Incident Vertebral Fractures among Children with Rheumatic Disorders 12 Months Post-Glucocorticoid Initiation: a National Observational Study

Abstract

Objectives

To determine the frequency of incident vertebral fractures (IVF) 12 months after glucocorticoid (GC) initiation in children with rheumatic diseases and to identify children at higher risk.

Methods

Children with rheumatic diseases initiating GC were enrolled in a prospective observational study. Annual spine radiographs were evaluated using the Genant semi-quantitative method. Spine areal bone mineral density (aBMD) was measured every 6 months. Clinical features, including cumulative GC dose, back pain, disease and physical activity, calcium and vitamin D intake, and spine aBMD Z-scores were analyzed for association with IVF.

Results

Seven (6%) of 118 children (95% Confidence Interval 2.9 to 11.7) had IVF. Their diagnoses were: juvenile dermatomyositis (n = 2), systemic lupus erythematosus (n = 3), systemic vasculitis (n = 1) and mixed connective tissue disease (n = 1). One child was omitted from the analyses after 4 months because of osteoporosis treatment for symptomatic IVF. Children with IVF received on average 50% more GC than those without (p=0.030), had a greater increase in body mass index (BMI) at 6 months (p=0.010), and had greater decrements in spine aBMD Z-scores in the first 6 months (p=0.048). Four (67%) of 6 children with IVF and data to 12 months had spine aBMD Z-scores less than −2.0 at 12 months compared to 16% of children without IVF (p=0.011).

Conclusions

The incidence of VF 12 months following GC initiation was 6%; most children were asymptomatic. Children with IVF received more GC, had greater increases in BMI and greater declines in spine aBMD Z-scores in the first 6 months.

INTRODUCTION

Compromised bone strength is increasingly recognized as an important consequence of childhood rheumatic diseases. Such patients have multiple risk factors for impaired skeletal health including the inflammatory process itself, delayed growth and development, decreased weight bearing and physical activity, muscle dysfunction, suboptimal nutritional status and medications, especially glucocorticoids (GC) (1–3).

A number of studies have documented bone mass reductions in children with rheumatic diseases (1–7). These decrements in bone mass may increase the risk of fractures in childhood, and potentially later in adulthood as a result of suboptimal accrual of peak bone mass (3, 4). More definitive evidence of bone fragility in children with rheumatic diseases is provided by studies documenting vertebral and extremity fractures. Studies have reported vertebral fracture (VF) prevalence of 10 to 34% (6, 8–12). Although these findings help to establish the magnitude of this problem during the course of the illness, clinically relevant questions remain unanswered regarding the frequency and timing of atraumatic incident (i.e. new) VF (IVF) in relation to GC dose and duration.

Through our national observational study of STeroid-associated Osteoporosis in the Pediatric Population (STOPP), we have prospectively studied a cohort of children with GC-treated rheumatic diseases. We have previously documented a VF prevalence of 7% in our cohort within 30 days of GC initiation (2). The aims of the present study were to determine the frequency and characteristics of IVF 12 months after GC initiation and to identify patients at higher risk of fractures.

PATIENTS AND METHODS

Patients and study design

Patients were recruited through the STOPP research initiative. Children from age one month to 17 years were enrolled (N = 136) between January 1 2005 and December 31 2007 in 10 participating Canadian children’s hospitals. Children were enrolled within 30 days of first-time GC treatment for inflammatory rheumatic disorders and grouped according to “related” disease categories as follows: juvenile dermatomyositis (JDM), systemic lupus erythematosus (SLE) and related conditions including mixed connective tissue disease and other overlap syndromes, juvenile idiopathic arthritis (JIA) (excluding systemic arthritis), systemic arthritis, systemic vasculitis (excluding Henoch-Schonlein purpura and Kawasaki Disease), and “other conditions” including juvenile scleroderma (both systemic and localized).

Children were excluded from the study if GC had previously been used for treatment of the underlying disease. Patients were also excluded if they received intravenous or oral GC for more than 14 consecutive days in the 12 months preceding study enrolment to treat any other medical condition (e.g. asthma), if they received prior medication for osteoporosis, or if they had received calcium or vitamin D supplementation that exceeded the Dietary Reference Intake for age recommended at the time of the study (13, 14). In addition, patients who developed symptomatic IVF during the observation period requiring osteoporosis treatment in accordance with the local standard of care were removed from the bone health natural history analyses from the time of osteoporosis intervention forward (with results presented separately). Finally, children who did not have spine radiograph data at 12 months were excluded. The clinical characteristics at study entry were compared for children with and without 12 month spine radiograph data.

Children were studied every 3 months for one year to allow accurate collection of the clinical data. The 12 month visit had a ±3 month window of time in which to collect the data. Each participating institution’s Research Ethics Board approved the study and informed consent/assent was obtained prior to enrolment.

