Low bone mineral density (BMD) has been recognized as a potential problem in both adults1–6 and children7–13 suffering from inflammatory bowel disease (IBD). The occurrence of low BMD shows a high variance in pediatric studies from different countries, ranging from 5%–70%. Recently, decreased BMD has been reported even in relation to other chronic diseases beginning early in life, such as insulin-dependent diabetes or cystic fibrosis.14–18
IBD is one of the most common chronic childhood diseases in the Western world, with an increasing incidence.19–21 About 20% of all patients with IBD develop the disease during childhood and adolescence.
The pathogenesis of disturbed bone mineralization in childhood IBD is regarded as multifactorial. Factors include cytokines released from the inflamed bowel due to the chronic inflammation,22 treatment with corticosteroids,23 low weight,24 vitamin D deficiency25 and genetics.26
In healthy individuals, BMD increases rapidly during puberty and reaches its highest level, peak bone mass, in lumbar spine at about age 16 in females and about age 20 in males.27 During development, BMD is susceptible to disturbances as inflammatory conditions. Decreased peak bone mass in young adulthood may predispose for development of osteoporosis later in life, which in turn may lead to osteoporosis-related fractures. The current available definition of osteoporosis provided by the World Health Organization is designed for adult women.28 According to the International Society for Clinical Densitometry (ISCD) 2007 Pediatric Official Positions,29 the diagnosis of osteoporosis in children and adolescents (males and females ages 5–19) requires the presence of both a clinically significant fracture history and low bone mineral content (BMC) or BMD, whereas low BMC or BMD is defined as a BMC or areal BMD Z-score that is less than or equal to −2.0, adjusted for age, gender, and body size, as appropriate. The term osteopenia should not be used in children and adolescents at all.
The aim of our study was to investigate BMD in a population of Swedish children and adolescents with IBD and to evaluate a number of factors that may influence BMD.
- Top of page
- MATERIALS AND METHODS
In the population of our IBD patients we found normal BMD values in the whole body scan (BMD mean Z-score 0.13, range −4.6 to +3.4 SD, P = 0.22). In contrast, decreased BMD values were found in the lumbar spine, especially for L2–L4 (Table 3), where the BMD mean Z-score was −0.78 SD for the entire group of patients (range −5.9 to +3.7 SD, P < 0.001). The individual BMD values of the lumbar spine of our patients are shown in Figure 1.
Figure 1. Bone mineral density (BMD) of the lumbar spine (L2–L4) in relation to chronological age for patients with inflammatory bowel disease, males (upper graph) and females (lower graph). Dashed line represents mean BMD, continuous line above +1 standard deviation (SD) and continuous line beneath −1 SD for age-matched controls.
Download figure to PowerPoint
Table 3. Bone Mineral Density at Different Skeletal Sites
|BMD whole body (g/cm2)||1.05||0.7-1.4||0.15|
|BMD Z-score whole body (SD)||0.13||−4.6-3.4||1.31|
|BMD spine L1-L4 (g/cm2)||0.93*||0.4-1.4||0.22|
|BMD Z-score spine L1-L4 (SD)||−0.73*||−4.4-3.7||1.54|
|BMD spine L2-L4 (g/cm2)||0.95*||0.4-1.4||0.22|
|BMD Z-score spine L2-L4 (SD)||−0.78*||−5.9-3.7||1.62|
BMD and Disease Category
Significantly lower BMD values of the lumbar spine (L2–L4) were found both in CD (BMD mean Z-score −1.1 SD, range −5.9 to +2.4 SD, P < 0.001) and in UC patients (BMD mean Z-score −0.8 SD, range −4.4 to +3.7 SD, P < 0.001). No statistically significant difference was found between the groups.
Decreased BMD with a BMD Z-score of the lumbar spine (L2–L4) < −1 SD occurred in 46.7% of the individuals with CD and 47.0% of those with UC. Low BMD with BMD Z-score of the lumbar spine (L2–L4) ≤ −2 SD was present in 26.7% of the patients with CD and in 24.1% of the UC patients.
In the IC group the BMD mean Z-score of the lumbar spine (L2–L4) was 0.07 SD (range −1.5 to 1.7 SD). Only 3 of 16 patients (18.7%) had a BMD Z-score < −1 SD and no patient had low BMD ≤ −2 SD. Because of the small number of patients, further statistical analysis of this group was not performed exclusively.
BMD and Gender
BMD Z-score of the lumbar spine (L2–L4) was significantly decreased both in boys (−1 SD, P < 0.001, 95% confidence interval [CI] −1.3 to −0.6) and in girls (−0.44 SD, P < 0.05, 95% CI −0.87 to −0.004). However, no significant gender difference was found.
