Dexmedetomidine (Precedex®) for Severe Alcohol Withdrawal Syndrome (AWS) and Alcohol Withdrawal Delirium (AWD): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Dexmedetomidine (Precedex®) for Severe Alcohol Withdrawal Syndrome (AWS) and Alcohol Withdrawal Delirium (AWD) (NCT01362205), a randomized phase 4 study. The primary outcome is The Length of ICU Stay Defined as the Time Between Randomization and ICU Transfer Orders..

2. Clinical Question

The investigators’ stated hypothesis covered several aspects of recovery in critically ill adults with severe AWS or AWD. They anticipated that integrating dexmedetomidine with usual therapy could reduce time to resolution of AWS or AWD, increase delirium-free and ventilator-free days during the first 28 days of hospitalization, reduce ICU and hospital length of stay, and improve neurocognitive and quality-of-life scores at hospital discharge. These are proposed effects, not conclusions established by the hypothesis itself. The registry record reports a primary endpoint and some secondary outcomes, but does not provide numerical results for every outcome named in the hypothesis.

The clinical comparison is therefore narrower than the full set of proposed benefits: the reported primary result concerns ICU transfer timing, and additional posted outcomes address cognition, mood, delirium, ventilation, and hospital stay. The registry record does not report whether the trial was designed to draw conclusions about each of these outcomes independently, nor does it give a reason for termination.

3. Trial Design

The registry describes a randomized, triple-masked, placebo-controlled, parallel-group phase 4 study. It lists Denver Health and Hospital Authority as the sponsor. Enrollment is recorded as 49, and the status is terminated. The recorded start is 2012-03 and primary completion is 2015-12. The reason for termination, the timing of the decision, and whether termination affected planned follow-up are not reported in the registry record.

Participants were assigned to dexmedetomidine or placebo. The arm summaries report 22 participants for dexmedetomidine and 27 for placebo. These are the group sizes attached to the posted results; the registry record does not explain whether every enrolled participant contributed to each outcome or whether any participants were excluded from analyses. It also does not report the randomization sequence, allocation concealment procedures, or the exact roles of the people included in the triple masking.

The active medication was described as titrated according to selected components of the MINDS score, with the goal of using the minimum amount needed and a maximum infusion rate of 1.4 μg/kg per hour. The placebo was administered in equal volume per hour to the active study medication. The official title and summary specify lorazepam rescue. Further details about rescue criteria, other usual care, treatment duration, and adherence are not reported in the registry record. The posted description of the dexmedetomidine regimen is incomplete, so it does not permit a full reconstruction of the treatment protocol.

The primary endpoint was the length of ICU stay, defined as time between randomization and ICU transfer orders, with a stated timeframe of up to 28 days in hours. Secondary outcomes listed by the registry include average MINDS score, CAM-ICU-negative days, ventilator-free days, hospital length of stay, and discharge scores on several cognitive and psychological assessments. The record identifies 11 outcome measures and 10 posted secondary outcomes. The detailed numerical summaries in the available record cover only some of these outcomes.

The registry does not report the planned sample-size justification, an enrollment target distinct from the recorded enrollment, an interim-analysis plan, or a full statistical analysis plan. Those omissions limit assessment of how the completed trial compared with its original design.

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 Length of ICU Stay Defined as the Time Between Randomization and ICU Transfer Orders., measured over up to 28 days in hours.

The primary endpoint is the time from randomization until ICU transfer orders, expressed in hours and described as ICU length of stay. It measures how long participants remained in the ICU before the order to transfer them was placed. The record states that observation extended up to 28 days. This definition depends on a clinical and administrative milestone rather than a direct assessment of withdrawal symptom resolution.

In general, time in intensive care is used as an indicator of resource use and the course of critical illness. It may be relevant when comparing management strategies because care in the ICU is intensive, and the timing of transfer can reflect a participant’s condition and readiness for a different level of care. However, transfer decisions can also depend on local practice, bed availability, and other clinical circumstances. These are general considerations, not documented features of this trial’s conduct.

