Colesevelam Versus Placebo in the Treatment of Nonalcoholic Steatohepatitis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Colesevelam Versus Placebo in the Treatment of Nonalcoholic Steatohepatitis (NCT01066364), a randomized phase 2 study. The primary outcome is The Primary Outcome Will be Improvement in Hepatic Steatosis by Liver MRI.

2. Clinical Question

3. Trial Design

This was a phase 2, randomized, parallel-group trial of colesevelam versus placebo for nonalcoholic steatohepatitis. The registry lists University of California, San Diego as the sponsor and records enrollment of 59 participants. The study status is completed. The listed study start was 2010-02, and primary completion was 2012-02.

Randomization means that assignment to study groups is intended to use chance rather than investigator choice. The registry identifies the allocation as randomized, but does not report the sequence-generation method, allocation concealment procedures, or any stratification factors. Those details are not reported in the registry record.

The design is listed as quadruple-masked. Masking is intended to limit the influence of treatment knowledge on participants’ experiences, clinical management, outcome assessment, or analysis, depending on who is masked. Which specific roles were masked and how masking was maintained are not reported in the registry record. The placebo was described as six tablets per day, identical to colesevelam; the colesevelam arm received 3.75 grams per day.

The primary outcome timeframe was 24 weeks. The registry lists insulin sensitivity measured by HOMA-IR, serum ALT and AST values, and lipid profiles as secondary outcomes. It records four outcome measures in total, while only one secondary outcome has posted arm-level results and a comparison. Analyses for the remaining listed secondary outcomes are not reported in the registry record.

The registry record does not report the trial’s original sample-size justification, its planned primary-analysis population, the detailed statistical analysis plan, or the approach to missing outcome data. These omissions limit how fully the posted estimates can be connected to the original design and analysis decisions.

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

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 The Primary Outcome Will be Improvement in Hepatic Steatosis by Liver MRI, measured over 24 weeks.

The primary endpoint was improvement in hepatic steatosis measured by liver MRI, expressed as percentage of fat. Liver fat is a continuous measurement: participants can have values across a range rather than being classified only as having or not having an outcome. The registry describes the comparison as testing whether colesevelam at 3.75 grams per day decreased liver fat compared with placebo in people with biopsy-proven nonalcoholic steatohepatitis.

A continuous measurement can retain information about the degree of liver fat and makes it possible to compare group averages. In the posted results, the outcome is summarized by a mean and standard deviation for each arm. The mean describes the center of the observed values, while the standard deviation describes their spread around that mean. These summaries do not show the full distribution of individual values or whether every participant changed in the same direction.

Liver MRI was the specified measurement method for this endpoint. In general, an imaging-based measure can quantify a feature of liver composition without reducing the result to a binary event. The registry record does not report the MRI acquisition or reading procedures, the precise calculation of the percentage-fat measure, or whether image readers were masked. The posted endpoint therefore identifies the intended measurement and timeframe, but not all technical details needed to assess its reproducibility.

The endpoint is a measure of liver fat, not a direct measure of every aspect of nonalcoholic steatohepatitis. A group difference on this measure should be interpreted as a comparison on the stated MRI outcome, rather than as a complete account of disease status or all outcomes relevant to participants.

7. Results Posted in the Registry

• Colesevelam Arm: mean 14.20 versus 17.90 on control; difference -3.70 (95% CI -7.79 to 0.39), standardized effect d = -0.53; z-test p = 0.076.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Insulin Sensitivity as Determined by HOMA-IR

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Colesevelam Arm: mean 7.60 versus 8.30 on control; difference -0.70 (95% CI -6.86 to 5.46), standardized effect d = -0.06; z-test p = 0.824.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Colesevelam Arm and the control. The interval gives the range of effects that remain compatible with the data.

The registry lists insulin sensitivity measured by HOMA-IR, serum ALT and AST values, and lipid profiles as secondary outcomes. Arm-level results and a comparison are posted for HOMA-IR, but results for serum ALT and AST and lipid profiles are not reported in the registry record. The available secondary evidence is therefore incomplete relative to the full set of listed measures.

