Pamidronate to Treat Osteogenesis Imperfecta in Children: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pamidronate to Treat Osteogenesis Imperfecta in Children (NCT00005901), a randomized phase 3 study. The primary outcome is Change in Bone Mineral Density in Response to Pamidronate.

2. Clinical Question

3. Trial Design

The registry describes a phase 3, randomized, parallel-group trial with no masking. The posted enrollment is 34. The trial record is marked completed, with a start date of 2000-06 and a primary completion date of 2014-05. The sponsor is the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD).

The study summary describes evaluating pamidronate in children with osteogenesis imperfecta, a genetic collagen disorder associated in the summary with weak bones and fractures. The broader planned comparison included pamidronate every 3 months or every 6 months, with or without growth hormone for eligible children. The summary describes children aged 2 to 4 years as eligible for random assignment between the pamidronate schedules, and describes growth hormone groups for children aged 4 years and older. The posted primary outcome comparison, however, is between pamidronate every 3 months for 3 years and every 6 months for 3 years.

Randomization means that assignment was intended to be determined by chance rather than by a clinician's choice. In general, random assignment can reduce systematic differences between groups at the start of a trial, although it does not guarantee that all characteristics will be balanced, especially in a small study. The registry identifies the trial as randomized, but the method used to generate the sequence and the allocation-concealment procedures are not reported in the registry record.

No masking was used according to the registry. This means the trial is described as open-label; the registry does not specify whether outcome measurement or analysis was conducted with any separate masking. The record gives the two outcome-group sizes and their results, but details such as adherence, treatment discontinuation, and the number assessed at each measurement time are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Analysis population for the primary outcome, as defined in the registry: All subjects received Pamidronate for 3 years.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Change in Bone Mineral Density in Response to Pamidronate, measured over Baseline vs. 12 months after first dose.

The primary endpoint is titled “Change in Bone Mineral Density in Response to Pamidronate.” It is measured as a Z-score, with the stated timeframe of baseline versus 12 months after the first dose. The registry says that bone density measurements were obtained using dual-energy X-ray absorptiometry (DXA), with a Hologic QDR 4500 densitometer and low density software package. It reports measurement precision of 0.011 SD. Raw measurements were converted to Z-scores using the manufacturer’s reference standards for age and pubertal status.

In general, a Z-score expresses a measurement relative to a reference distribution, in standard-deviation units. Using age and pubertal-status reference standards is relevant because bone measurements are interpreted against a reference that accounts for those characteristics. A Z-score therefore gives a standardized way to describe the measured bone density rather than reporting only a raw DXA value.

The posted group means are -3.54 for pamidronate every 3 months and -4.98 for pamidronate every 6 months. The registry also reports a mean difference for the comparison. Because the endpoint title refers to change while the timeframe is expressed as baseline versus 12 months, readers should interpret the result according to the registry’s endpoint label and time description; further details about the exact derivation of the analyzed value are not reported in the registry record.

7. Results Posted in the Registry

• Pamidronate Every 6 Months for 3 Years: mean -4.98 versus -3.54 on control; difference -1.44 (95% CI -2.34 to -0.54), standardized effect d = -1.05; z-test p = 0.002.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the mean difference represent?

The endpoint is a continuous measure, and the registry reports the mean and standard deviation for each group. A mean is the arithmetic average of the measurements represented in a group. The reported mean difference compares the group averages, expressed in the same Z-score units as the endpoint. For this comparison, the posted estimate is -1.44, with the every-6-months group compared against the every-3-months group.

In general, a difference between means summarizes a contrast in group location. It does not show the full distribution of participants' measurements: different distributions can have the same mean, and a group mean does not describe every participant. The standard deviations give additional information about within-group spread. They are 1.51 for the every-3-months group and 1.16 for the every-6-months group. Standard deviation is not the standard error of the mean and should not be read as uncertainty in the group-average estimate.

What does the confidence interval add?

A confidence interval gives a range of values compatible with the observed data under the method's assumptions. In general, it reflects both the estimated effect and the amount of statistical uncertainty around that estimate. A narrower interval indicates greater precision than a wider interval, for the same kind of estimate and method. The reported interval for the mean difference extends from -2.338 to an upper bound below zero. Its position entirely below zero indicates that the posted interval does not include no difference.

