Differences in bone mineral density (BMD) have been observed between adults with Down syndrome (DS) and the general population. The purpose of this article is to describe the prevalence of bone mass disorders in a cohort of adults with DS and their predisposing factors. We performed a cross-sectional study of 104 consecutively recruited adults with DS from an outpatient clinic of a tertiary care hospital in Madrid, Spain. We recorded epidemiological and anthropometric data, nutritional variables, coexisting clinical conditions, and laboratory variables. BMD was measured at the lumbar spine, total hip, and femoral neck using dual-energy X-ray absorptiometry. The prevalence of osteopenia ranged from 48% to 52%, and that of osteoporosis ranged from 19% to 22% depending on the site of measurement (femoral neck or lumbar spine, respectively). Age was the greatest risk factor associated for lower BMD, with similar bone mass accrual curve but with lower peak of BMD than the general population. We conclude that low bone mass is an extremely prevalent condition in adult patients with DS.
Down syndrome (DS) is the most frequent chromosomal disorder in live newborns and the most prevalent cause of congenital mental disability in Western countries (Roizen & Patterson, 2003). Its incidence ranges from 1/660 to 1/1000 newborns (Florez & Ruiz, 2006). Since the 1980s, life expectancy in the DS population has increased significantly (Khoshnood, De Vigan, Vodovar, Goujard, & Goffinet, 2004). In recent decades, medical advances, such as improvements in cardiac surgery, prevention of childhood infections, broader access to standard care, and better global psychosocial support, have enabled elderly adults with DS to reach the seventh decade of life (Bittles & Glasson, 2004; Yang, Rasmussen, & Friedman, 2002). In fact, the lifespan of adults with DS may be similar to that of the general population within the next generation (Kerins, Petrovic, Bruder, & Gruman, 2008). This “new” population of adults with DS represents unique clinical problems that differ from those of the pediatric population with DS or the general population (Morris & Alberman, 2009). However, data on the aging process in people with DS are lacking. The study of age-related processes, such as osteoporosis, is of the utmost importance (Hawli, Nasrallah, & Fuleihan, 2009).
Osteoporosis is a growing public health problem. More than 10 million Americans have osteoporosis, and a further 34 million have low bone mass and are, therefore, at increased risk for developing osteoporosis and fractures (Watts et al., 2010). About 80% of these individuals are women, most of them over 50 years of age. The prevalence of osteoporosis in Spain ranges between 25% in women aged 60 to 69 years and 40% in those aged 70–79 years, with age and menopause being the most important risk factors (Díaz Curiel et al., 2001).
DS has been proposed as an independent risk factor for the development of osteopenia and osteoporosis (McKelvey et al., 2013). Diverse studies have revealed altered bone mass in children and adolescents with DS (Baptista, Varela, & Sardinha, 2005; Guijarro, Valero, Paule, Gonzalez-Macias, & Riancho, 2008), whose bone composition differs from that of the general population with respect to its biomechanical resistance to fractures (Grimwood, Kumar, Bickerstaff, & Suvarna, 2000).
Preliminary studies adjusted for age and gender revealed lower bone mineral density (BMD) in adults with DS than in the general population (Kao, Chen, Wang, & Yeh, 1992; Sepúlveda et al., 1995). This difference may be caused by anomalous bone formation during childhood leading to reduced bone turnover (Angelopoulou, Souftas, Sakadamis, & Mandroukas, 1999). Suggested predisposing factors for osteopenia in this population include dietary deficiencies; a higher prevalence of celiac disease; the high prevalence of hypovitaminosis D; primary hypothyroidism and/or primary hypogonadism; the use of different medications such as antipsychotics, glucocorticoids, or mineral supplements; and a more sedentary lifestyle (McKelvey et al., 2013; Sakadamis, Angelopoulou, Matziari, Papameletiou, & Souftas, 2002). In addition, a very important factor that can negatively affect the bone mineralization in DS is hypotonia and inferior muscular strength, especially for low lumbar BMD (Angelopoulou et al, 2000). Furthermore, a higher incidence of falls has also been observed in adults with DS (Srikanth, Cassidy, Joiner, & Teeluckdharry, 2011).
However, these hypotheses are based either on studies that were conducted with a small sample size or data extrapolated from studies in the general population. The aim of this study is to describe the prevalence of BMD disorders in a large cohort of Spanish adults with DS. We also analyze the prevalence of diverse risk factors for osteoporosis in this DS sample.
