Bone mineral density, growth, pubertal development and other parameters in Brazilian children and young adults with sickle cell anaemia




To evaluate the occurrence of low bone mineral density (BMD) and its relationship with clinical and laboratorial characteristics in children and young adults with sickle cell anaemia living in Northeast-Brazil, and to assess the role of radiography in diagnosing low BMD.


Bone mineral density of lumbar spine was measured by dual energy X-ray absorptiometry (DXA) in 27 patients with Sickle cell anaemia (SCA) aged 7–28 years. Clinical history, calcium and calorie intake, laboratory measurements, anthropometrics and pubertal development were assessed, and X-rays were obtained. Z-scores and T-scores for weight, height, Body Mass Index (BMI) and BMD were calculated using age and gender matched reference data.


Mean lumbar spine BMD Z-scores and T-scores were −1.81 SD in boys and −0.80 SD in girls. BMD Z-scores were below -2 SD in 33.3% of girls and in 46.7% of boys. Low BMD (<−2 SD) occurred significantly more in patients with low height-for-age (P = 0.02), low weight-for-age (P = 0.001) and low BMI-for-age (P = 0.006). No significant relationships were found between BMD and other clinical and laboratory parameters. Radiography had a sensitivity of 75% and a specificity of 36% to detect low BMD, and was considered not useful in this context.


Patients with low height and/or low weight-for-age seem to be at high risk for developing low BMD.



Evaluer l'apparition d'une faible densité minérale osseuse (DMO) et son rapport avec les caractéristiques cliniques et de laboratoire chez les enfants et les jeunes adultes atteints de drépanocytose, vivant dans le nord-est du Brésil, et évaluer le rôle de la radiographie dans le diagnostic de la faible DMO.


La DMO de la colonne lombaire a été mesurée par absorptiométrie à rayons X en double énergie (DXA) chez 27 patients atteints de drépanocytose, âgés de 7 à 28 ans. L'histoire clinique, l'apport calorique et de calcium, les mesures de laboratoire, le développement anthropométrie et pubertaire ont été évalués, et les rayons X ont été obtenus. Les Z-scores et les T- scores pour le poids, la taille, l'indice de masse corporelle (IMC) et la DMO ont été calculés en utilisant des données de référence appariées pour l’âge et le sexe.


les Z-scores et T- scores moyens de la DMO lombaire étaient de −1.81 SD chez les garçons et −0.80 SD chez les filles. Les Z-scores de la DMO étaient inférieurs à -2 SD chez 33.3% des filles et 46.7% des garçons. Une DMO faible (<−2 SD) apparaissait significativement plus chez les patients avec une faible taille pour l’âge (P = 0.02), un faible poids pour l’âge (P = 0.001) et un faible IMC pour l’âge (P = 0.006). Aucune association significative n'a été trouvée entre la DMO et d'autres paramètres cliniques et de laboratoire. La radiographie avait une sensibilité de 75% et une spécificité de 36% pour la détection d'une faible DMO et n'a pas été jugée utile dans ce contexte.


Les patients avec une taille et/ou un poids pour l’âge faible semblent être à risque élevé de développer une faible DMO.



Evaluar la ocurrencia de una baja densidad mineral ósea (DMO) y su relación con las características clínicas y de laboratorio en niños y adultos jóvenes con anemia falciforme viviendo en el noreste de Brasil, y evaluar el papel de la radiografía en el diagnóstico de la baja DMO.


La DMO de la espina lumbar se midió mediante una absorciometría dual de rayos-X (ADX) en 27 pacientes con anemia falciforme (AF) y edades entre los 7–28 años. Se evaluaron la historia clínica, la ingesta de calcio y calorías, pruebas de laboratorio, antropométricas y el desarrollo puberal, y se tomaron los rayos X. Se calcularon los Z-scores t T-scores para el peso, estatura, Índice de Masa Corporal (IMC) y DMO utilizando la edad y el género pareados con datos de referencia.


La media de los Z-score y T-score para la DMO de la espina lumbar eran −1.81 DS en niños y −0.80 SD en niñas. Los Z-scores de DMO estaban −2 DS por debajo en un 33.3% de las niñas y en un 46.7% de los niños. Una DMO baja (<−2 SD) ocurría de forma significativa en pacientes con una baja talla-por-edad (= 0.02), bajo peso-por-edad (P = 0.001) y un bajo IMC-por-edad (P = 0.006). No había una relación significativa entre DMO y otros parámetros clínicos y de laboratorio. La radiografía tenía una sensibilidad del 75% y una especificidad del 36% para detectar una baja DMO, y en este contexto no se consideró útil.


