INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In recent years, the life expectancy of patients with cystic fibrosis has improved and it is now common for such patients to survive into their third decade (1). In adults, chronic obstructive pulmonary disease accounts for the majority of the morbidity and mortality from cystic fibrosis (2). For patients with advanced cystic fibrosis pulmonary disease, lung or heart-lung transplantation may be considered as a life-prolonging option (2). Rapid bone loss and painful osteoporotic fractures are common complications of liver and cardiac transplantation (3) and fractures have also been reported after lung transplantation (4, 5). Low bone mass prior to transplantation may place patients at increased risk for fracture after transplantation (6), a matter of some concern since adult patients with cystic fibrosis have been shown to have low bone mass compared with age- and sex-matched controls (7). The origin of the low bone mass in patients with cystic fibrosis is incompletely understood. Glucocorticoid therapy, maldigestion and malabsorption resulting in vitamin D deficiency and undernutrition, hypogonadism, physical inactivity, and the smaller skeletal size of patients with cystic fibrosis may all contribute to the observed osteopenia.

We have previously documented low bone mineral density (BMD) in a small number of patients with cystic fibrosis who were awaiting lung transplantation (10). In that report, we observed that bone density tended to be lower in patients with subnormal serum levels of 25-hydroxyvitamin D (25-OHD; 10). In this larger study, we examined the prevalence and pathogenesis of osteoporosis in 30 adults with cystic fibrosis and severe chronic pulmonary disease who were awaiting lung transplantation. The results of this study indicate that vitamin D deficiency plays a significant role in the skeletal demineralization of adults with cystic fibrosis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

Between January 1, 1993 and December 31, 1995, 30 caucasian men and women with cystic fibrosis were evaluated consecutively following acceptance to the lung transplantation waiting list. Individuals with the following disorders known to affect bone and mineral metabolism were not included in these analyses: renal insufficiency (serum creatinine > 2.5 mg/dl), primary hyperparathyroidism, multiple myeloma, thyrotoxicosis, or suppressed thyroid-stimulating hormone (TSH).

Control Population

BMD of patients with cystic fibrosis was compared with the normative database provided by the manufacturer (lumbar spine and radius). The proximal femur densities were compared with the National Health and Nutrition Examination Survey (NHANES) database (11- 13). In addition, BMD of patients was also compared with normal subjects recruited from the surrounding community. The normal group included 16 premenopausal women and 14 men, age 32 ± 2 yr (range: 18 to 55). Each patient was matched to a normal subject of the same gender, age (± 3 yr), and race.

Study Design

Historical information about current and past exposure to glucocorticoids and tobacco, pharmacologic regimen, menarchal age, reproductive status, and family history of osteoporosis was obtained from all patients. Information was also sought regarding nonvertebral fractures. Such fractures were excluded from fracture prevalence data if they resulted from severe trauma (motor vehicle accidents). Historical fracture data were confirmed by review of the medical record or the radiographic report. Pulmonary function tests were obtained by chart review.

Each patient was evaluated with BMD of the lumbar spine, femoral neck, total hip, and nondominant radius by dual-energy X-ray absorptiometry (DXA) using a QDR-1000 bone densitometer (Hologic, Inc., Waltham, MA). Measurement of lumbar spine BMD represented the average of three vertebrae, usually the second, third, and fourth. If one of these vertebrae was fractured, Ll was analyzed instead. In our laboratory, the reproducibility of the QDR-1000 densitometer using an anthropomorphic spine is 0.51%. The short-term in vivo coefficient of variation is 0.68% for the lumbar spine (L2-4) and 1.36% for the proximal femur. Bone density was expressed as g/cm² and as T and Z scores. T and Z scores compare individual BMD determinations to the average BMD of young (age 30) and age-matched, respectively, normal populations of the same gender and represent the number of standard deviations by which an individual differs from the mean of the control population. According to criteria defined by the World Health Organization (WHO), T scores more than 2.5 standard deviations below the mean ( 2.5) represent osteoporosis while T scores between 1.0 and 2.5 represent low bone mass (14).

To evaluate the effect upon BMD of the smaller bone size of the patients with cystic fibrosis, bone mineral apparent density (BMAD; g/cm³) of the spine, femoral neck, and radius was calculated by dividing the bone mineral content (BMC) by a reference bone volume as described by Marcus and colleagues (8, 15), with modifications (10). This expression of bone mass reduces the influence of bone size on bone density measurements and permits comparison of BMD among persons of differing heights.