This article describes the results of a cross-sectional study that included 104 adults with DS who were consecutively selected from patients at the Adult Down Syndrome Outpatient Clinic of the Department of Internal Medicine, at Hospital Universitario de La Princesa in Madrid, Spain. The recruitment period ran from January 2009 to December 2011. The study was conducted in accordance with the provisions of the Declaration of Helsinki and Good Clinical Practice guidelines, and the local institutional review board (IRB) approved the study protocol and waived the need for informed consent due to the retrospective design. All data were treated in the strictest confidence, according to the most recent Spanish legislation on data protection.
To be included, all participants had to be over 16 years of age and the presence of DS had to be documented by karyotype or typical phenotype. In the initial outpatient clinic attention, a comprehensive study of the health status of each subject was performed. This included a general blood test and, in most cases, a dual-energy X-ray absorptiometry (DXA) assay to study the bone status. Data for this study were obtained from medical records from the patients' initial evaluation, which was always performed by the same physician (MF). Most of these patients came to the outpatient clinic for comprehensive assessment, with no specific health problem, because the clinic is a national reference center for people with DS.
Most of the adults selected for the study were home based and belonged to upper middle class. Those of more advanced age lived with siblings or in specialized residences.
Measures and Covariate
We recorded epidemiological and anthropometric variables like age (years), gender, height (cm), weight (kg), and body mass index (kg/m2). Height was measured to the nearest centimeter using a standardized stadiometer, and weight was measured on a calibrated balance scale to the nearest 0.5 kg. Body weight categories (normal weight, overweight, and obesity) were defined according to accepted BMI cut-off values (Tsigos et al., 2008).
We also analyzed the following clinical variables that were obtained through systematic review of medical records: tobacco or alcohol use, thyroid disorders, menopausal status, primary hypogonadism, hearing or visual deficits (ophthalmological and audiometric testing were performed as part of routine clinical practice at the first visit), and bone loss-related medication that can influence BMD development (including anxiolytics, antidepressants, antipsychotics, levothyroxine, corticosteroids, vitamin D supplements, oral contraceptives, and hormone replacement therapy).
We investigated any prior clinical bone fractures by either reviewing their medical records or conducting medical questioning. Osteoporotic and non-osteoporotic fractures were defined based on the Spanish Society of Rheumatology criteria (Pérez Edo et al., 2011). An osteoporotic fracture or fragility fracture is conditioned by low-impact trauma. A fall from a standing or sitting position is, therefore, included in this concept, and fractures that occur as a consequence of sports or accidents are excluded from the definition. The most frequent and relevant osteoporotic fractures are those of the proximal femur, the spinal column, and the distal forearm (Pérez Edo et al., 2011).
An important tool (FRAX) to predict fracture risk at 10 years has been developed based on the use of clinical risk factors, with or without bone mineral density tests, and has been applied to the study of the general population. This tool has not been validated yet in DS. FRAX has proven useful in predicting fracture only in certain models of risk. In our study, it was calculated in all subjects for whom it was clinically appropriate, particularly in postmenopausal women or men aged over 50 (Kanis, Johnell, Oden, Johansson, & McCloskey, 2008).
All subjects underwent a DXA evaluation using a densitometer. In our experience, the in vivo variation coefficients for total hip and lumbar spine (LS) measurements with this device are around 1%. Osteoporosis and osteopenia were defined according to internationally accepted criteria using the t-score and Z-score (Kanis & Kanis, 1994). T-score compares the patient's BMD with the mean BMD of the young, healthy population (i.e., between 25 and 30 years) of the same sex and race, which is when peak bone mass is reached in the general population. Z-score is relative to healthy subjects of the same age and sex. Normal BMD is defined when t-score is −1 or greater, osteopenia when t-score is between < −1 and > −2.5, and osteoporosis when t-score is ≤ −2.5. Low BMD is defined when Z-score is ≤ −2 (Watts et al., 2010).
