Effects of Rosiglitazone on the Metabolic Phenotype of Impaired Glucose Tolerance in Youth: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effects of Rosiglitazone on the Metabolic Phenotype of Impaired Glucose Tolerance in Youth (NCT00413335), a randomized na study. The primary outcome is Mean Percent Change From Baseline in Whole-body Insulin Sensitivity.

2. Clinical Question

The endpoint is a measure of insulin sensitivity calculated from insulin and glucose values obtained during an oral glucose tolerance test and from corresponding fasting values. The trial also lists mean percent change in adiponectin as a secondary outcome. Numeric results for that outcome are not reported in the registry record.

3. Trial Design

The registry describes a randomized, parallel-group study comparing rosiglitazone with placebo. It records enrollment of 21 participants: 12 in the active arm and 9 in the placebo arm. The study is marked completed. Its listed phase is NA, so a clinical development phase is not specified.

The overall masking designation is quadruple. The placebo arm description separately says participants were randomized double-blind. The record does not identify which groups or roles were masked, or explain the difference between these descriptions. The masking details therefore cannot be interpreted more specifically from the registry record.

Both arm descriptions include baseline metabolic tests and imaging. These include an oral glucose tolerance test, a hyperinsulinemic-euglycemic clamp, and imaging or body-composition assessments. Participants were followed every 2 weeks according to the descriptions. Both arms mention imaging at 2 months and a visit at 12 weeks. The active-arm description says all tests were repeated at 4 months, whereas the placebo-arm description says all tests were repeated at 2 months. The registry does not clarify whether this reflects a difference in protocol, a recording error, or another explanation.

The primary endpoint’s timeframe is 4 months. Its registry results are reported as a mean percent change from baseline, with standard deviations. The record does not provide a protocol, a sample-size justification, details of the randomization procedure, allocation concealment, or a fuller schedule of assessments. Although the registry describes the study as randomized, those additional design details are not reported in the registry record.

The listed sponsor is Yale University. The study start is recorded as 2005-11 and primary completion as 2008-09. These dates describe the registry record and do not supply additional information about trial conduct or analysis.

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: Obese children and adolescents with positive risk factors for Type 2 Diabetes (T2DM) were screened by using a standard oral glucose tolerance test (OCTT).

Overall Study

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 Mean Percent Change From Baseline in Whole-body Insulin Sensitivity, measured over 4 months.

The primary endpoint is mean percent change from baseline in whole-body insulin sensitivity, measured at 4 months. Insulin sensitivity describes how responsive the body is to insulin. The registry defines its whole-body insulin sensitivity index using insulin and glucose values from the oral glucose tolerance test together with corresponding fasting values.

Combining measurements from the glucose tolerance test and fasting state gives the endpoint a broader metabolic basis than a single glucose reading. The registry’s description identifies the index as WBISI and specifies that it is calculated from glucose and insulin values. The endpoint is therefore a derived continuous measure, rather than a direct count of events or a simple laboratory value.

A percent change from baseline expresses a follow-up measurement relative to the participant’s baseline measurement. A group mean summarizes those changes across participants in that arm. The standard deviation describes how dispersed the participant-level changes are around the group mean; it is not a confidence interval and does not itself quantify uncertainty in the group comparison.

The endpoint can describe the direction and spread of the recorded changes within each arm. A treatment comparison, however, needs an estimate of the difference between arms and an account of its uncertainty. The registry does not report a confidence interval or inferential test for this primary comparison. The outcome values alone therefore do not establish how precisely the groups differ.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant’s result is represented by a numeric percent change, and the registry summarizes each arm using a mean and standard deviation. This type of summary describes both a typical value, through the mean, and variation among individual values, through the standard deviation. Neither statistic alone provides a direct estimate of how different the randomized groups are.

What does a group mean tell us?

A mean is the arithmetic average of the observed values in a group. For this endpoint, it summarizes the percent changes in whole-body insulin sensitivity recorded for participants in an arm. It does not show whether all participants had similar changes or whether the mean was influenced by a small number of unusually large or small values. The standard deviation adds information about spread, but it does not show the full distribution, identify outliers, or describe the uncertainty around the mean.

Here, the active-arm mean is 87.4 and its standard deviation is 77.0. The placebo-arm mean is 63.1 and its standard deviation is 61.6. The group means can be read alongside their standard deviations as descriptive summaries. A reader should not treat the standard deviations as confidence limits or use them as a substitute for a reported comparison analysis.

Why is a between-group estimate needed?

In a randomized study, the principal treatment comparison is generally framed as a contrast between the groups’ outcomes. An estimate of that contrast describes the observed separation; a confidence interval describes the range of values reasonably compatible with the data and the statistical method used. A p-value, when reported, summarizes how compatible the data are with a specified null hypothesis under the analysis assumptions. These concepts are general statistical explanations. The registry does not report a confidence interval, p-value, or contrast estimate for this endpoint.