Twenty-six unselected patients had a standard set of anteroposterior and lateral radiographs of the thoracic and lumbar spine to detect undiagnosed compression fractures. Radiographs were analyzed by one of us (RBS) and vertebral fractures (including wedge, biconcave, and compression deformities) were classified according to the method of Eastell and coworkers (16).

In 20 unselected patients, blood was obtained in the fasting state and measured for serum indices of mineral metabolism. Serum total calcium, phosphorus, albumin, alkaline phosphatase activity, creatinine, and blood urea nitrogen (BUN) were measured by standard automated techniques (Technicon Instruments, Tarrytown, NY). All serum calcium concentrations were corrected for serum albumin (17). Serum osteocalcin, a marker of bone formation and turnover (18), was measured by radioimmunoassay (19). Intact parathyroid hormone (PTH) (1-84) was measured by immunoradiometric assay (20). Serum 25-OHD and 1,25-dihydroxyvitamin D [1,25(OH)₂D] were measured after extraction, by radiobinding (21) and radioreceptor assays (22) respectively. Serum bone specific alkaline phosphatase, another marker of osteoblast function, was measured using a solid phase, two-site immunoradiometric assay (Hybritech Inc., San Diego, CA) that specifically measures the skeletal isoenzyme of alkaline phosphatase. In 21 unselected patients, a 24-hr urine collection was analyzed for calcium by atomic absorption spectrophotometry and for creatinine using standard autoanalyzer techniques. Hydroxypyridinium crosslinks of type 1 collagen (pyridinoline and deoxypyridinoline; 23), were measured by high-performance liquid chromatography (24). Samples (serum and urine) were obtained when evaluated for transplantation and thus were collected throughout the year.

Statistical Methods

All data are presented as mean ± SEM. Associations between various demographic, anthropomorphic, and biochemical parameters and BMD were assessed using single and multiple regression analysis, Student's paired and unpaired t tests, and factorial analysis of variance as appropriate (StatView; Abacus Concepts, Inc., Berkeley, CA).

This study was approved by the institutional review board of Columbia-Presbyterian Medical Center. Written informed consent for participation was obtained from all subjects. The study was conducted in the Irving Center for Clinical Research and the Cardiopulmonary Transplant Unit of Columbia-Presbyterian Medical Center.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Characteristics of the Patients

The average age of the 30 patients was 30 ± 2 yr (range, 17-52 yr). The 16 women were premenopausal (Table 1). All were ambulatory at the time of the evaluation. Although body mass index (BMI; kg/m²) did not differ by gender, body weight (percent of ideal) was lower in men.

Pulmonary function tests were available in 29 patients. The mean forced expiratory volume in one second (FEV₁), a measure of the severity of pulmonary dysfunction and predictor of mortality in patients with cystic fibrosis (2), was markedly decreased at 0.8 ± 0.1 L (25 ± 2% of predicted values). FEV₁ (percent of predicted value) was  30% in 25 of the 29 patients. Forced vital capacity (FVC; percent of predicted values) was significantly lower in men.

The mean age of menarche was slightly delayed and five women reported menarche at age 15 or older. Three women were taking oral contraceptives because of amenorrhea or irregular menses; the remaining women had regular menses. Serum testosterone was available in seven men. The mean concentration was at the low end of the normal range. Three men with subnormal concentrations had normal serum luteinizing and follicle-stimulating hormones, suggesting hypothalamic dysfunction as the cause of the hypogonadism.

Four patients were receiving elemental calcium (1,000 mg daily) and 24 were receiving 400 to 800 IU of vitamin D in oral multiple vitamin preparations. Ten subjects had diabetes mellitus and were receiving insulin. All but two patients were receiving pancreatic enzyme supplements. Three subjects had a past history of cigarette smoking. The average exposure was 12.5 pack-years.

Bone Mineral Density

The prevalence of low bone mass and osteoporosis is shown in Figure 1. T scores between 1.0 and 2.5 (low bone mass) were present in 38% at the lumbar spine, 53% at the femoral neck and total hip, and 54% at the radius. Densitometric osteoporosis (T scores below 2.5) was present in 41% at the lumbar spine, 23% at the femoral neck and total hip, and 7% at the radius. By World Health Organization (WHO) criteria, only 21% had normal lumbar spine BMD, 23% had normal femoral neck and total hip BMD, and 39% had normal radial BMD.

View larger version (41K): [in this window] [in a new window]   Figure 1.   Frequency distribution of bone mineral density measurements of the lumbar spine (A), total hip (B), femoral neck (C ), and 1/3 site of the radius (D) in adult patients with cystic fibrosis. The measurements are expressed relative to average values of a young normal population of the same gender (T scores). T-score measurements falling between 1.0 and 2.5 SD below the mean (hatched bars) indicate low bone mass, whereas more than 2.5 SD below the mean (cross-hatched bars) indicate osteoporosis.

