Helicobacter Pylori Eradication in Non-diabetic Non-alcoholic Steatohepatitis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Helicobacter Pylori Eradication in Non-diabetic Non-alcoholic Steatohepatitis (NCT01654549), a randomized phase 2 study. The primary outcome is Liver Fat Content.

2. Clinical Question

The trial asks whether adding Helicobacter pylori eradication to a lifestyle-modification program is associated with a different liver-fat outcome than lifestyle modification alone in non-diabetic participants with non-alcoholic fatty liver disease. Because both arms received the lifestyle program, the randomized contrast concerns the addition of eradication therapy, not eradication therapy versus no lifestyle intervention.

The record describes the primary outcome as change in liver fat content from baseline to the end of the study. It also lists liver function tests, fasting glucose, lipid profile, HOMA-IR, and anthropometric measurements among the secondary outcomes. These are outcomes the study aimed to evaluate; the primary outcome figures are the focus of the posted results. Further details of the clinical question, including a stated hypothesis or a minimum clinically important difference, are not reported in the registry record.

3. Trial Design

This was a randomized, double-masked, parallel-group phase 2 trial sponsored by Tehran University of Medical Sciences. The registry lists an enrollment of 40 and a completed status. The recorded study start was 2012-04, and primary completion was 2012-09. The reported primary-outcome summaries include 19 participants in the eradication group and 20 in the lifestyle-modification group.

The eradication arm received quadruple antibiotic therapy for two weeks and the calorie-restriction diet and programmed physical activity described as aiming to obtain ideal body weight. The comparator arm received the lifestyle program alone. Thus, the intended contrast was an add-on intervention against a shared lifestyle program. The registry labels the intervention as a drug and describes the study model as parallel.

The primary endpoint was liver fat content, expressed as a percentage, at baseline and at 8 weeks, described as 6 weeks post-treatment. The record also posts change summaries for each group. The outcome description says liver fat content was calculated using a formula based on metabolic syndrome status, type 2 diabetes status, fasting serum insulin, AST, and the AST/ALT ratio. It does not describe a direct imaging measurement in the endpoint text.

The registry lists six secondary outcomes: serum alanine aminotransferase, serum aspartate aminotransferase, fasting serum glucose, serum lipid profile, HOMA-IR, and anthropometric measurements. The record indicates seven outcome measures in total, while the details of one measure beyond the primary and named secondary outcomes are not established by the information reported. Randomization procedures, allocation concealment, the masking arrangements, adherence, and participant flow are not reported in the registry record. The original sample-size rationale and the trial's planned statistical analysis are also not reported.

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

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 Fat Content, measured over 8 weeks (6 weeks post-treatment).

The primary endpoint is liver fat content, measured as a percentage, with the stated primary outcome being change from baseline to the end of the study. The timeframe is 8 weeks, identified as 6 weeks post-treatment. The registry posts group means and standard deviations at baseline and at 8 weeks, as well as summaries labelled as change for each group.

The endpoint description says the percentage was calculated from metabolic syndrome status, type 2 diabetes status, log fasting serum insulin, log fasting serum AST, and log of the AST/ALT ratio. This matters for interpretation: the registry describes a formula-derived measure rather than specifying a direct liver-fat measurement method. The formula includes diabetes status even though the study population is described as non-diabetic; how the formula was applied to participants is not reported in the registry record.

A continuous endpoint can retain information across the range of observed values rather than reducing results to a yes-or-no category. Reporting a mean and standard deviation describes the group average and the spread of participant values. A baseline-to-follow-up change can describe within-person movement, while comparing the change between randomized groups is the contrast relevant to the added intervention. The registry gives within-group paired-test results and a between-group test of the change summaries, but it does not report a detailed model for the primary endpoint or an adjusted treatment-effect estimate.

