Published ahead of print on February 1, 2007, doi:10.1164/rccm.200606-776OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200606-776OC
Low Bone Mineral Density in Young Children with Cystic Fibrosis1 Service de Pédiatrie Générale, Centre de Ressources et de Compétence en Mucoviscidose, and 2 Laboratoire d'Explorations Fonctionnelles, Hôpital Necker-Enfants Malades, Paris, France Correspondence and requests for reprints should be addressed to I. Sermet-Gaudelus, M.D., Ph.D., Service de Pédiatrie Générale, 149 rue de Sévres, 75015 Paris, France. E-mail: isabelle.sermet{at}nck.aphp.fr
Rationale: Low bone mineral density (BMD) is a frequent problem for adult patients with cystic fibrosis (CF). Only limited information is available for young patients. Objectives: The aim of this study was to evaluate BMD of children with CF younger than 6 years. Methods: BMD was measured at the lumbar spine (LS) after adjustment for height, sex, and pubertal status in 25 children with CF younger than 6 years, 53 prepubertal children aged 6 to 10 years, and 36 adolescents aged 11 to 18 years. Nutritional status, body composition, pulmonary disease severity, corticosteroid usage, dietary calcium, caloric intake, and vitamin D status were evaluated as potential correlates of BMD. Measurements and Main Results: The mean LS z score in the youngest group was significantly lower than normal (–0.96; SEM, 0.3). It did not differ significantly from that of children aged 6 to 10 years (–0.91; SEM, 0.2) or adolescents (-1.4; SEM, 0.2). LS z score was positively correlated with fat-free mass in multiple regression analysis. LS z score was less than –1 in 34% of the patients with mild pulmonary disease and normal nutritional status. Conclusions: These data suggest that the origin of CF bone disease in early childhood may be independent of nutritional status or disease severity.
Key Words: bone mineralization • cystic fibrosis • fat-free mass • cystic fibrosis transmembrane conductance regulator
Deficient bone mineral density (BMD) is becoming an increasingly important clinical issue in adult patients with cystic fibrosis (CF). Contributing factors include malnutrition, chronic infection, vitamin D deficiency, hypogonadism, delayed puberty, and reduced physical activity (1). The pathogenesis of low BMD in individuals with CF remains uncertain and bone histomorphometry studies have, thus far, been limited to adults with usually severe CF disease (2). Only limited information is available about bone status in growing children, especially the youngest children (3). Defining bone mineral status in this population might provide new insight into the underlying pathogenesis of BMD deficits in CF, identify age at onset, and help optimize preventive treatment strategies during this critical developmental period. The aim of this study was to determine the spectrum of BMD in a cohort of children younger than 6 years and compare it with BMD in older children. We evaluated the correlation between BMD and clinical status, with special emphasis on fat-free mass (FFM), disease severity, vitamin D status, and caloric and calcium intake.
Patients This study recruited 114 children and adolescents aged 2 to 18 years from our outpatient clinic at Necker-Enfants Malades Hospital. CF diagnosis was based on elevated sweat chloride levels and/or two CFTR gene mutations. Puberty was evaluated with Tanner staging (4). Chronic colonization with Pseudomonas aeruginosa, dose and duration of oral and inhaled corticosteroid usage, and vitamin D supplementation were reviewed. We also recorded the following additional data for children younger than 6 years: birth weight, age at diagnosis, weight and height at diagnosis, and weight at 2 years after diagnosis. The study was conducted in accordance with the World Medical Association Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects, as part of the annual routine assessment. Disease severity was assessed when BMD was measured. It was defined by the number of antibiotic courses in the preceding year and by pulmonary function (i.e., FEV1 and FVC, expressed as percentage of predicted values for age and sex).