Clinical data

Demographic and puberty data were obtained as previously described (2). Height, weight, and body mass index (BMI, weight (kg) divided by height (meters) squared) raw values were transformed into age- and gender-matched Z-scores according to the United States Centers for Disease Control and Prevention National Center for Health Statistics normative database (15); for children under 2 years of age, BMI Z-scores were calculated according to the World Health Organization child growth standards (16). In addition, the children were assessed every 3 months for presence of back pain since the last visit, calcium and vitamin D intake was assessed by a validated food frequency questionnaire and supplemental calcium and vitamin D intake were also recorded (14). Physical activity was assessed quarterly using the Habitual Activity Estimation Scale (2, 17). Additionally, a physician global assessment of disease activity was determined quarterly on a 10 centimeter visual analog scale (VAS) by the patients’ attending rheumatologists (where 0 cm = inactive disease and 10 cm = extremely active disease) as previously described (2, 18–20). At 12 months, the spine was palpated for tenderness over the posterior spinous processes (T4 to L4) and history of back trauma was recorded.

Quantification of GC and methotrexate exposure

Each patient was seen quarterly to collect height and weight to determine body surface area (calculated as the average over the observation period) and records of all GC received. From this the dose of systemic GC therapy (oral and intravenous, including pulse therapy) was converted into prednisone equivalents and results were expressed in 3 ways (21–23): 1) cumulative GC dose, defined as the total amount of GC in prednisone equivalents (mg/m²) received during the observation period; 2) GC dose intensity, defined as the cumulative dose in prednisone equivalents (mg/m²) divided by the number of days actually taking GC during the observation period; and 3) average GC dose, defined as the cumulative dose in prednisone equivalents (mg/m²) divided by the total number of days in the observation period. Methotrexate exposure was expressed as whether children had received the drug or not and as a cumulative dose to the 12-month follow-up.

VF assessment

Lateral thoracolumbar spine radiographs were carried out at enrolment and 12 months. The Genant semi-quantitative method for vertebral morphometry (24) was performed as previously described (2). Vertebral bodies were first assigned a severity score: grade 0 (normal), grade 1 (mild), grade 2 (moderate), or grade 3 (severe). The morphometric grading corresponded to the extent of the reduction in height ratios when the anterior vertebral height was compared to the posterior height (anterior wedge fracture), the middle height to the posterior height (biconcave fracture), and the posterior height to the posterior height of the adjacent vertebral bodies (crush fracture). The scores corresponded to the following reduction in height ratios: grade 0: 20% or less; grade 1: > 20 to 25%; grade 2: > 25 to 40%; grade 3: 40%. Grade 0 was considered to be normal and higher grades were considered to be a fracture.

An IVF was defined as a new VF at 12 months in a vertebral body that was normal at study entry, or worsening of an existing VF (progression from a grade 1 or more, to a higher VF grade).

Lumbar spine (L-Spine) bone mineral density (BMD) by dual-energy-x-ray absorptiometry

Bone mineral density (BMD) was measured in the anterior-posterior direction at the lumbar spine (LS) (L1–L4) using either Hologic machines (QDR 4500, 3 centers; Discovery, 2 centers; Delphi, 1 center) or Lunar Prodigy (4 centers) at study enrolment and 6 and 12 months. Machines were cross-calibrated as previously described (2). Data were converted to Hologic units and Z-scores were generated using the Hologic 12.4 normative database. In vivo precision for L-spine areal BMD (aBMD) was available in 8 of 10 centers and ranged from 0.003 to 0.017 gm/cm².

Bone age and second metacarpal morphometry

Radiographs of the left hand and wrist for bone age were read independently and second metacarpal morphometry was obtained and analyzed using the methods previously described (2).

Statistical analyses

All analyses were conducted using SPSS 18.0. Categorical variables were summarized using frequency and percentage. Normally distributed continuous variables were summarized using the mean ± SD. Non-normally distributed continuous variables were summarized using the median and range. The 95% Confidence Intervals (95% CIs) for the proportion of patients with vertebral deformities were calculated using the Wilson score method (25). Z-score variables were compared against the healthy average (Z-score = 0.0) using 1-sample Student’s t-test to assess whether the patient population significantly differed from the normal reference values. Results for height, weight, BMI, L-spine aBMD and disease activity by VAS were compared using paired Student’s t-test at the time of study enrolment and 12 months. Children with IVF were compared to those without using Wilcoxon Mann-Whitney and Fisher’s exact tests. Presented p-values are two-sided and a p-value less than or equal to 0.05 was considered significant.

RESULTS

Characteristics of the natural history cohort

A total of 117 (86%) of the 136 children who were enrolled within 30 days of GC initiation completed the bone health natural history assessment through to 12 months. One additional child was included in the natural history analyses to 3 months, but was excluded thereafter because she received osteoporosis therapy (intravenous pamidronate) for symptomatic IVF at 4 months post- GC initiation; this child is described separately at the end of the Results. Eighteen other children were excluded from all analyses: 10 did not have spine radiographs at 12 months, 2 were lost to follow-up and 6 withdrew consent. At study enrolment, the clinical characteristics of these 18 children did not differ significantly from the natural history cohort (data not shown).