A confidence interval is not a range containing every individual participant's result. Nor should it be interpreted as a probability statement about the true value after the data have been observed. Its interpretation depends on the analysis method and assumptions. The specific model or test used to produce the registry comparison is not reported in the registry record.

How should the p-value be understood?

The unadjusted p-value is 0.002. In general, a p-value measures how unusual data at least as incompatible with a null hypothesis as the observed data would be, if that null hypothesis and the analysis assumptions held. It is not the probability that the observed result arose by chance alone, and it does not quantify the size or practical importance of the difference. The registry does not report an adjusted p-value or a multiplicity procedure for this outcome.

“Unadjusted” matters because a p-value may be interpreted differently when many outcomes, time points, or comparisons are tested. Adjustment methods can account for some forms of multiple testing, but no such method is specified for the posted comparison. The record also does not report a full statistical analysis plan, so the intended hierarchy of outcomes or comparisons is not available for assessment.

Why does spread matter?

In general, over-dispersion means that observed variability is greater than a model expects. When a method underestimates variability, standard errors can be too small and intervals can be too narrow. The posted outcome is continuous and comes with group standard deviations; the registry does not describe an over-dispersion assessment or identify a model that would make such an assessment relevant. The principle is general, not a claim that over-dispersion occurred in this trial.

Measurements may also be correlated, for example when repeated measurements are made on the same participant. Accounting for within-participant correlation is important in analyses that use multiple measurements per person. The record gives a baseline-versus-12-month timeframe but does not report whether the analysis modeled repeated measurements, analyzed a change score, or used another approach. It is therefore not possible to identify the exact handling of repeated measurements from the registry entry.

What can and cannot be inferred about the analysis?

The posted record gives group summaries, a mean difference, a confidence interval, and an unadjusted p-value. It lists no analysis entries. No specific test, regression model, covariate adjustment, missing-data method, or analysis population is reported in the registry record. Those details should not be inferred from the presence of a confidence interval or p-value alone.

Randomization supports a comparison of assigned groups in principle, but the estimate's interpretation also depends on which participants were included and how outcomes were handled. An intention-to-treat analysis generally keeps participants in their assigned groups regardless of treatment received, while a per-protocol analysis may focus on participants who followed specified treatment conditions. These are general analysis concepts; the registry record does not state which population was used for this result.

Because no masking was used, participants and study personnel were not masked according to the registry's design description. In general, lack of masking can affect behavior, co-interventions, assessment, or reporting, depending on the outcome and trial procedures. The record does not describe whether any outcome assessors or analysts were separately masked, so those details remain unknown.

10. Confidence Intervals and Precision

The reported mean difference is -1.44, with a 95% confidence interval extending from -2.338 to an upper limit below zero. The interval is entirely on the negative side of zero. This indicates that, under the method used to form the interval, zero is not among the values in the reported range. The unadjusted p-value is 0.002, which is consistent with the interval's exclusion of zero, although the p-value and interval convey different aspects of the result.

The interval gives a sense of the uncertainty around the estimated group contrast. It does not remove uncertainty, and its width should be read in the context of the small outcome groups: 19 subjects in the every-3-months group and 15 in the every-6-months group. The standard deviations, 1.51 and 1.16, describe variability among the measurements within each group, not the precision of the contrast itself.

Precision is not the same as importance. A confidence interval can help readers judge which effect sizes remain compatible with the data, but the registry does not provide a threshold for a clinically important difference. The measurement description reports DXA precision of 0.011 SD; that instrument-related precision statement is distinct from the confidence interval for the between-group mean difference. The analysis method, completeness of follow-up, and handling of missing data are not reported in the registry record, so they cannot be incorporated into a fuller evaluation of precision.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total event counts. They are illustrations based on the observed mean difference, rather than the trial's original sample-size calculation. For the full observed effect, the illustration gives 15 participants per arm for 80% power and 19 participants per arm for 90% power. For planning for 75% of the observed effect, it gives 26 participants per arm for 80% power and 34 participants per arm for 90% power. For planning for 50% of the observed effect, it gives 57 participants per arm for 80% power and 76 participants per arm for 90% power.

