Biochemical Markers of Bone Turnover Associated with Calcium Supplementation in Children with Juvenile Rheumatoid Arthritis

Abstract

OBJECTIVE

To determine the effects of calcium supplementation on bone physiology in corticosteroid-free juvenile rheumatoid arthritis (JRA) children by measuring serum and urinary bone-related hormones, minerals, bone formation and resorption markers.

METHODS

In this double-blind trial, patients were randomized to receive daily oral supplementation of 1000 mg of calcium and 400 IU vitamin D or placebo and 400 IU vitamin D for 24 months. The physiologic effect of calcium supplementation on bone physiology was followed periodically using markers of bone turnover.

RESULTS

198 patients met inclusion criteria and were followed in the study. At baseline there were no differences in markers of bone turnover between the two groups. Subjects with <4 active joints had higher serum calcium and higher PTH. Individuals receiving calcium with <4 active joints had lower osteocalcin. At follow-up1,25 (OH)2 vitamin D, PTH, osteocalcin and urine phosphorus were lower in the calcium supplementation group. Hypercalciuria noted with urine Ca/Creatinine were not noted on 24-hour urine studies.

CONCLUSIONS

Markers of bone physiology were significantly decreased in children with JRA receiving calcium supplementation. The physiologic changes were noted as early as 12 months into calcium supplementation. The hypercalciuria noted on spot testing did not correlate with further evaluation nor did it lead to renal pathology. These findings suggesting that the calcium supplementation met physiologic needs and caused an increased calcium loss in urine.

Juvenile rheumatoid arthritis (JRA) is the most common chronic rheumatic disorder of childhood. The prevalence is estimated to be anywhere between 10–90 cases per 100 000 children under the age of 16(1). Approximately 15–50% will have problems achieving normal bone mineral content. In older literature, 15–26% of children with JRA demonstrated radiologic findings of osteoporosis such as long bone and vertebral crush fractures (2–4). In 65 adults with a history of JRA, 43% had decreased bone mineral density (BMD) of the lumbar spine and 53% had decreased BMD in the hip (5). In recent studies up to 50% JRA patients will have osteopenia (5–7). These studies were cross-sectional and most included patients on glucocorticoids or a history of long-standing corticosteroid use. The effects of glucocorticoids on bone accretion are well-documented (8–12). A recent study evaluated glucocorticoid-free JRA patients and noted that approximately 30% of these individuals had osteopenia (13). JRA persists into adulthood in 30–50% of patients (14). It is hoped that the newer biologic therapies that result in excellent disease control in many children with JRA and the decreased need for corticosteroid therapy will result in less frequent osteopenia. However, attention to bone mineralization will continue to be an issue in pediatric rheumatology since more than 90% of peak bone mass is laid down by the second decade of life (15).

Calcium supplementation (16–18), diets with increased calcium (19) and physical activity (20, 21) have been demonstrated to improve bone mineral density in healthy children. Supplementation with vitamin D as monotherapy does not increase BMD in JRA (22). However, in a small, cross over study in children with a variety of rheumatic diseases with documented osteopenia, the combination of 1000 mg of calcium and 400 IU vitamin D resulted in significantly elevated BMD (23). In a large, prospective, randomized, placebo controlled, double blind study in children with JRA, 1000 mg of calcium carbonate and 400 IU of vitamin D per day for 24 months produced a small, but significant increase in BMD compared to placebo and 400 IU vitamin D per day (24). A significant increase in BMD was seen in a randomized trial to assess the effectiveness of a behavioral intervention to increase dietary intake in children with JRA (25).