Bone Mineral Assessment BMD is influenced by bone size and may be underestimated in smaller individuals (5). Accordingly, the fact that many of the children with CF in our study were smaller than average for their age may bias our findings. To compensate for differences in body size, BMD evaluation was based on height-adjusted rather than chronological age (5). Lumbar spine (LS) BMD z score was thus calculated with reference to a population of healthy Parisian children and adolescents matched for statural age, sex, and puberty status. Our control database included 317 children and adolescents aged 1 to 20 years, with normal nutritional status, living in the same geographic area, with no known cardiac, renal, or hepatic disorders (6–8). None had a history of skeletal abnormality or immobilization for more than 1 week. None had received antiepileptic drugs, corticosteroids, or calcium supplements. Most had received one oral dose of 5 mg vitamin D2 at the end of the preceding winter. Sixty-six control subjects were younger than 6 years. Longitudinal data were available for 90 subjects and showed normal growth velocity. In the following analysis, we considered only LS z score adjusted for statural age. Moderately low BMD was defined by a z score between –1 and –2 standard deviations (SDs), low BMD by a z score between –2 and –3 SDs, and a severe low BMD by a z score below –3 SDs.
Anthropometric Assessment Mean daily caloric and calcium intake was determined by a dietitian from a 3-day diary, and expressed as percentages of the recommended dietary allowance (RDA) for age. Caloric intake was expressed as a ratio to 120% of the RDA, in accordance with standard recommendations for CF care (12). Calcium intake was compared with recommended calcium requirements (13).
Laboratory Measurements Serum 25[OH]D (DiaSorin, Stillwater, MN), PTH (Allegro Intact PTH; Nichols Institute, San Clemente, CA), IGF-1 (IGF-I RIA-CT; CIS Bio International, Gif sur Yvette, France), and total OC (Osteocalcin RIA-CT; CIS Bio International) were all measured by radioimmunoassay. Intra- and interassay coefficients of variation were less than 10%. IGF-1 and OC concentrations were expressed as z scores calculated from our laboratory's normal values for age, sex, and pubertal stage. Vitamin D insufficiency was defined in accordance with recommendations by a 25[OH]D concentration of more than 30 ng/ml (3, 14). Our normal values for serum PTH were 10 to 46 pg/ml (15).
Statistical Analysis
Clinical Characteristics This study included 25 children younger than 6 years, 53 prepubertal children aged 6 to 10 years, and 36 adolescents aged 11 to 18 years. Table 1 summarizes and compares the clinical characteristics of the three patient groups. In the group of the children younger than 6 years, seven patients were aged 2 to 4 years. Mean %IWH in the youngest group was within normal range for 23 of 25 children. Body composition was available for 22 children. The mean z score for FFM (FFM z score) was significantly below normal, with values below –2 for 10 children. Half of the normal-weight patients (%IWH > 85%) had FFM z scores below –2. FFM z score was significantly correlated with respiratory function parameters such as FEV1 (r = 0.41, p < 0.0001). The FM z score was greater than 0 in 16 patients. Fifteen patients were able to perform lung function tests. Eleven had values exceeding 85% of predicted values for age and sex.
%IWH, FFM, and FM z score values decreased significantly with age (Table 2). Pulmonary disease severity increased significantly, mainly in adolescents, as shown by the higher proportion of patients colonized with P. aeruginosa, and the deteriorating FEV1 and FVC values. Among the normal-weight patients (%IWH > 85%), 60% had a FFM z score of less than –2. FFM was highly correlated with FEV1 (r = 0.409, p < 0.0001) and FVC (r = 0.4, p < 0.0001). Puberty was delayed in girls as shown by the later age for first menarche in our patients at 14.3 (SEM, 0.5) years compared with 12.2 (SEM, 1) years in our healthy control population. Age of puberty was 15 (SEM, 0.6) years in boys compared with 13.9 (SEM, 0.8) years in our control population.
Dietary data were available for analysis for 61 subjects (Table 1). Anthropometric values did not differ significantly between those with and without dietary evaluations. Mean caloric intake was 90 (SEM, 3)% of the recommended RDA. Mean daily calcium intake was above the recommended daily intake for age (13) and increased significantly in adolescents. Vitamin K supplementation was prescribed for the entire population (between 5 and 10 mg/wk). Fifty-five patients had received inhaled corticosteroids for longer than 6 months and 15 had received oral corticosteroid therapy for longer than 3 months.
