Osteoporosis International

, Volume 16, Issue 9, pp 1024–1035

Generalized low bone mass of girls with adolescent idiopathic scoliosis is related to inadequate calcium intake and weight bearing physical activity in peripubertal period

Authors

  • Warren T. K. Lee
    • Department of Orthopaedics & Traumatology, 5/F, Clinical Sciences BuildingPrince of Wales Hospital
    • Jockey Club Centre for Osteoporosis Care and ControlThe Chinese University of Hong Kong
  • Catherine S. K. Cheung
    • Department of Orthopaedics & Traumatology, 5/F, Clinical Sciences BuildingPrince of Wales Hospital
  • Yee Kit Tse
    • Centre for Epidemiology and Biostatistics, Faculty of MedicineThe Chinese University of Hong Kong
  • Xia Guo
    • Department of Rehabilitation SciencesHong Kong Polytechnic University
  • Ling Qin
    • Department of Orthopaedics & Traumatology, 5/F, Clinical Sciences BuildingPrince of Wales Hospital
    • Jockey Club Centre for Osteoporosis Care and ControlThe Chinese University of Hong Kong
  • Suzanne C. Ho
    • Department of Community and Family Medicine, Faculty of MedicineThe Chinese University of Hong Kong
    • Centre of Research and Promotion of Women’s Health, School of Public HealthThe Chinese University of Hong Kong
  • Joseph Lau
    • Centre for Epidemiology and Biostatistics, Faculty of MedicineThe Chinese University of Hong Kong
    • Department of Orthopaedics & Traumatology, 5/F, Clinical Sciences BuildingPrince of Wales Hospital
    • Jockey Club Centre for Osteoporosis Care and ControlThe Chinese University of Hong Kong
Original Article

DOI: 10.1007/s00198-004-1792-1

Cite this article as:
Lee, W.T.K., Cheung, C.S.K., Tse, Y.K. et al. Osteoporos Int (2005) 16: 1024. doi:10.1007/s00198-004-1792-1

Abstract

Generalized low bone mass has been well documented in patients with adolescent idiopathic scoliosis (AIS). However, studies linking calcium-intake (CA), weight-bearing physical-activity (PA) and bone mass of AIS are lacking. We aimed to study the relationship between CA, PA and bone mass in AIS girls and compared to those of healthy non-AIS controls during the peripubertal period. Newly diagnosed AIS girls (n=596) aged 11–16 years with Cobb angle ≥10° were recruited to compare with age-matched healthy girls (n=302) in a cross-sectional study. Anthropometric parameters, pubertal status, CA and PA were assessed. Areal bone mass of lumbar spine and femoral neck, and volumetric bone mass of distal radius and tibia were determined by dual-energy X-ray absorptiometry and peripheral quantitative computed tomography, respectively. The results showed that weight and body mass index (BMI) of AIS were lower than the controls (P<0.05). Corrected height and arm span of AIS were longer than those of controls from 13 years onwards (P<0.02). Median CA of AIS was <410 mg/day across the ages and did not differ from the controls (P=0.063). Median PA of AIS (1.6 h/day) was lower than the controls (1.8 h/day) (P=0.025). Bone mass of AIS was on average 6.5% lower than controls across the ages (P<0.05). CA and PA were significantly correlated with bone mass of AIS (P<0.04). Multivariate analysis showed that AIS in girls was associated with lower bone mass, and that both CA and PA were independent predictors of bone mass in AIS. In conclusion, AIS girls were found to have lower body weight and BMI, longer segmental lengths and generalized low bone mass. Inadequate calcium intake and weight-bearing physical activity were significantly associated with low bone mass in AIS girls during the peripubertal period. The importance of preventing generalized osteopenia in the control of AIS progression during the peribubertal period warrants further study.

Keywords

Adolescent idiopathic scoliosisBone mineral contentBone mineral densityCalcium intakeGirlsPhysical activity

Introduction

Adolescent idiopathic scoliosis (AIS) is a complex three-dimensional deformity of the spine occurring mostly in girls between 10 and16 years old [1,2]. In Hong Kong, the prevalence rate of AIS was approximately 4.5% among the school children population (unpublished report, Student Health Service, Department of Health, 2003). The etiology of AIS is still largely unknown [3,4]. For rapidly progressive curves, complicated surgical correction with instrumentation and permanent spinal fusion would be required. Severe scoliotic deformity affecting the thoracic region may deteriorate if untreated, progressively leading to impairment of cardiopulmonary function, and increased mortality rate [3]. Generalised low bone mass and osteopenia in both axial and peripheral skeleton in AIS have been reported in the literature [5–7] and from our group [8–10]. Abnormal histomorphometric bone cell activity has been found in AIS bone biopsies [11]. We have also shown that low bone mass in AIS patients is likely to persist through to adulthood [12]. There is a growing concern that adolescents with idiopathic scoliosis may have a lower peak bone mass, thereby increasing the risk of developing osteoporosis and related complications in later life [10,12].

Adequate calcium intake and weight bearing physical activity have been proven to be important lifestyle factors in determining bone mineral acquisition in childhood and adolescence in interventional studies [13–21]. Whether calcium intake and physical activity may contribute to generalised low bone mass in AIS has not been adequately investigated. There has been limited research linking calcium intake and physical activity to bone mass in AIS [5,7]. The objective of the current study was to explore the relationship between the two lifestyle factors (calcium intake and physical activity) and bone mass in adolescent girls with AIS and to compare them with those of healthy girls of similar age.

Materials and methods

Newly diagnosed AIS patients aged between 11 and 16 years old were enrolled at our scoliosis clinic. Diagnosis of AIS was confirmed by clinical and standardized X-ray examination. Scoliosis was diagnosed when the Cobb angle of the spine was ≥10° [1,2]. Patients receiving any forms of prior treatment for scoliosis were excluded from the study. Healthy girls of similar age range were recruited from local schools to serve as controls. All normal controls were also clinically examined to rule out any hidden scoliosis before entering the study. Subjects with a history of congenital deformities, neuromuscular disease, endocrine disease, skeletal dysplasia, connective tissue abnormalities or mental retardation were excluded from the study. Informed consent was obtained from the parent. The study was approved by the Clinical Research Ethics Committee of the university and the hospital.