Clinical features of the cohort at 12 months are shown in Table 1. The characteristics of the 12 month cohort did not differ significantly from the full cohort at study entry with respect to gender, ethnicity, concordance between bone age and chronologic age, and distribution of prepubertal to pubertal Tanner stages (data not shown). Rheumatic disease activity as measured by the 10-cm VAS showed a decrease at 12 months (median 0.5 cm, range 0 to 6.3) compared with study enrolment (median 6.3, range 0 to 10.0; p <0.001). Compared to enrolment, height Z-scores at 12 months were lower and weight Z-scores were higher, resulting in increased BMI Z-scores (Figure 1); these Z-scores at 12 months were all significantly different compared to the normative reference data (Table 1).

Table 1.

Description of children with rheumatic disorders and bone health natural history data to 12 months post-GC initiation^#

Clinical Characteristics at 12 months	Overall Study Cohort with Data to 12 Months	JDM	JIA	SLE and Related Conditions	Systemic Arthritis	Systemic Vasculitis	Other Conditions
	N=117	N=27	N=22	N=20	N=21	N=16	N=11
Demographic Data
Female, n (%)	74 (63)	16 (59)	14 (64)	17 (85)	12 (57)	8 (50)	7 (64)
Age, median (range) years	11.0 (2.3–17.9)	8.3 (2.8–16.0)	13.1 (4.7–17.9)	14.5 (6.0–17.2)	6.6 (2.3–16.2)	14.1 (5.5–17.9)	9.0 (4.4–17.5)
White ethnicity, no. (%)	92 (79)	24 (89)	19 (86)	9 (45)	18 (86)	11 (69)	11 (100)
Anthropometry
Height Z-score, mean ± SD	−0.2 ± 0.9 *	−0.4 ± 0.9	−0.4 ± 1.0	−0.3 ± 0.8	−0.2 ± 0.9	0.1 ± 1.0	0.2 ± 0.9
Weight Z-score, mean ± SD	0.4 ± 1.3 *	0.2 ± 1.4	0.1 ± 1.5	0.5 ± 1.2	0.5 ± 1.0	1.0 ±1.4	0.4 ± 0.9
BMI Z-score, mean ± SD	0.7 ±1.2 *	0.7±1.2	0.3±1.5	0.8 ±1.0	0.8 ±1.2	1.0 ± 1.3	0.5 ± 0.9
Pubertal stage, no. (%)
Stage 1	51 (48)	15 (65)	7 (33)	3 (16)	15 (75)	5 (36)	6 (67)
Stage 2–5	55 (52)	8 (35)	14 (67)	16 (84)	5 (25)	9 (64)	3 (33)
Bone age, median (range) years	10.8 (1.8–18.0)	7.2 (1.8–6.8)	13.4 (3.8–17.0)	14.5 (5.8–17.0)	6.0 (1.8–17.0)	14.3 (4.2–18.0)	8.8 (4.0, 18.0)
Rheumatic disease activity
Disease activity (10 -cm VAS), median (range)	0.5 (0.0–6.3)	0.2 (0.0–3.4)	1.3 (0.0–3.0)	0.4 (0.0–5.9)	0.2 (0.0–6.3)	0.4 (0.0–5.5)	1.1 (0.0–2.2)
Days since diagnosis, median (range)	393 (343–3408)	386 (344–714)	413 (357–3408)	385 (354–596)	385 (343–512)	389 (358–459)	444 (370–2321)
Lumbar Spine BMD
L-spine aBMD Z-score, mean ± SD	−0.8 ± 1.2*	−1.5 ± 1.1	−0.5 ± 1.2	−0.4 ± 1.1	−0.8 ± 1.2	−1.3 ± 1.2	0.0 ± 1.0
L-spine aBMD Z-score less than −2.0, no. (%)	21 (19)	10 (39)	1 (5)	2 (10)	4 (20)	4 (27)	0 (0)
Vertebral Fractures
Patients with incident vertebral fracture, no. (%) (95% CI)	6 (5) (2.4–10.7)	2 (7) (2.1–23.4)	0 (0) (0–14.9)	3 (15) (5.2–36.0)	0 (0) (0–15.5)	1 (6) (1.1–28.3)	0 (0) (0–25.9)
GC Treatment
Days between GC initiation and 12 month spine radiograph, mean ± SD	389 ± 25	391 ± 26	390 ± 23	393 ± 26	386 ± 31	386 ± 19	387 ± 23
Days in receipt of GC, median (range)	342 (4–454)	368 (93–444)	129 (4–454)	378 (167– 446)	362 (70– 442)	367 (206– 419)	75 (9–258)
Cumulative GC dose, mean ± SD mg/m2	6369 ± 5146	10258 ± 6330	1554 ± 1206	6422 ± 3441	5677 ± 5213	6824 ± 2480	7021± 4502
Average GC dosage, mean ± SD mg/m2/day	16 ± 13	26 ± 16	4 ± 3	16 ± 9	15 ± 14	18 ± 7	18 ± 11
GC dose intensity, median (range) mg/m2/day	17 (1–921)	23 (8–71)	7 (1–705)	17 (6–42)	15 (3–58)	21 (9–33)	54 (8–921)
MTX Treatment
Cumulative MTX dose median (range) mg/m2	470 (0–2977)	796 (0–1426)	620 (0–1527)	0 (0–414)	253 (0–2977)	0 (0–565)	936 (0–1594)
Patients who received any MTX, no. (%)	74 (63)	26 (96)	20 (91)	2 (10)	13 (62)	3 (19)	10 (91)