Prior studies of laboratory markers of bone mineralization and physiology in children with JRA have demonstrated varying results (23, 26–30). Reed et al found that children with active rheumatic disease had decreased levels of osteocalcin (osteocalcin) (27) but ionized calcium, vitamin D and parathyroid hormone (PTH) were not reduced. In a study of children with juvenile chronic arthritis (JCA) at baseline and after one year of follow-up the alkaline phosphatase, phosphate and calcium levels were normal (31). Systemic JRA patients had decreased 25-(OH)-D and 1,25-(OH)2-D levels at study entry and at one year follow-up. The 25-(OH)-D levels were decreased in the polyarticular group at baseline and at one year follow-up in the other JCA subsets. In another study, all the children with JRA demonstrated normal calcium, phosphate and OC levels, but above normal alkaline phosphatase levels (30). Falcini and colleagues observed that controls and JRA children had similar levels of Ca, phosphorus, Mg, alkaline phosphatase, PTH and vitamin D (32). A subset of children with active JRA had lower carboxyterminal telopeptide of type I collagen and OC compared to inactive JRA children. These studies reveal the variability of markers of bone turnover at baseline and in short-term follow-up. A few studies have also shown hypercalciuria in children with JRA (26, 27, 33). Overall the data on markers of bone turnover in JRA with or without calcium supplementation are not consistent and controversial at best. The vast majority of these studies included subjects with varying exposure to corticosteroid therapy.

The objective of this study was to determine the effect of calcium supplementation on bone physiology in children with JRA by measuring serum and urinary bone-related hormones (parathyroid, vitamin D’s), minerals (calcium, phosphorus), bone formation markers (osteocalcin, skeletal alkaline phosphatase) and bone resorption markers (urinary calcium/creatinine and pyridinoline crosslinks).

Materials and methods

Subjects

Study subjects were recruited from 6 pediatric rheumatology clinical centers. Patient enrollment and study evaluations took place at the study coordinating center at the General Clinical Research Center, Cincinnati Children’s Hospital Medical Center (CCHMC). The CCHMC Institutional Review Board (IRB) approved the study. Informed consent and assent was obtained as per CCHMC IRB guidelines prior to performing any study related investigation.

To be eligible for the study subjects had to be 6–18 years old, meet the American College of Rheumatology classification criteria for JRA (34), and had to be free of systemic corticosteroid therapy in the 3 months prior to enrollment. Patients were excluded from the study if they were taking calcium supplements, calcium containing antacids, or oral contraceptives. In addition, individuals who smoked, had ever been or were pregnant, had a chronic illness that affected growth or bone mineralization (e.g. Trisomy 21, inflammatory bowel disease, thyroid disease), or fasting random urinary calcium/creatinine ratio > 0.2 were excluded from the study. Exclusion of patients with these factors was done to allow for a non-confounded documentation of the physiologic effects of calcium supplementation in children with JRA.

The study enrolled 198 children with JRA (141 females, 57 males) 6 to 18 years old. Past and current diets were assessed with the Youth and Adolescent Food Frequency Questionnaire (YAQ) and a 3-day diet diary (35). Physical activity was documented with a validated standardized three-day physical activity diary based on a modification of the 7-day Stanford Physical Activity Recall Record (36, 37). Study visits occurred every 6 months for 2 years during the double-blind placebo-controlled intervention trial. Study visits also occurred 6 and 18 months after the end of the intervention trial.

Randomization of Calcium Intervention

Block randomization was utilized with stratification according to study center to assign study subjects in a double blind fashion to either the placebo group or the calcium supplementation group. The randomization schedule was developed and maintained at the coordinating center (CCHMC). The 500 mg calcium carbonate tablets (Tums®) and placebo tablets were matched for size, appearance, texture and taste (provided by GlaxoSmithKline). . Two tablets were given as a single dose for 1000 mg oral elemental calcium as CaCO₃ in the active arm and 0 mg CaCO₃ in the placebo arm. Both groups received adequate vitamin D in the form of one chewable multivitamin tablet (Vi-Daylin Multivitamin Chewable®, Ross Products Division of Abbott Laboratories) containing 400 IU vitamin D (100% RDA for vitamin D). Study medications were dispensed at 6 month intervals during study visits to CCHMC. The CCHMC Research Pharmacy was responsible for storage and blinded distribution of study medications to maintain a double-blind standard.

Clinical Assessment

At each study visit, growth and development were assessed with a general medical history, general and joint physical exam. The same pediatric rheumatologist (DJL) performed the history, general and rheumatologic physical exam in all patients at all study visits. Growth parameters measured were height (to nearest 0.5cm via Harpenden stadiometer), weight (to nearest 0.1kg), body mass index (BMI) and date of onset of menses. Tanner pubic hair and breast sexual stage exams were performed by the same study personnel throughout the trial (JR, DJL). Tanner exams were performed by a same sex examiner.

