Vitamin D3 placebo administered orally every 60 days plus calcium placebo twice daily.
A.2.c. Entry and randomization visit:
All eligible subjects will have the initial study entry evaluation performed at the SLRHC Pediatric Body Composition Unit by virtue of their participation in the cross-sectional study (Section D.2.f). As the qualifying serum 25(OH)D level will be known prior to the cross-sectional study visit, study participants eligible for the randomized clinical trial can be randomly assigned (see Methods) during the cross-sectional study visit to one of the two treatment arms; in such cases, the cross-sectional visit will also function as the entry/baseline visit. Each subject will be given the first dose of vitamin D3/placebo and the first 30 day supply of calcium/placebo will be dispensed. females of child bearing capacity
A.2.d. Follow-up Visits:
Participants will be seen monthly during the first year of the study and every two months during the second year. Entry/baseline, 12-month and 24-month visits will take place at the SLRHC Body Composition Unit. All other visits will take place at the site where usual care is received. Vitamin D3/placebo will be administered during study visits every 2 months. At monthly visits during the first year and every 2 months during the second year after randomization, a 30 or 60 day supply of calcium/placebo will be dispensed. Adherence to calcium supplementation will be assessed at each visit.
A.2.e. 12 and 24 month study visits:
These visits will take place at the SLRHC Body Composition Unit. At each visit, the following tests will be performed:
Medical history, including medications Pregnancy test for females of child bearing capacity Height and weight Assessment of pubertal stage Assessment of dietary intake of calcium and vitamin D Assessment of physical activity Determination of TBBMC, TBBBMD, spinal BMC and BMD Bone age radiograph
Randomization of HIV-infected children to treatment arms will be performed by computer-generated random numbers (SAS version 8.2 for Windows, SAS Institute, Cary, NC). Randomization will stratify by sex and age (e.g., > or < 14 years) in order to ensure that treatment groups are balanced with respect to these variables. Block randomization will be used to ensure that the number of subjects in each arm remains approximately similar over time. A randomization sequence will be generated by the study biostatistician and communicated to the research pharmacist who will dispense active or placebo medication as subjects are enrolled.
Adherence to calcium/placebo will be assessed at each visit through pill counts and use of blister packages. Adherence to vitamin D/placebo will be ensured by supervised administration during defined study visits.
A.3.c. Measurement of bone mineral status in children and adolescents:
Bone mineral status in HIV-infected subjects will be determined by whole body and lumbar spine DXA scans. These will be performed at the SLRHC Pediatric Body Composition Unit using the Hologic Delphi-A Scanner.
Measurement of bone mineral content in grams by DXA correlates well with total body calcium (Ellis 1996). To date, this laboratory has provided bone mineral and body composition data using DXA in more than 13,000 subjects for more than 50 research projects and clinical trials.
The cv for bone mineral density in human subjects in our laboratory is 0.8%. The Hologic Delphi
A.4.d. Assessment of skeletal age:
Skeletal age will be determined by standardized bone age radiographs of the left wrist which will be performed at the SL site Department of Radiology on the same day as the DXA scans. The sex-specific standards for bone age using plain films of the hand and wrist from the Atlas of Greulich and Pyle will be used as the standard reference (Greulich and Pyle 1959). A single pediatric radiologist, who will be blinded to the age and treatment status of the subjects, will read all bone age radiographs.
Follow up bone age radiographs will not be performed on subjects found on entry to have achieved skeletal maturity (i.e. epiphyseal fusion).
Maturation has profound effects on bone mineral accrual. Prior to the appearance of the physical signs of puberty the only available method of assessing relative maturity or "biological age" is to measure skeletal or "bone age" by standardized x-rays. Skeletal age continues to be of use, even after the appearance of the physical signs of puberty, because the rate of progression of bone maturation with puberty varies between individuals. It is particularly variable in children with chronic disease whose general health status and rates of growth and pubertal progress may "catch-up" or "catch-down" in response to changes in treatment or in disease activity. In children with chronic disease, bone age may be a better indicator of maturational status than chronological age, size, or pubertal stage. Bone age therefore an important co-variate in this protocol.