^#GC=Glucocorticoid; DM = Dermatomyositis; JIA = Juvenile idiopathic arthritis (excluding systemic arthritis); SLE = Systemic lupus erythematosus; BMI=Body mass index; VAS=Visual analogue scale; aBMD= areal bone mineral density; MTX= methotrexate; L-spine = Lumbar spine; aBMD = Areal bone mineral density; 95% CI = 95% confidence interval; SD=Standard deviation; MTX=Methotrexate

^*Significantly different compared to normative reference data (Z-score of 0)

Figure 1.

Mean z-scores for height, weight, body mass index (BMI) and lumbar spine (LS) areal bone mineral density (aBMD) Z-scores at study entry and 12 months.

IVF

Of the 117 children with natural history data to 12 months post-GC initiation, 6 (5%; 95% CI, 2.4 to 10.7%) had a total of 7 IVF (median age 10.9 years, range 8.4 to 17.2 years, n = 4 girls). Five children had a single IVF and 1 had 2 IVF. All of the IVF were new fractures in previously normal vertebral bodies with no significant back trauma having occurred in any of the children. None of the children with prevalent VF at enrolment developed IVF. Only children with JDM, SLE, and systemic vasculitis developed IVF (Table 1). Three children (50%) had mild (grade 1) fractures as the worst grade and 50% had moderate fractures. Five fractures (71%) were thoracic and 2 were lumbar; 6 of 7 IVF were anterior wedge fractures. Examples of IVF that were representative of those seen at 12 months are shown in Figure 2.

Figure 2.

Examples of incident vertebral fractures at 12 months. A (I) shows a normal spine radiograph at study entry (within 30 days of glucocorticoid initiation) in an 11 year old girl with systemic lupus erythematosus. At 12 months (II), this child manifested a grade 1 incident vertebral fracture at L1. B (I) shows a normal spine radiograph at study entry in a 16 year old girl with systemic lupus erythematosus. At 12 months (II), this child was identified as having a grade 2 incident vertebral fracture at T11.

Comparisons of the clinical characteristics between children with and without IVF in the natural history cohort are shown in Tables 2 and 3. There was no difference between the two groups for age, gender, pubertal status or disease activity. Both groups had similarly low measures of disease activity at 12 months according to the VAS (p=0.253) and few had back pain by self-report or by palpation, regardless their IVF status over the 12 month period. Specifically, 37% of those without fractures reported back pain versus 33% of children with IVF (p=1.00). No differences existed in second metacarpal morphometry, physical activity or calcium and vitamin D intake (data not shown).

Table 2.

Comparison of children with and without incident fractures at 12 months post-initiation of GC^#

