Effect of Diabetic Medications on Bone Metabolism: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effect of Diabetic Medications on Bone Metabolism (NCT00467285), a registered phase not stated study. The primary outcome is Changes in BMD at Femoral Neck.

2. Clinical Question

The study, titled “Effect of Diabetic Medications on Bone Metabolism,” is listed with the official title “Effect of Thiazolidinediones on Skeletal Health.” Its stated rationale is that skeletal effects of thiazolidinediones observed in animals might also be relevant to humans. The study hypothesis concerns whether pioglitazone treatment may be associated with reduced bone mineral density. That hypothesis is a research question, not a finding.

The planned scope described in the summary extends beyond the femoral neck. It mentions bone mineral density at the lumbar spine, femoral neck, and 0.33 radius, as well as bone turnover markers. The listed secondary outcomes include CTx, osteocalcin, and changes in CTx at follow-up. The registry record does not report the numerical results needed to assess the size or precision of the observed group differences for these measures.

3. Trial Design

The registry record describes a completed study sponsored by the VA Office of Research and Development. The listed start was 2006-10, and the primary completion date was 2009-09. Enrollment is recorded as 96. The study compares people with type 2 diabetes who were taking pioglitazone with people who were not taking it. Allocation is not stated, and the registry explicitly cautions that the arms are not a like-for-like control comparison.

The group descriptions list 140 subjects with type 2 diabetes on pioglitazone and 140 subjects with type 2 diabetes not on pioglitazone. The primary outcome description instead identifies 28 subjects on pioglitazone and 64 subjects not on pioglitazone. Those counts do not align with the enrollment figure and group descriptions as presented. The registry record does not explain the differing counts, so they should not be treated as interchangeable descriptions of one analysis population.

The primary outcome is a continuous measure: percentage change in femoral-neck bone mineral density over 6 months. Its stated summary measure is the mean, with standard deviation as the dispersion measure. The registry lists 7 outcome measures and 1 analysis, but does not describe the analysis procedure in enough detail to establish how the outcome measures were handled or how the analysis population was selected.

This is an observational comparison of medication-use groups in the registry description, not a documented randomized comparison. People taking and not taking pioglitazone may differ in ways other than medication use. Since allocation and adjustment details are not stated, the record does not support interpreting group differences as a causal effect of pioglitazone. The registry record also does not report the protocol, the reasons for the group-size differences, or a sample-size justification for the study’s own design.

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.

Before assignment: Young male veterans with type 2 diabetes with creatinine less than 1.4, aged less than 55 years

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: 28 subjects on pioglitazone and 64 subjects not on pioglitazone

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 Changes in BMD at Femoral Neck, measured over 6 months.

The primary endpoint is change in bone mineral density at the femoral neck, expressed as a percentage of change and assessed over 6 months. Bone mineral density is a continuous measurement of skeletal density. Expressing change as a percentage places the change relative to the measurement scale, while retaining the outcome as a quantitative measure that can be summarized by a mean.

The femoral neck is one of the skeletal locations named in the study’s outcome description. The broader description also includes the lumbar spine and 0.33 radius, along with bone turnover markers. The femoral-neck outcome is useful for the study question because it provides a defined skeletal measure to compare between the pioglitazone-use groups at the stated follow-up period. The registry record does not explain why this location was selected as the primary outcome over the other listed skeletal measures.

The record specifies a mean and standard deviation as the outcome summary, but does not report either group’s numerical result. It therefore does not show the direction or magnitude of the femoral-neck change, the variability among participants, or the precision of a comparison. A percentage-change endpoint also depends on the measurements and analysis used to derive it; those details are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous and is specified as a mean percentage change in femoral-neck bone mineral density. A mean describes the central value of a set of measurements. The standard deviation describes how spread out individual measurements are around that mean. Together, these summaries can characterize the group outcome, but they do not by themselves show the precision of a comparison between groups. The registry record gives the requested summary type but not the actual mean or standard deviation.

What does a confidence interval communicate?

In general, a confidence interval gives a range of values compatible with the data and the method used to calculate it. It helps distinguish an estimate’s apparent size from its precision: a narrower interval indicates greater precision under the assumptions of the analysis, while a wider interval indicates greater uncertainty. The registry record does not report a confidence interval for the primary outcome, nor does it provide the numerical group estimates needed to interpret one. A missing interval cannot be reconstructed from the outcome definition alone.

What would a group comparison need to address?

In general, a comparison of means assesses how the average outcome differs between groups, while accounting for variability and the amount of information available. The precise method depends on the design, the measurement structure, and the analysis population. Here, the registry lists a mean and standard deviation and describes one analysis, but the actual statistical test, model, covariates, and handling of repeated measurements are not reported in the registry record. It is therefore not possible to attribute a particular comparison method to the study.