Analytic measurements of hormones and vitamin D were performed to assess their association with lower BMD. Levels of thyroid hormones (thyroid-stimulating hormone and free T4), 25[OH] vitamin D, intact parathormone, testosterone, estradiol, progesterone, follicle-stimulating hormone (FSH), and luteinizing hormone (LH) were automatically determined using Elecsys and Cobas Roche analyzers. The variation coefficients of the parathormone and vitamin D assays in these analyzers are 1.3% to 2%. Several diagnostic categories based on serum vitamin D levels were used. Vitamin D insufficiency was defined as plasma 25[OH]vitamin D levels between 20 and 29 ng/mL; vitamin D deficiency was defined as plasma 25[OH] vitamin D levels between 10 and 19 ng/mL; and severe vitamin D deficiency was defined as plasma 25[OH]vitamin D levels <10 ng/mL (Dawson-Hughes et al., 2005). All participants had undergone a fasting blood test after a minimum 10-hour overnight fast in the last 6 months before their inclusion in the study.
See Figure 1 for bone mineral mass acquisition adjusted by age with a scatter diagram.
Descriptive analyses are presented as absolute frequencies (percentages) and means ± standard deviation (SD). A χ2 test (with Yates correction when applicable) was used to calculate statistical significance in the comparison of categorical variables. Interval scales variables were analyzed using the t-test. A linear multiple regression was planned simultaneously for the identification of relevant risk factors (vitamin D deficiency, hypothyroidism, use of drugs predisposing to bone loss, BMI, hypogonadism, gender, age, and menopausal status) for the development of BMD disorders. Based on all mean BMD values, a bone mass development curve was plotted using a linear interactive graph that associated mean BMD values with age by linear correlation. All statistical tests were 2-tailed. Statistical significance was set at p < 0.05. All data were processed using SPSS software (SPSS 15.0.0).
General Clinical and Anthropometric Characteristics
The study sample comprised of 104 adults with DS. The baseline clinical and anthropometric characteristics of the population of our study are presented in Table 1. The gender distribution was even, and the mean age was 35 (SD = 11) years. Only 41 subjects (39%) had normal weight and none were underweight. The prevalence of overweight and obesity was high (31% and 30%, respectively).
With regard to the different risk factors for the development of osteoporosis, 74 adults with DS (71%) had inadequately low vitamin D levels. Six subjects presented with concomitant secondary hyperparathyroidism, all as a result of vitamin D deficiency. Forty-seven adults (45%) had primary hypothyroidism, and all of them were receiving levothyroxine at the time of study entry. Eight adults were diagnosed with primary hypogonadism based on biochemical parameters (8/90; 9%). Only 10 adults (9.6%) had a previous fracture; nine of them were due to high energy/impact falls and only one was considered an osteoporotic fracture (distal radius). It is noteworthy that a high number of patients had visual impairment (76%) or hearing loss (16%). Fifteen subjects had both deficits at a time (14%). No relevant unhealthy habits (e.g., alcohol, smoking) were reported in these patients.
Table 1 also shows similar baseline characteristics in both sexes except for BMI, which was higher in women, who also had higher rates of hypothyroidism.
The most frequently prescribed drug groups are listed in Table 2. Out of the 74 individuals with inadequate vitamin D levels, 71 (94%) were receiving calcium and vitamin D supplements, and only 1 received concomitant bisphosphonates. Twenty-three adults with DS (22%) received psychiatric medication, such as neuroleptics (16 subjects), antidepressants (21 individuals), or both (14 adults). The differences in BMD between these subjects and the total sample were not statistically significant. Only 3 subjects received corticosteroids as part of their usual treatment, and they seemed to have a lower BMD than the rest of population (femoral neck BMD, 0.594 g/cm2 vs. 0.713 g/cm2); however, the difference was not statistically significant, probably owing to the small number of patients.
Mean densitometric values are shown in Table 3. Thirty subjects (29%) presented low lumbar spine BMD (LS-BMD) values for their age and gender. Nineteen males (37%) had inadequate LS-BMD, compared with only 11 females (21%; p = 0.08). Average lumbar Z-score was significantly lower in men (–1.84 [1.16] vs. –1.29 [1.01]; p = 0.02). Sixteen adults with DS (15%) had low femoral neck BMD (FN-BMD) values, 19% in males and 11% in females, although the differences were not statistically significant. The average femoral neck Z-score was lower in men (–1.29 [1.06] vs. –1.01 [1.06]; p = 0.09). Seven subjects (6.7%) had low total hip BMD, only one was female (11.5% in males vs. 2% in females; p = 0.05), and the average total hip Z-score was also significantly lower in men (–0.80 [0.85] vs. –0.47 [0.83]; p = 0.04).