The posted primary summaries are medians with interquartile ranges. The dexmedetomidine median was 79.2 hours, with an interquartile range from 43.9 to 134.6 hours. The placebo median was 104.9 hours, with an interquartile range from 63.7 to 183.1 hours. An interquartile range describes the middle half of the observed values; it is a measure of spread, not a confidence interval for the median. The registry does not post a confidence interval for the between-group difference in this endpoint.

Because the outcome is time from randomization to transfer orders, it should not be interpreted as time to discharge from hospital, time to resolution of withdrawal, or the duration of ventilation. Those are distinct outcomes. The registry record does not explain how participants still in the ICU at the end of the stated timeframe were handled in the analysis.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary endpoint is a duration measured from randomization to an ICU transfer order. Duration outcomes often have an asymmetric distribution: many participants may have relatively short stays while a smaller number have long stays. A mean can be strongly affected by such long observations. The median is less sensitive to extreme values and identifies the middle of the observed distribution. The interquartile range describes the spread across the middle portion of the data. These are general statistical properties; the registry does not report a plot or numerical distributional diagnostics for this endpoint.

What does the reported rank test assess?

The sponsor’s reported primary analysis used the Wilcoxon (Mann-Whitney) test. In general, this method compares the ranks of observed outcomes across groups rather than comparing only their means. It can be useful for skewed measurements and does not require the same normal-distribution assumption as a conventional two-sample t-test. Its interpretation is not automatically a test of the difference between medians. A simple median interpretation is most defensible when the group distributions have comparable shapes and differ mainly in location. The registry record does not report whether that condition was evaluated.

The reported p-value is 0.245. A p-value describes how compatible the observed data are with a specified null model, assuming the analysis model and its conditions hold. It is not the probability that the null hypothesis is true, and it does not measure the probability that a treatment effect is absent. A result that does not cross a conventional significance threshold can reflect a small effect, imprecision, or both. Without an effect estimate and confidence interval for the primary contrast, the reported test does not show which effect sizes remain compatible with the data.

Why distinguish medians, ranges, and confidence intervals?

The arm summaries give medians and interquartile ranges. The interquartile range is a descriptive summary of observed variation among participants; it does not quantify uncertainty in the estimated median. A confidence interval for an effect estimate instead describes uncertainty in the estimated contrast under the chosen method. These measures answer different questions. The registry record reports no confidence interval for the primary between-group comparison, so readers should not treat the interquartile ranges as if they were confidence limits.

The primary test result is reported by the sponsor, whereas the arm summaries provide the observed values in each group. Since the test is rank-based, its p-value need not be recoverable from the two medians and interquartile ranges alone. Individual-level observations, or further distributional information, would be needed to reproduce the rank comparison. A separate unadjusted recalculation from posted summary statistics would not substitute for that analysis.

How should recalculated secondary comparisons be read?

For some discharge outcomes, the posted means and standard deviations allow simple, unadjusted comparisons of group means. A mean difference is the difference between the two reported group means, with the ordering specified by the comparison. A confidence interval describes uncertainty around that estimated difference under the assumptions of the calculation. If an interval includes zero, the results are compatible with no mean difference as well as with differences in either direction at the stated confidence level. If an interval does not include zero, that does not establish clinical importance or remove the need to consider multiple testing.

These recalculations are based on summary statistics, not participant-level data. They are unadjusted and do not account for covariates, missing observations, or the relationships among outcomes. They should therefore be distinguished from a sponsor’s own prespecified analysis, if one is reported. No sponsor-reported comparisons are listed for the posted discharge assessments. The recalculated results are informative descriptions of the available summaries, but they cannot establish that those analyses were planned in advance.

What does over-dispersion mean in general?

For outcomes counted as events, over-dispersion occurs when the observed variability exceeds what a simple Poisson model expects for its mean. A model that ignores over-dispersion can underestimate uncertainty and produce confidence intervals that are too narrow. This general issue is relevant when interpreting event-count analyses, but the registry’s primary outcome is a duration, and the posted primary method is a rank test. The record does not report an over-dispersion assessment or an event-count model for the primary endpoint.