The design also matters. Random assignment can help balance measured and unmeasured characteristics across groups on average. In this registry record, allocation is not stated, and the record flags the arms as not a like-for-like control comparison. In general, nonrandomized group differences can be affected by confounding: a factor related both to medication use and to skeletal outcomes may contribute to an observed association. Without reported adjustment details, the extent to which such factors were addressed is unknown.

Why does variability matter?

In general, greater variability in a continuous outcome makes a group mean less precise when other factors are unchanged. Standard deviations help describe that variability. They are not confidence intervals and do not directly show the uncertainty in a difference between group means. The registry record’s stated dispersion measure is the standard deviation, but the numerical values are absent, so the spread of the outcome cannot be assessed.

Over-dispersion is a general concern when the observed variability is greater than a statistical model assumes. If a model understates variability, its uncertainty estimates may be too small. Whether over-dispersion is relevant depends on the data and model; neither the model nor diagnostics are reported in the registry record. It would be inappropriate to claim that over-dispersion was present or that any particular correction was used.

How should the planning calculations be read?

The planning assumptions describe an illustration computed from an observed effect, rather than the study’s own design justification. They specify a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation, with the target effect defined as a stated fraction of the observed effect. In general, such calculations ask how much information might be needed to detect a chosen effect under stated assumptions. They are not a measure of the study’s observed power, do not establish that the study was designed that way, and do not replace reporting the observed estimate and its uncertainty. Numerical planning results are not reported in the registry record.

For continuous outcomes, a planning calculation depends on assumptions about the effect and variability, as well as the significance threshold, allocation, and statistical method. If those assumptions differ from the actual data or design, the illustration may not reflect the study’s operating conditions. The record does not report the study’s own sample-size justification, the assumptions used when planning enrollment, or whether any target was revised.

What does the outcome list imply about multiple testing?

In general, assessing several outcomes or comparisons can increase the chance of finding at least one apparently unusual result by chance if each is tested separately. Statistical plans may address this through a prespecified hierarchy, adjustment, or another strategy. The registry lists 7 outcome measures and 1 analysis, but does not report a multiplicity strategy, the status of each outcome as primary or secondary beyond the named primary outcome, or a testing hierarchy. No correction method should be inferred from the number of listed measures alone.

The registry’s description is enough to identify the endpoint type, stated summary, and follow-up period. It does not provide the analysis specification or numerical results required to independently evaluate the estimate, its uncertainty, or the assumptions of a statistical model. Those missing elements limit both statistical interpretation and reproducibility.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Piogliazone", "No Pioglitazone")
n   <- c(28, 64)
m   <- c(-3.2, 0.25)
sp  <- c(0.25, 0.24)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Piogliazone  28  -3.200  0.250  -3.297  -3.103
CARM  No Pioglitazone  64  0.250  0.240  0.190  0.310

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 7 outcome measures, including femoral-neck bone mineral density change and secondary measures such as CTx, osteocalcin, and changes in CTx at follow-up. Considering multiple outcomes can raise multiplicity concerns because repeated statistical testing can increase the chance of at least one apparently positive finding by chance. The registry record does not report whether a multiplicity adjustment, outcome hierarchy, or other strategy was prespecified. It also does not report numerical results for the listed secondary outcomes.

Missing-data procedures are not reported in the registry record. The record does not state whether outcome data were incomplete, how missing measurements were handled, or whether any sensitivity analyses were used. It also does not describe the analysis population beyond the primary outcome population counts. Without those details, the relationship between enrollment, group descriptions, and participants included in the primary outcome analysis remains unclear.

There is a notable difference among the counts stated in the registry: enrollment is 96; the group descriptions state 140 subjects in each group; and the primary outcome population is described as 28 on pioglitazone and 64 not on pioglitazone. The record does not explain these differences. It is not possible to determine from the registry record whether they reflect different study stages, eligibility definitions, outcome availability, or another reason.

The record does not report a statistical analysis plan, prespecified missing-data assumptions, or analysis-population definitions. These omissions prevent assessment of whether the reported analysis, if any, followed a prespecified approach or how sensitive it might be to incomplete data. The registry’s completed status does not resolve these questions.

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

The registry record describes a study of bone mineral density and bone turnover markers in people with type 2 diabetes who were or were not taking pioglitazone. Its primary outcome is percentage change in femoral-neck bone mineral density over 6 months, summarized by a mean and standard deviation. The record does not report those numerical results. It therefore does not establish the direction, magnitude, or precision of the primary outcome comparison.

The study question is relevant to the stated concern about skeletal health and thiazolidinedione use, but a study question and a hypothesis are not results. Likewise, the list of outcome measures does not show whether any measurement changed, whether groups differed, or how much uncertainty surrounded an estimate. Those conclusions require posted outcome values and an interpretable analysis.