BMD and Calciotropic Metabolism Data
Compared with the youngest group, LS z score remained stable among the children aged 6 to 10 years (–0.91 [SEM, 0.2]) and decreased, although not significantly, in the adolescents (–1.4 [SEM, 0.16]). The proportion of patients with a value of less than –2 did not vary significantly among the three groups but tended to increase in the adolescent group (28% in the children < 6 yr, 22% in the 6–10-yr group, and 34% of the adolescents). Calcium and phosphorus were within normal range for all patients. Ninety percent had 25[OH]D serum levels below the recommended threshold of 30 ng/ml, and 9% had elevated PTH (> 46 pg/ml). There was no relation between the 25[OH]D level and the season of sampling. Adolescents had significantly lower 25[OH]D levels than the other two groups. All the patients included in the study received vitamin D2 supplementation of 2,000–4,000 units per day. Although mean OC level was lower than normal, OC z scores varied substantially, ranging from –3.9 to +2.8. OC z score was correlated with FEV1 (r = 0.32, p = 0.009) and FFM (r = 0.24, p = 0.03) but not with the IGF-1 z score, 25[OH]D, calcemia, or PTH.
Correlates of Bone Mass LS z score was significantly correlated with FVC, FEV1, and FFM z score in the univariate analysis (Table 4, Figure 3). FFM was normal in patients with normal BMD (FFM z score = –0.4 [1] in patients with an LS z score > 1 vs. –3.1 [SEM, 0.2] in those with an LS z score < 1; p = 0.009, Mann-Whitney test). When these positive correlations were tested in a stepwise multivariate regression model, only the FFM z score remained correlated with LS z score (multiple correlation coefficient = 0.35, p = 0.006; partial correlation coefficient for the FFM z score = 0.14, p = 0.01).
CF lung disease was minimal (FEV1 > 75% of predicted) and nutritional status was good (%IWH > 85%) in 41 patients with a mean age of 8.9 (SEM, 0.6) years. The mean LS z score of these patients was –0.5 (SEM, 0.2). Fourteen had LS z scores below –1 and 5 below –2. These patients did not differ from the other patients for 25[OH]D level, OC z score, or IGF-1 z score. According to the univariate analysis, the LS z score in this group was not correlated with body composition or respiratory disease severity (Table 5).
This large study in patients with CF across a wide age range from 2 years through puberty shows that low BMD can develop in early childhood. Although the values were normalized for statural age to correct for the patients' smaller size, mean BMD at the LS was significantly lower than expected in the children younger than 6 years. Our data also attest to a defect in bone mineralization in a subset of patients with mild disease and normal nutritional status. This suggests that the bone disease in CF may be independent of nutritional status or disease severity. This is the first study to focus on bone status in patients with CF younger than 6 years, including children as young as 2 years old. Although most of these children had normal nutritional status and mild pulmonary disease, half had z scores lower than –1 and nearly one-third lower than –2. These percentages did not differ significantly from those of the children between 6 and 10 years and the adolescents but the proportion of low BMD did increase in the adolescent group. Studies of prepubertal children with CF generally have included older children in the analysis and have reported discordant findings (16–28). These discrepancies may in part reflect methodologic differences, such as sensitivity of the densitometric technique, the skeletal site studied, the control population used for normalization, and the method used for adjustment. Most studies use DEXA, which provides a measure of areal BMD, representing the integral (cortical and trabecular) mineral content of the bone envelope of the skeleton region scanned divided by its projected bone area. Because depth, and therefore volume, of the scanned bone is not measured, areal BMD fails to distinguish between changes in mineral density and bone size in growing children (5). This is an important consideration because CF may impair both growth and mineralization of the skeleton. As many of the adolescents included in the study had a reduced height and a delayed puberty, we compared the measurements with those of our local control subjects matched for statural age, sex, and pubertal status. Our results should therefore not be due to shorter stature or delayed puberty. This approach gives a "corrected" value of the BMD z score, as shown by the significantly higher values of the z scores adjusted for statural age, compared with chronological age in our study (data not shown). Haslam and colleagues and Gronowitz and coworkers also reported lower BMD in children with CF compared with reference values matched for chronological age (20, 22). Similarly, Bianchi and colleagues also found a mean reduction of LS z score to –1.6 in 55 children aged 3 years to puberty, and confirmed these data even after adjusting for vertebral volume (28).