Los pacientes con una baja estatura y/o bajo peso-por-edad parecerían tener un alto riesgo de desarrollar una baja DMO.


Sickle cell anaemia (SCA) is an autosomal recessive hereditary disease. Sickle-shaped red blood cells cause vascular occlusion, leading to tissue ischaemia and infarction, known as vascular occlusive crisis (VOC). Furthermore, SCA is characterised by a chronic haemolytic anaemia due to the shortened lifespan of the sickle-shaped cells (Bookchin & Lew 1996). Previous studies have shown that children with SCA have lower bone mineral density (BMD) than reference values of healthy children and healthy controls with identical age, length, pubertal development and fat free mass (Brinker et al. 1998; Soliman et al. 1998; Buison et al. 2005; Lal et al. 2006; Chapelon et al. 2009). Although poor growth status, altered body composition, delayed skeletal and sexual maturation and nutritional deficiencies have long been recognised in children with SCA, bone deficits cannot be explained completely by these factors (Platt et al. 1984; Reed et al. 1987; Barden et al. 2002). The exact mechanism causing low BMD in SCA patients is probably multifaceted. Complications in SCA such as VOC's, osteonecrosis and osteomyelitis have a direct effect on bone structures and areas of infarcted bone are high-risk sites for osteomyelitis (Keeley & Buchanan 1987; Resar et al. 1996; Smith 1996). Furthermore, repeated episodes of decreased blood flow and bone marrow hyperplasia may result in loss of bone trabeculae (Sadat-Ali et al. 1994). This may lead to stress fractures, orbital compression, dental problems, vertebral collapse and articular surface sclerosis (Emodi & Okoye 2001; Ganesh et al. 2001; Naran & Fontana 2001; Bahebeck et al. 2002; Demirbas et al. 2004). Approximately, half of the SCA patients will develop epiphyseal osteonecrosis by the age of 35 years (Styles & Vichinsky 1996).

Delayed puberty is known to have an effect on the accretion of peak BMD (Finkelstien et al. 1996). Furthermore, children with SCA are often malnourished, due to the chronic hyper-metabolic state they experience. Calcium and vitamin D are nutritional factors known to have an important role in bone formation (Borel et al. 1998). There is scant information about the relationship between BMD and growth, puberty, and other clinical and laboratory characteristics in SCA patients, with a lot of controversy in the literature (Soliman et al. 1998; Buison et al. 2005; Lal et al. 2006; Chapelon et al. 2009). Gaining more insights into the processes and influencing factors in bone formation and resorption in patients with SCA in different stages of life may help in finding interventions for preserving skeletal health. More information about the relationship between poor growth and maturation and decreased BMD may help in finding out when to start possible interventions. To our knowledge, there are no studies performed on BMD in SCA patients in Brazil. The main objective of the current study was therefore to evaluate BMD in children and young adults with SCA living in the Northeast region of Brazil and to assess the relation between low BMD and age, sex, pubertal development, growth, haemoglobin level, serum calcium, disease severity and calcium intake. Furthermore, we investigated if there is a role for simple radiography in diagnosing low BMD.


Study subjects

This uncontrolled comparative cross-sectional study was conducted between March and September 2011 in children and young adults attending the outpatient Paediatric Haematology service of the Federal University Hospital in Aracaju, Northeast-Brazil. This is a regional reference centre for treatment of SCA, which predominantly provides assistance to patients from low socio-economic stratum. A total of 27 patients with homozygous SCA, confirmed by haemoglobin electrophoresis, participated in the study. The participants (7–28 years) had no co-morbidities such as chronic osteo-muscular, congenital or other disease not associated with SCA. Participants did not take systemic corticosteroids for more than 7 days before the start of the study and had never used supplemental vitamin D, calcium, zinc or growth hormone analogues. Pregnant females were excluded since participants were exposed to radiation. Informed consent was obtained from the participants, and if younger than 18 years from the parent or guardian. The protocol of the study was approved by the Ethical Committee of Human Studies of the Federal University of Sergipe.