Without those comparison statistics, the difference in the group means cannot be judged for precision using the registry results. Random assignment supports a comparison in the trial design, but it does not remove sampling uncertainty, and it does not supply a missing estimate or analysis. The registry’s arm-level summaries are not enough to reconstruct every possible comparison method, particularly without participant-level data or a stated analysis plan.

Why does variability matter?

Variation among participants affects how precisely a group mean or between-group contrast can be estimated. Larger spread relative to the information available from the participants generally makes estimates less precise. This is a general principle. The standard deviations are broad relative to their respective means, but the registry does not provide distribution plots or participant-level observations that would support a more detailed assessment of the data’s shape.

Some continuous outcomes can have skewed distributions or unusually influential observations. In such settings, an analysis based only on means and standard deviations may not convey all relevant features. The registry does not report distribution checks, transformations, robust methods, or alternative summaries. It is not possible to infer that any particular method was used or needed.

What does the planning illustration mean?

The registry’s sample-size illustration is based on the observed effect and states assumptions of a two-sided significance level of 0.05, equal allocation, and a normal approximation. It considers detecting a stated fraction of the observed effect. Such figures are planning illustrations: they answer a hypothetical question about the number of participants per arm under specified assumptions. They are not the number enrolled in this trial, do not establish that the original trial was designed around those assumptions, and are not observed power.

A planning calculation based on an observed effect can be especially uncertain when that effect comes from a small study. It should not be mistaken for evidence that a future study will obtain the same effect or precision. The registry does not report the trial’s own sample-size justification, so the illustration cannot be treated as the original design rationale.

What cannot be inferred from the posted analysis?

The registry records zero analyses for the outcomes and does not provide an inferential primary comparison. That does not establish that no analysis was conducted outside the registry; it means the analysis is not reported in the registry record. The record also does not identify covariate adjustment, handling of repeated observations, a model for the outcome, or a strategy for missing data. No method should be attributed to the investigators without that information.

The outcome’s statistical unit and calculation matter as well. The registry says the index is calculated from glucose and insulin measurements obtained during the oral glucose tolerance test and corresponding fasting values. The endpoint is thus a derived measure. The registry does not report analytical validation details or explain how measurement quality and missing component values were addressed.

These limitations do not invalidate the descriptive means. They define what can be concluded from them: the registry reports arm-specific average changes and their dispersion, while leaving the uncertainty and method for the between-arm comparison unreported.

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("Active Arm (Rosiglitazone)", "Inactive Arm (Placebo)")
n   <- c(12, 9)
m   <- c(87.4, 63.1)
sp  <- c(77, 61.6)
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  Active Arm (Rosiglitazone)  12  87.400  77.000  38.477  136.323
CARM  Inactive Arm (Placebo)  9  63.100  61.600  15.750  110.450

11. Multiplicity, Missing Data and Analysis Populations

The registry lists five outcome measures and identifies mean percent change from baseline in adiponectin as a secondary outcome. Numeric results for adiponectin are not reported in the registry record. The names and results of the other outcome measures are also not fully described in the posted outcome information.

Multiplicity concerns arise when several outcomes or comparisons are assessed. If many statistical tests are conducted, the chance of at least one apparently notable result can increase unless the analysis plan addresses that issue. This is a general statistical explanation. The registry does not report a multiplicity strategy, a hierarchy of outcomes, or a detailed analysis plan. It is not possible to determine whether any adjustment or prespecified prioritization was used.

The record does not describe missing-data methods, the number of participants contributing to each outcome, or the analysis population. It therefore does not say whether the reported means include everyone randomized, only participants with complete measurements, or another defined group. It also does not explain how participants with missing endpoint measurements were treated.

These details matter because missing outcomes can affect the estimated group summaries and the comparability created by randomization. This is a general concern, not evidence that data were missing in this trial. The registry does not provide enough information to establish whether missing data occurred or how they were handled. No interim-analysis plan, stopping rule, or other analysis-planning detail is reported in the registry record.

12. 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.

For serious adverse events, the registry lists 12 participants at risk in the rosiglitazone arm and 9 at risk in the placebo arm. The number affected is recorded as 0 in each arm. These figures mean that no participants are recorded as having a serious adverse event in the posted results. The registry lists no serious-event terms.

The record also gives 12 and 9 participants at risk for other adverse events, with 0 affected in each arm. These are counts of participants recorded as affected, not a complete description of every event, its severity, duration, or relationship to study treatment. The adverse-event definitions, observation period, collection methods, and attribution process are not reported in the registry record.

Deaths are not reported in the registry record. The zero affected counts should not be interpreted as proof that no adverse experiences occurred outside the recorded categories or observation process. They describe what is posted for this small study, and do not establish the absence of risk.