Clinical Characteristics	Children without Incident Vertebral Fractures	Children with Incident Vertebral Fractures	pa
	N=111	N=6
Demographic data
Female, n (%)	70 (63)	4 (67)	1.000b
Age at 12 months, median (range) years	11.0 (2.3 to 17.9)	10.9 (8.4 to 17.2)	0.578
Anthropometry, mean
Height Z-score at 12 months, mean ± SD	−0.2 ± 0.9	−0.3 ± 0.5	0.863
Δ Height Z-score study entry to 6 months, mean ± SD	−0.2 ± 0.7	−0.2 ± 0.2	0.879
Δ Height Z-score 6 to 12 months, mean ± SD	−0.1 ± 0.4	0.0 ± 0.3	0.949
Weight Z-score at 12 months, mean ± SD	0.4 ± 1.3	0.4 ± 0.8	0.911
Δ Weight Z-score study entry to 6 months, mean ± SD	0.3 ± 0.6	1.1 ± 0.7	0.009ψ
Δ Weight Z-score 6 to 12 months, mean ± SD	−0.2 ± 0.4	−0.6 ± 0.6	0.056
BMI Z-score at 12 months, mean ± SD	0.7 ± 1.2	0.7 ± 0.8	0.951
Δ BMI Z-score study entry to 6 months, mean ± SD	0.5 ± 0.8	1.4 ± 0.8	0.010ψ
Δ BMI Z-score 6 to 12 months, mean ± SD	−0.3 ± 0.5	−0.7 ± 0.7	0.205
Pubertal stage at 12 months (Tanner Stage 1), n (%)	50 (50)	1 (20)	0.365b
Bone age, median (range) years	10.8 (1.8 to 18.0)	10.0 (7.0 to 17.0)	0.822
Rheumatic disease characteristics, median (range)
VAS score at 12 months	0.5 (0.0 to 6.3)	0.0 (0.0 to 2.2)	0.253
Δ VAS score, study entry to 6 months	−4.4 (−9.6 to 3.0)	−2.5 (−7.6 to 1.9)	0.182
Δ VAS score 6 to 12 months	−0.2 (−4.9 to 5.8)	−0.2 (−1.7 to 0.0)	0.978
Days since diagnosis	393 (343 to 3408)	409 (362 to 447)	0.528
L-spine aBMD
L-spine aBMD Z-score at 12 months, mean ± SD	−0.8 ± 1.2	−1.7 ± 1.1	0.060
Δ L-spine aBMD Z-score study entry to 6 months, mean ± SD	−0.4 ± 0.5	−0.8 ± 0.5	0.048ψ
Δ L-spine aBMD Z-score 6 to 12 months, mean ± SD	0.1 ± 0.4	0.3 ± 0.6	0.674
L-spine aBMD Z-score less than −2.0 at 12 months, no. (%)	17 (16)	4 (67)	0.011ψ
Vertebral Fractures at Study Entry
Patients with vertebral fractures at study entry, no. (%)	9 (8)	0 (0)	1.000 b
Back Pain
Back pain during 12 months of follow up, anytime after study entry, no. (%)	41 (37)	2 (33)	1.000 b
Patients with spine pain by palpation at 12 months, no. (%)	4 (4)	1 (20)	0.192 b
GC Treatment
Days between GC initiation and 12 month spine x-ray radiograph mean ± SD	388 ± 25	406 ± 27	0.073
Days in receipt of steroids, median (range)	342 (4 to 454)	383 (264 to 446)	0.088
Cumulative GC dose, mean ± SD mg/m2	6203 ± 5181	9452 ± 3447	0.030ψ
Average GC dosage, mean ± SD mg/m2/day	16 ± 13	24 ± 9	0.044ψ
GC dose intensity, median (range) mg/m2/day	16 (1 to 921)	30 (7 to 34)	0.166
MTX Treatment
Cumulative MTX dose, median (range) mg/m2	503 (0 to 2977)	158 (0 to 1111)	0.560
Patients who received any MTX, no. (%)	71 (64)	3 (50)	0.668b

^#GC= Glucocorticoid, SD=Standard deviation, BMI=Body mass index, VAS=Visual analogue scale, L-spine = Lumbar spine, aBMD= areal bone mineral density, GC=Glucocorticoid, MTX=Methotrexate

^aStatistical significance determined by Mann-Whitney U test for independent samples

^bStatistical significance determined by Fisher’s Exact test

^ψStatistically significant at P ≤ 0.05

Table 3.

Distribution of glucocorticoid (GC) and methotrexate (MTX) exposure for children with and without incident vertebral fractures during the 12 month follow-up period

	0 to 6 months			6 to 12 months
Parameters	Children without Incident Vertebral Fractures	Children with Incident Vertebral Fractures	pa	Children without Incident Vertebral Fractures	Children with Incident Vertebral Fractures	pa
GC Quantification
Days between GC initiation & 12-month spine radiograph, mean ± SD	201 ± 21	193 ± 13	0.311	187± 29	212 ± 21	0.024ψ
Days in receipt of GC, median (range)	182 (4 – 245)	192 (176 – 209)	0.233	168 (1 – 252)	199 (56 – 237)	0.226
Cumulative GC dose median (range) mg/m2	4570 (15 – 25525)	8450 (2079 – 10864)	0.025ψ	1156 (14 – 13006)	1291 (229 – 4804)	0.904
Average GC dose, median (range) mg/m2/day	21 (0.1 – 123)	43 (10 – 60)	0.026ψ	6 (0.1 – 64)	6 (1 – 21)	0.694
GC dose intensity, median (range) mg/m2/day	23 (1 – 922)	43 (10 – 60)	0.093	7 (1 – 903)	7 (2 – 21)	0.628
MTX quantification
Cumulative MTX dose, median (range) mg/m2	236 (0 – 1007)	107 (0 – 508)	0.673	270 (0 – 2608)	52 (0 – 611)	0.558
Patients who received any methotrexate, no. (%)	67 (60)	3 (50)	0.683 b	69 (62)	3 (50)	0.674 b

^aStatistical significance determined by Mann-Whitney U test for independent samples

^bStatistical significance determined by Fisher’s exact test

^ψStatistically significant at P ≤ 0.05

Analysis of GC exposure demonstrated that those with IVF received, on average, 50% more steroid compared to those without fractures as assessed by cumulative (p=0.030) or average (p=0.044) GC dose, and nearly double as assessed by GC dose intensity (p=0.166) (Table 2). The greater GC exposure in those with IVF occurred primarily in the first 6 months of treatment as shown in Table 3. Those with IVF gained more weight resulting in a higher BMI, especially at 6 months (p=0.010); this difference disappeared by 12 months. Additionally, children with IVF had a greater decline in L-spine aBMD Z-score in the first 6 months of GC therapy (p = 0.048) (Table 2).