The HOMA-IR comparison has a lower mean in the colesevelam arm than in the placebo arm, in the same general direction as the primary comparison. Its confidence interval is broad and includes zero, so it does not supply clear, precise corroboration of the primary result. It should be considered as a separate secondary measurement rather than as confirmation of the MRI finding.

Secondary endpoints can help describe whether a pattern extends across different measures, but they typically carry less weight than a prespecified primary endpoint when the analysis plan and multiplicity procedures are unavailable. Examining several outcomes creates more opportunities for chance findings. If many estimates are considered, an isolated small p-value may be less persuasive unless the analysis accounts for the number of comparisons or the outcome hierarchy. No multiplicity-adjusted secondary analyses are reported in the registry record.

The registry identifies four outcome measures in total, but only one secondary outcome has posted analyses. The reason other listed outcomes lack results is not reported in the registry record. Without a full accounting of planned and reported outcomes, readers cannot determine whether the available secondary results represent the complete set of analyses or whether other results were omitted. The secondary record thus provides limited context and should not be used to overcome the uncertainty around the primary comparison.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Insulin Sensitivity as Determined by HOMA-IR
• Serum ALT and AST Values
• Lipid Profiles

9. Statistical Interpretation

10. Statistical Methods Explained

For the primary endpoint, the registry reports a continuous outcome summarized as a mean and standard deviation in each randomized group. This is a common way to describe measurements such as percentage of liver fat. The mean gives a group-level center, and the standard deviation indicates how dispersed individual measurements were around that center. Neither summary alone reveals the shape of the data, outliers, or how measurements changed within individual participants.

What does the mean difference represent?

The comparison is expressed as a mean difference for colesevelam relative to placebo. It is calculated in the direction colesevelam minus placebo, so a negative value corresponds to a lower group mean in the colesevelam arm. The estimate of -3.7 is in the outcome’s percentage-of-fat units. It is an absolute difference on that measurement scale, not a relative percentage change. The registry does not report a baseline-adjusted primary estimate, so the displayed comparison should not be described as an adjustment for baseline values.

How should the confidence interval be understood?

A confidence interval gives a range of values compatible with the data and the method used to estimate uncertainty. In general, a narrower interval indicates greater precision, while a wider interval indicates less precision. Here, the posted interval extends from -7.787 to just above zero. It includes zero, which represents no mean difference on this scale. Thus the data are compatible with no difference and with negative differences of varying magnitude. A confidence interval is not a guarantee that the true value lies within its endpoints, and it does not express the probability that a particular effect is true.

What does the p-value say?

The unadjusted p-value is 0.076. In general, a p-value describes how unusual data at least as incompatible with a specified null hypothesis would be, if that null hypothesis and the statistical model were true. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. The registry’s sample-size illustration specifies a two-sided significance level of 0.05 for its planning calculations, but that is a planning assumption for the illustration. The trial’s original significance threshold and primary analysis plan are not reported in the registry record. The p-value should therefore be read as a description of the posted comparison, not as a complete account of the trial’s planned inferential decision.

The p-value and confidence interval address related aspects of uncertainty, but neither indicates whether the observed difference is important in practice. The interval emphasizes the range of plausible mean differences on the outcome scale; the p-value summarizes compatibility with a particular null model. Interpretation depends on the design, the reliability of the endpoint, the analysis population, and whether the estimate was obtained as planned. Several of those details are not reported in the registry record.

Why does variation matter?

The standard deviations were 7.7 in the placebo arm and 6.3 in the colesevelam arm. These describe variation among the observations, not uncertainty in the group means. More variability makes it harder to estimate a group difference precisely when the number of observations is limited. The standard errors reported for the arm means also reflect precision of those means; they are not additional measures of the spread of individual outcome values.