Anthropometry and puberty

The weight of subjects in light clothes was measured using a digital weighing scale (Soehnle, Germany). Standing height without shoes was measured using a wall-mount stadiometer (Technique Services Unit, The Chinese University of Hong Kong). For scoliotic patients, corrected height was derived from Bjure’s formula (Log y=0.011x−0.177, where y is the loss of trunk height (cm) due to the deformed spine and x is the greatest Cobb angle of the primary curve) [18]. Body mass index (BMI) was determined by dividing weight (kg) by the square of the uncorrected height (m2). Pubertal status (breast development) of the subjects was physically examined in accord with Tanner’s staging [23].

Calcium intake and physical activity

Habitual dietary intakes including calcium, protein and energy were assessed by a food frequency questionnaire (FFQ) [23]. Portion size was quantified by the subject using household measures (spoons, tablespoons, glasses, and bowls, etc). Photos of common food items were collected in an album to assist description of portion size. Dietary intakes were determined using the Foodworks nutrient analysis software (ver 2.1, Xyris Software, Australia Pty Ltd, Highgate Hill, QLD, Australia). Nutrient compositions of common local foods were also included in the nutrient database. The FFQ used in the present study was validated against a 6-day dietary record in a group of healthy school children (n=67) aged 13–17 years [25]. There was good agreement in calcium intake between the two methods (r=0.725, P<0.05). To test the short-term reproducibility of the FFQ in assessing calcium intake, a random sample of 30 subjects were asked to fill in the FFQ and the FFQ was repeated in 3 months’ time. There was no significant different in mean±SD calcium intake at the two time points assessed 3 months apart (603±264 mg versus 507±229 mg/day, P=0.15, paired t-test). The results showed that the FFQ was valid and reproducible to assess dietary intakes of calcium in Chinese children and adolescents.

Habitual weight-bearing physical activity was assessed by a standardized quantitative physical activity questionnaire adapted from Slemenda et al. [26]. Types of physical activity were classified into either weight bearing or non-weight bearing. Weight-bearing activities were further categorised into either high or moderate impact types, in order to assess the impact of weight bearing activity on BMD. Reproducibility of the method was assessed by repeated activity assessments at 6-month intervals in 118 children and adolescents. There was good within-person agreement for weight bearing total activity (r=0.66, P<0.05) and for most individual activities (r=0.4–0.77, P<0.05). In addition, correlations between mothers’ and children’s reports of hours spent on total activities (r=0.66, P<0.05) and of hours spent on watching television (r=0.6, P<0.05) were also strong [26]. Hence, the questionnaire is reliable in assessing quantitatively the amount of physical activity among children and adolescents. Table 1 shows some examples of high and medium impact types of weight bearing activities. Each type of weight bearing activity is expressed as number of hours per day. The sum of hours spent on various types of weight bearing activity was used to summarize daily activities for each subject.
Table 1

Examples of physical activities: weight bearing activities of high and medium impact types, and non-weight bearing activities

Classification

Examples of activities

Weight bearing activities

Overall (high and medium impact)

Warm-up exercise, walking, fast walking, gymnastics, ten-pin bowling, skate-bike, tai-chi; jogging, running, soccer, basketball, badminton, dancing, tennis, squash, kung-fu, weight lifting, rope-skipping, jumping, track and field events, etc.

High impact

Jogging, running, soccer, basketball, badminton, dancing, tennis, squash, kung-fu, weight lifting, rope-skipping, jumping, track and field events, etc.

Medium impact

Warm-up exercise, walking, fast walking, gymnastics, ten-pin bowling, skate-bike and tai-chi, etc.

Non weight bearing activities

Swimming, rowing, cycling and diving, etc.

Measurement of bone mineral content (BMC) and bone mineral density (BMD)

Dual energy X-ray absorptiometry

Femoral neck BMC (FNBMC) and BMD (FNBMD), greater trochanter BMC (GTBMC) and BMD (GTBMD), and Ward’s triangle BMD (WTBMD) of the non-dominant proximal femur, and lumbar spinal BMC (LSBMC) in antero-posterior position were evaluated by dual-energy X-ray absorptiometry (DXA) (XR-36; Norland Corp., Fort Atkinson, Wisc., USA). The scoliotic curvature of the spine in AIS patients may present difficulties in measuring the spinal BMD reliably; to minimize this problem, the spine was pre-scanned once, a reference line was drawn to join the highest points of the iliac crests, which usually passes between the third and fourth lumbar spinal processes. On that reference line, a rectangle was erected to include L2–L4, and this was defined as the scan area [10]. Furthermore, a previous study showed that the projected spinal bone area varies with the degree of rotational deformity of the scoliotic spine. This will result in underestimation of the lumbar spinal BMD [19]. Hence, lumbar spinal BMC adjusted for projected spinal bone area will be presented in the present study. All measurements were performed on the non-dominant side.

The principles of DXA, accuracy and precision of the technique have been reported previously [10–12]. The radiation that the subject actually absorbed was low (<30 μSv per scan) [27]. Quality assurance was performed by daily calibration against the standard phantoms provided by the manufacturer. The in-vivo precision of using DXA in measuring axial and appendicular BMC and BMD of the subjects was 1.1–3.7%.