L-spine aBMD

A modest decrease in L-spine aBMD Z-score for the entire cohort at the 12 month follow-up is shown in Table 1 and Figure 1 (p<0.001 compared to study entry and also to the healthy average). Low bone mass for chronologic age, defined as a L-spine aBMD Z-score less than −2.0 (26), was found in 19% of patients at 12 months, and in 67% of those with IVF (Tables 1 and 2) compared to 16% of those without (p=0.011). aBMD Z-scores below 0 (the mean for chronological age) at 12 months were present in all of the children with IVF and 78% of those without IVF. Two of 6 children with IVF showed an increase in L-spine aBMD Z-score from baseline to 12 months; this was similar to the 34% of children without IVF who had a rise in L-spine aBMD Z-scores.

Patient with multiple VF requiring bisphosphonate therapy

When the patient treated with intravenous pamidronate was considered along with the children in the natural history cohort, the IVF rate was 6% (95% CI, 2.9–11.7%). This 7 year old girl with mixed connective tissue disease developed multiple symptomatic IVF 4 months post-GC initiation (Figure 3). At enrolment her spine radiograph was normal and her L-spine aBMD Z-score was −1.1. Analysis of her bone health natural history data to 3 months demonstrated GC exposure very similar to the other 6 children with IVF (cumulative GC dose to 3 months was 6,838 mg/m² in this patient, versus 7,186 mg/m² (range 1,613, 9,224) for the other children with IVF and 3,466 mg/m² (range 14.8,19,349) for patients without IVF). She did not differ significantly from the natural history cohort with respect to calcium and vitamin D intake, disease activity, or physical activity. Notable clinical features of this patient included a dramatic increase in her BMI from baseline to 3 months (Δ BMI Z score 3.1 versus mean ± SD 1.1 ± 0.4 for IVF patients and 0.6 ±0.7 for children without IVF), as well as the development of marked cushingoid features.

Figure 3.

Spine radiographs from a 7 year old girl with mixed connective tissue disease. At study entry (A), her spine radiograph showed no signs of vertebral fractures; however, she manifested multiple, painful vertebral fractures during the observation period (B). This patient was excluded from the bone health natural history analyses because she required bisphosphonate therapy to treat the symptomatic fractures; however, a brief description of her skeletal phenotype is provided in the Results.

DISCUSSION

Prospective longitudinal data on bone health outcomes in children with rheumatic diseases are scant, although cross-sectional studies have demonstrated reduced bone mass as well as fragility fractures in this population (4–6, 8–12, 27). Our study provides novel insights into the pace and nature of spine fragility in children with rheumatic disorders treated with GC. The 6% VF incidence was calculated using the standard adult definition of an IVF, with the patient as the study unit, not the individual vertebrae (28). This definition established a priori included both de novo as well worsening of existing VF, since prior VF has been associated with the risk for IVF (29); however, it is not known whether this is the case in children. Incidentally, all of the IVF detected in our cohort were de novo fractures while none were worsening of existing fractures and so this particular risk factor could not be explored. Ongoing follow-up of our cohort will allow us to assess whether the number of children with IVF will indeed increase over time and whether previous fractures will predict future fractures in children.

There was an excellent retention rate (86%) of children in this cohort. When we examined disease subgroups, we found IVF rates of 7% (95% CI, 2.1 to 23.4) in JDM, 15% (95% CI, 5.2 to 36.0) in SLE and 6% (95% CI, 1.0 to 28.3) in systemic vasculitis. Several other reports concur that children with fractures typically have more severe underlying diagnoses such as SLE, JDM, systemic JIA or systemic vasculitis (8, 9). The relative contributions of systemic inflammation severity, the need for high dose GC and other potential risk factors for poor bone health remain undetermined in these disease groups (1, 7–10).

Other publications, based largely on cross-sectional analyses, cite pediatric VF rates ranging between 10 to 34% in rheumatic disorders (6, 8–12). These studies had higher VF rates than we found in our cohort, likely because of their longer period of follow-up, usually about 5 years, as well as selection bias with continued follow-up of children with more severe disease and evaluation of symptomatic children. Previous studies have also reported higher rates of fracture per child, one study with an average of 2.9 fractures per child (8), and one with an average of 3.3 fractures per child (9). Varonos et al. (9) conducted a retrospective study of 23 children with GC-treated JIA (including 19 with systemic arthritis) who had at least one VF and compared these patients to similar-aged children with JIA also treated with GC but who had no history of VF. In this study, the mean time from starting GC to vertebral collapse, as documented on routine yearly spine radiographs, was 2.7 years (range 8 months to 5.5 years). Only a quarter of children who developed VF did so within one year of starting GC. This suggests that our study of the IVF rate at 12 months may have detected only those patients with the greatest predisposition to bone fragility, manifesting as IVF early in the course of disease treatment.