A.3.e. Assessment of pubertal status:
Standardized physical examination performed by study investigators (Drs. Stephen Arpadi and Mary Horlick) after explanation of the procedure and rationale to subjects and family members to determined the level of sexual maturation. These will be performed during study visits to the SLRHC Pediatric Body Composition Unit using the method of Tanner (Tanner 1970, Tanner 1971). Repeat assessments will not be performed on subjects determined on entry to have experienced sexual maturation (i.e. Tanner Stage 5).
Because of the normal wide variability in timing and pace of puberty, standardized serial assessment of pubertal status of study participants will be performed to help distinguish between the effects of puberty and the effects of study treatments, increased age, height and weight on bone mineral accrual. Measurement of this variable is potentially of great importance in studies of children with chronic illnesses such as HIV-infection where delays in sexual maturation are common (de Martino 2001).
Changes in bone mineral mass by DXA at the time of puberty and its accompanying growth spurt have been reported in both males and females (Glastre 1990, Gilsanz 1991, Rico 1993, Mora 1999). However, the reference standards generated by DXA manufacturers and recent longitudinal studies are age- and sex-specific only (Bachrach 1999). Reference standards based on age or puberty only are not helpful in evaluating patients with disorders, such as HIV infection, which affect growth, timing of puberty, and may independently affect bone mineral acquisition (Leonard 1999, Thearle 2000, Arpadi 2002).
A.3.f. Procedure for assessment of pubertal status:
The evaluation of puberty in this project will follow exactly the protocol of the BMDCS project. The impact of pubertal stage on bone mineral accrual will be explained to all participants at the entry study visit, so that they and their families will understand the purpose of the physical examination for pubertal status. The pediatric endocrinologist (MH), following the protocol established for the multicenter BMDCS project, will evaluate pubertal stage. For girls, breast development will be assessed by visual criteria, but also by palpation in the earliest stages of puberty (Marshall and Tanner 1971, Biro 1992). For boys, testicular size will be assessed using the ovoids of the Prader orchidometer (Zachman 1974, Biro 1995). For both boys and girls, pubic hair development will be assessed by the criteria of Tanner (Marshall and Tanner 1970, 1971).
A.3.g. Procedure for measuring height and weight:
Standing height will be measured to the nearest 0.1cm using a wall-mounted stadiometer (Holtain, Crosswell, Wales). Three separate measurements will be obtained at each time point specified and the results averaged. Weight will be obtained with subjects in hospital gown and measured to the nearest 0.1 kg (Weight Tronix, New York, NY).
A.3.h. Assessment of modifiable behavioral factors that influence bone mineral accrual:
We will use a food frequency questionnaire (FFQ) for dietary data collection to provide an estimate of nutritional intake for this study. The availability of standardized information on dietary intake also will permit investigation of the relation between food/nutrient intake and bone accretion during growth in HIV-infected children.
Weight-bearing physical activity will be determined using the published self-report tool of Slemenda et al (1991). Although several studies support the validity of measures of children's self-reports of physical activity, only Slemenda's study reported the use of self-report of childhood weight-bearing physical activity (Sallis 1993, Slemenda 1991).
A.3.i. Procedures for dietary intake and physical activity questionnaires:
Following the protocol established for the multicenter BMDCS project, subjects will complete a FFQ (with the help of their parent or guardian if they are under 14 years). The questionnaire will take about 20 minutes and will be completed at the time of the visit. A picture containing various serving sizes will be included with each FFQ. Subjects are asked to indicate which picture best matches their serving size. Subjects will be reminded to indicate how much they usually ate - not how much was served.
All participants will be administered a questionnaire asking them to recall the number of hours during the past week that they were involved in each of several physical activities. The investigator will prompt the participants to recall their activity as accurately as possible. The child may ask the parent to help. About ten minutes is needed to complete the questionnaire.