Peripheral quantitative computed tomography (pQCT)

Non-dominant distal radius and distal tibia were measured with a high precision multislice peripheral quantitative computed tomography (pQCT) (Densiscan 2000; Scanco Medical, Bassersdorf, Switzerland) with an effective X-ray energy of 40 keV and a local radiation dose of <50 μSv for a four-slice scanning program. A standard phantom measurement was performed daily which was reported to have a long-term reproducibility of 0.3% [10,28]. The forearm or lower leg was positioned in a radiolucent splint anatomically fitted for the subject during the scanning. The average volumetric BMD (vBMD) of the trabecular bone in a core volume (central 50% of the total bone area) of the distal radius (RtBMD) and distal tibia (TtBMD) and integral vBMD of both the cortical and trabecular bone within the total bone volume of the distal radius (RiBMD) and distal tibia (TiBMD) were evaluated. Cross-sectional areas of the scanned radius (RiCSA) and tibia (RiCSA) were also obtained from the average of the four-slice program. The integral trabecular and cortical BMC of the distal radius (RiBMD) and the distal tibia (TiBMC) was determined by multiplying RiBMD and RiCSA, and TiBMD and TiCSA, respectively. Detailed descriptions on technical operation of the pQCT device can be found elsewhere [29,30]. In our densitometry laboratory, the short-term precisions of using pQCT to evaluate vBMD at the distal radius and distal tibia have been performed in 11 AIS girls aged 12–16 years. The coefficients of variation (CV) of repeated measurements at the distal radius (RtBMD, RiBMD and RiCSA) were 1.65%, 1.52% and 2.94%, respectively, while the CVs for the distal tibia (TtBMD, TiBMD and TiCSA) were 0.99%, 0.82% and 1.41%, respectively. An accredited technician with ISCD (International Society of Clinical Bone Densitometry) certification was responsible for using DXA and pQCT devices to measure bone mass at our clinical densitometry laboratory throughout the study.

Statistical analysis

Normality of the data was tested by one-sample Kolmogorov–Smirnov test. Group comparisons were tested by two-tailed Student’s t-test with normally distributed data; for skewed data, non-parametric Mann–Whitney U-test was applied. Data were summarized as either mean±SD or median and inter-quartile range (IQR). Some variables were not normally distributed and were therefore natural log transformed before Pearson’s correlation analysis. BMD was adjusted for age, and BMC was adjusted for age and bone area to correlate with calcium intake and physical activity by using Pearson or Spearman correlation. Multiple regression analysis was used to identify independent variables that significantly predicted the variation of BMD at various sites of measurements after controlling for potential confounding factors. Level of significance was set at P≤0.05. SPSS Version 11 (SPSS, Chicago, Ill., USA) was used for statistical analysis.

Results

Subject characteristics

A total of 898 adolescent girls participated in the present study (596 AIS patients and 302 controls). The age ranges for the AIS and control groups were 10.5–16.3 years and 10.7–16.3 years, respectively. There was no significant difference in mean±SD age of AIS girls and controls (13.3±1.2 years versus 13.3±1.2 years, P=0.981). When subjects were sub-divided into four age groups for data analysis, there was also no significant difference in age between AIS and controls within these four sub-groups (Table 2). Mean±SD Cobb angle and 95% confidence interval of the mean Cobb angle of the scoliotic girls were 26.3±7.9 degrees (95% CI: 25.6–26.9 degrees).
Table 2

Age, anthropometric parameters, physical activity, calcium and nutrient intakes of the AIS and control groups by age (mean±SD)

Age groups (years)

≤12

13

14

≥15

AIS (n=163)

Control (n=83)

P-value

AIS (n=153)

Control (n=100)

P-value

AIS (n=171)

Control (n=60)

P-value

AIS (n=109)

Control (n=59)

P-value

Age (years)

11.7±0.5

11.8±0.5

0.237

13.0±0.3

13.0±0.3

0.543

14.0±0.3

14.0±0.3

0.542

15.0±0.4

15.1±0.5

0.081

Weighta (kg)

37.8 (32.9–42.8)

39.4 (35.4–45.8)

0.012

41.0 (37.8–45.6)

43.1 (38.3–49.3)

0.044

42.4 (39.3–46.7)

44.2 (40.8–48.5)

0.118

44.5 (40.2–48.9)

45.3 (41.6–52.3)

0.137

BMIa

16.3 (15.2–17.8)

17.3 (15.7–19.4)

0.002

17.1 (15.8–18.5)

18.0 (16.2–19.7)

0.006

17.4 (15.9–18.9)

17.8 (16.6–20.3)

0.049

17.3 (16.1–18.8)

18.7 (16.9–20.8)

0.002

Uncorrected heightb (cm)

150.6±7.4

151.3±6.8

0.477

155.4±5.9

154.8±5.5

0.395

157.5±5.5

156.4±4.5

0.198

160.1±5.8

158.0±5.3

0.025

Corrected heightc (cm)

152.0±7.5

151.3±6.8

0.466

156.7±5.9

154.8±5.5

0.011

158.9±5.5

156.4±4.5

0.002

161.4±5.8

158.0±5.3

<0.001

Calciuma (mg/day)

368 (230–558)

269 (212–418)

0.021

349 (235–525)

321 (223–484)

0.202

340 (208–467)

355 (219–717)

0.168

407 (241–621)

309 (234–577)

0.242

Proteina (g/day)

59.7 (45.1–82.8)

54.7 (43.0–75.6)

0.375

67.1 (52.3–94.3)

57.2 (40.1–88.7)

0.173

63.3 (48.0–88.6)

65.7 (43.9–102.1)

0.962

68.5 (50.4–100.5)

79.5 (57.4–92.8)

0.992

Energya (kcal/day)

1426 (1176–1792)

1291 (1115–1729)

0.357

1637 (1241–2076)

1357 (1108–2044)

0.274

1488 (1175–1870)

1689 (1129–2285)

0.424

1686 (1217–2098)

1929 (1412–2305)

0.275

Weight bearing physical activity (h/day)

Overall

1.6 (1.0–2.4)

2.0 (1.3–2.8)

0.007

1.9 (1.1–2.7)

1.8 (1.2–2.9)

0.624

1.5 (0.9–2.7)