Z scores, which compare BMD of the patients with that of individuals of the same age, were also analyzed because patients below age 30 may not have achieved peak bone mass. Mean T and Z scores did not differ significantly at any site (Table 2). Z scores below 1.0 were considered to represent low bone mass. Analysis of BMD measurements by Z score rather than T score did not alter the categorization of lumbar spine or radial BMD in any patient. Femoral neck and total hip BMD changed from low bone mass to normal in two patients, age 46 and 52 yr.

BMD was markedly lower in men and women with cystic fibrosis than in normal matched control subjects at the spine and hip (Figure 2). In contrast, radial BMD was similar in women with cystic fibrosis and normal women and only slightly reduced in men. The percent reduction compared with normal subjects was greatest at the total hip and femoral neck (24 ± 3% and 20 ± 4%, respectively), intermediate at the lumbar spine (17 ± 3%), and least apparent at the radius (4 ± 2%) which was less demineralized than the other sites (p  0.003).

View larger version (31K): [in this window] [in a new window]   Figure 2.   Bone mineral density (g/cm2) of the lumbar spine (A), total hip (B), femoral neck (C ), and proximal radius (D) in adult men and women with cystic fibrosis (black bars) and healthy control subjects (white bars). Data are expressed as mean ± SEM.

BMAD was also lower in patients than control subjects. Lumbar spine BMAD measured 0.135 ± 0.01 g/cm³ in women with cystic fibrosis and 0.158 ± 0.01 g/cm³ in normal women (p < 0.0001). Similarly, lumbar spine BMAD was also lower in men with cystic fibrosis than in normal men (0.131 ± 0.01 versus 0.148 ± 0.01 g/cm³; p < 0.0001). BMAD was also lower at the femoral neck and radius (data not shown).

Significant associations between BMD, age, and anthropomorphic variables (Table 2) varied by site measured. Lumbar spine BMD was directly associated with age, weight, and BMI but not with height. In contrast, femoral neck BMD was directly associated only with BMI and total hip with weight and BMI. At the radius, positive associations were observed between weight and height, but there was no association with age or BMI.

Menarchal age was not correlated with BMD. In the seven men with measurements of serum testosterone, there was a near significant, direct association with femoral neck T score (r = +0.615; p = 0.14). BMD was similar in patients with and without diabetes mellitus (data not shown). No associations between BMD and pulmonary function were observed.

Vitamin D Deficiency

The majority of subjects (80%) were taking 400 to 800 IU of vitamin D daily in the form of ADEK vitamins (Scandipharm, Inc., Birmingham, AL). However, mean serum 25-OHD in the 20 patients with measurements, was at the low end of the normal range (16 ± 2 ng/ml; normal: 10 to 52). Eight patients (40%) had levels below 10 ng/ml and seven patients (35%) had levels between 11 and 20 ng/ml, defined by Peacock and colleagues as the vitamin D insufficient range (25). Serum 25-OHD was the same in the four patients not taking vitamin D supplements.

BMD was lower in patients with low 25-OHD (Figure 3A) at the lumbar spine (0.774 ± 0.02 versus 0.913 ± 0.04 g/cm²; p = 0.01) and total hip (0.648 ± 0.04 versus 0.811 ± 0.04 g/cm²; p = 0.01), and tended to be lower at the femoral neck (0.600 ± 0.04 versus 0.704 ± 0.03 g/cm²; p = 0.07). T scores (Figure 3B) were significantly lower at the lumbar spine, total hip, and radius and tended to be lower at the femoral neck (2.361 ± 0.39 versus 1.436 ± 0.30; p = 0.07).

View larger version (29K): [in this window] [in a new window]   Figure 3.   Bone mineral density of the lumbar spine, total hip, femoral neck, and the proximal radius in patients with low serum concentrations of 25-hydroxyvitamin D ( 10 ng/dl; black bars) and those with borderline or normal serum concentrations of 25- hydroxyvitamin D (> 10 ng/dl; white bars). In Panel A, BMD is expressed in raw units (g/cm2) and in Panel B, as T scores.

Linear regression analysis of serum 25-OHD and T scores at the femoral neck (r = +0.417; p = 0.07) and total hip (r = +0.435; p = 0.055) supported the significant relationship between BMD and vitamin D stores. Univariate linear regression analysis of 25-OHD against lumbar spine BMD was not significant. However, stepwise forward linear regression analysis of anthropomorphic variables, pulmonary function, and indices of mineral metabolism retained only weight and serum 25-OHD in the model as independent predictors of lumbar spine BMD [r = +0.859; multiple r² = 0.737; F(3,81) = 11.212; p = 0.0048].

