Insulin Resistance in Non-alcoholic Fatty Liver Disease: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Insulin Resistance in Non-alcoholic Fatty Liver Disease (NCT00252499), a randomized na study. The primary outcome is Liver/Spleen Ratio at 6 Months.

2. Clinical Question

3. Trial Design

The registry describes a randomized, triple-masked, parallel-group study. It lists three treatment arms: a placebo arm, a rosiglitazone arm, and a fenofibrate arm. The intervention descriptions specify matching placebo for both drugs in the placebo arm, rosiglitazone with fenofibrate placebo in one active-treatment arm, and micronized fenofibrate with rosiglitazone placebo in the other. No combined active-treatment arm is listed.

The enrollment entry is 13. The primary outcome summaries use generic arm labels: Arm 1 has n = 4.0, Arm 2 has n = 2.0, and Arm 3 has n = 6.0. The record identifies Arm 1 as the control, but does not explicitly associate Arm 2 and Arm 3 with the named regimens. These outcome counts should therefore be read as the posted arm-specific counts, not as a description of the complete analysis population.

The study start is recorded as 2005-10, and primary completion as 2010-08. Its status is TERMINATED. The phase is listed as NA, and the sponsor is the US Department of Veterans Affairs. The record does not state why the study was terminated, whether termination affected planned enrollment or follow-up, or what led to the dates recorded.

Randomization and masking describe features of the design, but the record does not report the allocation sequence method, concealment procedures, or who was masked. It also does not describe the duration or details of treatment beyond the listed regimens and the six-month outcome timeframe. The registry record does not report the protocol, sample-size justification, interim analyses, or a detailed statistical analysis plan.

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: Subjects underwent a screening visit prior to randomization to ensure that they qualified for the study.

Overall Study

Reasons for not completing this period:

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 Liver/Spleen Ratio at 6 Months, measured over 6 months.

The primary endpoint is the liver/spleen ratio at 6 months. It is a continuous measure derived from non-contrast CT: liver density is compared with spleen density using Hounsfield units. The registry describes the ratio as an estimate of liver fat and notes that it has previously been correlated with liver fat quantification by magnetic resonance spectroscopy.

The measurement procedure is described in some detail. Ten separate measurements, equally distributed throughout the liver and spleen, were obtained and the Hounsfield units averaged. For subjects with more than one slice through the liver and spleen, values across all slices were averaged. The endpoint is therefore a ratio-based imaging measure, rather than a direct report of a clinical event or a laboratory concentration.

In general, a continuous endpoint can retain information across a range of observed values, and a group mean can describe the central level of such a measure. The liver/spleen ratio is used as an imaging-based measure related to liver fat. Its statistical interpretation depends on how the measure is collected and analyzed, as well as on the uncertainty in the posted estimates. The registry reports means and standard errors for the arms, but does not report confidence intervals or an estimable two-arm comparison for this endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is continuous, and the posted statistic is a mean with standard error. The outcome is the liver/spleen ratio at 6 months. The mean summarizes the observed endpoint level within an arm, while the standard error describes the precision of that estimated mean. In general, a standard error is not the same as the standard deviation: the standard deviation describes variation among observations, while the standard error concerns uncertainty in the estimated mean. The registry reports standard errors, not standard deviations.

What can the arm summaries tell us?

Arm 1 is identified as the control and has a mean of 0.85 with standard error 0.08. Arm 2 has a mean of 0.96 with standard error 0.25, and Arm 3 has a mean of 0.6 with standard error 0.17. Their listed endpoint counts are n = 4.0, n = 2.0, and n = 6.0, respectively. These are descriptive summaries. A comparison requires an estimate of the contrast between groups and its uncertainty, usually reported with an inferential statistic or interval. The registry says a two-arm comparison for this continuous endpoint is missing or not estimable, and provides no such estimate.

Why is a confidence interval useful?

In general, a confidence interval presents a range of values compatible with the data under a specified statistical procedure. It can help readers judge the precision of an estimated effect and whether the data are consistent with a wide range of possible effects. No confidence intervals are reported for the endpoint means or for a comparison. The standard errors alone do not provide the full result of a between-arm analysis, and no interval should be reconstructed from them without further assumptions and a defined comparison.

What does randomization contribute?

In general, random assignment aims to make treatment groups comparable with respect to both measured and unmeasured characteristics on average. It does not guarantee that small groups will have similar characteristics, nor does the word randomized describe how assignment was generated or concealed. The registry identifies the study as randomized, but the sequence-generation method, concealment method, and baseline balance are not reported. The posted arm-level means therefore cannot be assessed for baseline comparability from these registry entries.

How should masking be understood?