1.7 (0.9–3.4)

0.711

1.6 (1.1–2.5)

1.8 (1.2–2.5)

0.601

High impact

0.4 (0.2–0.7)

0.6 (0.4–0.8)

0.003

0.5 (0.3–0.8)

0.6 (0.4–0.9)

0.681

0.5 (0.3–0.7)

0.6 (0.3–0.9)

0.329

0.5 (0.2–0.8)

0.6 (0.3–0.9)

0.509

Medium impact

1.0 (0.6–1.7)

1.2 (0.7–2.0)

0.07

1.1 (0.6–2.0)

1.0 (0.6–2.1)

0.751

1.0 (0.5–2.0)

1.0 (0.5–2.1)

0.840

1.0 (0.6–1.8)

1.1 (0.6–1.9)

0.770

aMann-Whitney U-test was used to test the difference in BW. Medians and inter-quartile ranges were presented

bComparison between apparent height of the AIS and controls

cComparison between height of the controls and AIS with correction for trunk loss

Anthropometry

Weight and BMI, uncorrected and corrected height of AIS and controls by age are depicted in Table 2. Weight of AIS at ≤12 years (P=0.012) and 13 years (P=0.044) was significantly lower than those of controls. BMI of AIS was significantly lower than that of controls across the ages (P=0.049–0.002). Uncorrected height of AIS was almost similar to that of the controls, except at age ≥15 years, at which time AIS girls were significantly taller than the controls (P=0.025). From age 13 years onwards, the corrected height of the AIS girls was significantly taller than that of controls (P=0.011–0.001).

Table 3 summarizes anthropometric parameters by pubertal staging. Similarly, significantly lower body weight was also found in AIS at breast stage I (P=0.015), stage II (P=0.005) and stage IV (P<0.001) when compared with the controls. BMI of AIS girls was significantly lower than that of controls from stages II–IV (P=0.038–0.001). Uncorrected height of AIS girls was not significantly different from that of controls except at breast stage I (P=0.001), whereas corrected height of AIS girls was significantly taller than that of the controls across the ages (P=0.007–0.001).
Table 3

Anthropometric parameters of the AIS and control groups by Tanner staging (breast stage) (mean±SD)

Tanner staging

I

II

III

IV

V

AIS

Control

P-value

AIS

Control

P-value

AIS

Control

P-value

AIS

Control

P-value

AIS

Control

P-value

Weighta (kg)

28.9 (26.5–31.6)

31.4 (30.0–38.3)

0.015

37.6 (34.1–41.1)

40.3 (36.3–43.7)

0.005

41.7 (38.3–46.5)

42.6 (39.3–47.6)

0.108

44.0 (40.1–47.8)

47.5 (43.4–51.0)

<0.001

47.1 (42.4–50.8)

47.9 (42.5–56.4)

0.629

BMIa

14.3 (13.2–16.1)

14.5 (13.4–16.6)

0.566

15.7 (14.8–16.8)

17.0 (15.4–18.5)

<0.001

17.2 (16.0–18.7)

17.8 (16.2–20.2)

0.038

17.7 (16.4–19.2)

19.5 (17.9–20.9)

<0.001

18.6 (16.9–19.6)

19.0 (17.1–22.5)

0.175

Uncorrected heightb (cm)

142.2±6.5

149.8± 6.2

0.001

153.6±6.9

153.0±6.9

0.538

155.8±6.7

155.3±5.2

0.464

157.8±5.6

156.4±5.8

0.052

159.3±5.6

156.4±5.6

0.057

Corrected heightc (cm)

143.4±6.5

149.8±6.2

0.005

155.1± 6.9

153.0±6.9

0.033

157.2±6.6

155.3±5.2

0.006

159.2±5.6

156.4±5.8

<0.001

160.7±5.8

156.4±5.6

0.007

aMann-Whitney U-test was used to test the difference in BW. Medians and inter-quartile ranges were presented

bComparison between apparent height of the AIS and controls

cComparison between height of the controls and AIS with correction for trunk loss

Dietary intakes

There was no significant difference in median calcium intake between the AIS and controls groups (361 mg/day, IQR: 230–532 mg/day versus 319 mg/day, IQR: 220–494 mg/day; P=0.063). After grouping the subjects into the four age groups, there was no significant difference in median calcium intake between the AIS group and controls except at age ≤12 years, at which calcium intake of AIS was 36.8% higher than that of controls. In fact, the median calcium intakes of the AIS and control groups during adolescence reach only 36% and 32%, respectively, the Chinese “Dietary Reference Intake (DRI), 1000 mg/day” for calcium [31]. Low calcium intake was due to lower consumption of milk and milk products among the subjects.

There was no significant difference in median protein intake of the AIS and controls (65.9 g/day, IQR: 49.5–90.5 g/day versus 62.3 g/day, IQR: 43.1–88.0 g/day; P=0.229). Median energy intake of the AIS group did not differ from that of controls (1503 kcal/day, IQR: 1214–1889 Kcal/day versus 1467 Kcal/day, IQR: 1127–2067 kcal/day; P=0.803). When the AIS and controls were categorized into the four age groups, there were no significant differences in median protein and energy intakes between the AIS and control groups within the four age groups (Table 2).

Weight bearing physical activity

AIS girls spent significantly less time on overall weight bearing physical activities (1.6 h/day, IQR: 1.0–2.6 h/day versus 1.8 h/day, IQR: 1.2–2.8 h/day; P=0.025) and high impact weight bearing physical activity (0.5 h/day, IQR: 0.3–0.8 h/day versus 0.6 h/day, IQR: 0.3–0.9 h/day; P=0.006), whereas there was no significant difference in medium impact weight bearing physical activity between the AIS and control groups (1.0 h/day, IQR: 0.6–2.0 h/day versus 1.1 h/day, IQR: 0.6–2.0 h/day; P=0.183). After categorizing the subjects into the four age groups, AIS spent significantly less time on overall (P=0.007) and high impact (0.003) weight bearing physical activity than those of controls at age ≤12 years (Table 2).