Indices of mineral metabolism were normal and did not differ by serum 25-OHD level (Table 3). Serum phosphorus was directly associated with serum 25-OHD (r = +0.471; p  0.04). Serum intact PTH was normal in all but two patients, one with normal and one with low 25-OHD. Urinary calcium tended to be lower (p = 0.08) and serum 1,25(OH)₂D to be higher (p = 0.09) in the subjects with low 25-OHD. Deoxypyridinoline, a marker of bone resorption, was elevated in five subjects, but did not vary by serum 25-OHD.

Glucocorticoid Therapy

Eleven patients had exposure to glucocorticoids, defined as  7.5 mg of oral prednisone or its equivalent for > 3 mo. Regimens varied considerably and included daily and alternate day therapy and high-dose pulses (40 to 60 mg) with tapers. Patients with occasional brief exposure and those managed solely with steroid inhalers were not included in this analysis.

Mean duration of glucocorticoid use was 1.5 yr (range: 3 mo to 4 yr). BMD did not differ at any site (data not shown). There was a near-significant, inverse association between duration of glucocorticoid use and femoral neck BMD (r = 0.580; p = 0.06).

Fractures

Vertebral fractures were present in 5 of the 26 subjects (19%). A history of previous atraumatic fracture was given by 41% of the 27 subjects in whom historical information could be confirmed. Rib, wrist, and digit fractures were most common. BMD was the same in patients regardless of prevalent vertebral fractures or history of nonvertebral fractures.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of this study confirm that adults with cystic fibrosis and severe obstructive pulmonary disease have significantly reduced BMD compared with a normal population of the same race, age, and gender. BMAD, which minimizes the effect of bone size on BMD, was also lower in patients with cystic fibrosis than in control subjects. BMD was lowest at the hip and spine, while the proximal radius, composed predominantly of cortical bone, was least severely affected. Similarly the prevalence of densitometric osteoporosis, defined by WHO criteria, was most common at the lumbar spine and least common at the radius. Bone density was most strongly associated with age, weight, and BMI. Despite supplementation with 400 to 800 units of vitamin D in 80% of patients, serum concentrations of 25-OHD were insufficient (11-20 ng/ ml) or frankly subnormal ( 10 ng/ml) in 75% of the patients. Those patients with levels  10 ng/dl had lower BMD and T scores than those with insufficient or normal values. Significant associations were not observed between BMD and pulmonary status, diabetes mellitus, or glucocorticoid therapy.

Osteoporosis and fractures are among the medical complications that have emerged in association with the increased life span of patients with cystic fibrosis (7). Our observations are consistent with our previous results in a smaller number of patients (10) and with other reports in the literature (7- 9, 26). The first report of demineralization in children with cystic fibrosis in 1979 (27) documented significant radial and ulnar demineralization in 33% of their patients, similar to the 27% of our subjects who had T or Z scores < 2.0 at the radius. Moreover, BMD was relatively normal in children below age 13, suggesting that the mineral deficit was associated with adolescence. Most, although not all (28) studies of children and adolescents with cystic fibrosis published since have observed BMD to be below normal (29).

Bone demineralization has also been reported in adults with cystic fibrosis (7, 26, 32). Grey and coworkers measured total body, lumbar spine, and proximal femur in 16 men and women with cystic fibrosis ranging in age from 17 to 42 (7). None had fractures and mineralization was not as severely reduced as in our population. Bachrach and colleagues (8) evaluated 22 adults with cystic fibrosis, clinically similar to the patients reported herein. Their application of BMAD measurements to these patients established that the deficit in bone mineral could not be accounted for by smaller bone size. They also observed that age, weight, and BMI predicted bone mass. In a larger study (49 patients) that included both children and adults (31), Bachrach and colleagues reported mean lumbar spine and femoral neck BMD Z scores to be reduced to a similar degree to those reported herein, although only 53% of their patients had significantly reduced BMD. In contrast, Aris and colleagues (9) found osteoporosis or low bone mass to be almost universal in their cohort of 70 adult patients with end-stage cystic fibrosis.