Disease severity was determined based on the number of joints with active arthritis (defined as joint swelling or, in the absence of joint swelling, limitation of motion with pain on motion and/or tenderness). Additional evaluations included: a) the Juvenile Arthritis Functional Assessment Report (JAFAR), a measure of physical function (theoretical range 0 best, 46 worst) (38), b) parent assessment of the child’s overall well being measured on a 10 cm visual analog scale (VAS) with `very poor' at the left end (scored 0) and `very well' at the right end (scored 100), c) articular severity score as per prior studies (39), d) ACR Functional Class, and e) physician assessment of disease activity (1 - Active, 2 - Partial Remission, 3 - Total Remission).

Laboratory Methods

Blood and urine samples were collected during study visits (0, 12, and 24 months) and during the 18 month open follow-up. Blood was drawn in the morning after the patients had an overnight fasting period, which served to minimize diurnal effects. The blood was allowed to clot at room temperature and then serum was separated, aliquoted and frozen at −70°C. A 24-hour urine sample was collected to evaluate the calcium physiology (e.g. calcium, phosphorus, creatinine, pyridinoline) and evaluate for hypercalciuria. Fasting random urine samples were collected at each visit to determine Ca/creatinine to assess hypercalciuria.

Serum samples were analyzed for total calcium (Ca), ionized calcium (iCa), 25-hydroxyvitamin D (25(OH)D), 1,25-dihydroxyvitamin D (1,25(OH)₂D), osteocalcin (OC), phosphorus, bone-specific isoenzyme of alkaline phosphatase (sALP), magnesium, carboxyterminal propeptide of human type I collagen (PICP), and PTH. Urine samples were analyzed for calcium, creatinine, cyclic 3’, 5’ adenosine monophosphate (cAMP), deoxypyrdinoline (D-PYD), pyridinoline (PYD), phosphorus (P), and magnesium.

The NIH funded CCHMC Clinical Research Center Bone and Mineral Laboratory performed laboratory analysis for the urine and blood samples obtained. All blood and urine samples, except PTH and ionized Ca, were analyzed in batches. Laboratory testing methods as described in Lovell et al (24)

Bone Mineral Density

Total BMD and BMC were determined utilizing DXA (Hologic 2000, Waltham, MA) as described in Lovell et al (24).

Compliance

Significant effort was made in this study to maximize compliance with taking the calcium and vitamin D supplements so as to be able to measure effectively the true physiologic effect of calcium supplementation. Study medications were given in a single daily dose. During the initial visit the reasons and methods for measuring compliance were reviewed with parents and patients. Families received verbal and printed instructions describing the study and study medications. Families were contacted by the study coordinator (who remained blinded to treatment group assignment) by phone throughout the study to encourage improved compliance and encourage continued participation. Parents and patients received compliance counseling at study entry and at each study visit.

Several methods to measure compliance were used in the study. Pill counts were conducted at each visit. The Medication Event Monitoring System (MEMS) Cap (Aprex Corporation) method has been shown to be a reliable indicator of medication dosing (40–42). The MEMS Caps provided a visual display of the number of times the study medication bottle had been opened in a 24-hour period and the number of hours since it was last opened. Data was stored in the MEMS Caps of the date and time of all pill bottle openings between study visits. This longitudinal data was downloaded, shared and discussed with the study subject and parents at each study visit. MEMS monitoring and pill counts allowed for determination of concordance between pill counts and number of bottle openings.

