Modulation of Pharmacologically Induced Alcohol Craving in Recently Detoxified Alcoholics: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Modulation of Pharmacologically Induced Alcohol Craving in Recently Detoxified Alcoholics (NCT00605904), a randomized phase 2 study. The primary outcome is Alcohol Craving Rating in Response to Saline Infusion.

2. Clinical Question

3. Trial Design

The registry record describes a randomized, triple-masked, parallel-group phase 2 study of acamprosate and placebo in the context of alcohol-craving research. Enrollment was 37, and the study status is completed. The sponsor was the National Institute on Alcohol Abuse and Alcoholism. The registry record lists a start date of 2008-01 and a primary completion date of 2010-09.

Participants were assigned to acamprosate or placebo. The arm descriptions specify three tablets taken orally three times daily for a minimum of two weeks; the acamprosate tablets were 333 mg each, with a stated total dose of 999 mg. The brief summary describes subsequent intravenous saline, mCPP and yohimbine infusions. It also describes admission to the NIH Clinical Center for about 35 days, participation in an alcohol treatment program, and return-to-hospital breathalyzer and urine screening after passes. These are descriptions in the brief summary, not details of the statistical analysis.

The primary endpoint was alcohol-craving rating 180 minutes after saline infusion. Its analysis included only participants who completed all three infusion types: 12 in the acamprosate group and 13 in the placebo group. Thus, the posted primary comparison represents this completion-defined subset, rather than all 37 enrolled participants. The registry record does not report why participants were excluded from this analysis or whether the subset was defined before data collection.

The record lists three outcome measures but no posted secondary outcome results. It does not report the original sample-size justification, planned analysis model, or other protocol-level statistical design assumptions. Randomization and masking are recorded, but further operational details about those procedures are not reported in the registry record.

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.

Baseline Data Collection and Dosing

Infusion Sessions

Analysis population for the primary outcome, as defined in the registry: The analyses included only those subjects who completed all three types of infusions (saline, meta-Chlorophenylpiperazine, and yohimbine)

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 Alcohol Craving Rating in Response to Saline Infusion, measured over 180 minutes after the start of the infusion.

The primary endpoint is the mean alcohol-craving rating on the Penn Alcohol Craving Scale 180 minutes after the start of saline infusion. The scale is a five-item self-administered instrument covering the frequency, intensity and duration of thoughts about drinking, as well as the ability to resist drinking. Its single score ranges from 0 to 30, with 30 representing the maximum amount of craving on the scale.

A craving scale provides a continuous measure of participants' reported experience, rather than a simple count of whether craving was present or absent. Comparing group means describes the average ratings in the two assigned groups at the stated assessment time. The saline condition is the setting for this primary endpoint; the outcome does not itself measure craving after yohimbine or mCPP. The registry record says the analysis included only participants who completed all three infusion types.

As general context, a standardized assessment can make a subjective experience more consistently measurable across participants. The result remains a rating on the instrument: it is not, by itself, a measure of drinking behavior, treatment retention, or longer-term outcome. The registry record does not report the scale's measurement properties in this study, nor does it identify a threshold for interpreting a particular change as meaningful.

7. Results Posted in the Registry

• Acamprosate: mean 1.70 versus 1.77 on control; difference -0.06 (95% CI -1.93 to 1.80), standardized effect d = -0.03; z-test p = 0.948.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is a continuous outcome, and the registry record reports a mean for each group. For this kind of measure, a group mean describes the average rating among the participants included in that group. The acamprosate mean is 1.704 and the placebo mean is 1.766. The posted comparison expresses the difference as acamprosate minus placebo, giving -0.062. The negative sign follows from the ordering of the groups in that contrast; it does not mean that every participant's rating was lower with acamprosate.

What does the standard error describe?

The arm-level spread values are standard errors: 0.715 for acamprosate and 0.627 for placebo. As general statistical context, a standard error describes the sampling uncertainty around an estimated mean. It is not the same as the spread of individual participants' ratings. The registry record labels these values as standard errors and does not report the underlying standard deviations for each arm. The pooled standard deviation used in the posted comparison is 2.367, and the reported standardized effect is -0.026.

How should the mean difference be read?