7. Results Posted in the Registry

• Liver Fat Content in H.Pylori Eradication at 8 Weeks: mean 7.72 versus 8.38 on control; difference -0.66 (95% CI -3.46 to 2.14), standardized effect d = -0.15; z-test p = 0.644.
• Liver Fat Content in Lifestyle Modification at Baseline: mean 8.50 versus 8.38 on control; difference 0.12 (95% CI -3.20 to 3.44), standardized effect d = 0.02; z-test p = 0.944.
• Liver Fat Content in Lifestyle Modification at 8 Weeks: mean 7.29 versus 8.38 on control; difference -1.09 (95% CI -4.08 to 1.90), standardized effect d = -0.23; z-test p = 0.475.
• Liver Fat Content Change in H.Pylori Eradication: mean 0.66 versus 8.38 on control; difference -7.72 (95% CI -9.75 to -5.69), standardized effect d = -2.42; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports t-tests for the primary outcome comparisons. A two-sided t-test evaluates whether the data are inconsistent with a null hypothesis of no difference in means in either direction. The registry reports baseline and 8-week between-group tests, paired t-tests within each arm, and a two-sided test comparing the two change summaries. The reported p-values are 0.94, 0.78, below 0.01 for each paired test, and 0.10 for the comparison of changes. The registry does not report a fuller statistical analysis plan, including whether any tests were designated as primary for inference or whether covariate adjustment was planned.

What does a mean difference represent?

A mean difference subtracts one group mean from another in a stated order. Its sign depends on which mean is subtracted from which, so the estimate should be read alongside the group labels and means. A confidence interval gives a range of values compatible with the estimate under the analysis assumptions. If it includes zero, the data do not rule out no difference at the confidence level used. An interval that extends to both positive and negative values also leaves uncertainty about which group has the higher average.

For the 8-week comparison, the recalculated mean difference is -1.09, with a 95% confidence interval from -4.083 to 1.903. The estimate is based on the posted group means and standard deviations and is an unadjusted recalculation. The interval includes zero. The sponsor's reported two-sided test for the same follow-up groups gives p = 0.78, which differs from the recalculated p-value of 0.475. Because the discrepancy flag is true, the sponsor's own result takes precedence over the recalculation for interpretation of that comparison.

Why distinguish paired and between-group tests?

A paired t-test uses the association between repeated measurements from the same participants to test whether the average within-person change differs from zero. It addresses change in one group. It does not directly compare change in the intervention arm with change in the comparator arm. The two-sided test comparing the change summaries addresses that separate question. The registry reports p < 0.01 for the within-group tests and p = 0.10 for the between-group comparison of changes. These results can coexist: both groups can show evidence of within-group change without the evidence showing a difference between their changes.

For an add-on intervention in a randomized trial, the between-group contrast is the central comparison for assessing whether outcomes differed between assigned strategies. Comparing groups at follow-up can be informative, while comparing changes can account descriptively for their starting values. A baseline comparison is not evidence of a treatment effect because it precedes the intervention. The registry reports the baseline comparison as p = 0.94; this is a check on the observed starting values, not a test of the intervention's effect.

What is the role of standard deviation?

The standard deviation describes variation among participant measurements around a group mean. It is not the same as the standard error, which describes uncertainty in an estimated mean. The posted standard deviations for baseline and 8-week liver fat content range by arm and visit: 4.47 and 4.34 in the eradication arm, and 6.03 and 5.06 in the lifestyle arm. The spread of observed values is relevant to the precision of group averages and comparisons. The registry does not report individual participant data, so the association between each participant's baseline and follow-up values cannot be reconstructed from these group summaries alone.

In general, repeated measurements on the same participant are correlated. A paired analysis can use that within-person relationship, whereas an analysis treating measurements as independent would not. The registry identifies paired t-tests for the within-group comparisons, but the details of the calculation of change summaries and the exact analysis population are not reported.

Why can recalculation differ from the registry analysis?

A simple recalculation from rounded group means and standard deviations may not reproduce a sponsor analysis that used different precision, a different analysis population, a different statistical procedure, or other details not reported in the record. Those possibilities cannot be confirmed for this trial. The posted recalculations are unadjusted comparisons from summary statistics, not replacements for the sponsor's analysis. The discrepancy flag is true, so the sponsor's own analysis takes precedence.

There is also an estimand issue in the recalculated change rows. The listed recalculated comparisons use the baseline mean of 8.38 as one value and a within-group change summary of 0.66 or 1.21 as the other. That subtraction does not compare the intervention arm's change with the comparator arm's change. The registry's own between-group test of the change summaries, p = 0.10, is the relevant posted test for that contrast. The recalculated values should not be interpreted as an effect estimate for the randomized comparison in change.