Generalized low BMD and BMC in AIS

For vBMD, except for RtBMD at age ≤12 years, RtBMD of AIS at all ages were significantly lower than that of controls (P=0.046–0.025). Similarly, TtBMD of the AIS group was also significantly lower than that of controls across the ages (P=0.003–0.001). Table 4 shows the comparisons of BMD and BMC between AIS and controls by age. All the volumetric BMD and BMC of AIS, except for RiBMDC at age ≤12 years, was significantly lower than those of controls. For aBMD, except for LSBMC and FNBMC at age ≤12 years and GTBMC at age 13 years, all BMC of AIS as measured by DXA were significantly lower than those of controls (P=0.05–0.001). Table 4 showed that all the areal BMD of AIS except for FNBMD at age ≤12 years and 14 years were significantly lower than those of controls.
Table 4

Comparisons of bone mineral density (BMD) and bone mineral content (BMC) of distal radius, distal tibia, lumbar spine and proximal femur between AIS and controls by age (mean±SD). RiBMD distal radial trabecular and cortical BMD; RiBMC total radial BMC; TiBMD distal tibial trabecular and cortical BMD; TiBMC total tibial BMC; LSBMD lumbar spine BMD; FNBMD femoral neck BMD; GTBMD greater trochanter BMD; WTBMD Ward’s triangle BMD

Age groups

≤12

13

14

≥15

AIS

Control

P-value

AIS

Control

P-value

AIS

Control

P-value

AIS

Control

P-value

Volumetric BMD and BMC

RiBMD (mg/cm3)

436±80

450±81

0.202

458±85

484±85

0.020

481±84

527±91

0.001

518±100

584±111

<0.001

RiBMC (mg)

0.077± 0.014

0.082± 0.017

0.005

0.087±0.016

0.094±0.018

0.003

0.095±0.017

0.103±0.018

0.003

0.103±0.019

0.114±0.019

0.001

TiBMD (mg/cm3)

410±61

428±66

0.037

429±72

467±74

<0.001

443±69

479±67

0.001

463±71

516±81

<0.001

TiBMC (mg)

0.25±0.04

0.27±0.05

0.001

0.27±0.04

0.29±0.05

<0.001

0.28±0.04

0.30±0.05

0.004

0.30±0.05

0.33±0.05

<0.001

Areal BMD and BMC

LSBMD (g/cm2)

0.67 ±0.11

0.71 ±0.12

0.018

0.75±0.11

0.80±0.13

0.004

0.78±0.10

0.85±0.12

<0.001

0.82±0.10

0.92±0.12

<0.001

FNBMD (g/cm2)

0.68±0.10

0.71±0.10

0.069

0.73±0.11

0.76±0.10

0.012

0.75±0.10

0.77±0.10

0.076

0.77±0.11

0.85±0.14

<0.001

GTBMD (g/cm2)

0.57±0.09

0.59±0.09

0.050

0.60±0.09

0.63±0.09

0.024

0.59±0.08

0.63±0.09

0.001

0.61±0.09

0.67±0.12

<0.001

WTBMD (g/cm2)

0.57±0.10

0.61±0.10

0.021

0.61±0.11

0.64±0.11

0.047

0.61±0.11

0.65±0.10

0.019

0.63±0.12

0.71±0.14

<0.001

Similarly, when subjects were categorized by pubertal groups, most volumetric and areal BMD and BMC of AIS girls were significantly lower than those of controls across the pubertal groups (Table 5). With respect to the percentage differences in BMD and BMC between AIS and controls across the four age groups, from age 13 years onwards, volumetric BMD at the distal tibia was on average 8.4% lower than that of the controls, areal BMD at the proximal femur of AIS was on average 6.7% lower than that of the controls, whereas axial and peripheral BMC of the AIS group was on average 6.9% lower than that of controls. In short, significantly lower bone mass was found in AIS girls when compared with those of controls from 13 years to ≥15 years of age, and that the disparity in bone mass between the AIS and control groups increases with age.
Table 5

Comparisons of bone mineral density (BMD) and bone mineral content (BMC) of distal radius, distal tibia, lumbar spine and proximal femur between AIS and controls by Tanner’s staging (mean±SD). RiBMD distal radial trabecular and cortical BMD; RiBMC total radial BMC; TiBMD distal tibial trabecular and cortical BMD; TiBMC total tibial BMC; LSBMD lumbar spine BMD; FNBMD femoral neck BMD; GTBMD greater trochanter BMD; WTBMD Ward’s triangle BMD

Tanner’s staging

I

II

III

IV

V

AIS

Control

P-value

AIS

Control

P-value

AIS

Control

P-value

AIS

Control

P-value

AIS

Control

P-value

Volumetric BMD and BMC

RiBMD (mg/cm3)

430±64

455±81

0.295

432±69

450±72

0.079

467±90

501±98

0.003

498±91

544±107

0.001

533±122

589±121

0.106

RiBMC (mg)

0.06±0.01

0.07±0.01

0.007

0.08±0.02

0.08±0.02

0.070

0.09±0.02

0.10±0.02

0.018

0.10±0.02

0.11±0.02

<0.001

0.10±0.02

0.11±0.02

0.064

TiBMD (mg/cm3)

388±51

403±51

0.396

402±55

427±60

0.003

437±69

471±77

<0.001

455±73

499±73

<0.001

467±67

531±77

0.003

TiBMC (mg)

0.20±0.03

0.23±0.03

0.006

0.25± 0.04

0.27±0.04

0.002

0.27±0.05

0.29±0.05

0.002

0.29±0.04

0.32± 0.06

<0.001

0.30±0.03

0.33±0.05

0.010

Areal BMD and BMC

LSBMD (g/cm2)

