Ursodiol in Treating Patients With Barrett Esophagus and Low-Grade Dysplasia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Ursodiol in Treating Patients With Barrett Esophagus and Low-Grade Dysplasia (NCT01097304), a na phase 2 study. The primary outcome is Reversal of Oxidative DNA Damage as Assessed by Changes in 8-hydroxy-2' -Deoxyguanosine (8OHdG) Immunostaining.

2. Clinical Question

3. Trial Design

This is a completed phase II study with a single-group model. The registry lists enrollment of 36 and identifies the National Cancer Institute as sponsor. The study began in 2010-04, reached primary completion in 2014-04, and has completed status. These record details describe the trial’s administrative history; they do not indicate how many participants contributed data to the primary analysis.

The treatment description specifies ursodiol by mouth twice daily for 6 months, in the absence of disease progression or unacceptable toxicity. The registry also lists laboratory biomarker analysis. Masking is listed as none, and allocation is listed as NA. There is no randomized comparison in the described design, so the within-participant baseline and post-intervention measurements are the central comparison for the primary endpoint.

The primary measure is a continuous tissue biomarker: the percentage of strongly or moderately stained nuclei for 8OHdG, assessed from baseline to 6 months. Three secondary outcome descriptions concern changes in gastric bile-acid composition and cell proliferation in Barrett esophagus epithelium. The registry lists four outcome measures and four analyses, but does not report the numerical findings for these measures.

The analysis description excludes participants with fewer than 500 total nuclei in longitudinally sectioned crypts opening to the lumen. The number excluded, the number analyzed, and the handling of participants who did not complete the post-intervention assessment are not reported in the registry record. As a result, enrollment should not be read as the primary analysis sample size.

4. Participant Flow and Accounting

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

Overall Study

Analysis population for the primary outcome, as defined in the registry: participants with fewer than 500 total nuclei in longitudinally sectioned crypts opening to the lumen were excluded from analysis

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 Reversal of Oxidative DNA Damage as Assessed by Changes in 8-hydroxy-2' -Deoxyguanosine (8OHdG) Immunostaining, measured over Baseline to 6 months.

The primary endpoint is the change from baseline to 6 months in 8OHdG immunostaining, described in the outcome title as reversal of oxidative DNA damage. The listed unit is the percentage of strongly or moderately stained nuclei. The outcome description also refers to the percentage of positively stained nuclear area, so the registry record uses two related but not identical descriptions of how the staining measure is expressed. A more detailed operational definition is not reported in the registry record.

In general, a tissue biomarker can provide a measurable biological response to an intervention without directly measuring a patient’s symptoms, function, or later clinical outcomes. Here, the endpoint is a continuous measure summarized by a mean and standard deviation. A baseline-to-follow-up comparison asks whether the measured biomarker changed during the intervention period among participants included in the analysis. The registry’s endpoint label gives the biological rationale, but it does not report the observed mean change or its uncertainty.

The record specifies a paired t-test for the primary analysis and says observed results would be reported with corresponding confidence intervals. The actual estimate, confidence interval, and test result are not reported in the registry record. Thus, the endpoint’s intended role is clear, but the posted record does not show what change was observed or how precisely it was estimated.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is continuous, and the registry describes it using a mean and standard deviation. It also specifies a paired t-test at a one-sided 0.05 significance level and says results would be reported with confidence intervals. The record does not include the numerical analysis results. The explanations below describe the statistical ideas generally; they do not add findings to the registry record.

What does a paired t-test assess?

In general, a paired analysis uses two measurements from the same participant, such as a baseline value and a later value. It focuses on the within-participant differences rather than treating the two sets of measurements as independent groups. A paired t-test evaluates whether the average difference is compatible with a specified null value, under its assumptions. For this study, the stated time points are baseline and 6 months, and the registry describes the target as change during intervention.

Pairing can account for the fact that people may start with different biomarker levels: the comparison is based on each person’s change. It does not, however, create a control group. In general, a within-person change can still reflect influences other than the intervention, including time-related changes or measurement variability. The study design cannot rule out those explanations by itself.

A paired t-test also relies on assumptions about the distribution and independence of the participant-level differences. In general, the test is most interpretable when the paired differences are appropriately measured and their distribution is reasonably compatible with the model. The registry record does not report diagnostic checks, transformations, or alternative analyses for the primary endpoint.

What does the stated significance level mean?

The primary description specifies a one-sided significance level of 0.05. In general, a one-sided test evaluates evidence in a prespecified direction rather than treating departures in either direction as equally relevant to the test. The endpoint is framed as reversal, but the detailed directional hypothesis and any decision rule are not reported in the registry record. A significance level is a threshold used in a test; it is not the probability that a conclusion is true, nor does it measure the magnitude or clinical importance of a change.