The mean difference is an estimate on the original scale: -0.062 units. Its interpretation depends on the arm order and the scale, which ranges from 0 to 30. A mean difference condenses the two sample means into one contrast, but it does not describe how ratings vary across individuals or whether the groups overlap. The comparison also does not show whether a participant's craving changed from before infusion, because a baseline outcome is not reported in the primary endpoint details.

What does the confidence interval add?

A confidence interval communicates uncertainty around an estimated effect. In general, intervals generated by a statistical procedure reflect how estimates would vary across repeated samples under its assumptions. The posted interval is -1.926 to 1.802. It includes zero, so the observed data do not narrowly locate the difference on one side of zero. The interval also allows differences appreciably larger in either direction than the point estimate. The confidence level and the exact method used to calculate the interval are not reported in the registry record; no additional method should be presumed.

What does the p-value say?

The unadjusted p-value is 0.948. In general, a p-value is calculated relative to a null hypothesis and an analysis model. It describes the compatibility of the observed result with that null model, rather than the probability that the null hypothesis is true. It is not a measure of effect size, precision, or clinical importance. Because the value is explicitly unadjusted, it should not be described as accounting for multiple comparisons. The registry record does not report a multiplicity adjustment, the full statistical model, or whether the test was one-sided or two-sided.

Why does over-dispersion matter in general?

For count outcomes, over-dispersion means that observed variation exceeds what a chosen count model expects. Ignoring it can make uncertainty estimates too small and tests too optimistic. That issue is a general statistical consideration, not a reported feature of this endpoint. The primary measure is continuous and is reported as a mean with standard error. The record does not identify an over-dispersion assessment or a count model, so over-dispersion should not be attributed to this analysis.

What population does the estimate describe?

The registry describes the analyzed population as participants who completed all three infusion types: saline, meta-chlorophenylpiperazine (mCPP), and yohimbine. The arm results list 12 acamprosate participants and 13 placebo participants. The estimate therefore pertains to this analysis subset. Random assignment is recorded for the trial, but the record does not give details on how incomplete infusion participation was handled, whether the analysis was intention-to-treat, or whether other participants had missing primary outcomes.

The registry record lists no analysis records and no posted secondary outcome results. It also does not report the original analysis plan or the method used to derive the confidence interval and p-value. Statistical interpretation should therefore stay with the posted descriptive estimates and comparison, without attributing unreported modeling choices to the study.

10. Confidence Intervals and Precision

The primary comparison estimates a mean difference of -0.062 units between the acamprosate and placebo groups. The confidence interval extends from -1.926 to 1.802. This interval is important because the point estimate alone gives no indication of how uncertain the estimate is. The interval covers zero and values in both directions, so the observed comparison does not precisely determine whether the underlying mean is lower in one group or the other.

In general, a confidence interval is a range generated by a statistical method to express uncertainty around an estimate. Its interpretation depends on the method and confidence level. The interval limits and the level are not identified beyond the posted range in the registry record; the confidence level is not reported. The interval should not be read as containing every plausible value with certainty, nor as a range of individual participants' ratings.

The analysis includes 12 acamprosate participants and 13 placebo participants, a subset of the total enrollment of 37. The registry record does not report why the analysis population is smaller than enrollment, how missing data were treated, or whether the subset selection could affect the comparison. These gaps limit assessment of how representative the analyzed participants are of everyone enrolled.

The unadjusted p-value of 0.948 is consistent with substantial uncertainty about a group difference, but it does not replace the interval or tell how large a difference is compatible with the data. The interval gives the more direct view of precision and direction: the estimate is near zero, while the range includes differences on either side. The confidence interval method and adjustment for multiple testing are not reported.

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 sample-size figures are planning illustrations based on the observed effect, not the trial's original design justification. All figures are participants per arm, not total events. The calculations use a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning row uses the stated fraction of the observed effect as the effect to detect. The registry record does not report whether the original trial used these assumptions.

For planning around 100% of the observed effect, the illustration gives 22,871 participants per arm for 80% power and 30,618 participants per arm for 90% power. For 75% of the observed effect, it gives 40,659 participants per arm for 80% power and 54,431 participants per arm for 90% power. For 50% of the observed effect, it gives 91,482 participants per arm for 80% power and 122,469 participants per arm for 90% power.