The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. In plain terms, the 80% power column gives the estimated participants per arm for a design with an 80% chance of detecting the specified effect under those assumptions, if that effect is real. The 90% column gives the corresponding planning quantity for a 90% chance. These are design calculations, not a description of what happened in this trial.

Smaller effects require more participants to detect under the same planning assumptions, which is why the listed sample sizes increase when the target is a smaller fraction of the observed effect. These figures are planning illustrations, not observed power. The original trial's sample-size justification and design assumptions are not reported in the registry record.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -1.440; pooled standard deviation σ = 1.368.
2. n = 2 × (1.96 + 0.842)² × 1.368² / -1.440² = 14.2, rounded up to 15 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -1.44                      # observed difference in means
sigma <- 1.36793868649147                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(5901)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 15 (80% power), 19 (90% power)
 75% of observed effect: n per arm = 26 (80% power), 34 (90% power)
 50% of observed effect: n per arm = 57 (80% power), 76 (90% power)
power.t.test: 15.2 
Simulated power at n = 15 per arm: 0.797 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry posts one primary outcome comparison in the results and lists no posted secondary outcome results. The broader study summary names several areas of interest, but the record does not provide results for those additional outcomes in this entry. It also does not report a multiplicity procedure, such as a prespecified hierarchy or adjustment across outcomes, time points, or comparisons. The unadjusted p-value should therefore be understood as unadjusted for any such multiplicity.

In general, examining many outcomes or comparisons can increase the chance of finding at least one result that appears statistically notable even when there is no corresponding effect. Methods for controlling that risk depend on the planned outcome structure. The registry record does not describe the planned statistical hierarchy or whether any analyses beyond the posted primary comparison were conducted.

The result gives group sizes, but the analysis population is not specified. The record does not say whether the analysis included all randomized participants, only those with complete measurements, or another defined group. It also does not report how many participants had measurements at baseline and at 12 months, whether any outcomes were missing, or how missing values were handled. Those details matter because the analyzed group can differ from the group originally randomized if follow-up is incomplete.

The record does not report a statistical analysis plan, an interim-analysis plan, or rules for stopping or modifying analyses. It is therefore not possible to determine from the registry record whether these elements were prespecified. No assumptions about them should be made from the posted estimate or p-value.

14. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events for two safety groups. In one row, 19 participants were at risk and 0 were recorded with a serious adverse event. In the other row, 15 participants were at risk and 0 were recorded with a serious adverse event. No serious adverse-event terms are listed. Death counts are not reported in the registry record.

For other adverse events, the row with 19 participants at risk records 0 affected, while the row with 15 participants at risk records 1 affected. The registry labels the groups “Active Comparator: 1” and “Active Comparator: 2”; it does not provide interval-specific labels for these safety rows in the posted safety data. The counts describe the events recorded in the registry, not necessarily every event that occurred during the study.

In general, an absence of recorded events does not establish that no event occurred, and small groups provide limited information about uncommon events. The record does not supply further definitions, observation periods, severity details for the other adverse event, or a narrative accounting of safety reporting. These figures should be read as registry counts rather than as a complete safety profile.

15. What the Result Does, and Does Not, Mean

The posted result describes a difference in the primary bone mineral density Z-score outcome between two pamidronate schedules. The reported mean was -3.54 in the group receiving pamidronate every 3 months for 3 years and -4.98 in the group receiving it every 6 months for 3 years. The estimated mean difference was -1.44, with a confidence interval entirely below zero and an unadjusted p-value of 0.002. These figures describe the reported outcome and comparison; they do not establish the significance of other outcomes named in the study summary.

The result does not, by itself, establish whether the observed difference is clinically important. The registry does not provide a threshold for a meaningful difference in this Z-score outcome. Nor do the posted data establish effects on fractures, bone strength, growth, functional ability, or other outcomes described in the broader summary. Results for secondary outcomes are not posted in the registry record.