0.56±0.07

0.60±0.10

0.107

0.68±0.10

0.73±0.11

0.002

0.76± 0.11

0.80±0.12

0.006

0.80±0.10

0.89±0.12

<0.001

0.79±0.06

0.93±0.14

<0.001

FNBMD (g/cm2)

0.58±0.05

0.65±0.10

0.017

0.68±0.10

0.72±0.09

0.005

0.74±0.11

0.75±0.10

0.293

0.76±0.10

0.82±0.11

<0.001

0.77±0.06

0.86±0.15

0.003

GTBMD (g/cm2)

0.50±0.04

0.55±0.07

0.066

0.56±0.08

0.59±0.07

0.002

0.60±0.09

0.62±0.08

0.053

0.61±0.09

0.66±0.09

<0.001

0.60±0.06

0.71±0.14

<0.001

WTBMD (g/cm2)

0.50±0.06

0.57±0.08

0.004

0.56±0.11

0.61±0.10

0.003

0.61±0.11

0.63±0.10

0.142

0.63±0.11

0.68±0.11

0.001

0.62±0.07

0.73±0.16

0.001

Calcium intake in association with BMD and BMC

There was a significant correlation of the natural log transformed calcium intake between age-adjusted BMD, and age and bone-area adjusted BMC in the AIS group: RiBMD (r=0.105, P=0.023), TtBMD (r=0.101, P=0.028), TiBMC (r=0.106, P=0.023), LSBMD (0.119, P=0.009), LSBMC (r=0.117, P=0.011), FNBMD (r=0.096, P=0.037), FNBMC (r=0.104, P=0.024), GTBMD and GTBMC (r=0.105, P=0.022). However, calcium intake was only correlated with RiBMD (r=0.137, P=0.037) and RiBMC (r=0.151, P=0.022) in the control group.

Physical activity in association with BMD and BMC

There was also significant correlation of overall weight bearing activity between almost all age-adjusted BMD, and age and bone-area adjusted BMC in the AIS group: RiBMD (r=0.127, P=0.003), TtBMD (r=0.113, P=0.007), TiBMD (r=0.133, P=0.002) and TiBMC (r=0.106, P=0.023), LSBMD (r=0.092, P=0.029), FNBMD (r=0.110, P=0.009), FNBMC (r=0.112, P=0.008), GTBMD (r=0.095, P=0.023), GTBMC (r=0.110, P=0.009) and WTBMD (r=0.090, P=0.033). However, such a significant correlation was only found in RtBMD (r=0.158, P=0.012) and RiBMC (r=0.151, P=0.022) of the control group. Furthermore, high impact weight bearing activity was correlated with almost all BMD and BMC (r=0.082 to 0.142, P=0.05–0.001), except for RiBMD, RiBMC. Medium impact weight bearing activity was correlated with RiBMD (r=0.12, P=0.004), TtBMD (r=0.091, P=0.032), TiBMD, (r=0.125, P=0.003) and TiBMC (r=0.114, P=0.008).

Multivariate analysis

Multivariate analysis was undertaken to examine the contribution of group (AIS or control), anthropometric measurements, age or pubertal staging, calcium intake and overall weight bearing activity to the variation of BMD and BMC. Table 6 summarizes the results of the four multiple regression models using volumetric BMD and BMC as dependent variables. Group (AIS or controls), calcium intake and overall weight bearing activity among others were independent predictors on RiBMD (R-square=24.5%), RiBMC (R-square=57.1%), TiBMD (R-square=28.1%) and TiBMC (R-square=66.6%). Table 7 depicts the results of the five multiple regression models using areal BMD and BMC as dependent variables. Again, group (AIS or controls) and calcium intake, among others, were independent predictors of LSBMD (R-square=50.7%), LSBMC (R-square=74.8%), FNBMD (R-square=43.8%), FNBMC (R-square=62%) and GTBMC (R-square=34.8%). Physical activity could predict the variation of femoral neck BMC and BMD, but not those of the lumbar spine. The R-squares in these models did not vary very much if pubertal development was substituted for chronological age in the regression models, indicating that age and pubertal staging had similar effects in predicting bone mass. Group (AIS=1, controls=0) gives a negative the unstandardized coefficient (B) among the regression models, implying that being an AIS patient independently and inversely predicts the variation of bone mass. In summary, multiple regression analysis confirms that AIS patient status, calcium intake, weight bearing exercise, age/puberty, weight and height were significant independent predictors of most of the variation in BMD and BMC. Further analysis showed that calcium intake had no interaction with bone mass of the AIS and controls and neither did weight bearing physical activity (P>0.05); the results implied that the effect of calcium or weight bearing physical activity on the variation of bone mass was not different between the AIS and healthy non-AIS girls. Based on the regression models in Table 6 and 7, one may predict the increment bone mass of bone mass for each 100 g increase in calcium intake and 30 min increase in weight bearing physical activity on a daily basis by multiplying the amount of calcium intake or weight bearing physical activity increase with the unstandardized coefficient (B) in each regression model (Table 8). It seems that for each 100 g/day increase in calcium intake is more effective than 30 min/day increase in weight bearing physical activity to promote bone mass in AIS.
Table 6

Independent variables of grouping (AIS or controls), calcium intake, weight bearing exercise, age, weight and height in predicting volumetric BMDs in multiple regression analyses. RiBMD distal radial trabecular and cortical BMD; RiBMC total radial BMC; TiBMD distal tibial trabecular and cortical BMD; TiBMC total tibial BMC

Dependent variables

RiBMD

RiBMC

TiBMD

TiBMC

Unstandardized B

P-value

Unstandardized B

P-value

Unstandardized B

P-value

Unstandardized B

P-value

Group (AIS/control)

-22.3

0.001

−0.005

<0.001

−23.3

<0.001

−0.015

<0.001

Calcium intake (mg/day)

0.029

0.003

4.6×10-6

0.004

0.016

0.028

1.5×10-5

<0.001

Overall weight bearing activity (h/day)