In plain language, the 80% power column describes the planned number per arm under the stated assumptions for a design with an 80% chance of detecting the chosen effect if that effect were the true one. The 90% power column describes the corresponding planning target for 90% power. These are planning quantities, not observed power. The smaller the effect a study aims to detect, the more participants are needed under otherwise fixed assumptions. Here, the starting point is the observed effect in a small analysis subset, so the planning figures should be read as illustrations rather than as recommendations or as evidence about what the original designers planned.

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 = -0.062; pooled standard deviation σ = 2.367.
2. n = 2 × (1.96 + 0.842)² × 2.367² / -0.062² = 22,870.5, rounded up to 22,871 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 <- -0.0620000000000001                      # observed difference in means
sigma <- 2.36652215929511                     # 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(605904)
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 = 22,871 (80% power), 30,618 (90% power)
 75% of observed effect: n per arm = 40,659 (80% power), 54,431 (90% power)
 50% of observed effect: n per arm = 91,482 (80% power), 122,469 (90% power)
power.t.test: 22871.4 
Simulated power at n = 22,871 per arm: 0.795 (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. Multiplicity, Missing Data and Analysis Populations

The registry record lists three outcome measures, but it reports no posted secondary outcome results and no analysis records. The available comparison concerns the primary endpoint: craving rating after saline infusion. The intervention context also includes mCPP and yohimbine infusions, but outcome comparisons for those conditions are not posted in the record.

Multiplicity arises when several hypotheses or outcomes are tested. As general statistical context, conducting multiple tests can increase the chance of obtaining at least one apparently unusual result by chance. Researchers may address this through a prespecified testing order or an adjustment to significance thresholds or p-values. The record does not report whether such a strategy was planned or used. The p-value attached to the posted comparison is explicitly unadjusted, so it should not be described as corrected for multiple testing.

The registry identifies the primary analysis population as participants who completed all three infusion types. It reports 12 participants in the acamprosate group and 13 in the placebo group for this outcome. However, the record does not report how many enrolled participants failed to complete the infusions, why they did not complete them, or how missing observations were handled. It also does not state whether the analysis followed an intention-to-treat approach, whether any imputation was used, or whether sensitivity analyses were performed.

These omissions matter because restricting an analysis to completers can produce a group that differs from the randomized population. That is a general concern, not evidence that bias occurred in this study. Without the reasons for non-completion and the missing-data methods, readers cannot assess how the completion-based population might affect the estimate. The original statistical analysis plan, prespecified hierarchy of outcomes, and multiplicity procedures are not reported in the registry record.

14. 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 reports serious adverse events for 12 participants at risk in the acamprosate group and 13 at risk in the placebo group. No serious adverse events are reported as affecting participants in either group. The record lists no serious adverse-event terms.

Other adverse events are reported as affecting 11 of 12 participants in the acamprosate group and 9 of 13 in the placebo group. These counts are distinct from the serious adverse-event counts. They indicate how many participants were recorded as affected, not how many separate events occurred. The registry record does not provide event descriptions, severity details for these other events, or an assessment of whether they were related to the study intervention.

Deaths are not reported in the registry record. The available counts do not establish that every possible adverse event was captured or that the groups differed in safety. The group sizes are small, and the registry information does not describe the ascertainment or coding procedures. These figures are a record-level summary, not medical advice or a complete safety assessment.

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

The posted primary result describes mean craving ratings 180 minutes after saline infusion among participants who completed all three infusion types. The means were 1.704 for acamprosate and 1.766 for placebo, with a mean difference of -0.062. The confidence interval extends from -1.926 to 1.802, and the unadjusted p-value is 0.948. Together, these statistics show a point estimate close to zero with considerable uncertainty about the size and direction of the underlying difference.

The result does not establish that the treatments are equivalent. A comparison that does not precisely distinguish a difference from zero is not the same as a study designed to show that any difference is small enough to be unimportant. The registry record does not report an equivalence margin or a non-inferiority margin. The confidence interval permits differences in either direction, and the record does not identify a threshold for clinical importance on the scale.

The finding is specific to the saline-infusion endpoint and the analysis subset. It does not answer whether craving ratings differed after yohimbine or mCPP, because results for those conditions are not posted. It also does not establish effects on drinking, longer-term recovery, or any other outcome not reported for this comparison. The brief summary describes the study's rationale and procedures, but the primary result should not be extended beyond its measured endpoint and population.