The FFQ, BLOCKIDS (NutritionQuest, University of Berkeley, Berkeley, CA), is comprised of foods identified by NHANES III as important for this age and demographic group. Of relevance to this project, measurements include dietary intake of Vitamin D and calcium. The FFQ has been validated in African American children ages 8-10 years (Block 2000) for some macro and micronutrients (e.g. correlation for calcium = 0.64). Additional validation studies are planned as part of the BMDHC project.
A.3..j. Procedures for biochemical analyses:
General laboratory tests will be performed in the clinical chemistry laboratories of study sites including CBCs, lymphocyte subsets (CD4, CD8), HIV viral load, renal and liver function tests. Sera for specialized biochemistries (PTH, vitamin D metabolites, OC, BSAP, NTx) is aliquoted, frozen immediately after collection, and transferred to SLRHC where they will be stored at -70 degrees. The frozen samples will be analyzed in the Core laboratory of the Irving Center for Clinical Research at Columbia University. All samples from individual subjects will be analyzed in batches in the same assay to reduce inter-assay variability. All of the planned assays are currently in place in the Core Laboratory.
Serum: Kits for intact PTH, vitamin D metabolites, and osteocalcin will be purchased from Corning-Nichols Laboratory (San Juan Capistrano, CA).
Parathyroid hormone: PTH is currently being measured by a chemiluminescent method that is now known to detect the 7-84 fragment of PTH as well as the intact form (1-84)of the molecule (72). Intra-assay and inter-assay variability are 3.4% and 5.6% respectively. The normal adult range for this assay is 10-65 pg/ml.
Vitamin D metabolites: After extraction of serum samples, 25OHD will be measured in a radio-binding assay. The intra-assay and inter-assay variability are 7.5% and 9.6% respectively. The normal range is 9-52 ng/ml. Serum 1,25(OH)2D will be assayed in a radioreceptor assay. The normal adult range is 15-60 pg/ml. The intra-assay and inter-assay variability are 7.6% and 9.8% respectively.
Osteocalcin: Osteocalcin is measured by RIA according to the method of Gundberg et al. The intra-assay and inter-assay variability are 4.3% and 5.7% respectively. The normal range is 3.4-11.7 ng/ml for adult men, 2.4-10.0 ng/ml for adult premenopausal women.
Bone Specific Alkaline Phosphatase (BSAP): BSAP is measured by a solid phase, two-site immunoradiometric assay that specifically measures the skeletal isoenzyme of alkaline phosphatase in serum. Kits are purchased from Hybritech Incorporated (San Diego, CA). The intra-assay and inter-assay variability are 4.2% and 7.2% respectively. The normal range is 8.0 - 16.6 ng/ml for adult women and 7.2 - 15.8 ng/ml for adult men.
Serum N-telopeptide (Osteomark, Ostex, Seattle, WA) is measured using a competitive-inhibition enzyme-linked immunosorbent assay that utilizes microwells as the solid phase onto which NTx has been adsorbed. Assay values are reported in nmoles Bone Collagen Equivalents (BCE) per liter. The normal range is 6.2 - 19.0 in adult women and 5.4 to 24.2 in adult men. The intra-assay and inter-assay variability are 4.6% and 6.9% respectively.
A.3.l. Immunologic and virologic measures of HIV disease:
CD-4 +lymphocyte number and CD4% values for subjects will be obtain from results performed in Clinical site laboratories as part of patient care. Quantitative viral load assays -HIV RNA determined by polymerase chain reaction method performed as part of clinical care at collaborating clinical sites will be collected used for this study.
A.3.m. Procedure for safety monitoring and toxicity management:
Urine pregnancy test will be administered to all females of child-bearing capacity at each SLRHC visit and immediately before any radiological studies are performed. All females of child bearing capacity will be questioned if they are pregnant prior to urine testing and DEXA scanning.