BMD and Age
While analyzing the age groups with t-test regarding gender, no significant differences were found in BMD Z-score (L2–L4) of either age groups or the total sample (results shown in Table 4). Boys tended to have lower BMD in all age groups except the youngest. As a result, we correlated BMD to age separately for males and females. We found a weak negative correlation of lower BMD to higher age in the male group (r = −0.21, P < 0.05), but not in the female group (r = −0.04, P = 0.78).
Table 4. Bone Mineral Density in Relation to Gender and Age
|Age Group||BMD Mean Z-score L2-L4 in Females||BMD Mean Z-score L2-L4 in Males||BMD Mean Z-score L2-L4 in All Patients|
|6-11 years||−0.67 (n = 12, SD 1.4)||−0.32 (n = 13, SD 0.5)||−0.48 (n = 25, SD 1.0)|
|12-14 years||−0.12 (n = 19, SD 1.5)||−0.89 (n = 26, SD 1.6)||−0.56 (n = 45, SD 1.6)|
|15-16 years||−0.16 (n = 12, SD 1.7)||−0.78 (n = 25, SD 1.7)||−0.58 (n = 37, SD 1.7)|
|17-19 years||−1.26 (n = 8, SD 1.6)||−1.47 (n = 29, SD 1.9)||−1.43 (n = 37, SD 1.8)|
|Total||−0.44 (n = 51, SD 1.5)||−0.97 (n = 93, SD 1.6)||−0.78 (n = 144, SD 1.6)|
BMD and Bone Age
Bone age was obtained from 141 patients. While comparing the BMD mean Z-score (L2–L4) for these patients, using bone age instead of chronological age in the reference, no major differences were seen (Table 5). However, in 27 of these patients bone age was delayed by at least 1 year (19.1%). Those with delayed bone age had lower BMD mean Z-score of the lumbar spine (L2–L4) than the patients with normal bone age (−1.4 SD versus −0.6 SD). While comparing the BMD mean Z-score (L2–L4) for the subgroup of those patients with delayed bone age, using bone age instead of chronological age as the reference age, BMD mean Z-score of the lumbar spine (L2–L4) increased from −1.4 SD to −0.81 SD.
Table 5. Comparison of Bone Mineral Density Corrected for Chronological Age or Bone Age
|Gender and Number||BMD Mean Z-score (L2-L4) Corrected for Chronological Age and Gender||BMD Mean Z-score (L2-L4) Corrected for Bone Age and Gender|
|Females (n = 50)||−0.41||−0.63|
|Males (n = 91)||−0.96||−0.77|
|All patients (n = 141)||−0.76||−0.72|
BMD and BMI
There was a stronger positive correlation between BMD mean Z-score of the whole body and BMI in females (r = 0.65, P < 0.001) than in males (r = 0.31, P < 0.05). In line with this, the correlation coefficients for BMD mean Z-score of the lumbar spine (L2–L4) to BMI were 0.46 in girls (P < 0.001) and 0.10 (not significant, P = 0.37) in boys.
BMD and Body Fat
In this analysis we found a stronger positive correlation between BMD of the whole body and total body fat in females (r = 0.64, P < 0.001) than in males (r = 0.32, P < 0.05). The correlation coefficients for BMD lumbar spine (L2–L4) to total body fat were 0.61 in girls (P < 0.001) and 0.19 (not significant, P = 0.07) in boys.
BMD and Lean Body Mass
Analyzing the correlation between BMD of the whole body and lean mass, we found a strong positive correlation in males (r = 0.90, P < 0.001) as well as in females (r = 0.79, P < 0.001). The correlation coefficients for BMD lumbar spine (L2–L4) to lean mass were 0.86 in boys (P < 0.001) and 0.70 in girls (P < 0.001).
Possible Factors Associated with Lower BMD
Performing a multiple regression model with BMD lumbar spine (L2–L4) as the dependent variable, male gender and treatment with azathioprine were associated with lower BMD (Table 6). In contrast, age and BMI showed a positive correlation with BMD. Neither treatment with prednisolone, disease category, nor disease duration turned out to represent risk factors for lower BMD in this model.
Table 6. Predictive Factors for Low Bone Mineral Density in a Multiple Regression Model with Bone Mineral Density as the Dependent Variable
|Treatment with azathioprine||−0.20||<0.001|
- Top of page
- MATERIALS AND METHODS
To the extent of our knowledge, this is the first population-based study of BMD in children and adolescents with IBD. In concordance with other studies the lowest BMD values were found in the lumbar spine, where the mean Z-score was decreased to almost −1 SD compared to healthy controls. Almost 50% of our patients had decreased BMD with BMD Z-scores < −1 SD, whereas low BMD with BMD Z-scores ≤ −2 SD was present in one-fourth.