Based on lumbar spine t-score values, 48% of adults with DS were diagnosed with osteopenia and 22% with osteoporosis. Only 30% had a normal lumbar t-score. Based on femoral neck t-score values, 52% were diagnosed with osteopenia, 19% with osteoporosis, and 29% had normal femoral neck t-score. The risk factors considered in linear multiple regression were vitamin D deficiency, hypothyroidism, use of drugs predisposing to bone loss, BMI, hypogonadism, gender, age, and menopausal status. Among them, age and menopause were the only variables associated with lower BMD (R = 6.5 [p < 0.001] and R = 7.5 [p < 0.02], respectively).
The FRAX tool was used to estimate the long-term risk of an osteoporotic fracture in a subsample of 18 subjects (4 men and 14 women). The mean FRAX score was 3.2 (1.3)% for major osteoporotic fractures and 1.2 (1)% for hip fracture. In the gender-stratified analysis, male adults with DS presented a significantly higher FRAX score for hip fracture than women [2.3 (1.4)% vs. 0.9 (0.7)%; p < 0.05], although the overall FRAX score for major osteoporotic fracture was similar in both sexes [4.0 (1.5)% vs. 3.0 (1.2)%; p = 0.14].
Development of Bone Mass in our Sample of Adults With DS
Finally, we analyzed the bone mineral mass acquisition adjusted by age (Figure 1), which shows, in a scatter diagram, the BMD development curve in our sample. A progressive increase in BMD was noted until approximately 25–28 years of age, with a continuous decline thereafter. The peak femoral neck BMD was reached between 21 and 29 years, and its mean value was 0.832 (0.12) g/cm2. Given the similarity of the raw data and the total sample size, gender distribution data were analyzed together.
The correlation Pearson test between age and femoral neck BMD was −0.65 (−0.76 to −0.52 95% CI; p < 0.0001). That is, the FN-BMD decreased significantly with age. Interestingly, the values in both sexes were similar.
In our study, young adults with DS show a high prevalence of low bone mass–related disorders. In our sample, the prevalence of osteopenia and osteoporosis was close to 50% and 20%, respectively. This situation could be related in part to the lower peak of bone mass acquisition observed in these subjects. BMD of DS population was similar in both sexes in all ages, unlike the general population, in which BMD is lower in women (Díaz Curiel et al 1997). However, when looking at Z-score (which compares BMD with that of healthy subjects of the same age and sex) male DS patients presented lower Z-scores than DS females. An explanation for these results could be that postmenopausal women are under-represented in our study, so the higher prevalence of relatively young women in the sample preserved the protective estrogen effect with regard to males.
On the other hand, females with DS had higher BMI and high prevalence of hypothyroidism, but these differences did not affect BMD when the multivariate analysis was adjusted by gender. Of note, the majority of the subjects had normal thyroid function at the moment of inclusion in the study.
Our results show that young adults with DS have similar densitometric values to those observed in a cohort of Spanish postmenopausal women (Díaz Curiel et al., 2001). McKelvey et al. (2013) had already reached similar conclusions in a much smaller study of adults with DS (n = 30), where 10 subjects (30%) and 6 subjects (20%) with DS presented inappropriately low Z-scores at the lumbar spine and femoral neck, respectively, compared with the 30 (29%) and 16 subjects (15%) in our study. The average BMD values in our study sample were slightly lower than those observed by Angelopoulou et al. (2000), LS-BMD 0.926 (0.07) g/cm2 and FN-BMD 0.979 (0.12) g/cm2 in young adults with DS. However, these discrepancies could be due to a lower average age in the Angelopoulou and colleagues' series than in the present study.
Other studies consider that the smaller bone size in DS could magnify the effect of lower areal BMD. In these studies, as in ours, male gender was a risk factor for low BMD (Guijarro et al, 2008). In our study, we have used the areal BMD, which could be underestimating the true BMD value in this population, due to the small size of patients with DS. It is possible that other parameters such as volumetric (v) BMD, trabecular bone score (TBS), or comparison with a standardized curve of DS population may be better than the t-score and Z-score to estimate fracture risk in these subjects. TBS is a bone assessment tool that offers information about a bone's solid volume fraction and mean solid thickness, in addition to providing information about fracture risk. vBMD is a good tool to estimate the bone strength in growing subjects or small size people, because it represents the three dimensions of the space. However, at the moment, areal BMD is the only validated parameter to estimate the risk of fracture in the general population.