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

The registry results show arm-specific means for percent change in whole-body insulin sensitivity: 87.4 for rosiglitazone and 63.1 for placebo, with standard deviations of 77.0 and 61.6, respectively. The active-arm mean is numerically higher. The posted figures describe the observed summaries, but they do not provide a confidence interval or formal comparison that would show the precision of the between-group result.

The numbers do not by themselves establish that rosiglitazone caused a difference. A randomized design is relevant to causal comparison, but interpretation also depends on how outcomes were analyzed, whether participants were included as planned, and how missing values were treated. Those analysis details are not reported in the registry record. The unresolved difference between the endpoint timeframe and the placebo-arm testing description adds uncertainty about the assessment schedule.

The results also do not establish that there was no difference. Absence of an inferential comparison is not the same as evidence that the groups were equivalent. The registry does not report an equivalence or noninferiority framework, a prespecified margin, or a confidence interval that could be used to assess such a claim.

The safety entries record zero participants affected by serious adverse events in both arms and zero participants affected by other adverse events in both arms. These entries are limited to the categories and information posted in the registry. They should not be read as a broad guarantee about safety, particularly because definitions and observation details are not reported.

The appropriate statistical reading is therefore bounded: the registry presents descriptive outcome and adverse-event counts for a small randomized study, while leaving important comparative and analytic information unreported. The figures can inform understanding of what was recorded, but they do not support a definitive claim that the intervention works or does not work.

14. Why This Trial Matters Statistically

This trial illustrates the difference between reporting group summaries and reporting a treatment comparison. Means and standard deviations can communicate the observed outcome levels and their spread, but they do not replace a contrast estimate with an uncertainty measure. For a randomized trial, readers need those comparative statistics to judge how precisely the observed groups differ.

The example also highlights the importance of clear outcome timing. The primary endpoint is assigned a 4-month timeframe, while the placebo-arm description says all tests were repeated at 2 months. Such ambiguity makes it harder to connect the outcome label to the assessment schedule. Clear, consistent reporting helps readers understand which observations contribute to each endpoint.

Small enrollment and wide variation are another statistical consideration. A study with 12 participants in one arm and 9 in the other has limited information for estimating a group contrast, and variability can make that estimate less precise. This general point does not determine the result; it explains why uncertainty reporting is particularly important.

Finally, a sample-size illustration based on an observed effect is not the same as a prespecified design justification or observed power. Keeping those concepts distinct prevents retrospective planning calculations from being mistaken for evidence of a study’s original statistical design. The registry record leaves the original sample-size rationale and analysis plan unreported.

15. 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 registry does not provide a formal statistical comparison for the primary endpoint. It reports group means and standard deviations but no between-arm estimate, confidence interval, or p-value. The precision and inferential strength of the apparent difference therefore cannot be evaluated from the posted outcome statistics.

The enrolled groups are small, with 12 participants in the rosiglitazone arm and 9 in the placebo arm. The standard deviations are sizable, and no participant-level data or distributional summaries are reported. The influence of individual observations and the shape of the outcome distributions cannot be assessed from the registry record.

Outcome timing is not fully clear. The primary endpoint is listed at 4 months, the active-arm description says all tests were repeated at 4 months, and the placebo-arm description says all tests were repeated at 2 months. The record does not reconcile those statements. The overall masking designation is quadruple, while the placebo-arm description uses the term double-blind; the details and relationship between those descriptions are not reported.

Eligibility criteria, baseline characteristics, missing-data handling, analysis populations, and the statistical model are not reported in the registry record. The record also does not provide the trial’s sample-size justification, a multiplicity plan, or a detailed protocol. Five outcome measures are recorded, but only one secondary outcome is named and its numeric results are not reported. Together, these omissions restrict assessment of the completeness, comparability, and precision of the posted findings.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Effects of Rosiglitazone on the Metabolic Phenotype of Impaired Glucose Tolerance in Youth — NCT00413335.
• Registry record last updated 2013-07-22; status last verified 2013-07.
• Results first posted 2013-04-19.
• 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.

19. Record Summary

NCT00413335 is a completed randomized, parallel-group study of rosiglitazone versus placebo in youth with impaired glucose tolerance. The registry records enrollment of 21 participants, with 12 assigned to the active arm and 9 to placebo. The overall masking designation is quadruple.

The primary endpoint is mean percent change from baseline in whole-body insulin sensitivity at 4 months. The registry reports arm means of 87.4 and 63.1, with standard deviations of 77.0 and 61.6. It does not report an inferential comparison, confidence interval, or p-value. Mean percent change in adiponectin is listed as a secondary outcome, without numeric results.

Serious adverse events and other adverse events are each recorded as affecting 0 participants in both arms. Deaths are not reported in the registry record. The statistical analysis plan, missing-data methods, sample-size justification, and several other design details are also not reported.