For continuous outcomes, a mean-based comparison is most informative when the mean is a useful description of the distribution and the analysis assumptions are reasonable. The registry does not report distributional checks, transformations, or model diagnostics. A standard deviation cannot establish that the measurement distribution was symmetric or that unusual observations had no influence. Nor does the registry report whether a model adjusted for other variables.

What does randomization contribute?

Random assignment is intended to make the groups comparable on average at the start of a trial, including on factors that were not measured. It supports a causal interpretation when assignment is implemented as planned, outcome data are adequately observed, and the analysis respects the randomized groups. The registry record does not report baseline characteristics, adherence, treatment discontinuation, or the analysis population. It is therefore not possible from the posted information alone to assess the degree to which those conditions were met.

The posted group summaries and comparison support a descriptive statement about the observed arms. A causal statement requires more than the numerical contrast: it depends on trial conduct and analysis choices. The registry identifies randomization and masking, but the detailed procedures and analysis decisions are not reported in the registry record. This is why the estimate is best read with its interval and with attention to the limitations of the available record.

11. Confidence Intervals and Precision

The primary mean-difference estimate is -3.7, with a confidence interval from -7.787 to just above zero. The interval crosses zero, so the observed data do not pin the comparison to a single direction with high precision. Values near zero remain compatible with the reported uncertainty, as do more negative values. The interval is therefore more informative than the point estimate alone: it shows that the estimate has meaningful uncertainty on the percentage-of-fat scale.

Precision is influenced by the amount of information available, including the number of observed outcomes and the variability among participants. The posted primary summaries include 22 participants in the placebo arm and 24 in the colesevelam arm, with standard deviations of 7.7 and 6.3, respectively. The record lists 59 enrolled participants, but does not report why the primary arm counts are lower than enrollment or how missing primary outcomes were handled. That missing context matters when considering the precision and representativeness of the estimate.

The interval is an uncertainty range under the method used; it is not a range of individual participant outcomes. The standard deviations describe individual-level spread, whereas the confidence interval describes uncertainty about the group comparison. The registry does not report an adjusted interval, a model specification, or analysis diagnostics. The posted comparison should consequently be interpreted as a summary estimate with the reported uncertainty, rather than as a complete description of all possible sources of uncertainty.

For the secondary HOMA-IR comparison, the reported interval also spans zero and is wide relative to its estimate. That result does not sharpen the primary estimate. Taken together, the posted intervals support a cautious reading: the available summaries leave substantial uncertainty about the magnitude of any group difference.

12. 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 sample-size estimates are planning illustrations based on the observed effect, not the original design assumptions for this trial. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The planning effect is set to a stated fraction of the observed effect. For these continuous outcomes, the figures are participants per arm, not a total across both arms and not total events.

The 80% power column gives the number of participants per arm estimated to provide an 80% chance of detecting the specified effect under those planning assumptions, if that effect and the assumptions were correct. The 90% power column describes the corresponding planning target for a 90% chance. These are design calculations, not statements about the probability that the trial result is true or about the likelihood that the completed study detected an effect.

The estimates include planning for 100%, 75%, and 50% of the observed effect. Planning for a smaller effect requires more participants per arm because a smaller separation between group means is harder to distinguish from ordinary sampling variation. The registry record does not report the assumptions used to select the original enrollment target or whether these later planning illustrations match the trial’s design process.

These estimates are not observed power. Observed power calculated from the study’s estimated effect would largely restate the uncertainty already conveyed by the estimate and its confidence interval. The planning values are more usefully read as conditional illustrations of how sample-size requirements change under specified assumptions, rather than as evidence about the result itself.

13. 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 = -3.700; pooled standard deviation σ = 7.003.
2. n = 2 × (1.96 + 0.842)² × 7.003² / -3.700² = 56.2, rounded up to 57 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 <- -3.7                      # observed difference in means
sigma <- 7.00318109536984                     # 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(1066364)
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 = 57 (80% power), 76 (90% power)
 75% of observed effect: n per arm = 100 (80% power), 134 (90% power)
 50% of observed effect: n per arm = 225 (80% power), 302 (90% power)
power.t.test: 57.2 
Simulated power at n = 57 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.

14. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when multiple outcomes, comparisons, timepoints, or analytical choices are considered. In general, the more statistical tests conducted, the greater the chance that at least one result may appear unusual by chance alone. Prespecified testing hierarchies or adjustment procedures can help control that risk, depending on the study’s goals. The registry lists several secondary outcome areas in addition to the primary MRI measure, but does not report a multiplicity strategy or a complete analysis plan.

The registry reports an unadjusted p-value for the primary comparison and an unadjusted secondary comparison for HOMA-IR. Whether the trial’s planned primary test used the same model, whether any adjusted analyses were conducted, and how secondary tests were to be interpreted are not reported in the registry record. The posted secondary findings should therefore be treated as exploratory context unless the intended testing framework is otherwise established.

Missing data are another important uncertainty. The record lists 59 participants enrolled, while the primary arm summaries include 22 and 24 participants. Secondary HOMA-IR summaries include 25 participants in each arm. The reasons these outcome-specific counts differ, whether all randomized participants are represented, and how missing measurements or withdrawals were handled are not reported in the registry record. The record also does not report whether any imputation, sensitivity analysis, or missing-data model was used.

The analysis population is not identified. The registry does not say whether the primary estimate follows an intention-to-treat approach, includes only participants with observed measurements, or uses another population definition. This matters because excluding participants after randomization can affect the comparability created by random assignment, particularly if missingness is related to treatment or outcome. Without information on analysis populations and missingness, the direction or size of any resulting influence cannot be inferred.

Baseline adjustment, covariate selection, model diagnostics, and handling of protocol deviations are also not reported in the registry record. These omissions do not invalidate the posted summaries, but they limit assessment of whether the displayed comparison reflects the trial’s prespecified analysis and whether alternative reasonable analyses would give a similar result.

15. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent non-serious adverse event terms

The adverse-event record lists 25 participants at risk for serious adverse events in each arm. It reports zero participants affected by a serious adverse event in the placebo arm and zero in the colesevelam arm. No deaths are listed in either arm. The registry does not list serious adverse event terms; specific serious-event descriptions are therefore not reported in the registry record.

For other adverse events, the record lists 2 affected participants among 25 at risk in the placebo arm and 1 among 25 at risk in the colesevelam arm. These counts are descriptive. The registry does not provide event-level detail, severity, relatedness, timing, or statistical comparisons for these other events. The counts should not be interpreted as a comprehensive characterization of safety.

Zero reported serious adverse events in these groups does not establish that serious events cannot occur, and the small number of participants limits the ability to characterize uncommon events. The posted safety information should be read as what was recorded for the trial, not as evidence that a treatment is free of risk.

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

The primary result describes a difference between group averages in liver fat measured by MRI after 24 weeks. The colesevelam group had a lower posted mean than the placebo group, and the comparison estimate was negative when expressed as colesevelam minus placebo. The confidence interval extends across zero, leaving uncertainty about the size and direction of the underlying mean difference.

This finding does not establish that colesevelam caused the observed difference. Randomization supports a causal comparison in principle, but interpretation also depends on implementation, follow-up, outcome availability, and analysis. The registry identifies randomization and masking but does not report enough detail on these aspects to fully evaluate them. The trial’s planned analysis, missing-data handling, and analysis population are not reported in the registry record.

The MRI outcome concerns liver fat, not every aspect of nonalcoholic steatohepatitis or every outcome important to participants. A result on this measure cannot, by itself, establish effects on other listed outcomes such as liver enzymes, lipid profiles, or insulin sensitivity. The registry reports one secondary endpoint analysis and does not report results for the other listed secondary outcomes. Nor does the primary mean comparison show how many individuals experienced a decrease or whether the group distributions overlapped.

The enrolled sample was 59, and the primary arm summaries represent fewer participants than that enrollment total. The record does not explain the difference or report missing-outcome procedures. Consequently, the posted estimates may not provide a complete account of outcomes for everyone enrolled.