Lumbar spine DXA and simple radiography

Bone densitometry of the lumbar spine (L1–L4) was assessed in a private radiology clinic, using a DXA scan (GE Lunar DPX NT). DXA has shown to be the best investigation method to assess BMD in adults and children because of the precision and low radiation exposure. The results were expressed as Z-scores for patients to 20 years and T-scores for patients of 20 years and older, by comparison with healthy reference values matched for age, ethnicity and gender (Khan et al. 2004). Z-scores and T-scores below -2 standard deviations (SD) were classified as low BMD, and Z-scores and T-scores from -2 SD and above were classified as normal BMD. Radiography was obtained in 19 patients. X-rays of each patient were taken of the femur, dorsal and lumbar spine. An experienced radiologist (T.O.F.), who was blinded to any clinical information, evaluated the X-rays as normal bone density or osteopenia. For data analysis, results were considered as osteopenia if at least 2 of the 3 X-rays were evaluated as osteopenia.

Growth status and maturation

Height was assessed using a stadiometer (Tonelli 120A, Brazil) and weight was assessed using a mechanic scale (Welmy, Mechanical Platform Scale Ref. 104 A, Brazil). Body Mass Index (BMI) was calculated in kg/m2. Height, weight and BMI were compared with the Centres for Disease Control and Prevention (CDC) 2000 reference growth standards, age and gender-specific Z-scores were calculated (Kuczmarski et al. 2002). Previous research has shown that reference growth charts of the CDC can be used for Brazilian children (Silva et al. 2010). Pubertal status was assessed by physical examination in patients to 18 years old, a self-assessment pictorial questionnaire with descriptions of the five developmental stages as described by Tanner was used in older patients (Marshall & Tanner 1969, 1970). Patients were classified as pre-pubertal when in Tanner stage 1 and pubertal when in Tanner stages 2–5. Pubertal development was considered normal if both of the measured values were found normal; P (pubic hair) and G (genital development) in boys, P (pubic hair) and M (breast development) in girls. Pubertal development was considered delayed if at least one of the two values was found delayed. For all data-analysis on pubertal development patients older than 20 years were excluded, since they already reached adulthood. Children younger than 10 years were excluded from normal or delayed pubertal development analysis, since it is normal at this age if pubertal development is still absent.

Haematological and dietary measurements

Blood samples were obtained in 25 patients to measure complete blood count and serum calcium. Haemoglobin level was calculated from the mean value of the last three measurements prior to enrolment. Standard methods were used for determination of the measurements (Clinical Laboratory of the University Hospital, Sergipe). Dietary intake information was obtained from 23 of the 27 patients through two 24-h recall standardised questionnaires, with portion booklets to estimate serving size. Nutrient analysis was performed by standardised methods to determine total calorie and total calcium intake. Calorie intake was compared with the estimated energy requirement for low active children (Food & Nutrition Board, Institute of Medicine. Energy 2002). Calcium intake was compared with adequate dietary reference intakes (Ross et al. 2011). Clinical information was obtained through a structured questionnaire and review of the medical records. A modified El-Hazmi score was used to classify the clinical disease severity (El Hazmi 1992). An El-Hazmi score ≤2 was considered mild; 3–5 as moderate and ≥6 as severe SCA.


Quantitative data were described as means and SD. Data were analysed by descriptive statistics with the use of comparative tests. Chi-square and Fisher exact tests were used to assess differences in categorical variables, and independent student's t-tests were used to analyse differences in continuous variables. Differences were considered significant if P-values were < 0.05. All analyses were performed using SPSS 19.0 (SPSS Inc, Chicago, IL, USA).


Study characteristics

Twenty-seven Brazilian children and young adults with SCA (15 males and 12 females) were evaluated, with a mean age of 14.9 years (range 7–28 years) (Table 1). Sickle cell anaemia was associated with poor growth; mean height-for-age, weight-for-age and BMI-for-age values where all below normal values in males and females. There were six pre-pubertal and 17 pubertal patients; the remaining four patients were adults (>20 years). Twelve of the 19 patients between 10 and 20 years had delayed pubertal development. Twenty-five patients had a moderate to severe disease severity (nine severe and 16 moderate) and two patients had a mild disease severity. All patients had a calcium intake below recommended dietary allowance (RDA). All patients except for one had a total calorie intake below Estimated Energy Requirement (EER). Mean values were normal for serum calcium in all patients. Haematological measures reflected SCA diagnosis, with reduced haemoglobin in all patients (mean value of total cohort 8.06 g/dl) and reduced haematocrit concentrations (mean value of total cohort 0.22 L/L). No significant gender differences were observed for age, height, weight, pubertal developmental stage, disease severity, laboratory measurements, calorie and calcium intake or radiography results. Clinical history revealed that one patient had avascular necrosis of the hip and one patient had experienced osteomyelitis.