In the present study CD and UC patients had equally low BMD values. In accordance with our results, von Tirpitz et al4 showed in a study of adults with IBD low BMD in both CD (n = 125) and UC (n = 15) but could not find any significant difference between the groups. Previously studies have reported varying data regarding low BMD in children and adolescents with IBD. A study from Gokhale et al10 showed that patients aged 5 to 18 years with CD (n = 58) had significantly lower BMD than UC patients (n = 37). Boot et al9 reported the same from a smaller sample size. Low BMD values in CD patients had also been reported by Semeao et al,23 who found that in a sample of 119 individuals with CD (age 5 to 25 years) 70% had BMD Z-scores ≤ −1 SD and 32% BMD Z-scores ≤ −2 SD. In another sample of 43 pediatric CD patients, Herzog et al13 found that 44% of the patients had BMD Z-scores < −2 SD.
Another finding in our study was that BMI as well as total body fat and lean mass were positively correlated with BMD. This positive correlation of BMD and BMI has been reported previously by other authors.10, 11 Lean mass has been reported to be a strong predictor of BMD in healthy children.38 Accordingly, in our IBD patients BMD and lean mass as well as total body fat showed a strong positive correlation despite chronic disease, which per se might influence body composition.
In the present study treatment with azathioprine turned out to be a factor associated with lower BMD. In Sweden almost all patients are treated according to the national treatment protocol for pediatric IBD which is in line with international recommendations.39 Azathioprine is initially given as a supplement in a “step-up strategy” in relapsing or steroid-dependent disease. Azathioprine is not known to have an effect in itself on bone turnover. It is used in children with more active disease and we consider it therefore a marker of disease course severity. This assumption is in agreement with Semeao et al,23 who reported that treatment with 6-mercaptopurin (6-MP), a major metabolite of azathioprine, was a significant risk factor for lower BMD. They suggested that the use of 6-MP was an indicator of worsening disease severity, as it is used in refractory CD. Recapitulating, we think that azathioprine is a useful surrogate marker for disease course severity.
The influence of treatment with prednisolone on BMD in pediatric IBD patients remains controversial. Walther et al11 stated that severe low BMD with Z-score < −2 SD is present before steroid treatment. A number of case reports support this.40, 41 Other authors postulated that cumulative doses of prednisolone >5 g23 or >10 g42 were a risk factor for lower BMD. In our study treatment with prednisolone turned out not to be a risk factor. However, we chose not to calculate cumulative doses, because it was difficult to assess these doses accurately from the patient files retrospectively.
A methodological issue lies in the nature of DXA. DXA does not reflect a real density but a ratio of bone mineral content over an area, which may lead to an underestimation of BMD in pediatric subjects who are growth stunted, like some of our IBD patients. One attempt to overcome these difficulties was to correct BMD values for bone age. Herzog et al13 showed while performing this procedure that the figure for patients with abnormal BMD Z-score < −2 SD fell from 44% to 26%–30%. In the present study we also corrected BMD for bone age, which did not have any clear impact on the results for the whole group of our patients (Table 5), but there was a limited impact in those with delayed bone age. In other studies the areal BMD was transformed mathematically into volumetric BMD.11 Peripheral Quantitative Computed Tomography (pQCT) provides the possibility to measure the “true” volumetric BMD but the method was not available at the time we conducted our study.
As previously mentioned, we used the Lunar material as pediatric DXA reference data and did not recruit our own control group. We considered the Lunar material as sufficient as it contains a large number of measurements of Caucasian volunteers from different countries, inter alia also from Europe. Moreover, the relevance of the Lunar reference data was shown by Eriksson et al43 in a Swedish population from the area of Gothenburg, in which part of Sweden our study was also conducted.
The clinical significance of decreased BMD in pediatric IBD patients is still debated. BMD changes dramatically during childhood and adolescence, reaching its highest amount, peak bone mass, in puberty. It is accepted that reduced peak bone mass is a main contributing factor to osteoporosis and may subsequently lead to osteoporosis-related fractures in adulthood. The question to be answered is whether the occurrence of IBD during adolescence contributes to decreased peak bone mass and to what extent this contributes to the described consequences later in life. The clinical significance of low BMD is always measured in terms of fractures in adult studies. Besides a few case reports,40, 41 there is only a single pediatric IBD study addressing this question. Persad et al.44 found no statistically significant difference in the prevalence of fractures in IBD patients compared with their normal siblings, which was an unexpected finding. Here would be longitudinal data of interest, following the pediatric onset IBD patients into adulthood.
In summary, in this cross-sectional population-based study we showed that low BMD was present in a considerably large proportion of children with IBD. Furthermore, male gender, low BMI, and treatment with azathioprine, as a probable marker for disease course severity, turned out as potential risk factors for low BMD. However, it is not possible from this study to conclude whether there is a risk for reduction in peak bone mass or increased rate of fractures. For this purpose longitudinal studies are needed.