6.81

0.001

0.0006

0.049

4.22

0.006

0.002

0.003

Age (years)

33.0

<0.001

0.0059

<0.001

20.1

<0.001

0.006

<0.001

Weight (kg)

3.29

<0.001

0.001

<0.001

4.07

<0.001

0.004

<0.001

Uncorrected height (cm)

-4.81

<0.001

−0.0002

0.025

−3.98

<0.001

−2.5×10-5

0.924

R-square (%)

24.5

57.1

28.1

66.6

Table 7

Independent variables of grouping (AIS or controls), calcium intake, weight bearing exercise, age, weight and height in predicting areal BMDs in multiple regression analyses. LSBMC lumbar spine BMC; FNBMD femoral neck BMD; FNBMC femoral neck BMC; GTBMD greater trochanter BMD; GTBMC greater trochanter BMC

Dependent Variables

LSBMD

LSBMC

FNBMD

FNBMC

GTBMD

Unstandardized B

P-value

Unstandardized B

P-value

Unstandardized B

P-value

Unstandardized B

P-value

Unstandardized B

P-value

Group (AIS/control)

−0.044

<0.001

−1.80

<0.001

−0.023

0.001

−0.064

0.001

−0.026

<0.001

Calcium intake (mg/day)

2.6×10−5

0.014

0.001

0.014

2.7×10−5

0.006

8.6×10−5

0.001

2.3×10−5

0.009

Overall weight bearing activity (h/day)

0.003

0.186

0.106

0.207

0.004

0.044

0.012

0.040

0.003

0.075

Age (years)

0.030

<0.001

0.942

<0.001

0.015

<0.001

0.045

<0.001

0.002

0.354

Weight (kg)

0.009

<0.001

0.299

<0.001

0.008

<0.001

0.024

<0.001

0.007

<0.001

Uncorrected height (cm)

6.8×10−5

0.923

−0.077

0.013

−0.0004

0.554

0.001

0.467

−0.001

0.026

R-square (%)

50.7

74.8

43.8

62.0

34.7

Table 8

Prediction of bone mass increment for daily increases in calcium intake and weight bearing physical activity by 100 mg and 30 min, respectivelya

Predicted increment in bone mass per:

100 mg calcium intake daily

0.5 Hr weight bearing physical activity daily

RiBMD (mg/cm3)

2.9

3.405

RiBMC (mg)

4.6×10−4

3×10−4

TiBMD (mg/cm3)

1.6

2.11

TiBMC (mg)

1.5×10−3

1×10−3

LSBMD (mg/cm2)

2.6×10−3

1.5×10−3

LSBMC (g)

0.1

0.053

FNBMD (mg/cm2)

2.7×10−3

2×10−3

FNBMC (g)

8.6×10−3

6×10−3

GTBMD (mg/cm2)

2.3×10−3

1.5×10−3

aCalculation based on results of multiple regression analyses in Tables 6 and 7

Further analysis was attempted to group scoliotic curve patterns of the patients into thoracic curves according to the King’s classification (King’s III and King’s V) [32] and lumbar curve, in order to examine if there were any differences in the anthropometric parameters, bone mass, dietary intakes or physical activity of AIS between thoracic curve types and lumbar curve types. A total of 173 AIS girls with thoracic curves (King’s III, n=136, King’s V, n=37) were compared with 54 AIS girls with lumbar curve type. Statistical analysis showed that there were no significant differences in anthropometry, bone mass, dietary intakes or physical activity in AIS girls with thoracic or lumbar curve type (P>0.05).

Discussion

To our knowledge, this is the first large scale cross-sectional study involving 596 AIS girls and over 300 healthy girls aged 11–16 years to investigate the relationship between calcium intake, physical activity and bone mass and compared with those of age-matched normal controls. Results from the present study demonstrated that axial and appendicular bone mass of AIS was significantly lower than that of age-matched controls, even after controlling for confounding factors. Furthermore, the disparity in bone mass between AIS and controls becomes greater in the older age groups than in the younger age groups. Multivariate analysis also indicated that the AIS patient was associated with a significantly lower bone mass. These observations agreed with findings from earlier studies [5,6,33] and our recent reports [8–12] that generalized osteopenia is associated with the development of AIS.

Until recently, it has been uncertain as to which factors might contribute to low bone mass in AIS during pubertal years. Studies on calcium intake and physical activity of AIS are scanty in the literature. There have been two previous studies documenting the relationship between bone mass and habitual calcium intake in AIS patients. One earlier study found that there was no significant difference in calcium intake between AIS (n=44) and non-AIS girls (n=44) [5]. Another 39-week follow-up study also found that calcium intake was not significantly correlated with BMD at the spine or hip [7]. The findings of these earlier studies were not conclusive because of the small sample size. Results from the present large-scale study, however, have revealed the important roles of the two lifestyle factors (calcium intake and weight bearing physical activity) in bone mineral acquisition among AIS girls during peripubertal years. Calcium intakes in both the AIS and control groups were relatively low reaching less than 40% of the Dietary Reference Intake, which reflects the typical non-milk based dietary practice among the Hong Kong Chinese. The level of calcium intake in AIS was not sufficient to meet the requirement for rapid bone mineralization during the adolescent growth spurt [31]. Although calcium intake of AIS did not differ from that of the controls during adolescent years except at age ≤12 years, calcium intake was significantly correlated with axial and appendicular BMD and BMC in the AIS group. In addition, multivariate analysis showed that after controlling for confounding factors, calcium intake was still an independent predictor of the variation of axial and appendicular BMD and BMC. The results indicated that calcium intake was an important lifestyle factor for the promotion of bone mass in AIS.