Statistical Analysis

In order to determine the effect of calcium supplementation on bone physiology in the study subjects, the analyses were done using an intention-to-treat (ITT) approach. The ITT population was defined as all randomized subjects who received at least one dose of study medication. Study group characteristics at base line and follow up visits, rates of loss to follow-up, and laboratory measurements were compared between the Ca and placebo group using Chi-square test or Fisher’s exact test for categorical variables, student T-test for continuous scale variables and Wilcoxon rank sum test for non-normally distributed continuous variables. The primary study outcomes for this study were: bone physiology measures from the blind phase (24 month) of intervention, including serum and urinary bone-related hormones, minerals, bone formation markers and bone resorption markers. These parameters at baseline were assessed for variability due to demographic characteristics (gender, race, Tanner stage, and age groups) using one-way analysis of variance (ANOVA). Subjects were dichotomized based on number of active joints (≤4 active joints and >4 active joints) to evaluate the associations between the bone physiology measures and disease severity. Generalized estimation equations (GEE) were used to analyze this longitudinal data in order to assess the effect of the intervention (calcium vs. placebo) on the bone physiology. The GEE analyses included the subject’s baseline value on the study parameters: baseline disease activities, study visit, and intervention group effect. Potential interaction effects were also tested. A two sided p value ≤ 0.05 was considered significant. Analyses were performed with SAS, Version 9.1 (SAS Institute Inc., Cary, NC, USA).

Results

Demographics

The baseline characteristics of the study population (198 subjects) are summarized in Table 1. The age ranged from 6 to 18 years, with a mean of 11.7 years. Baseline characteristics between the active and placebo groups were not significantly different except for gender and Tanner stage distribution. The placebo group had more males and prepubertal subjects than the calcium group (p values 0.02 and 0.05, respectively). Measures of disease activity (articular severity score, ACR Functional Class, and physician assessment of disease activity), methotrexate use, and disease duration were evenly distributed between the groups. The mean methotrexate dose at baseline (95% CIs) for the Ca group was 4.90 mg/week (3.39, 6.41) and for the placebo group was 4.88 mg/week (3.43, 6.33). There were no differences in physical activity or diet (calcium and vitamin D intake) between the groups at baseline.

Table 1.

Baseline Demographics of the children enrolled in the JRA Calcium Trial

Characteristic	Calcium (n = 103)	Placebo (n = 95)	P value
Sex, no. (%)
Female	81 (78.6)	60 (63.2)	0.02a
Mean age, years (SD)	11.8 (3.1)	11.6 (3.4)	0.60b
Ethnicity (%)			0.39c
Caucasian	95 (92.2)	90 (94.7)
Black	6 (5.8)	5 (5.3)
Other	2 (1.9)	0 (0.0)
JRA onset type, N (%)			0.51a
Pauciarticular	49 (48.0)	52 (55.3)
Polyarticular	39 (38.2)	33 (35.1)
Systemic	14 (13.7)	9 (9.6)
Mean duration of JRA, years (SD)	5.21 (3.75)	5.94 ( 3.62)	0.11b
JRA activity			0.96a
≤ 4 joints	68 (66.0)	63 (66.3)
>4 joints	35 (34.0)	32 (33.7)
Joints with active arthritis Mean (SD)	7.11 (11.3)	7.44 (13.9)	0.87b
ANA positive, N. (%)	48 (46.6)	54 (56.6)	0.17a
Tanner Stage, N (%)			0.05a
1	34 (35.4%)	47 (52.8%)
2	17 (17.7%)	8 (9.0%)
3	12 (12.5%)	4 (4.5%)
4	15 (15.6%)	11 (12.4%)
5	18 (18.8%)	19 (21.4%)
Methotrexate use, N. (%)	37 (35.9)	35 (36.8%)	0.89b
Mean dose, mg/week (95% CI)	4.90 (3.39,6.41)	4.88 (3.43,6.33)	0.88b
Number of joints with active arthritis, mean (median)	7.1 (2)	7.4 (2)	0.79d
Number of swollen joints, mean (median)	3.9 (1)	4.4 (1)	0.64d
JAFAR score, mean (median)	2.6 (0)	2.2 (0)	0.49d
VAS, mean (median)	8.2 (9)	8.3 (9.1)	0.79d

JRA = juvenile rheumatoid arthritis, ANA = antinuclear antibody, SD = standard deviation, CI = confidence interval, JAFAR = Juvenile Arthritis Functional Assessment Report, VAS = 10 cm visual analog scale