Finally, p-values describe how compatible data are with a specified null model; they are not the probability that the null hypothesis is true, and they do not measure clinical importance. Estimates and confidence intervals, the design, and the limitations in reporting all matter when interpreting the evidence. In this trial, the reported sponsor comparison of changes and the imprecision of the available follow-up comparison are more informative than isolated within-arm p-values for understanding the randomized contrast.

10. Confidence Intervals and Precision

The recalculated 8-week mean difference is -1.09, with a 95% confidence interval from -4.083 to 1.903. This interval includes zero and extends in both directions. It is compatible with a lower average in the eradication group as well as a higher average, given the comparison's direction. The point estimate alone should not be treated as a precise estimate of the difference.

A confidence interval expresses uncertainty around an estimate under the assumptions of the analysis. It is not a range containing 95% of individual participant values, nor does it assign a 95% probability to the true effect after the data are observed. A wider interval indicates less precision than a narrower interval for the same outcome and scale. Here, the registry's reported means and standard deviations come from groups of 19 and 20 participants, and the recalculated interval for the follow-up comparison remains broad.

The sponsor's own reported 8-week comparison is p = 0.78, while the unadjusted recalculation gives p = 0.475. The discrepancy flag is true, so the sponsor's analysis takes precedence. The registry does not provide a sponsor confidence interval or a sponsor estimate for this comparison, so the precision of the sponsor's own estimated contrast cannot be assessed from the reported analysis details.

The registry reports p = 0.10 for the between-group comparison of change summaries, but does not report a confidence interval or a numerical between-group change estimate in that analysis. The absence of that interval limits direct assessment of the range of change differences consistent with the sponsor's analysis. The recalculated change comparisons are not valid substitutes for it, because they compare baseline with within-group change rather than comparing the two groups' changes.

11. 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 planning estimates are participants per arm, not total events. They use a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The assumed effect to detect is a stated fraction of the observed effect. The registry record does not report the trial's own original sample-size assumptions, so these figures are planning illustrations based on the observed effect rather than a reconstruction of the original design.

The 80% power column gives the estimated number of participants per arm under those assumptions for an 80% chance of detecting the specified effect if that effect is present. The 90% power column gives the corresponding estimate for a 90% chance. These are design calculations for future studies; they are not probabilities that this trial's result is correct. This is not observed power.

The figures depend strongly on the assumed effect size. Planning for 75% of the observed effect requires more participants than planning for 100%, and planning for 50% requires still more. For example, the follow-up comparison estimates 303 participants per arm for 80% power when planning for 100% of the observed effect, 538 for 75%, and 1,209 for 50%. At 90% power, the corresponding figures are 405, 720, and 1,618 participants per arm. Those estimates are illustrations, not recommendations, and they do not account for any trial-specific design features beyond the listed assumptions.

The extreme estimates attached to the recalculated change rows should not be interpreted as credible requirements for replicating the randomized comparison. Those rows use a subtraction of baseline mean and within-group change rather than a between-group change contrast. The sample-size calculations inherit that mismatch. The sponsor's reported between-group change test is p = 0.10, but the registry does not report a numerical effect estimate or interval for that contrast from which a better grounded planning calculation could be read.

12. 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 = -0.660; pooled standard deviation σ = 4.405.
2. n = 2 × (1.96 + 0.842)² × 4.405² / -0.660² = 699.4, rounded up to 700 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 <- -0.660000000000001                      # observed difference in means
sigma <- 4.40547954256968                     # 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(1654549)
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 = 700 (80% power), 937 (90% power)
 75% of observed effect: n per arm = 1,244 (80% power), 1,665 (90% power)
 50% of observed effect: n per arm = 2,798 (80% power), 3,746 (90% power)
power.t.test: 700.4 
Simulated power at n = 700 per arm: 0.798 (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.

13. Analyses Reported by the Sponsor

The registry reports the sponsor's tests as two-sided t-tests for the baseline and 8-week group comparisons, paired t-tests for within-group changes, and a two-sided t-test for the change summaries. The sponsor reported p = 0.94 at baseline, p = 0.78 at 8 weeks, p < 0.01 for each within-group paired test, and p = 0.10 for the between-group comparison of changes.