Table 1. Characteristics of the total sickle cell anaemia cohort by gender
 Male N = 15Female N = 12P-value
  1. EER, estimated energy requirement in kcal/day; RDA, recommended dietary allowance.

Age (years, mean ± SD) 14.7 ± 4.615.2 ± 6.20.81
Height (cm, mean ± SD)152 ± 15146 ± 170.34
Height for age Z-score (mean ± SD)−1.14 ± 1.41−0.97 ± 0.920.71
Weight (kg, mean ± SD)40.6 ± 11.138.1 ± 13.70.61
Weight for age Z-score (mean ± SD)−1.10 ± 1.07 −1.10 ± 0.450.99
BMI (kg/m2, mean ± SD) 17.3 ± 1.8 17.4 ± 3.00.91
BMI for age Z-score (mean ± SD)−0.87 ± 0.69 −0.80 ± 0.560.79
Pre-pubertal N (%)2/13 (15)4/10 (40)0.20
Delayed pubertal development N (%)8/11 (73)4/8 (50)0.30
Disease severity: Severe N (%)5 (33)4 (33)1.00
BMD Z-score (g cm2, mean ± SD)−1.81 ± 1.81−0.80 ± 2.350.22
Radiography: Osteopenia N (%)3/9 (33)3/10 (30)0.88
Haemoglobin (g/dl, mean ± SD)8.15 ± 1.607.95 ± 1.190.71
Serum calcium (mg/dL, mean ± SD)9.07 ± 0.329.33 ± 0.350.07
Calorie intake (kcal/day, mean ± SD)1731 ± 4581545 ± 3300.30
Adequacy calorie intake (% of EER, mean ± SD)76 ± 2178 ± 150.85
Calcium intake (mg/day, mean ± SD)467 ± 252380 ± 2070.36
Adequacy calcium intake (% of RDA, mean ± SD)39 ± 2335 ± 210.65

Bone mineral density

Low BMD was present in 11 patients (41%) of the total cohort; this was in four of the females (33%) and seven of the males (47%). Mean BMD Z-score was −1.81 SD in males and −0.80 SD in females (Table 1). The observed differences in gender and age for BMD were not statistically significant. The mean height-for-age Z-scores, weight-for-age Z-scores and BMI-for-age Z-scores were significantly lower in the low BMD group than in the normal BMD group (Table 2). Boxplots of mean height-, weight- and BMI-for-age are shown in Figure 1. Although more patients in the low BMD group had delayed pubertal development than in the normal BMD group (67% vs. 33%), this was not statistically significant. Furthermore, no association was found between pre-pubertal or pubertal developmental status and BMD (Table 2 and Figure 2). There were also no associations found between BMD and disease severity, number of VOC's, mean haemoglobin level, serum calcium concentration and calorie or calcium intake (Table 2). There was no significant relationship found between BMD obtained with densitometry and bone density obtained with radiography. Mean BMD in patients diagnosed with osteopenia on the radiographs was higher than mean BMD in patients diagnosed with normal bone density on the radiographs (1.20 SD and −1.67 SD respectively), this was not statistically significant. Calculated sensitivity and specificity for radiography (with bone densitometry considered as the gold standard) were 75% and 36% respectively.

Table 2. Bone Mineral Density divided into low (<−2SD) and normal (≥−2 SD) in relation to possible influencing factors
 Low BMD N = 11Normal BMD N = 16P-value
  1. EER, estimated energy requirement in kcal/day; RDA, recommended Dietary Allowance.