The child who received bisphosphonate treatment during our study was the exception and demonstrated numerous, painful IVF early in her disease course. A few similarly affected children, each with 10 fractures, have been reported (8, 9). Our patient with multiple IVF received a similar cumulative GC dose compared to the other 6 children with 1 to 2 fractures each; however, she had a greater increase in BMI in the first 3 months of GC treatment compared with the other 6 children with IVF. Her early, marked cushingoid features raise the possibility of individual genetic susceptibility to GC.

The location and severity of IVF were consistent with findings from other pediatric studies, with a bimodal mid-thoracic and thoracolumbar junction distribution (2, 6, 8, 9). Typically most of the burden was in the mid-thoracic spine and the majority of the fractures were mild to moderate with anterior wedge morphology (2, 6, 8, 9). A proposed explanation for this location relates to different mechanical forces along the spine associated with its normal kyphosis (30).

Despite the robust size of our cohort and the acquisition of prospective longitudinal data, our ability to derive predictors of fracture risk was limited by the small number of patients with IVF. Those with IVF received more GC in the first 6 months of treatment (i.e. higher GC intensity early on), had a somewhat higher Δ BMI at 6 months and also worsening of their L-spine aBMD Z-score over the same time period. We were not able to delineate the relative contributions of disease severity or cumulative GC dose to IVF. This is made challenging by the fact that GC dose may be a marker of an underlying disease variable such as the degree of systemic inflammation. The cumulative GC dose was strongly associated with prevalent VF in a study of 94 children with a variety of rheumatic diseases followed at a single centre (8). There are, however, conflicting reports in the literature on the correlation of GC dose with the prevalence of VF and reduced bone mass in children with rheumatic diseases (6, 8). This may, in part, reflect the heterogeneity of the disease populations studied, as well as the varied duration of diseases and their treatment in the various reports.

Based on observations in adults with VF, there is a suggestion of a threshold for lower BMD predisposing to fractures (31). Similarly, in children with rheumatic diseases a threshold Z-score was observed for volumetric L-spine BMD of −1.8 predicting fractures (11). In our study, we noted that 67% of those with IVF had a BMD Z-score less than −2.0 compared to 16% of those without IVF. Given the small number of children with IVF in this study, we were unable to determine if a precise spine BMD threshold exists.

On average we found a modest decrease in the L-spine aBMD Z-score over 12 months post-GC initiation. Data from most longitudinal studies are not directly comparable to ours because the commencement of their follow-up occurred on average 2 years following GC initiation (5, 7, 32). Within the whole cohort, 19% had L-spine aBMD Z-scores less than −2.0, which has been referred to as “low bone mass for chronologic age” in the literature (10, 26). In particular, approximately one quarter of children with juvenile JDM and systemic vasculitis had a L-spine aBMD Z-score less than −2.0 at 12 months, as did approximately 20% of those with systemic JIA. Ongoing longitudinal follow-up of our cohort is essential to help understand the significance of these spine BMD findings, particularly in relation to IVF.

Within the IVF group, only a small number complained of back pain or tenderness, which is consistent with previous studies reporting that many children who experience IVF are asymptomatic, or if symptoms are present they are mild and may be transient (6, 12). This is in direct contrast to our cohort at enrolment, when back pain was found to be highly associated with prevalent VF (2). Gender was not associated with IVF nor was the use of methotrexate. The role of methotrexate as osteotoxic or osteoprotective has not been determined, as there are conflicting reports regarding its association with IVF (8, 11).

There are limitations to our study which deserve consideration. First, since back pain was only assessed on a 3 monthly basis, it is possible that the frequency of transient pain may have been underestimated. Second, daily divided GC dosing was not captured, and therefore we could not determine if the frequency of the dosing in a 24 hour period played a role in the development of IVF (9). Moreover, we did not ascertain precise data on ancillary medications such as disease modifying anti-rheumatic drugs other than methotrexate that might have provided additional proxy data on the severity of the underlying illness. It is also possible that the use of potent biologic agents may have significantly reduced the risks of IVF, by the inhibition of cytokines known to promote bone resorption (33). This cohort of children is quite heterogeneous in regard to underlying diagnoses, potentially limiting our power to see differences. Finally, serum 25-hydroxyvitamin D levels were not available; the deliberate rationale for excluding this parameter has been outlined previously (2).

In summary, the incidence of VF at 12 months post-GC initiation in children with rheumatic diseases was 6% and most children were asymptomatic. One child was more severely affected with multiple painful fractures. Overall, there was a suggestion that children with IVF received more GC during the observation period and had greater increases in BMI and declines in L-spine aBMD Z-scores in the first 6 months of GC treatment. The longitudinal nature of this study with further follow-up should provide much needed insights into the future risk for VF in this context.

Significance and Innovations.

• This prospective pediatric study provides novel insights into the timing and nature of spine fragility following glucocorticoid initiation in children with rheumatic diseases.

• A sub-set of children with rheumatic conditions has the potential to develop incident vertebral fractures within 12 months of glucocorticoid initiation; those who did typically manifested a single asymptomatic incident vertebral fracture.