^aChi-square test

^bStudent-T test

^cFisher’s exact test

^dKruskal-Wallis test

Compliance

Comparison of compliance as measured by the MEMS Smart Caps demonstrated an overall high rate of study medication compliance in both the calcium and placebo groups. The overall compliance for the Ca and placebo subjects was 84.6% and 86.5% respectively. The correlation between pill counts and MEMS was 0.6111. The drop-out rates were not significantly different between the two groups during the study period (Figure 4). By the end of the 2 year study, approximately 28% of the patients had dropped out. Subjects who dropped were compared to the remaining subjects at each visit in order to ensure group comparability over time. No evidence was found to suggest informative drop-out (i.e. subjects dropping out of the study did not differ from those continuing in the study on values of study outcomes at the time of discontinuation from study).

Figure 4.

Lost to Follow-up

Number of patients lost to follow-up at 6 month visits (0, 6, 12, 18 and 24 months). By 24 months 28% of patients had dropped out. The rate of drop-out between the between the CALCIUM and groups was not significantly different. No evidence to support informative dropout was noted between the groups.

Bone Mineral Density

Logarithmic transformation was used in the longitudinal analyses of Ca/Cr urine and bone mineral density (BMD) (Figure 1 & Figure 3). There were no significant differences in the total BMD and bone turnover variables at baseline between the groups. Figure 1 demonstrates how the average total BMD increased linearly over time in both study groups. Overall the Ca group increased slightly faster than the placebo group. This difference is most evident at the end of the second year. As seen in Figure 1, Ca supplementation increased total body BMD by a small but statistically significant amount over placebo.

Figure 1.

Total Body BMD Profile Changes in Total Body bone mineral density (BMD) mean plot for ITT data (167 subjects) over 24 month period. BMD taken at baseline, 6, 12, 18 and 24 months. Increased total BMD noted in all patients (--▲--), but more notable in calcium (--♦--) group compared to placebo () group.

Figure 3.

Changes in urine markers of bone turnover over time. Mean plot for ITT data (167 subjects) *Urine Ca/Cr Geometric mean was used

Laboratory Results

At baseline 16.7% of the study subjects (calcium and placebo) had an elevated sALP. By the 24 month, 68.3% of subjects had an elevated sALP. One patient had an elevated PTH at baseline and at the 24 month follow-up. One other subject had an elevated PTH at 24 month. No patients had low serum Ca levels at study entry. Seven patients had elevated serum Ca, two at the 12 month and five at the 24 month. Thirteen patients had low serum Ca, seven at the 12 month and six at the 24 month. Four (2%) JRA subjects had decreased 25(OH) vitamin D levels: four at enrollment and one of these again at the 12 month. All remaining patients had normal 25(OH) vitamin D levels at the 24 month visit. No subjects with abnormal serum Ca had abnormal 25(OH) vitamin D and vice-versa. One subject had an elevated serum creatinine which was normal on repeat studies. Abnormal creatinine levels at baseline normalized upon repeat analysis and remained within normal at subsequent visits for all subjects. Using urine Ca/Cr >0.2 as a measure of hypercalciuria, ten (5%) of the subjects had an elevated Ca/Cr ratio at baseline. During the treatment phase 93 (47%) patients had an elevated urine Ca/Cr. Of these patients 17 (9%) had more than one occurrence of an elevated urine Ca/Cr. In contrast, 24-hour urine collections as a measure of hypercalciuria did not reveal elevated urine calcium levels in any subject. No subject demonstrated kidney stones or calcium related hematuria during the study. Osteocalcin levels were below normal for all subjects during first 18 months. In the last visit two subjects had OC levels within normal. Published reference ranges for the OC assay used in this study were utilized to evaluate OC levels (43).

Disease Activity

The Spearman correlations between the number of active joints and the biochemical markers of bone turnover at the baseline did not demonstrate any significant linear association between the total number of active joints and the bone turnover markers. The number of active joints was then dichotomized (≤4 active joints and >4 active joints). At baseline serum calcium levels were higher in subjects with <4 joints compared to subjects with ≥ 4 active joints (9.96+/− 0.7 in lower disease activity vs. 9.75+/−0.33 in higher disease activity group, Student-t; p = 0.013). In GEE analysis over 24 months, the lower disease activity group demonstrated a higher PTH (p = 0.03) compared to the higher disease activity after taking into account the intervention effect and patients’ baseline PTH. The GEE analysis also demonstrated that the lower disease activity group, who received calcium supplementation had a decreased OC (p = 0.009), and an increased total BMD (p = 0.003) compared to subjects with lower disease activity who did not receive supplementation received placebo. No such findings were noted in the higher disease activity (>4 active joints) subjects regardless of calcium or placebo intervention.