The result also does not demonstrate that every participant experienced a change equal to the group mean. Group averages conceal individual variation, and the standard deviations show that measurements varied within both groups. The confidence interval concerns uncertainty around the group contrast, not the range of participant outcomes.

Randomization is a useful design feature for comparing groups, but the posted record leaves important interpretive details unresolved: the analysis population, missing-data handling, baseline group characteristics, adherence, and the precise statistical model are not reported. The trial had no masking according to the registry. These features do not erase the observed estimate, but they inform how cautiously it should be interpreted and how much context is available for evaluating it.

This statistical reading is not medical advice. It is a description of the registry result, its uncertainty, and the limits of what the posted information can establish.

16. Why This Trial Matters Statistically

This trial illustrates how a randomized comparison can be summarized using a continuous outcome. Rather than reducing the result to a yes-or-no judgment, the reported group means, standard deviations, mean difference, confidence interval, and p-value describe different parts of the evidence. The means locate the groups on the outcome scale, the standard deviations describe within-group spread, and the interval expresses uncertainty about the estimated contrast.

The example also shows why the endpoint definition and timing matter. The registry identifies a bone mineral density Z-score outcome and gives a baseline-versus-12-month timeframe. It describes the DXA method and the age- and pubertal-status reference standards used to convert measurements into Z-scores. These details help explain what the outcome represents, while the record's lack of detail about the exact analysis of repeated measurements limits reconstruction of how the result was calculated.

The reported groups are small, making precision and analysis-population details especially relevant when reading the estimate. The confidence interval gives a range around the comparison, but the registry does not state how missing observations were handled, which participants were included in the analysis, or what model produced the interval. Statistical reporting is most informative when effect estimates are accompanied by these design and analysis details.

Finally, the planning illustrations show that sample-size requirements depend on the effect a future study aims to detect and on its chosen power and significance assumptions. They are not a retrospective measure of the trial's success. Distinguishing planning quantities from observed results, and distinguishing statistical evidence from clinical importance, are central habits of careful trial interpretation.

17. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The outcome groups are 19 and 15 subjects, and the registry does not report baseline characteristics for these groups. It is therefore not possible to assess their comparability on measured prognostic factors from this record. Randomization is reported, but the sequence-generation method and allocation concealment are not described.

The trial is described as having no masking. The registry does not say whether assessors or analysts were masked separately, nor does it explain how the absence of masking may have been addressed in outcome collection. The record also does not report adherence, withdrawals, treatment discontinuation, or the number of participants contributing measurements at each time point.

The endpoint title refers to change in bone mineral density, and the timeframe is baseline versus 12 months after the first dose. The record does not specify in detail whether the analyzed value was a change score, a comparison of follow-up values adjusted for baseline, or another formulation. It also does not report the statistical test or model, covariate adjustment, missing-data method, or analysis population.

The registry record contains no posted secondary outcome results and no analysis entries. Multiplicity handling, the full statistical analysis plan, and interim-analysis procedures are not reported in the registry record. The posted safety counts are also limited: serious-event terms are absent, death counts are not reported, and the safety-group labels do not specify the dosing intervals. These limitations restrict how fully the result and safety record can be evaluated, without changing the values that are posted.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Pamidronate to Treat Osteogenesis Imperfecta in Children — NCT00005901.
• Registry record last updated 2016-03-01; status last verified 2016-02.
• Results first posted 2016-03-01.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

21. Record Summary

This completed phase 3 trial is recorded as randomized, parallel-group, and unmasked, with enrollment of 34. It concerns children with osteogenesis imperfecta and compares pamidronate given every 3 months with pamidronate given every 6 months in the posted primary result. The primary outcome is titled “Change in Bone Mineral Density in Response to Pamidronate,” measured as a Z-score at the baseline-versus-12-month timeframe.

The reported group means are -3.54 for 19 subjects receiving pamidronate every 3 months and -4.98 for 15 subjects receiving it every 6 months. The posted mean difference is -1.44, with an interval that remains below zero, and the unadjusted p-value is 0.002. The registry does not report secondary outcome results, analysis-population details, or the statistical model. Safety entries record no serious adverse events in either listed group and one other adverse event in the group with 15 participants at risk.