The design description further limits what an observed difference could mean. Medication use defines the comparison groups, allocation is not stated, and the registry warns that the arms are not a like-for-like control comparison. In general, such a comparison can be influenced by differences between the people in each group. The record does not report enough information about adjustment or participant characteristics to assess this possibility. A group association would not, by itself, demonstrate that pioglitazone caused a change in bone density.

The registry lists enrollment as 96, group descriptions of 140 subjects in each arm, and a primary outcome population of 28 and 64. Because the record does not reconcile these figures, no single analyzed sample size can be inferred from them. The absence of the primary outcome estimates and the unexplained counts make the precision and representativeness of any comparison difficult to judge.

Accordingly, the registry information supports a careful account of the study’s subject, outcome, follow-up, and design limitations. It does not support a conclusion that pioglitazone either increased or decreased bone mineral density, nor does it establish the absence or presence of an association. Those judgments are not possible from the numerical results reported in this record.

13. Why This Trial Matters Statistically

This study raises a statistical issue common to clinical research: the distinction between an outcome question and evidence that can answer it. The registry names a continuous primary endpoint, specifies its summary measure, and gives a follow-up period. Yet without the mean, standard deviation, and comparison estimate, readers cannot assess the magnitude or precision of the result. Clear outcome definitions are valuable, but they cannot substitute for the numerical findings.

The registry also illustrates why group counts and analysis populations matter. Enrollment is recorded as 96, the group descriptions state 140 subjects in each group, and the primary outcome population is described as 28 and 64. These statements are not reconciled in the record. Transparent reporting of who entered the study, who contributed outcome data, and how participants were analyzed is necessary for readers to understand the denominator behind an estimate and assess its relevance.

Another central issue is comparison design. The study contrasts medication-use groups, but allocation is not stated and the record warns that the groups are not a like-for-like control comparison. Statistical adjustment may sometimes address measured differences, but it cannot be assumed without a reported method, and unmeasured differences can remain. The design therefore places limits on causal interpretation, regardless of whether a numerical association was observed.

Finally, the listed outcome measures invite attention to multiplicity and analysis planning. Multiple outcomes can create interpretive challenges, especially when the prespecified hierarchy or adjustment approach is unknown. The registry record does not report those methods. This example shows why statistical reporting needs to connect design, participant flow, endpoint definitions, analysis choices, and estimates with uncertainty. Without those elements, a completed registry entry may document what was asked without showing what the data indicate.

14. 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 most important limitation for interpreting this study record is the absence of numerical results for the primary outcome. Although the registry specifies percentage change in femoral-neck bone mineral density, a mean, and a standard deviation, it does not report the values or a comparison estimate. The direction, size, and precision of the primary finding therefore cannot be assessed. Numerical results for the listed secondary outcomes are also not reported in the registry record.

The participant counts are not consistent across the record’s descriptions: enrollment is 96, the group descriptions state 140 subjects in each group, and the primary outcome population is 28 on pioglitazone and 64 not on pioglitazone. The record provides no explanation for the differences. This prevents a clear account of the analyzed population and makes it difficult to understand how complete the outcome data were.

The comparison is nonrandomized as described, with allocation not stated and the registry explicitly noting that the arms are not a like-for-like control comparison. The record does not report participant characteristics, an adjustment method, or details that would allow assessment of confounding. Thus, even if a group difference were available, its causal interpretation would be limited.

The statistical methods are also incompletely described. The record does not report the model or test, the approach to missing data, a multiplicity strategy, or the study’s own sample-size justification. It does not explain why the femoral neck was chosen as the primary site or provide enough detail to assess how percentage change was derived. These are limits of the registry record; they should not be filled by assumptions about the protocol or analysis.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Effect of Diabetic Medications on Bone Metabolism — NCT00467285.
• Registry record last updated 2026-08-13; status last verified 2026-07.
• Results first posted 2015-01-14.
• 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.

18. Record Summary

NCT00467285 is listed as a completed study titled “Effect of Diabetic Medications on Bone Metabolism,” with the official title “Effect of Thiazolidinediones on Skeletal Health.” The sponsor is the VA Office of Research and Development. The listed start was 2006-10, the primary completion date was 2009-09, and enrollment is recorded as 96.

The study compares people with type 2 diabetes taking pioglitazone with people not taking pioglitazone. Its primary outcome is percentage change in femoral-neck bone mineral density over 6 months, summarized by a mean and standard deviation. The numerical primary outcome results are not reported in the registry record. The record lists 7 outcome measures and 1 analysis.

Group counts differ across the registry descriptions, and the record does not explain the discrepancy. Allocation is not stated, and the registry notes that the arms are not a like-for-like control comparison. The available record therefore documents the research question and broad outcome definitions, but does not permit evaluation of the primary result’s direction, size, or precision.