• One child developed multiple, painful incident vertebral fractures 4 months following glucocorticoid initiation, prompting intervention with an intravenous bisphosphonate.

• Children with incident vertebral fractures tended to receive higher doses of glucocorticoid therapy, particularly in the first 6 months of treatment, and also had greater increases in body mass index Z-scores and declines in lumbar spine areal bone mineral density Z-scores.

Abbreviations

aBMD

Areal bone mineral density

BMI

Body mass index

CI

Confidence interval

JDM

Juvenile dermatomyositis

GC

Glucocorticoid

IVF

Incident vertebral fracture

JIA

Juvenile idiopathic arthritis

STOPP

STeroid-associated Osteoporosis in the Pediatric Population

SLE

Systemic lupus erythematosus

VAS

Visual analogue scale

VF

Vertebral fracture

The Canadian STeroid-associated Osteoporosis in the Pediatric Population (STOPP) Consortium (a pan-Canadian, pediatric bone health working group)

Co-ordinating Center

Children’s Hospital of Eastern Ontario, Ottawa, Ontario: Leanne M. Ward^#,*,§ (Study Principal Investigator), Janusz Feber^*,§ (Nephrology), Jacqueline Halton^*,§ (Oncology), Roman Jurencak (Rheumatology), MaryAnn Matzinger (Radiology, Central Radiograph Analyses), Johannes Roth (Rheumatology), Nazih Shenouda^§ (Radiology, Central Radiograph Analyses)

Ottawa Hospital Research Institute, Ottawa Methods Centre Ottawa, Ontario: David Moher^*,§, Tim Ramsay

Participating Centers

Alberta Children’s Hospital, Calgary, Alberta: David Stephure (Site Principal Investigator), Reinhard Kloiber (Radiology), Victor Lewis (Oncology), Julian Midgley (Nephrology), Paivi Miettunen (Rheumatology)

British Columbia Children’s Hospital, Vancouver, British Columbia: David Cabral^* (Site Principal Investigator), David B. Dix (Oncology), Kristin Houghton (Rheumatology), Helen R. Nadel (Radiology)

British Columbia Women’s Hospital and Health Sciences Center, Vancouver, British Columbia: Brian C. Lentle^§ (Radiology)

Brock University, Faculty of Applied Health Sciences, St. Catharines, Ontario: John Hay^§ (Physical Activity Measurements)

Children’s Hospital, London Health Sciences Centre, University of Western Ontario, London, Ontario: Robert Stein (Site Principal Investigator), Elizabeth Cairney (Oncology), Cheril Clarson (Bone Health), Guido Filler (Nephrology) ^§, Joanne Grimmer (Nephrology), Keith Sparrow (Radiology)

IWK Health Center, Halifax, Nova Scotia: Elizabeth Cummings (Site Principal Investigator), Conrad Fernandez (Oncology), Adam M. Huber^§ (Rheumatology), Bianca Lang^*,§ (Rheumatology), Kathy O’Brien (Radiology)

McMaster Children’s Hospital, Hamilton, Ontario: Stephanie Atkinson^*,§ (Site Principal Investigator), Steve Arora (Nephrology), Ronald Barr^§ (Oncology), Craig Coblentz (Radiology), Peter B. Dent (Rheumatology), Maggie Larche (Rheumatology), Colin Webber^* (DXA Methodology),

Montréal Children’s Hospital, Montréal, Québec: Celia Rodd^§ (Site Principal Investigator), Sharon Abish (Oncology), Lorraine Bell (Nephrology), Rosie Scuccimarri (Rheumatology)

Shriners Hospital for Children, Montréal, Québec: Frank Rauch^*,§ (Co-Chair, Publications and Presentations Committee and Ancillary Studies Committee), Francis Glorieux^*

Ste. Justine Hospital, Montréal, Québec: Nathalie Alos^* (Site Principal Investigator), Josée Dubois (Radiology), Caroline Laverdière (Oncology), Véronique Phan (Nephrology), Claire Saint-Cyr (Rheumatology)

Stollery Children’s Hospital, Edmonton, Alberta: Robert Couch^* (Site Principal Investigator), Janet Ellsworth (Rheumatology), Claire LeBlanc (Rheumatology), Maury Pinsk (Nephrology), Kerry Siminoski^§ (Radiology), Beverly Wilson (Oncology)

Toronto Hospital for Sick Children, Toronto, Ontario: Ronald Grant^* (Site Principal Investigator), Martin Charron (Radiology), Diane Hebert (Nephrology)

Université de Sherbrooke, Department of Family Medicine, Sherbrooke, Québec: Isabelle Gaboury^*,§ (Biostatistics)

Winnipeg Children’s Hospital, Winnipeg, Manitoba: Shayne Taback^§ (Site Principal Investigator), Tom Blydt-Hansen (Nephrology), Sara Israels (Oncology), Kiem Oen (Rheumatology), Martin Reed (Radiology)