Monitoring for hypercalcemia, hypercalciuria, and hypocalemia: In order to maintain blinding of subject treatment assignment, laboratory reports will be reviewed by Dr. Shane, a Co-investigator, located at New York-Presbyterian Hospital, which is physically apart from any of the clinical research facilities. This is also the location for the GCRC where specialized laboratory assays will be performed.
Serum calcium and albumin and spot morning fasting urine will be monitored on a monthly basis during the first year on study, and every 2 months during the second year. These assays will be performed through Quest Diagnostics Laboratory as convenience samples (i.e., run when collected).
Clinical or laboratory abnormalities considered by the investigator to be related to study drug will be managed as follows:
If hypercalcemia (>10.5 mg/dl, albumin-corrected) develops, the dose of calcium will be reduced by 250 mg/day and the levels rechecked one week later. If the abnormality does not resolve, weekly stepwise reductions in oral calcium will be followed until a normal level is achieved or supplemental calcium is discontinued. Vitamin D3 or placebo will be withheld until serum calcium returns to normal.
Subjects with persistent abnormality will be referred to the treating physician. Vitamin D3 or placebo will be withheld until serum calcium returns to normal.
If severe hypercalcemia develops (>12.0 mg/dl), all calcium supplements and Vitamin D/placebo will be discontinued immediately and the subjects will be withdrawn from the study.
Hypercalceria If hypercalciuria develops (urinary calcium /creatinine ratio >0.25 on spot urine specimens), the subject will have a 24 hour urine collection performed for confirmation. For confirmed hypercalciuria (>4mg/kg/24 hr), the dose of calcium will be reduced by 250mg and a spot urine repeated one week later. If the abnormality does not resolve, weekly stepwise reductions in calcium will be followed until hypercalceria resolves or supplemental calcium is discontinued. If necessary, vitamin D3 or placebo will be withheld until urine calcium returns to normal.
If hypocalcemia develops (< 8.4 mg/dl, corrected for albumin), blood will be drawn for Blood UreaNitrogen, serum electrolytes, creatinine, 25-OHD, PTH, Creatinine, Magnesium, and Albumin.
A subject with confirmed hypocalemia associated with vitamin D deficiency (25-OHD<10 ng/dl and PTH >73 will be withdrawn from the study and referred to their medical provider for treatment and withdrawn from the study
Monitoring for vitamin D deficiency
During the first year on study, serum for measurement of 25-OH and iPTH will be drawn monthly and every 2 months during the second study year. Each subject's specimens will be frozen immediately after collection, and transferred to SLRHC where they will be stored at -70 degrees. At one year intervals the each subjects' samples will be assayed in "batched " fashion. On the basis of these annual determinations, subjects with persistent vitamin D deficiency (25-OHD<10ng/l and PTH>73) that lasts throughout both winter and summer months, will be withdrawn from the study and referred to their medical provider.
We anticipate low incidence of these particular abnormalities. In a prior study in which the same dose of vitamin D3 (100,000 IU) was administered every 60 days for 6 months, Guillemant et al. reported no instances of hypercalcemia in French adolescent males, ages 13-16, with seasonal vitamin D deficiency (Guillemant 2000). In addition, in an on-going study conducted by Dr. Shane (Prevention of Osteoporosis After Cardiac Transplantation, ROI AR 46124-04), among the 74 subjects receiving the metabolically active form of Vitamin D, calcitriol (0.25 ug BID), there were only 4 episodes of urinary calcium excretion >400 mg/g creatinine. Besides being on calcitriol, these subjects were receiving corticosteroids, which also increase urinary calcium excretion. Corticosteroids are not a permitted co-administered medication in the present study.
We do not anticipate Vitamin D3 will have an adverse effect on HIV infection or antiretroviral medications. However, HIV-infected subjects who experience an otherwise unexplained decline in immune status, based on CD4 count or CD4%, or an increase in measure of HIV viral load believed by the treating physician to be possibly related to study medications will be withdrawn from study participation. Measurements of CD4 count, CD%, and HIV viral load will be measured at the clinical care sites as per standard of care. Results will be monitored by the treating physician.