Markers of Bone Turnover

Evaluation of the biochemical markers of bone turnover in the ITT subjects did not demonstrate significant differences in the sALP, type I procollagen, pyridinoline, deoxypyridinoline, serum Ca, serum phosphorus, magnesium, cAMP or 25(OH) vitamin D between the calcium and placebo groups. Statistical analysis revealed that there were overall effects of treatment for 1, 25(OH)₂ vitamin D, PTH, urine phosphorus, urine calcium, and urine Ca/Cr (Tables 2). Study subjects receiving calcium supplementation had higher urine calcium levels (p = 0.02) and higher Ca/Cr ratios (p < 0.0001). The active treatment group had significantly lower 1,25(OH)₂ vitamin D (p = 0.005), PTH (p = 0.004), urine phosphorus (p = 0.01), and OC (p = 0.02). Differences in the two groups on these bone markers were noted as early as the 12 month visit.

Table 2^*.

The Least Square Means (LSM) GEE models for Bone Marker

Variable	Calcium, mean(SD)	Placebo, mean(SD)	LSM of Calcium vs Placebo	P value
Vitamin D 1,25 (pg/ml)	44.35 (1.36)	50.81 (1.45)	−6.46 (1.97)	0.001
PTH (pg/ml)	20.05 (0.77)	23.53 (0.93)	−3.48 (1.20)	0.004
Phosphorus urine (mg/dl)	98.23 (5.29)	115.42 (4.33)	−17.19 (6.83)	0.01
Osteocalcin (ng/ml)	11.62 (0.36)	12.92 (0.43)	−1.30 (0.56)	0.02
Calcium urine (mg/dl)	25.31 (0.96)	22.06 (0.94)	3.25 (1.34)	0.015
Log (Ca/Cr urine)	−1.74 (0.06)	−2.05 (0.05)	0.31 (0.08)	<0.0001

^*GEE (generalized estimation equation) models were controlled by baseline outcome value and visit number. The effect of disease activity at baseline (≤4 active joints vs. >4 active joints), and whether the intervention group effect may differ by patients’ baseline disease activity were tested.

Discussion

Individuals with JRA are at an increased risk for problems with normal bone mineralization. Individuals with a history of JRA who have had inactive disease for prolonged periods have demonstrated significantly decreased BMD (5). Calcium supplementation has been demonstrated to increase bone mineral density in JRA subjects with previously identified osteopenia (23) and in children with JIA not previously treated with steroids (24). In order to better understand the effect of JRA and calcium supplementation on the underlying physiology of bone metabolism, we excluded individuals on medications (including corticosteroids) and those with conditions known to adversely affect bone mineral density.

Medication compliance plays an important role in the efficacy of medications. In fact, it is often a silent confounder in demonstrating the full effect of therapeutic interventions. In this trial, focused efforts were made to maximize and quantitate compliance to effectively measure physiologic changes as a result of calcium supplementation and placebo. By not quantifying the effects of compliance our results would reveal the averages of compliance and efficacy. The overall correlation between pill count and MEMS caps was significant but in those with discrepant scores it may reflect factors that each separately cannot account for. An example of such a factor is opening the bottle for a reason other than accessing the medication (44). Overall the compliance in both the treatment groups was excellent; especially considering that the treatment was directed towards prevention of an asymptomatic condition. These efforts to increase compliance are not usually performed in routine clinical practice and thus the results seen in this study in calcium intake and thus BMD most likely exceed those that will be achievable with more standard rates of compliance with taking calcium.

In previous studies hypercalciuria had been noted in a small number of patients with JRA. In the present study, approximately 5% of the study population was found to have hypercalcuria at baseline. Calcium supplementation, in this study, was not associated with kidney stones, hematuria or asymptomatic hypercalciuria on 24 hour urine collections.

