically diagnosed vertebral fractures: a population-based study in Rochester, Minnesota, 1985–89. J. Bone Miner. Res71992221227 |
| 24. | Bengner U., Johnell O., Redlund-Johnell I.Changes in incidence and prevalence of vertebral fractures during 30 years. Calcif. Tissue Int421988293296 |
| 25. | Kiel D. P., Hannan M. T., Genant H. K., Felson D. T.Prevalence and incidence of vertebral fractures in the elderly: initial results from the Framingham study. J. Bone Miner. Res91995S129 |
| 26. | Hall G., Spector T., Delmas P.Markers of bone metabolism in postmenopausal women with rheumatoid arthritis. Arthritis Rheum381995902906 |
| 27. | Niewoehner D., Erbland M.Systemic corticosteroids in COPD: an unresolved clinical dilemma. Chest1101996867869 |