Weight bearing activity was also low among both the AIS and control groups (<2 h/day). However, AIS as a group spent relatively fewer hours per day than controls on weight bearing activity. The number of hours spent on weight bearing activity was significantly correlated with bone mass in AIS girls. Furthermore, multivariate analysis also showed that after controlling for confounding factors, weight-bearing activity was an independent determinant on the variation of BMD. Again, being physically active was an important determinant for the enhancement of bone mass in AIS. Patients in the current study were recruited and interviewed during their first visits to our scoliosis clinic when idiopathic scoliosis was newly diagnosed. Technically, therefore, the timing of diagnosis should not have constituted bias in recalling history of physical activity.

Further analysis indicated that calcium intake and physical activity had no interaction with the bone mass of AIS and controls implying that calcium or weight bearing physical activity had the similar impact on the bone mass of AIS and controls. It may merit a further interventional study to examine the possible beneficial effects of calcium intake and physical activity on bone mass acquisition in AIS.

In addition, calcium intake and weight bearing physical activity virtually were not associated with bone mass in the control group. This might be due to the fact that the strength of univariate association of bone mass with calcium intake and physical activity is often weaker than that with anthropometric parameters or bone size in adolescents [18,34], and that the sample size of the control group was relatively smaller than that in the AIS group.

It is noteworthy that even though the magnitudes of calcium intake and weight bearing physical activity of the control group did not greatly exceed those of the AIS group, bone mass of AIS was found more than 6% lower than that of the controls. After controlling for the confounding factors in multivariate analysis, the AIS group still had significantly lower bone mass. It is logical to speculate that lower bone mass in AIS might be attributable to a faster anthropometric growth [35] and abnormally increased bone turnover [36] when compared to the age-matched controls during the peripubertal period. Faster growth during puberty may lead to increased bone dimensions, and failure of bone mineralization to catch up with escalated bone growth may result in apparent osteopenia [35,36]. Further investigation is necessary to delineate the relationship between anthropometric growth, bone turnover and bone mass in AIS.

The current study also revealed the poor lifestyle practices of AIS and non-AIS girls in Hong Kong: low calcium intake and inadequate participation in physical activity in these girls. Most AIS patients in the present study are underweight, with BMI <18.5. Poor dietary habits, replacing proper meals by snacks and skipping breakfast were common among both the AIS and control groups. The findings demonstrated poor eating habits among adolescent girls in Hong Kong. Health education and life-style interventional programs (nutrition and physical exercise) should be devised to improve body weight and nutritional status in both AIS patients and non-AIS adolescents at the critical period of pubertal growth spurt.

Although adolescent patients with idiopathic scoliosis have developed deformity of the spine, they would not develop pain at the back or the limbs at a wide range of curve severity. Hence, pain would not be a limiting factor for them to perform regular physical activity. Also, clinicians would not discourage them from performing any form of everyday physical activity. They could participate in any form of regular physical activity. In the literature, there have been no studies suggesting which types of physical activity are suitable or not for adolescents with AIS. Hence, pain will not be a concern for AIS girls in performing any form of weight bearing activity in an intervention study with programmed weight bearing physical exercise.

There are important research and clinical implications based on findings from the present study. The prevalence of low bone mass in AIS has been found higher than that in normal adolescent population [10], and the low bone mass in AIS seems to persist through to adulthood [9,33] and even in the elderly [37,38]. Therefore, promotion of bone mass could be of paramount importance in this group of patients to attain peak bone mass and to reduce the risk of osteoporosis later in life. Thirdly, in AIS patients requiring surgical procedures, information on preoperative bone mass status is helpful for surgeons because bone mass correlates well with bone strength, so that appropriate selection of implants and sites of fixation for spinal fusion can be decided. Fourthly, would early treatment of low in bone mass modify the progression of scoliosis and minimize the long-term problems associated with adult osteoporosis? These clinical implications are significant and in our opinion merit further research.

A random sample of about 350 patients and controls from the present cohort have been followed up for 3 years to examine physical growth, dietary intakes, physical activity and other related parameters that might predict the outcomes of bone mineral acquisition and spinal progression. We are planning to follow up these girls until skeletal maturity, in order to obtain more valuable information related to the development of the disease.

Results from the present study have demonstrated the importance of calcium intake and weight bearing physical activity on bone mass development in AIS. In healthy girls, interventional studies combined with calcium supplements and programmed physical activity have been shown to enhance bone mass during adolescence [20,22]. A controlled trial combined with calcium supplementation and programmed weight bearing physical training program is necessary to confirm the effects of these two lifestyle factors on bone mass acquisition in AIS. Consequently, it might help prevent the progression of the scoliotic deformity, possible osteoporotic fractures associated with low BMD in late adulthood, and other complications following the interaction between scoliosis and osteoporosis.

In conclusion, AIS patients aged 11–16 years had significantly lower bone mass than that of age-matched controls during pubertal period. Absolute calcium intake and weight bearing physical activity among AIS were low, and require timely health education and lifestyle interventional programs to improve body weight, nutritional status and bone mass. Multivariate analysis showed that the AIS patient was associated with significantly lower bone mass, and that both calcium intake and weight bearing physical activity were independent predictors on bone mass. Prevention of generalised osteopenia is as important as controlling spinal-progression in the management of AIS during peribubertal period. It merits a future interventional study with calcium supplementation and programmed physical training to confirm the beneficial effects of calcium and weight bearing physical activity on bone mass acquisition in AIS during peri-pubertal period.

Acknowledgements

We would like to express our gratitude to the patients and their parents, the Heads of school, students and parents from the participating schools; without their supports, our study would not have been successful. Thanks are also due to Mr. Jacky W.W. Chau, Miss Vivian Hung, Miss Christine Lee, Miss Catherine Li, Miss Sylvia Lam, Miss Elaine Au and Mr. Jason So for fieldwork assistance and data entry. Special thanks are also expressed to the Jockey Club Centre for Osteoporosis Care & Control, The Chinese University of Hong Kong for generous manpower support on this study. The study was supported by Research Grant Council (no. CUHK 4336/99M) & Health Service Research Grant (no. HSRF 921024), Hong Kong SAR.

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2005