The only factor that remained significantly associated with BMD in the multivariate analysis, independently of all other disease-related risk factors, was FFM, which is indirectly related to degree of inflammation. The mechanism underlying this finding may be the continuous inflammatory and protein catabolic state that simultaneously induces more severe pulmonary disease, greater bone resorption, and low muscle mass, as suggested by the association of both low bone mass and low FFM with high circulating levels of IL-6 and tumor necrosis factor- Interestingly, even when we limited BMD analysis to well-nourished patients with functionally mild pulmonary disease, 34% had LS z scores below –1, including patients younger than 6 years. Similar findings were reported by Gronowitz and associates and Bianchi and coworkers in their cohort of normally growing prepubertal children with mild lung disease (22, 28). In view of the high prevalence of osteopenia in these very young children, we studied contributing factors. We examined whether these patients also had a higher prevalence of undernutrition in early infancy. Although there is a trend to a lower BMD in the patients underweight at diagnosis, this difference was not significant and, 2 years after diagnosis, weight was normal in these patients. A definitive conclusion is, however, not possible because of the small sample. We cannot formally rule out the possibility that undernutrition in early infancy together with insufficient nutritional catch-up and inadequate bone growth set the conditions for later osteopenia. Altered bone remodeling may be present in young children as suggested by their significantly low OC levels, But there was no significant correlation between this marker and LS z score. Moreover, the underlying mechanism cannot be reliably determined from information about a single evaluation (35). Further study of this hypothesis is warranted for a definite conclusion. Puberty was delayed in girls as shown by the later age for first menarche in our patients, as previously reported (36, 37). Longitudinal studies of healthy adolescents show that up to 25% of the peak bone mass is acquired during the period of peak height velocity (38). Peak bone mass may be insufficient in our population because of the delayed onset of puberty. These findings may explain in part the lower bone mass in adolescents with CF. Steroid therapy, either oral or inhaled, was significantly associated with low BMD in our patients. Long-term steroid treatment diminishes the peak bone mass and induces osteoporosis in pediatric populations (28, 39). The principal mechanisms for the alteration of bone metabolism by steroids are the inhibition of osteoblast activity, intestinal calcium absorption, renal tubular reabsorption, and the suppression of gonadal and growth hormone secretion (39). Accordingly, our observations of a reduction of BMD beginning in early childhood, even in children with moderate CF, provide arguments for a primary bone CF disease. This hypothesis is further supported by data in CF knockout mice, which form less trabecular and cortical bone than their non-CF littermates (40). A recent study established that BMD in patients homozygous for the F508del mutation is significantly lower than normal and thus suggests that there may be a genetic link between CF and bone disease (41). Like two other studies (28, 42), we did not find such correlation. The relation between BMD and CFTR genotype, if any, may therefore be more complex. The question of other modifier genes involved in bone mineralization remains unresolved and requires extensive screening studies (42). An important issue is whether CFTR expression by itself or by interaction with other proteins is involved in Ca2+ homeostasis or bone metabolism (43). A very recent study by Aris and collaborators reports the presence of CFTR transcripts in osteoblasts from rat calvaria (44). Moreover, a cAMP-dependent Cl– current has been identified in rat osteoblasts (45). Although its pharmacokinetic properties differ from those of CFTR, this ion transport may play a role in bone mineralization. On the other hand, Bakouh and colleagues demonstrated a functional interaction between CFTR and the renal sodium-phosphate cotransporter NPTIIa, involved in phosphate reabsorption (46). These examples make it likely that CFTR, by direct link or through interaction with other proteins related to bone metabolism, may be a determinant factor in bone health. It thus might be that some patients with CF have a genetic background prone to osteoporosis from birth onward. Additional disease-related factors, such as low FFM, excessive inflammation, poor nutritional status, and low vitamin D levels, may then aggravate the initial bone defect and contribute to inadequate bone mineral accretion during adolescence
Conclusions
The authors thank Laetitia Parbaille for technical assistance.
Originally Published in Press as DOI: 10.1164/rccm.200606-776OC on February 1, 2007 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form June 12, 2006; accepted in final form February 1, 2007
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
(