In addition to age and anthropomorphic variables, we considered the influence of disease severity (pulmonary function), glucocorticoid use, and disordered mineral metabolism (vitamin D deficiency) as potential influences on BMD. Similar to Aris and colleagues (9) but in contrast to other reports (7, 26, 31, 33), we did not find that pulmonary status correlated with BMD. The absence of such a relationship was likely related to patient selection bias. Because all of the participants in our study and that of Aris and colleagues were referred for lung transplantation, pulmonary function was severely impaired. The mean FEV₁ of the patients in this study was only 25 ± 2% of predicted values and 86% had FEV₁ below 30% of predicted values. In contrast, other study groups were characterized by better and more variable pulmonary function (7, 26, 31, 33). Thus it may have been possible to detect an association between BMD and disease severity with a wider range of pulmonary dysfunction.

Glucocorticoids, frequently employed to manage airway obstruction in patients with cystic fibrosis, are well-known to cause bone loss and fractures (34). Similar to other investigators, we did not observe BMD to be significantly different in those patients with a history of glucocorticoid exposure (8, 31). In contrast, Aris and colleagues found cumulative prednisone dose to be one of the strongest predictors of BMD (9). This discrepancy may have been related to the fact that our steroid-treated patients tended to be older; thus, the deleterious effects of glucocorticoids on BMD may have been obscured by the positive association between age and BMD. However, the negative association between BMD and duration of glucocorticoid use observed in our patients, albeit not statistically significant, supports the results of Aris and colleagues and suggests that steroids adversely affect bone mass in cystic fibrosis.

Vitamin D deficiency, as evidenced by very low serum 25-OHD concentrations, was very common in this patient population, which was derived predominantly from the northeastern United States. Low serum 25-OHD concentrations were also associated with lower BMD, suggesting that vitamin D deficiency may play a significant role in the pathogenesis of demineralization in cystic fibrosis. Although not statistically significant, the pattern of differences in biochemical indices between patients with low and normal 25-OHD (that is, lower serum calcium and phosphorus, higher intact PTH and 1,25-dihydroxyvitamin D levels, and lower urinary calcium excretion) is suggestive of subclinical osteomalacia.

Vitamin D deficiency in association with cystic fibrosis was first described by Hahn and associates in 1979 (35). Low serum 25-OHD concentrations have since been reported in 23% to 75% of patients (7, 26, 31, 33). Two studies have observed that serum 25-OHD concentrations and BMD were directly associated (31, 35). In contrast, other investigators have found no difference in BMD between vitamin D replete and deficient patients (7, 26, 33). The reasons for the discrepancy between our findings and those of two of these reports (7, 8) may be related to the small number of patients with vitamin D deficiency (n = 5 and n = 4, respectively). In another report (33), the lack of relationship between vitamin D status and bone density may have been because the study took place in California where there is more exposure to sunlight. Patients with levels below 18 ng/ml (14 of 27) were considered vitamin D deficient relative to the normal population; only one patient had a level below 10 ng/ml (33).

It is important to note both that serum 25-OHD levels were low, notwithstanding vitamin D supplementation in 80% of the patients, and that serum 25-OHD concentrations did not differ between supplemented and unsupplemented subjects. Hahn and associates also reported that despite an average daily intake of vitamin D of 1,053 IUmore than 2.5 times the United States Recommended Daily Allowanceserum 25-OHD levels in patients with cystic fibrosis were 64% below those of normal control subjects receiving only 363 IU daily (35). Moreover, in a prospective study of vitamin D supplementation, 40% of patients with cystic fibrosis failed to achieve normal levels after 4 to 10 wk of 800 IU daily (30). Therefore, our observations and those of others clearly demonstrate that the widespread practice of prescribing 400 to 800 IU of vitamin D for patients with cystic fibrosis is frequently ineffective in maintaining normal body stores. It is clear that larger oral doses of vitamin D, parenteral vitamin D, or oral 25-OHD may be necessary to achieve serum 25-OHD concentrations in the mid- to high-normal range. Because renal function is generally normal in these patients, it should not be necessary to treat with 1,25(OH)₂D. Serum 25-OHD concentrations should be monitored to ensure the effectiveness of the selected regimen.

In summary, this cross-sectional study has confirmed that demineralization and fractures are common in adult patients with cystic fibrosis and pulmonary disease severe enough to consider lung transplantation. In view of the potential for further bone loss and fractures after lung transplantation, these results are of particular concern. Despite routine supplementation with parent vitamin D, low serum 25-OHD concentrations were also common and were associated with significantly lower BMD. Without histomorphometric studies of bone biopsies, it is not clear whether the demineralization reflects osteoporosis or osteomalacia. However, it is now obvious that vitamin D therapy must be individualized in order to ensure normal vitamin D stores. Further studies are necessary to establish the most effective regimen and to assess the effects of adequate vitamin D replacement on BMD in patients with cystic fibrosis.