Subjects who have decline in TBBMC z-score of >0.5 from baseline at the one year on-treatment DXA will be discontinued from study and referred for additional clinical evaluation by the treating physician. In addition, if low impact or repeated fractures occur, a clinical evaluation by the treating physician, safety officer and Study PI will determine if the subject should remain on study.Monitoring DXA scan results of study subjects will be performed by Dr. Horlick, a Co-Investigator and Director of the Pediatric Body Composition Laboratory at St. Luke's-Roosevelt.
Finally, because this is a placebo-controlled trial involving children, we have elected to include an interim data analysis to be performed on 12 month on-treatment data (Human Subjects, Section E.). A Data Safety and Monitoring Board comprised of experts in biostatistics, pediatric HIV disease, pediatric endocrinology, and pediatric bone acquisition will be established as per NIH guidelines to monitor study results for safety and efficacy, to establish criteria for early termination of the study in consultation with the Principal and Co-Investigators, and to determine if and when these criteria have been fulfilled.
A.3.n Data analysis and power calculations Specific Aim 1: To determine the effect of calcium and vitamin D3 supplementation on bone mineral accrual in HIV-infected children and adolescents receiving HAART.
The primary hypothesis of this study is that BMC and BMD will increase in HIV-infected subjects randomized to receive vitamin D3 plus calcium compared to subjects receiving placebo. We will test this hypothesis by comparing TBBMC, TBBMD, SBMC, and SBMD measured by total body and spine DXA scans performed at baseline and at 1 and 2 years post-randomization in the 2 study arms using an intention-to-treat analysis.
A repeated measures analysis of covariance will be used to analyze the bone accrual data. Differences in bone mineral content between the treatment and placebo groups will be compared at 1 and 2 years while controlling for any possible differences between the groups at entry to the trial. The between-group main effect (vitamin D3 and calcium versus placebo) will test whether there is an overall difference between the groups, while the within-group x time interaction effect will test whether the difference between the groups is constant over the two follow-up years. The magnitude of change in the vitamin D3 and calcium group will also be compared against normative data from the BMDCS to determine how many children in each group achieve levels within normal limits (95 % CI) using the algorithm presented in preliminary studies section C.5.
Based on our prior pilot data, the mean yearly change in TBBMC in 30 HIV-infected children between the ages of 7-14 years was 120 g with a standard deviation of 84. A difference in SD between boys and girls was noted. In order to maintain conservative estimates we chose SD observed in girls. This can be assumed to be the level of yearly change that will be observed in the placebo group. Based on data from the Pediatric Rosetta Project (NIDDK 37352), a cross-sectional study of >1300 children, the mean yearly change in TBBMC in healthy African-American and Hispanic children ages 7-14 years is 195 g.
The effect size (d) comparing a mean yearly change of 120 g in the placebo group versus a change of 195 g in the Vitamin D3 and calcium group is therefore 0.89. Twenty-one children in each group will be sufficient to detect an effect size of this magnitude with alpha = 0.05 and beta = 80%. If the sample size is increased to 30 children in each group, power will be sufficient to detect a difference of 61 gm between the vitamin D3 and calcium group and the placebo group with alpha = 0.05 and power = 80%. This is comparable to the difference observed in a clinical trial of healthy adolescent girls (mean age, 11.9 years) who received either calcium (110% of RDA) or placebo (Lloyd 1996).
Differences in the rate of sexual maturation during the study between treatment groups may "unbalance" the groups with respect to a factor of significance to the bone mass outcomes. To address this we have included a standardized measurement pubertal stage for use as a covariate in the analyses.
Specific Aim #2: Determine the effect of HIV infection and vitamin D and calcium supplementation on indices of mineral metabolism and markers of bone turnover.