Male subjects N (%)7 (64)8 (50)0.48
Age (years, mean ± SD)16.1 ± 3.0 14.1 ± 6.3 0.28
Disease severity: Severe N (%)3 (27)6 (38)0.58
No. of VOC's in last year (mean ± SD)0.82 ± 1.082.19 ± 3.080.12
Haemoglobin (g/dl, mean ± SD)

7.92 ± 1.06

N = 11

8.16 ± 1.63

N = 16

Serum calcium (mg/dl, mean ± SD)

9.20 ± 0.51

N = 11

9.18 ± 0.16

N = 13

Calorie intake (kcal/day, mean ± SD)1532 ± 3981710 ± 3970.32
Adequacy calorie intake (% of EER, mean ± SD)72 ± 1881 ± 160.25
Calcium intake (mg/day, mean ± SD)

381 ± 222

N = 10

454 ± 239

N = 14

Adequacy calcium intake (% of RDA, mean ± SD)34 ± 2039 ± 230.56
Radiography: Osteopenia N (%)6/8 (75)7/11 (64)0.60
Figure 1.

Weight-, height- and BMI for age values (Z-scores) in the low and normal BMD groups. All shown differences were significant. Each boxplot shows the median, inter-quartile range (box length) and extreme values within a category. Circles: Cases with values between 1.5 and 3 box lengths from upper or lower edge of the box.

Figure 2.

BMD values (Z-scores) in the different pubertal development groups. None of the shown differences was significant. Each boxplot shows the median, inter-quartile range (box length) and extreme values within a category. Circles: Cases with values between 1.5 and 3 box lengths from upper or lower edge of the box.


This was the first study on BMD in patients with SCA living in the Northeast of Brazil. Our results indicate a high prevalence of low BMD of the lumbar spine in children and young adults with SCA. We found that low BMD is more severe in patients with deficiencies in height-for-age, weight-for-age and BMI-for-age. The occurrence of low BMD in our patients was not explained by differences in age, gender, pubertal development, current calcium or calorie intake according to questionnaires, disease severity, number of VOC's or laboratory measurements.

In this study, the overall prevalence of low BMD in a group of general young Brazilian SCA patients was 41%, which is rather high. Prevalence of low lumbar spine BMD (Z-score <−2 SD) varies from about 19% in a French population similar to ours, to 56% in a United States population of SCA children with severe manifestations, which also might be an explanation for the relative high prevalence in this latter study (Lal et al. 2006; Chapelon et al. 2009). Differences in the observed percentages of low BMD furthermore may be explained by the limited sizes of the investigated populations and by the variety in study locations, ranging from the United States, to Europe and South-America. The increased incidence of low BMD in SCA patients may result in an increased risk on bone fractures and osteoporosis in later life. Few case reports on bone fractures in SCA patients are mentioned, such as vertebral collapse, stress fractures and orbital compression (Emodi & Okoye 2001; Ganesh et al. 2001; Naran & Fontana 2001; Bahebeck et al. 2002; Demirbas et al. 2004). So far, there is no available literature on follow-up and bone fractures in SCA patients with low BMD.

We found no relationship between BMD and gender in children and young adults with SCA, which is consistent with two other studies (Buison et al. 2005; Lal et al. 2006). Another study reported significantly lower BMD in pre-pubertal girls compared to pre-pubertal boys, but not in their pubertal group (Chapelon et al. 2009). In young adults, significantly lower BMD in males compared to females was reported (Miller et al. 2006). Furthermore, we found no relationship between BMD and age. A previous study found low BMD more frequently in older boys, without significant age differences in girls (Buison et al. 2005). Hence, there are no unequivocal indications for the existence of a certain age and gender group being at high risk for developing low BMD.

The observed significant relationships between low BMD and low height-for-age, weight-for-age and BMI-for-age we found are consistent with three other studies (Soliman et al. 1998; Buison et al. 2005; Sarrai et al. 2006). This makes it very likely that SCA patients with delayed growth parameters are at high risk to develop low BMD. Two studies of SCA children did not find significant relationships between BMD and anthropometric parameters, which is a minority (Lal et al. 2006; Chapelon et al. 2009). One study of Brazilian adults found no relationship between BMD and BMI (Baldanzi et al. 2011). Unfortunately, we did not have sufficient power to detect a relationship between delayed puberty and low BMD. This is probably the consequence of the limited sample size of our population in combination with the large distribution of the five Tanner stages. Nevertheless, the significant relationships between anthropometric data and low BMD we found seem to suggest that low BMD is associated with delayed growth and maturation, including pubertal development.