The study is identified as phase 2. The registry record does not report a sample-size justification or the original statistical assumptions used for design. The sample-size calculations associated with the observed effect are planning illustrations only and should not be treated as evidence of success or failure. Overall, the result is a limited estimate of a specified MRI measure in this trial, with appreciable uncertainty and important gaps in the registry’s reporting.

17. Why This Trial Matters Statistically

This trial illustrates why an estimated group difference should be read together with its confidence interval. The point estimate indicates the direction and size of the observed contrast, but the interval makes clear that the estimate is uncertain and remains compatible with no difference. Focusing on the point estimate alone would give an incomplete account of what the data can support.

It also illustrates the distinction between variation among participants and precision of a group comparison. The standard deviations describe spread in individual liver-fat measurements, while the confidence interval reflects uncertainty around the contrast between means. These are related but different statistical quantities. The outcome counts and variability help explain why conclusions should remain cautious, without implying that any one feature alone determines the result.

Registry completeness matters for interpreting clinical-trial statistics. The record identifies the design, primary endpoint, posted arm summaries, and a secondary comparison, but does not report the original analysis plan, missing-data approach, or full analysis population. It lists multiple secondary outcomes while only one has posted results. Such gaps make it harder to distinguish confirmatory findings from exploratory analyses and to assess whether the reported comparisons were selected from a larger set.

Finally, sample-size planning illustrations are not substitutes for the trial’s original design justification or for uncertainty in the observed estimate. They show how hypothetical planning targets vary with the effect assumed, while the confidence interval describes the precision of the completed comparison. Keeping those roles separate helps prevent planning calculations from being mistaken for evidence about the result.

18. 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 analysis counts are lower than the enrollment total, and the registry does not explain why. It also does not report the number randomized to each arm separately from the outcome-specific counts, participant flow, withdrawals, or missing primary measurements. This prevents a clear assessment of how representative the posted arm summaries are of all enrolled participants.

The registry does not identify the primary analysis population or state whether the comparison was adjusted for baseline liver fat or other covariates. Baseline outcome values are not reported in the posted summary. Without these details, the analysis cannot be assessed against a prespecified model or evaluated for sensitivity to baseline imbalance.

Technical MRI procedures, central reading, quality control, and the handling of readings are not reported in the registry record. The record also does not explain the precise calculation of the percentage-fat measure. These omissions limit assessment of measurement consistency and possible observer or measurement effects, even though the study is listed as quadruple-masked.

Only one secondary outcome has posted analyses, although the registry lists other secondary measures. No detailed multiplicity plan or complete analysis plan is reported. This makes it difficult to place the secondary result within the full set of intended tests. The registry also does not report adherence, treatment discontinuation, protocol deviations, or the statistical approach to missing data.

Finally, serious adverse events are recorded as zero in both arms, but the trial’s groups are small and the record offers little event-level detail. These safety summaries cannot characterize uncommon outcomes. The limitations reflect what is and is not documented in the registry record; they do not establish that an unreported procedure was or was not performed.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Colesevelam Versus Placebo in the Treatment of Nonalcoholic Steatohepatitis — NCT01066364.
• Registry record last updated 2020-07-21; status last verified 2020-07.
• Results first posted 2020-07-21.
• 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.

22. Record Summary

NCT01066364 is a completed phase 2 randomized, quadruple-masked, parallel-group trial sponsored by University of California, San Diego. It enrolled 59 participants and compared colesevelam at 3.75 grams per day with a placebo described as six identical tablets per day. The primary outcome was liver fat measured by MRI at 24 weeks in participants with biopsy-proven nonalcoholic steatohepatitis.

The posted primary summaries show mean liver-fat values of 17.9 for placebo and 14.2 for colesevelam, with a mean difference of -3.7 and a confidence interval extending from -7.787 to just above zero. The record also posts HOMA-IR results, but does not report analyses for the other listed secondary outcomes. Serious adverse events and deaths are listed as zero in both arms. Details of the original analysis plan, missing-data handling, and analysis populations are not reported in the registry record.