We hypothesize that indices of mineral metabolism are abnormal (low serum 25(OH)D and elevated PTH) and biochemical markers of bone formation and resorption are increased in HIV-infected children on HAART compared to normal controls and that indices of mineral metabolism and markers of bone formation and resorption will return toward normal in HIV-infected children and adolescents who are randomized to receive vitamin D and calcium supplementation. We will test this hypothesis first by comparing markers of bone turnover (osteocalcin, BALP and N-Tx) in HIV-infected and healthy children and adolescents enrolled in the cross-sectional study using independent samples t-test to compare means.
Longitudinal changes in bone turnover markers in HIV-infected children and adolescents receiving vitamin D and calcium supplement will also be compared to those receiving placebo. Levels of vitamin D metabolites and markers of bone turnover ((osteocalcin, BALP and N-Tx) taken at intervals will be analyzed using a mixed model regression to model trends in these measures over time to evaluate dependent inter-relationships between these variables.
Specific Aim #3: Evaluate vitamin D stores in HIV-infected children and adolescent receiving HAART.
We hypothesize that vitamin D stores are decreased in children with HIV infection compared to healthy controls.
We will test this by comparing serum levels of vitamin D metabolites and indices of mineral metabolism (25(OH)D 1,25(OH)2D) and PTH) in HIV-infected children and adolescents receiving HAART compared to healthy controls. Data for testing this will be obtained from the cross-sectional study. The HIV-infected group will consist of a sample of HIV+ children screened for recruitment into the study while the healthy group will consist of similar aged children recruited from the same neighborhood (including healthy siblings of the HIV positive children). Screening levels of serum vitamin D metabolites and PTH in the HIV-positive children will be compared with levels in the healthy children using an independent samples t-test.
The main variable of interest is 25(OH)D as it best reflects vitamin D stores. Based on prior pilot data, the mean (SD) expected 25(OH) among HIV-infected subjects is 21 (11) ng/ml. The sample size in the present study has 80% power to detect change in 25(OH) D of 8 ng/ml with alpha=0.05. Analysis of covariance based on a general linear regression model will then be used to determine whether any possible group differences remain after controlling for other factors such as race, sex, estimated dietary intake of Vitamin D and sunlight exposure (using month of study as surrogate for sunlight exposure).
Specific Aim #4: Evaluate if vitamin D stores are a determinant of bone mass in HIV- infected children and adolescents receiving HAART.
Our hypothesis is that among HIV-infected children and adolescents receiving HAART, vitamin D stores, as measured by serum 25(OH)D, is a significant determinant of bone mass.
We will assess the relationship these variables have to TBBMC in HIV-infected children and adolescents. This will be conducted through analysis of data obtained from a sample of at least 60 HIV-infected children and adolescent being screened for the enrollment in the longitudinal clinical trial. Within the HIV positive group, linear regression will be used to determine whether HIV-related factors including antiviral medications, markers of immunodeficiency (CD4count), HIV viral load, and levels of cytokines play a role in predicting serum vitamin D levels and bone mass. Additional co-variates that will be included in the model include age, sex, race, height, weight and pubertal stage.
These analyses are especially critical measurements for inclusion in this clinical trial and in the analyses as these will allow for development of potential mechanistic pathways for both the etiology of abnormal bone mass and intermediate response variables involved in amelioration.
Using multiple regression modeling we will assess the relationships of TNF, IL-6 vitamin D levels and TBBMC. It is estimated that 50% of the HIV positive children who are screened will meet the inclusion criteria for the trial and that approximately 100 to 120 HIV positive children will be available for the cross sectional comparison and so an equal number of healthy children will be also recruited. With a total of 200 children, the sample size will be large enough to add up to 18 covariates in addition to the grouping factor (HIV-positive versus healthy) and test for a 5% increase in R2, with alpha = 0.05 and power = 80 %. A sample of 100 HIV-positive children alone will allow for a regression analysis that includes up to 9 predictors which collectively explain at least 15% of the variance of the outcome variable (TBBMC, etc.) with alpha = 0.05 and power = 80 %.