The unadjusted recalculation for the 8-week comparison gives a mean difference of -1.09, a 95% confidence interval from -4.083 to 1.903, and p = 0.475. Its p-value differs from the sponsor's p = 0.78. The discrepancy flag is true, so the sponsor's own analysis takes precedence. The reason for the difference is not reported in the registry record; it should not be attributed to a particular method or data-handling choice without evidence.

The recalculated change rows also do not reproduce the sponsor's between-group change comparison. They use baseline mean 8.38 and a within-group change summary, rather than the two groups' change summaries as the contrast. Therefore, their estimates and associated planning illustrations should not be read as the treatment-versus-comparator effect in change. For that question, the registry's sponsor analysis reports p = 0.10.

14. Multiplicity, Missing Data and Analysis Populations

The record lists the primary outcome and six secondary outcomes, and reports five statistical analyses for liver fat content. It does not report a multiplicity adjustment, a hierarchy of secondary tests, or a strategy for controlling the chance of false-positive findings across outcomes and comparisons. It is therefore not possible to determine from the registry record whether the reported p-values were adjusted, or how the study's investigators intended to interpret multiple tests.

The primary endpoint is change in liver fat content. The listed secondary outcomes include liver enzymes, fasting glucose, lipid profile, HOMA-IR, and anthropometric measurements. Testing multiple outcomes or several contrasts can increase the chance that at least one result appears statistically notable by chance. That is a general statistical issue, not evidence that any particular reported result is false. The registry does not state a multiplicity plan, so the p-values should be read as the reported tests rather than as a documented family-wise analysis.

Missing-data handling is not reported in the registry record. The primary outcome summaries list 19 participants in the eradication group and 20 in the lifestyle group, while total enrollment is 40. The record does not explain the difference between enrollment and the number summarized, identify reasons for unavailable outcomes, describe imputation, or state whether an intention-to-treat analysis was used. It also does not describe a per-protocol population or other analysis populations.

These omissions matter because the results can depend on who contributed measurements and how missing observations were handled. The reported group sizes alone do not establish why a participant is absent from a summary or whether the analysis retained all randomized participants. The registry also does not report a detailed participant-flow account or sensitivity analyses for missing outcomes. Conclusions about the effect of missingness or the robustness of the estimates cannot be drawn from the posted information.

15. 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 0% reporting threshold.

The registry lists two safety groups: H. Pylori Eradication and No Intervention. Each lists 20 participants at risk for serious adverse events and 20 at risk for other adverse events, with zero affected participants in both categories in each group. The record's adverse-event group label No Intervention differs from the trial's described lifestyle-modification comparator, so the safety label should be read as it is recorded rather than assumed to give further detail about the comparator's care.

These figures mean that no serious or other adverse events are recorded as affecting participants in the listed groups. They do not establish that no events occurred outside the reported categories or that every participant had complete safety follow-up. Death counts are null in the record, which means they are not reported in the registry record; null should not be interpreted as zero deaths.

The registry provides no serious-event terms, and it does not give event descriptions, severity details, timing, or a fuller account of safety ascertainment. With 20 participants at risk in each listed group, the posted counts describe this small study's recorded safety data; they do not provide a precise estimate of uncommon harms. These are descriptive registry figures, not a basis for medical advice.

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

The registry shows that mean liver fat content was lower at 8 weeks than at baseline in both groups, and it reports paired-test p-values below 0.01 within each arm. These within-group findings describe changes over time under the assigned study conditions. They do not isolate the effect of adding Helicobacter pylori eradication, because participants in both groups received a lifestyle-modification program.

For the randomized contrast in change, the sponsor reports p = 0.10. The posted change means are 0.66 in the eradication group and 1.21 in the lifestyle group, but the registry does not provide a sponsor estimate and confidence interval for the difference between them. The recalculated change comparisons do not estimate that contrast, and the discrepancy flag means the sponsor's own analysis takes precedence. The reported results therefore do not establish a statistically clear between-group difference in change, nor do they establish equivalence.

The 8-week group means were 7.72 and 7.29, and the sponsor's reported comparison has p = 0.78. The unadjusted recalculated comparison has a 95% confidence interval from -4.083 to 1.903, which includes no difference and remains broad. That recalculation differs from the sponsor result and should not replace it. Together, the available estimates and tests leave uncertainty about the size and direction of any difference between strategies.