General statistical interpretation distinguishes evidence about an average group difference from individual responses. Group means and their interval do not show what happened to each participant, and they do not imply that participants within a group responded uniformly. Nor does a p-value give the probability that a treatment does or does not work. The posted figures are best read as an imprecise estimate of a specific comparison, rather than a broad conclusion about acamprosate or alcohol-craving treatment.

16. Why This Trial Matters Statistically

This trial illustrates how a narrowly defined laboratory-style endpoint can be separated from broader clinical outcomes. The primary question concerns a rating after saline infusion, while the study context includes additional infusion challenges. Keeping the outcome condition explicit prevents the saline comparison from being mistaken for results after yohimbine or mCPP, for which no comparisons are posted.

It also demonstrates why analysis-population details matter. Although enrollment was 37, the primary comparison includes participants who completed all three infusion types, with 12 in the acamprosate group and 13 in the placebo group. The registry does not explain the difference between enrollment and the analyzed population. A reader therefore needs to distinguish the randomized trial's overall enrollment from the subset contributing to this outcome.

The result is a useful example of reading an estimate together with its uncertainty. The mean difference is -0.062, while the confidence interval ranges from -1.926 to 1.802. Reporting the point estimate alone could give an overly definite impression; the interval shows that the estimate does not precisely determine the direction or size of the group difference. The unadjusted p-value is not a substitute for that assessment.

Finally, the planning illustrations show how sample-size requirements depend on the effect a future study aims to detect, under stated assumptions. Because the observed comparison is close to zero, those illustrations should not be confused with the original trial's rationale or with evidence of observed power. The original design assumptions are not reported. This case highlights the statistical value of separating observed results, uncertainty, analysis-population limits, and forward-looking planning.

17. 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 main limitation for interpreting the primary comparison is the analysis population. The registry specifies that only participants completing saline, mCPP and yohimbine infusions were included. Arm-level results include 12 acamprosate participants and 13 placebo participants, whereas total enrollment was 37. Reasons for non-completion or exclusion are not reported, nor is the handling of missing primary-outcome data. This makes it difficult to assess how well the analyzed subset represents all randomized participants.

The reported endpoint is one rating at 180 minutes after saline infusion. The registry does not post comparison results for craving after mCPP or yohimbine, despite including these infusion types in the study description. No secondary outcome results are posted. The primary outcome therefore cannot support conclusions about those other conditions or outcomes not reported in the record.

The comparison has limited precision: the mean difference is -0.062 and the interval extends from -1.926 to 1.802. The confidence level and calculation method are not reported. The unadjusted p-value is reported, but the record does not specify the full statistical model or any multiplicity strategy. The original sample-size justification and analysis plan are also not reported.

The registry describes a randomized, triple-masked, parallel study, but it does not give further details about randomization implementation or masking procedures. It also does not report a threshold for a meaningful change on the craving scale. These limitations are gaps in the registry information; they do not demonstrate that the trial lacked procedures that are not described.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Modulation of Pharmacologically Induced Alcohol Craving in Recently Detoxified Alcoholics — NCT00605904.
• Registry record last updated 2012-07-03; status last verified 2012-05.
• Results first posted 2012-07-03.
• 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.

21. Record Summary

NCT00605904 is a completed randomized, triple-masked, parallel-group phase 2 study of acamprosate versus placebo in people with alcohol dependence. Enrollment was 37. The registry lists a start date of 2008-01 and primary completion date of 2010-09, with the National Institute on Alcohol Abuse and Alcoholism as sponsor.

The primary endpoint was Penn Alcohol Craving Scale rating 180 minutes after saline infusion. The analysis included participants who completed all three infusion types: 12 assigned to acamprosate and 13 assigned to placebo. Their mean ratings were 1.704 and 1.766, respectively. The posted mean difference was -0.062, with a confidence interval from -1.926 to 1.802 and an unadjusted p-value of 0.948.

The registry lists three outcome measures, but no analysis records or secondary outcome results are posted. Details on missing-data handling, multiplicity, and the original sample-size rationale are not reported in the registry record.