The design is recorded as triple-masked. Masking can reduce the influence of expectations on treatment delivery, measurement, or reporting, depending on who is masked and how. The registry record does not specify the three masked parties or describe masking procedures. The label is useful design information, but it does not establish that masking was maintained throughout the study.

What is the role of the planning illustrations?

Planning illustrations based on the observed effect use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. They consider an effect to detect stated as a fraction of the observed effect. These are planning calculations, not observed power and not evidence that the trial was designed with those assumptions. The registry does not report its own sample-size justification or establish that the trial used these assumptions. The record also does not provide enough information in the registry entries to interpret a specific planning result as a formal design calculation.

What analysis details are absent?

The registry record does not report the statistical test or model for the primary outcome, whether any covariates were included, how missing observations were handled, or which analysis population was intended. It does not report an inferential comparison for the primary endpoint. The entry for analyses is 0. Accordingly, the arm means and standard errors should be read as posted descriptive results, not as a complete account of a treatment-effect analysis. There is no basis in the record to infer an adjusted result, an interim analysis, or a prespecified decision rule.

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("Arm 1", "Arm 2", "Arm 3")
n   <- c(4, 2, 6)
m   <- c(0.85, 0.96, 0.6)
sp  <- c(0.08, 0.25, 0.17)
sd  <- sp * sqrt(n)      # standard errors converted to SDs
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  Arm 1  4  0.850  0.160  0.595  1.105
CARM  Arm 2  2  0.960  0.354  -2.217  4.137
CARM  Arm 3  6  0.600  0.416  0.163  1.037

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome and five secondary outcomes. The secondary measures include change in alanine aminotransferase, change in the liver/spleen ratio by CT, change in peripheral insulin sensitivity, change in intra-abdominal fat area, and change in hepatic insulin sensitivity. The record reports six outcome measures in total, but the registry entries do not include results for the secondary outcomes.

When a study evaluates multiple outcomes or multiple comparisons, the chance of finding at least one apparently notable result can increase. In general, multiplicity procedures may be used to control error rates or to distinguish confirmatory from exploratory analyses. The registry record does not report whether multiplicity was considered, whether any adjustment was planned, or how the primary and secondary measures were prioritized in analysis. No conclusion about multiplicity control can be made from the number of listed outcomes alone.

The endpoint summaries include arm-specific counts, means, and standard errors, but the record does not state how missing measurements were handled, whether all randomized participants were analyzed, or which analysis population was used. The enrollment entry is 13, while the primary endpoint summaries report n = 4.0 for Arm 1, n = 2.0 for Arm 2, and n = 6.0 for Arm 3. The record does not explain the relationship between these entries, so they should not be used to infer the number or reasons for missing measurements.

The record also does not report imputation, sensitivity analyses, exclusions, or a missing-data assumption. Without those details, the potential influence of missing endpoint information on the observed summaries cannot be assessed. The two-arm comparison for the continuous primary endpoint is listed as missing or not estimable, and no analysis population or inferential procedure is described.

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.

Most frequent non-serious adverse event terms

The registry safety entries use generic arm labels. For serious adverse events, Arm 1 reports 0 affected among 5 at risk, Arm 2 reports 0 affected among 2 at risk, and Arm 3 reports 0 affected among 6 at risk. These are the serious-event counts recorded for the listed groups; they should not be interpreted as proof that serious events could not occur or as a complete safety profile.

For other adverse events, Arm 1 reports 3 affected among 5 at risk, Arm 2 reports 0 affected among 2 at risk, and Arm 3 reports 4 affected among 6 at risk. The registry record does not provide the event terms in the registry entries, nor does it clarify whether the counts represent participants with at least one event or another counting convention. Arm 1 is identified as the control, but the other generic labels are not explicitly linked to the named regimens in the outcome results.

Deaths are not reported in the registry record. The serious adverse event term list is empty. Event severity, timing, duration, and relationship to treatment are not described in these entries. The figures should therefore be read as limited recorded counts, not as a basis for comparing safety between named treatments.

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

The registry record describes a randomized, triple-masked, parallel-group study of rosiglitazone, fenofibrate, and placebo regimens in the context of fatty liver and insulin resistance. Its primary outcome is an imaging-based liver/spleen ratio at 6 months. The posted results give the mean and standard error for three generic arms, with Arm 1 identified as the control. This is useful descriptive information about the endpoint summaries that were recorded.

The results do not provide an estimable two-arm comparison for the primary continuous outcome. There is no reported p-value or confidence interval for a treatment contrast, and the generic outcome labels do not explicitly link the non-control summaries to the named regimens. Thus, the mean values should not be converted into claims that either named drug changed the outcome relative to placebo. The record also supplies no results for the listed secondary outcomes, so no finding about insulin sensitivity, liver enzymes, or other secondary measures can be inferred from those listings.