In spite of gender differences, we found the most relevant predisposing factors for osteopenia and osteoporosis to be age and menopause, which are also key to the development of these conditions in the general population. No other factors were significantly linked to the development of bone metabolism disorders in adults with DS. Of note, we found a high prevalence of vitamin D deficit, with three of every four subjects presenting inappropriately low vitamin D levels. However, these high percentages have also been reported in the general population in Spain (Calatayud, Jódar, Sánchez, Guadalix, Hawkins, 2009). Interestingly, most of our patients (94%) take supplements of vitamin D, which does not happen in the Spanish general population. Other predisposing conditions included a lower prevalence of hypogonadism than in previous studies. Although Hsiang et al. (1987) found a prevalence of hypogonadism of 30–40% in adults with DS (compared with 9% in the present study), this difference is probably due to the inclusion of patients with subclinical hypogonadism (normal testosterone levels with elevated LH/FSH).
The use of psychiatric medications (neuroleptics, antidepressants, or both) in our cohort was frequent, 22%. This percentage was similar to that of the general population, around 25% (Torres et al., 1993). Therefore, the authors do not believe that current medications could account for the high prevalence of bone mineral disorders observed in this population.
Regarding FRAX prognostic score, this tool revealed a very low mean risk of major osteoporotic and/or hip fracture (3.2% and 1.2% respectively), and values remained low despite the high prevalence of other classic risk factors such as frequent falls or visual or hearing impairment in adults with DS. However, FRAX underestimates the fracture risk in populations with special characteristics. Given our low mean results and the fact that the FRAX tool has not yet been validated in this population, we conclude that the utility of this index seems low in adults with DS.
Interestingly, the prevalence of osteoporotic fractures was much lower than in the general population stratum with similar densitometry values. It is possible that the high number of overweight patients and patients with short stature could have protected against osteoporotic fractures. More studies of bone microarchitecture with CT or biopsy would be necessary in the future to better analyze the risk of fractures in this population.
To our knowledge, this is the first study to propose a BMD acquisition curve in young adults with DS. Although the shape of the curve is similar to that of the general population, the average BMD peak values were significantly lower, possibly because of a lower bone formation-to-resorption ratio and a lower bone turnover rate. Although in the general population females have lower values than males, we did not find any significant gender difference in the BMD development in the DS population.
These findings have already been demonstrated in murine DS models and may have future therapeutic implications for adults with DS (Fowler et al., 2012). It is possible that high prevalence of weight-related disorders (Real de Asúa, Parra, Costa, Moldenhauer, & Suarez, 2014) and lower muscle strength among young adults with DS may be the determining factor of the lower peak bone mass (Angelopoulou et al., 2000).
Our study has several strengths. We collected relevant clinical, metabolic, and anthropometric variables and performed a multivariate analysis to correct our results for possible confounders. Our inclusion criteria were broad, and the gender distribution was homogeneous. Although we recruited the largest sample of patients with DS to date, these results couldn't be extrapolated to all adults with DS.
Our study is also subject to a number of limitations. As already mentioned, t-score is not the best parameter to measure BMD in this population, and more if v-BMD is not calculated. Thus, we consider that Z-score is a more appropriate parameter to draw conclusions in our sample.
Given its retrospective nature and that radiologic studies were not performed systematically (in order to avoid unnecessary radiation), the number of vertebral fractures may also have been underestimated. Moreover, the cross-sectional nature of the research design did not enable the evaluation of incidental fractures, which would allow for a better understanding of the process of BMD loss in this population. Finally, DS patients attending our unit could have more medical needs, theoretically leading to a selection bias, although most of them did not have a specific health problem.
To conclude, adults with DS present an inappropriately high prevalence of osteopenia and osteoporosis for their age, and also a high prevalence of the predisposing factors for the development of osteoporotic fractures. Development of BMD in adolescence and young age seems similar to that observed in the general population, albeit with lower peak values. Further prospective studies are needed to assess the true incidence of osteoporotic fractures in this population and the real impact of our findings in the quality of life of patients with Down syndrome.
The authors would like to thank to Eugenio Escolano, DXA technician, for his invaluable support in DXA measurements, and Mr. T. O'Boyle for his kind review of the final draft of the manuscript.
Part of this study was presented as an oral communication at the 11th Congress of the European Federation of Internal Medicine, Madrid, Spain (October 24–27, 2012).