The sample-size assumptions separately describe planning illustrations using a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effects expressed as fractions of the observed effect. Those are assumptions for planning illustrations, not a description of the actual single-group design. The record does not report numerical planning results or the observed effect needed to interpret them. Such illustrations are not observed power and should not be treated as evidence about the study’s outcome.

How should a confidence interval be read?

In general, a confidence interval gives a range of values compatible with the observed data under the statistical model and method used. Its width conveys precision: a narrower interval indicates a more precise estimate than a wider one, all else being equal. The interval’s location describes which directions and sizes of change remain compatible with the data. A confidence interval is not a guarantee that the true value lies in the displayed range, and it does not remove design limitations such as the absence of a control group.

The primary outcome description says confidence intervals would accompany observed results, but no interval is reported in the registry record. Consequently, the precision of the primary estimate cannot be evaluated. Nor can the one-sided test result substitute for an estimate and interval: the test addresses evidence against a null hypothesis, while the estimate and interval describe magnitude and uncertainty.

What do mean and standard deviation describe?

For a continuous measure, the mean describes the average value, while the standard deviation describes variation among the observations around that average. In a paired analysis, it is useful to distinguish the average of participant-level changes from the variability of those changes. The registry identifies the primary parameter as a mean and the dispersion as standard deviation, but it does not provide the observed values. It is therefore not possible to tell from the record how much the biomarker varied or how large the average change was.

General statistical interpretation should also distinguish variability in individual measurements from uncertainty in the estimated average. The standard deviation concerns variation in the data; a confidence interval concerns uncertainty about the estimate. They answer related but different questions. Without the reported estimate, dispersion, and interval, neither the average change nor its precision can be characterized.

What can the analysis population tell us?

The primary analysis description excludes participants with fewer than 500 total nuclei in longitudinally sectioned crypts opening to the lumen. In general, an analysis-population rule can affect which observations contribute to an estimate and how broadly the estimate applies. The registry does not report how many participants met this criterion, whether any other exclusions occurred, or how missing paired measurements were handled. Those details matter for understanding whether the analyzed participants represent everyone enrolled and whether the analysis could be affected by incomplete follow-up.

Overall, the methods description gives the planned paired test, significance level, outcome scale, and intended confidence intervals. It does not report enough results or analysis details to assess the estimate, precision, assumptions, or robustness of the primary analysis.

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("Treatment (Ursodiol)")
n   <- c(25)
m   <- c(1.62)
sp  <- c(13.13)
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  Treatment (Ursodiol)  25  1.620  13.130  -3.800  7.040

11. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary biomarker endpoint and three secondary outcome descriptions: changes in gastric bile-acid composition for two classes of bile acids, and change in cell proliferation in Barrett esophagus epithelium. The record gives no numerical results for these outcomes. It also does not report whether the analyses were treated as confirmatory or exploratory, or whether any adjustment for multiple outcome testing was planned or used. A multiplicity strategy is therefore not reported in the registry record.

In general, testing several outcomes can increase the chance of obtaining at least one apparently notable result by chance if each is interpreted as a separate confirmatory test. Statistical approaches to multiplicity can control or describe that issue, but the appropriate approach depends on the study’s aims and analysis plan. No particular approach should be inferred for this study from the list of outcomes alone.

The primary analysis population excludes participants with fewer than 500 total nuclei in the specified crypt sections. The number excluded and the number analyzed are not reported in the registry record. The record also does not report the number with missing baseline or follow-up measurements, reasons for missing data, or an imputation or sensitivity-analysis method. Those details are important because a paired analysis needs both measurements for a participant to calculate an individual change, unless a different method is used.

Enrollment is listed as 36, but the record does not say how many contributed to each endpoint analysis. It does not describe whether analysis followed an intention-to-treat principle or another defined population approach. These omissions limit assessment of the completeness of the analysis and the potential influence of exclusions or missing observations. No assumption about how missing values were handled is warranted.

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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hypotension (2.8%), ischemia/infarction (2.8%), atrial fibrillation (2.8%), enteritis (2.8%), barrett's esophagus with at least intramucosal carcinoma (2.8%).

For the treatment group, the registry lists 36 participants at risk for serious adverse events and 5 affected. For other adverse events, it lists 36 at risk and 31 affected. These are safety counts recorded in the registry; they are not comparative estimates because the study has no concurrent control group. The record does not provide a comparator-based assessment of excess risk.

The listed serious adverse event terms include hypotension, ischemia/infarction, atrial fibrillation, enteritis, Barrett's esophagus with at least intramucosal carcinoma, and pancreatitis. Each term is displayed with 2.8%. A term-specific percentage should not be read as the percentage of all serious events or as evidence that each term occurred in a separate participant. The registry does not explain whether participants could contribute more than one event term or provide further event-level detail.

The deaths field has no reported value. This is not equivalent to a reported count of zero; deaths are not reported in the registry record. The record also does not specify the safety observation window or provide enough detail to interpret timing, duration, severity beyond the serious-event classification, or relationship to treatment. These figures describe the posted safety categories and should not be used alone to establish causation.