The study's stated population is non-diabetic subjects with non-alcoholic fatty liver disease. The registry record does not report detailed eligibility criteria, participant flow, the missing-data approach, a multiplicity plan, or the original sample-size justification. The primary outcome is described as formula-derived liver fat content, and the endpoint text does not specify a direct measurement method. Interpretation should remain within those reporting limits. The result does not establish a clinical benefit or lack of benefit, and it does not support advice for individual care.

17. Why This Trial Matters Statistically

This trial illustrates why the comparison that matches the clinical question matters. The intervention was added to a lifestyle program that both arms received, so the relevant question is whether the groups differed, not simply whether either group's measurements changed over time. The paired tests show evidence of within-group change in both arms; the sponsor's between-group test of the change summaries gives p = 0.10. Those are different statistical questions and should not be conflated.

It also illustrates the value of estimates and intervals alongside p-values. The recalculated 8-week mean difference is -1.09, but its 95% confidence interval from -4.083 to 1.903 permits values in either direction. A non-significant test does not demonstrate that the strategies are equivalent, while a within-group significant result does not demonstrate that one strategy is superior to the other.

The record also makes the distinction between sponsor analyses and simple summary-statistic recalculations important. The sponsor's 8-week comparison gives p = 0.78, whereas the recalculation gives p = 0.475. The discrepancy flag directs interpretation to the sponsor result. Moreover, some recalculated change rows compare a baseline mean with a change mean, not one arm's change with the other arm's change. Recognizing the estimand prevents an arithmetic result from being mistaken for the randomized treatment contrast.

Finally, planning calculations depend on the assumed effect. Smaller target effects require more participants under the stated assumptions. Illustrations based on observed effects can be unstable, especially when the underlying contrast is imprecise or not the intended between-group estimand. The registry's lack of an original sample-size rationale, analysis plan, and missing-data details limits assessment of design choices and robustness.

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 registry record contains limited detail needed to assess the trial's conduct and the precision of its conclusions. It does not report the randomization sequence, allocation concealment, how double masking was implemented, adherence, or participant flow. The endpoint summaries include 19 and 20 participants despite enrollment of 40, but the reason for this difference is not reported.

The primary liver-fat endpoint is described as a formula using metabolic syndrome status, diabetes status, fasting insulin, AST, and AST/ALT. The record does not specify a direct measurement technique or explain how the formula was applied in this non-diabetic population. The clinical meaning and measurement properties of the derived endpoint cannot be assessed from the registry description alone.

The statistical reporting is also incomplete. The registry does not report the original sample-size justification, a prespecified analysis plan, an adjusted treatment estimate, a missing-data method, or a multiplicity strategy. The sponsor reports p-values but not a confidence interval or numerical between-group estimate for its comparison of change summaries. The recalculated results are unadjusted and, for change rows, do not represent the between-group change contrast. Because the discrepancy flag is true, the sponsor's own analysis takes precedence.

The study is described as phase 2 and has an enrollment of 40. The posted summaries therefore offer limited precision, as reflected in the broad recalculated interval for the follow-up comparison. Further population details and generalizability information are not reported in the registry record. Safety reporting lists zero affected participants in the stated categories but does not report death counts, event terms, or fuller ascertainment details. These reporting gaps constrain interpretation without showing that any particular unreported procedure or event occurred.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Helicobacter Pylori Eradication in Non-diabetic Non-alcoholic Steatohepatitis — NCT01654549.
• Registry record last updated 2018-06-28; status last verified 2018-04.
• Results first posted 2012-12-21.
• 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

NCT01654549 is a completed, randomized, double-masked, parallel-group phase 2 trial sponsored by Tehran University of Medical Sciences. It enrolled 40 non-diabetic subjects with non-alcoholic fatty liver disease. The comparison was Helicobacter pylori eradication with quadruple antibiotic therapy for two weeks plus lifestyle modification versus lifestyle modification alone.

The primary outcome was liver fat content as a percentage, assessed from baseline to 8 weeks, described as 6 weeks post-treatment. The sponsor reported p < 0.01 for within-group paired tests and p = 0.10 for the between-group comparison of changes. The sponsor's 8-week group comparison was p = 0.78. Recalculated results differ, and the registry record does not report the original sample-size rationale, a detailed analysis plan, or missing-data handling.