In general, a treatment-effect conclusion depends on more than a difference between descriptive means. It requires a defined comparison, a method for estimating uncertainty, and enough information about the analysis population and missing data to understand what the estimate represents. Those details are not reported for the primary endpoint. Nor does the registry report the reason for termination, a sample-size justification, or interim analyses.

The posted safety counts likewise do not establish comparative safety. They record no serious adverse events among the listed at-risk counts, along with counts for other adverse events, but do not give event terms or link every generic group label to a named regimen. The absence of reported serious events in these entries should not be interpreted as evidence of no risk.

Overall, the record supports a limited description of study design, endpoint measurement, and posted arm summaries. It does not support a definitive estimate of a treatment effect or a broad conclusion about the clinical implications of the interventions.

14. Why This Trial Matters Statistically

This trial is statistically informative as an example of the difference between a registered design and a complete inferential result. The registry identifies randomization, triple masking, and a parallel-group structure, and it defines a continuous imaging endpoint. Yet the primary endpoint entries are descriptive arm summaries, and the record states that a two-arm comparison is missing or not estimable. Design labels alone cannot replace a reported contrast and its uncertainty.

The endpoint also illustrates why measurement and analysis details matter. The liver/spleen ratio is derived from CT density measurements, with averaging across locations and, when applicable, slices. That description helps explain what the outcome represents. At the same time, interpretation of the mean requires knowing how participants were selected for the analysis, how missing measurements were handled, and how groups were compared. Those analysis details are not reported in the registry record.

The listed arm counts are limited, and the recorded endpoint summaries use generic labels. This creates a practical interpretive constraint: readers can describe the three means and standard errors, but cannot securely attribute the non-control summaries to the named treatment regimens from the registry entries. The study's terminated status adds context, but the reason for termination is not reported, so its effect on the planned analyses or collected data cannot be assessed.

Finally, the study lists several related secondary outcomes. In general, multiple endpoints can provide a broader view of a research question, while also making prespecification and multiplicity handling important. The record does not report how those issues were addressed, and no secondary outcome results are included in the registry entries. The central statistical lesson is that endpoint definitions and design features are valuable, but conclusions depend on transparent reporting of comparisons, uncertainty, analysis populations, and missing-data handling.

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 primary limitation for interpreting this specific record is that it does not report an estimable two-arm comparison for the continuous primary endpoint. Although the registry gives means and standard errors for three generic arm labels, it provides no confidence interval, p-value, or other inferential comparison. Arm 1 is identified as the control, but the record does not explicitly connect Arm 2 and Arm 3 to their named treatment regimens in the outcome results.

The outcome counts are n = 4.0, n = 2.0, and n = 6.0 across the three posted arms. The enrollment entry is 13, but the registry does not explain how these entries relate to the analysis population or whether endpoint observations were missing. It also does not describe exclusions, imputation, or sensitivity analyses. The reasons for termination are not reported, which limits understanding of whether the recorded endpoint data reflect the intended study conduct.

Additional details needed to evaluate the analysis are not reported in the registry record: the statistical test or model, covariate adjustment, analysis population, allocation sequence and concealment procedures, and the parties masked. The record gives no sample-size justification or interim-analysis information. It also does not report multiplicity handling for the five secondary outcomes, and their results are absent from the registry entries.

The safety entries are limited to event counts by generic arm label. Serious adverse event counts are recorded as zero affected in each group, but event definitions, terms for other adverse events, timing, and relationship to treatment are not described. Deaths are not reported in the registry record. These limitations preclude a broad assessment of comparative safety as well as a firm treatment-effect interpretation.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Insulin Resistance in Non-alcoholic Fatty Liver Disease — NCT00252499.
• Registry record last updated 2014-08-20; status last verified 2014-08.
• Results first posted 2014-04-11.
• 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

The registry record describes NCT00252499, titled “Insulin Resistance in Non-alcoholic Fatty Liver Disease.” It lists fatty liver and insulin resistance as conditions, randomization, triple masking, and a parallel-group model. The study is marked TERMINATED; its start is recorded as 2005-10 and primary completion as 2010-08. The phase is NA, and the sponsor is the US Department of Veterans Affairs.

Three arms are described: placebo, rosiglitazone with fenofibrate placebo, and micronized fenofibrate with rosiglitazone placebo. The primary outcome is the liver/spleen ratio at 6 months, measured by non-contrast CT and reported as a mean with standard error. Arm-specific means are 0.85 for Arm 1, 0.96 for Arm 2, and 0.6 for Arm 3. The record identifies Arm 1 as control but does not explicitly map the other generic labels to named regimens.

The registry lists five secondary outcomes but provides no results for them in the registry entries. It does not report an estimable two-arm comparison for the primary continuous endpoint, the reason for termination, or the detailed analysis and missing-data methods.