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

The study addresses whether a tissue biomarker changed during a course of ursodiol in a group of participants with Barrett esophagus or low-grade dysplasia. Its primary endpoint is the baseline-to-6-month change in 8OHdG immunostaining. The registry record describes the test intended for that comparison, but the observed primary result and its confidence interval are not reported. Thus, the posted information does not show whether the measured biomarker increased, decreased, or remained similar, nor how precisely any change was estimated.

Even if a biomarker change were reported, its interpretation would be limited by the single-group design. In general, a before-and-after difference does not by itself show that the intervention caused the difference. Without a concurrent control group, background change, other influences over time, and measurement variation cannot be separated from a treatment-associated change using the described design alone.

The endpoint is a tissue biomarker rather than a direct measure of cancer prevention or another clinical outcome. The registry summary describes chemoprevention as the broader motivation, but the listed primary outcome is 8OHdG immunostaining. The posted record does not establish that a biomarker change corresponds to a change in cancer risk or clinical outcomes. No such relationship should be inferred from the endpoint description alone.

The safety entries report affected counts for serious and other adverse events in the treatment group, but the lack of a concurrent comparison and limited event detail restrict interpretation. The record does not report deaths, rather than reporting a zero count. Taken together, the available information describes the study design and selected safety counts, while leaving the main efficacy estimate unavailable.

Accordingly, the registry record supports a description of what the trial set out to measure, not a conclusion about the size or direction of the primary biomarker response, its clinical significance, or treatment causality. Those judgments would require outcome results and supporting analysis details that are not reported in the registry record.

14. Why This Trial Matters Statistically

Statistically, this study illustrates the distinction between a carefully defined measurement question and an available result. The record specifies a continuous primary endpoint, a baseline-to-follow-up comparison, a paired t-test, and confidence intervals. Those details identify how the primary question was intended to be evaluated. Yet the primary estimate and its uncertainty are not reported in the registry record, so the planned analysis cannot be translated into an assessment of the observed change.

The single-group design also highlights the difference between measuring change and attributing change. Paired data can account for each participant’s starting measurement when estimating an average within-person change. In general, however, that feature does not replace a concurrent control group. The design can describe a temporal pattern in the measured biomarker, but the causal interpretation remains limited.

The analysis exclusion based on the number of nuclei makes the analysis population another important statistical issue. Enrollment is reported, but the number retained for the primary analysis is not. This gap prevents readers from assessing how much information supported the intended paired comparison or how exclusions may affect interpretation.

Finally, the record lists multiple biomarker outcomes without reporting a multiplicity strategy or the results. This matters because outcome selection, missing observations, analysis populations, and uncertainty are part of interpreting evidence, not merely technical details. The study is useful for recognizing those questions in a pilot biomarker study, while its registry record leaves the central numerical result unresolved.

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 outcome result is not reported in the registry record. No observed mean change, standard deviation, confidence interval, or test result is available, so the direction, magnitude, and precision of the main endpoint cannot be evaluated. Numerical results for the listed secondary outcomes are also not reported.

The design is single-group, without a concurrent control, and masking is listed as none. These features limit causal interpretation of any change during intervention. The registry gives an enrollment figure but not the number included in the primary analysis. It specifies an exclusion for participants with fewer than 500 total nuclei in longitudinally sectioned crypts opening to the lumen, but does not report how many were excluded or how missing measurements were handled.

The endpoint wording is not fully aligned: the title refers to the percentage of strongly or moderately stained nuclei, while the description refers to the percentage of positively stained nuclear area. The exact scoring definition and how the two descriptions relate are not reported in the registry record. Further details about eligibility, assessment procedures, analysis diagnostics, and sensitivity analyses are also not reported.

The registry lists secondary biomarkers but does not report a multiplicity approach. The safety record includes affected counts and serious-event terms, but does not provide a comparison group, event-level timing, or a reported deaths count. These limitations constrain interpretation of both the biomarker question and the safety entries; they do not determine what the unreported results would have shown.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Ursodiol in Treating Patients With Barrett Esophagus and Low-Grade Dysplasia — NCT01097304.
• Registry record last updated 2017-12-19; status last verified 2017-11.
• Results first posted 2015-05-15.
• 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

NCT01097304 is a completed phase II, single-group study sponsored by the National Cancer Institute. The registry lists enrollment of 36 and describes ursodiol by mouth twice daily for 6 months, unless disease progression or unacceptable toxicity occurred. The primary endpoint is change in 8OHdG immunostaining from baseline to 6 months, analyzed with a paired t-test according to the outcome description.

The record describes three secondary biomarker outcomes and reports safety counts for the treatment group. It does not report numerical primary or secondary outcome results, the primary analysis population size, or a deaths count. The design has no concurrent control group, limiting causal interpretation of any within-participant change. The posted record therefore describes the study question and planned analysis but does not provide the primary estimate or its precision.