We found no significant relationship between low BMD and calcium or calorie intake, according to the recall questionnaires. However, calcium and calorie intake in our study were insufficient among all study participants, particularly in the knowledge that SCA patients are in relative need for a high nutritional intake since they experience a chronic hyper-metabolic state (Borel et al. 1998). It is therefore expected that low BMD is partly a result of insufficient nutritional intake. Similar observations have been made in earlier studies. For example, mean vitamin D and calcium intake of children and young adults with SCA were below recommended levels, but did not show associations with BMD (Buison et al. 2005; Lal et al. 2006; Chapelon et al. 2009). Furthermore, we found no relationships between BMD and disease severity and serum calcium- and haemoglobin concentrations. Previous studies also did not find associations between low BMD and disease severity, serum calcium or haemoglobin concentrations (Soliman et al. 1998; Buison et al. 2005; Lal et al. 2006). In contrast, a study among 65 Brazilian adults found significant relationships between BMD and haemoglobin concentration and degree of haemolysis (Baldanzi et al. 2011).

The skeletal site assessed with DXA was the lumbar spine, which is the best skeletal site to assess BMD since it has a very large precision and few radiation exposure. We found no role for radiography in diagnosing low BMD. Previous research has shown that osteopenia is visible on X-rays with a reduction in BMD of at least 30%, with a wide inter-doctor variation of radiography findings (Garton et al. 1994). Our results confirmed a low specificity for radiography in diagnosing low BMD. It is, however, important to mention that densitometry should be interpreted with care, as alterations in growth, puberty and body composition occur in children. It has been suggested to compare BMD with children who are matched for height, Tanner stage, bone age, weight, lean body mass or chronological age (Khan et al. 2004). Unfortunately, we were not able to include healthy controls. So far, one study on bone mineral content in SCA children included healthy controls matched for age, height, weight and Tanner stage and found significantly lower bone mineral content in the SCA population (Buison et al. 2005).

A first limitation of our study was the small number of investigated patients. This was mostly due to the fact that patients had to travel long distances to visit the hospital and were therefore not willing to participate. A second limitation was that growth parameters were obtained at one moment and unfortunately no growth curves or bone age values were available. A third limitation was the lack of information on serum 25-hydroxyvitamin D, bone markers, the age of menarche and onset time of puberty. It is known that SCA patients experience more vitamin D deficiency than their healthy counterparts. It is therefore possible that vitamin D could have influenced our study results as it is an important factor of bone formation. Nevertheless, previous studies on BMD in SCA patients that included measurement of 25-hydroxyvitamin D found decreased vitamin D in almost all study participants, but no significant associations between vitamin D and BMD (Buison et al. 2005; Lal et al. 2006; Miller et al. 2006; Chapelon et al. 2009; Arlet et al. 2013). Four patients were older than twenty years and reached adult pubertal stage, and no information on their experienced puberty was recorded. A fourth possible limitation is the El-Hazmi severity score we used (El Hazmi 1992). This score was developed in Saudi Arabian SCA patients and has not yet been validated for Brazilian SCA patients. Nevertheless, we have used this score previously satisfactorily (Cipolotti et al. 2001). Furthermore, we investigated if there were other indications for the existence of a relationship between BMD and disease severity. We analysed if there was an association between BMD and number of VOC's separately, which was not the case. Finally, there was no nutritional intake information about the past, where the duration of malnutrition may be an important factor in bone formation. However, there was no reason to think that nutritional intake proportions had changed, since there were no experienced changes in social class. In addition, outcomes of these nutritional questionnaires were markedly decreased, especially in a SCA population in which patients experience a chronic hyper-metabolic state.


This study indicates a high prevalence of low BMD in children and young adults with SCA living in the Northeast of Brazil. Healthcare professionals should be aware of the fact that SCA patients with low height and/or low weight for age may be at high risk of developing low BMD. More research with long-term follow-up and healthy controls matched for gender and age is needed to determine possible implications of low BMD in SCA patients, such as fracture risk and osteoporosis in adulthood. Furthermore, the role of possible interventions for the prevention of low BMD, such as extra calorie intake, calcium and vitamin D supplementation to optimise peak bone mass should be investigated in randomised control trials.