In studies of healthy children and children with JRA, vitamin D deficiency has been noted. However, in this study population only 3 subjects were noted to have low levels of vitamin D 25(OH) at baseline. No study subject demonstrated hypocalcemia at any time during the study. Overall, the vitamin D and calcium levels were normal in the study population.

Disease activity using the number of active joints did not demonstrate a correlation with markers of bone turnover at baseline. However, dichotomization of the subjects based on disease activity revealed that the lower disease activity group (≤4 active joints) had higher serum calcium at baseline. Over the 24-month intervention period, the lower disease activity group that received calcium had a higher PTH, a markedly higher total BMD and lower osteocalcin (bone formation). Before intervention the lower disease activity subjects had higher serum calcium, but after calcium supplementation they had lower bone formation and more hormonal stimulation for calcium. Even without accounting for disease activity, we found that Ca supplementation increased total body BMD by a small but statistically significant amount over placebo.

Evaluation of the overall effects of calcium supplementation revealed that some of the markers of bone turnover were significantly different between the two groups. These differences were seen as early as 12 months into the study. Lower levels of PTH, vitamin D 1,25 (OH) and osteocalcin were seen in the calcium supplement group compared to the placebo group indicating that the physiologic needs had been more adequately met in the calcium intervention group. The lower PTH levels were associated with lower levels of phosphorus as was seen in the calcium group. The urine calcium excretion and urine Ca/Cr were slightly higher in the calcium group compared to the placebo group but both groups remained without evidence of clinical pathology. Since the physiologic calcium requirement had been more than adequately addressed, then there was a slight but not clinically important increase in the amount of calcium lost in the urine.

In the present study, the laboratory assessments indicated that the calcium supplementation as performed in this trial was associated with physiologic saturation of the bone mineralization process as early as 12 months into the trial. In accordance with these laboratory markers, the rate of bone mineralization was initially greater in the calcium group but by 18 months the rate of increase of BMD was similar between the two treatment groups. Lower disease activity patients had evidence of lower bone formation. The calcium intervention group also demonstrated less bone formation as early as 12 months into intervention.

More evidence has accumulated on the role of new and known markers of bone metabolism. The study population was fairly homogenous in demographic make-up. The majority of patients had mild to moderate JRA severity. JRA subjects on corticosteroids were excluded but methotrexate use was permitted. The latter treatment may play a role in bone metabolism and BMD which has yet to be clearly elucidated. Corticosteroid effects on bone physiology and subsequent BMD consequences are better understood. Thus this study may underestimate the effects of osteopenia and osteoporosis on bone physiology.

Previous studies have demonstrated reduced bone formation in JRA patients. We also demonstrated similar findings using markers of bone mineralization in JRA patients on calcium supplementation. The majority of patients in this study had mild to moderate disease and yet almost all patients had osteocalcin levels below the reference range throughout the study. These patients had increased calcium loss in urine and a reduced rate of bone turnover in particular bone formation. Subjects on calcium had increased BMD compared to placebo patients, even with evidence of decreased bone formation.

Dietary supplementation with calcium in this study population of JRA patients with mild to moderate disease provided a small but measurable increase in BMD and met biochemical needs, evident as early as 1 year. However, this increase was seen in these subjects in whom the baseline BMD measurement was normal. Effect of this small increase in BMD over a two year study may or may not be related to a change in the frequency of the clinically important event of bony fractures. Given the high frequency of normal BMD at baseline, normal progression seen in those on placebo and significant but small increase seen in BMD in those on Ca supplement, the authors do not recommend routine calcium supplementation for children with noncorticosteroid treated JRA. The use of corticosteroids has significant effect of bone mineralization these study results do not address that JRA population. In addition, future studies of bone mineralization in JRA that include subjects on the newer biologic anti-inflammatory agents, incorporate the new markers of bone metabolism, and other methods of assessing bone mineralization will be necessary to continue to provide timely information regarding the role of calcium supplementation in the care of children with JRA.

Figure 2.

Serum Markers of Bone Turnover Mean plot for ITT data (167 subjects)