How should planning illustrations be understood?

The sample-size illustrations are explicitly based on observed effects and are not the trial’s own design justification. They assume a two-sided significance level of 0.05, equal allocation, and a normal approximation; the illustrated effect is a stated fraction of the observed effect. In general, a planning calculation asks how many participants per arm would be needed to attain a specified power under its assumptions. It does not measure the power achieved by this completed trial, and it is not evidence that an observed result is true. Using an observed effect for planning can give an optimistic illustration because the observed estimate may be unstable, particularly in a small study. The registry record does not report the assumptions used when the trial was originally designed or a prospective sample-size calculation.

For duration outcomes, the listed sample-size illustrations use a total-event convention rather than participants per arm. That distinction reflects the calculation method, not the number of people who participated or an observed trial result. The posted planning material should be read as hypothetical design guidance only, not as an explanation for the study’s enrollment or termination.

10. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis is the Wilcoxon (Mann-Whitney) comparison of ICU duration between dexmedetomidine and placebo. Its p-value is 0.245. The record does not report a sponsor-estimated primary treatment contrast or a confidence interval for that contrast. The posted medians and interquartile ranges describe each group, but they do not supply the missing estimate and uncertainty for the rank-based comparison.

There is no recalculated primary comparison to set against the sponsor’s test. The recalculated comparisons available in the record concern selected secondary discharge measures and are based on group means, standard deviations, and sample sizes. They use a different endpoint and a different statistical approach from the primary rank test. They therefore cannot confirm, contradict, or replace the sponsor’s primary analysis.

The registry record lists no sponsor-reported comparison tests for those discharge measures. The recalculations are unadjusted and should be treated as descriptive statistical additions rather than as the sponsor’s analysis plan. No discrepancy between a sponsor calculation and a recalculation is identified for the primary result. Important details needed to assess the sponsor’s analysis more fully—including its handling of missing outcomes, the analysis population, and the rationale for the test—are not reported in the registry record.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 11 outcome measures, including 10 posted secondary outcomes. Testing many outcomes raises the chance that one or more results may appear statistically notable even when there is no true effect on those outcomes. In general, multiplicity procedures can control this problem by adjusting significance thresholds, organizing outcomes into a testing hierarchy, or defining a limited set of confirmatory comparisons. The registry record does not report whether any such procedure was prespecified or used. The available secondary recalculations are unadjusted and should not be interpreted as if they account for the full set of outcome comparisons.

The record identifies one reported primary analysis, using the Wilcoxon (Mann-Whitney) method. It does not report whether the primary analysis was prespecified in a statistical analysis plan, whether any alternative analyses were conducted, or whether interim analyses were planned. The absence of such details is not evidence that these procedures did or did not occur; they are not reported in the registry record.

Missing-data handling is also not reported in the registry record. The arm summaries give group sizes, but the record does not state whether those numbers represent the analysis population for every endpoint, whether all participants had an observed primary outcome, or how incomplete follow-up was handled. It does not report imputation, censoring rules, or sensitivity analyses for missing observations. These details matter because an analysis restricted to people with complete outcome data can be biased if completeness is related to prognosis or treatment.

The analysis population is not reported in the registry record. It is therefore unclear whether the primary comparison followed randomized assignment regardless of treatment received, included only participants who started study medication, or used another population definition. The record also does not state how protocol deviations, crossover, or rescue medication were incorporated into the analyses. Without those details, it is difficult to assess how closely the posted estimates preserve the benefits of randomization.

The trial’s terminated status is recorded, but its reason and the relationship between termination and outcome reporting are not reported in the registry record. The record does not establish that termination caused missingness or selective reporting. It does, however, leave these questions unresolved and limits the ability to evaluate completeness and prespecification.

12. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: respiratory failure (9.5%), atrial fibrillation (4.8%), cardiac arrest (4.8%), hepatitis alcoholic (4.8%), enterococcal infection (4.8%).

The registry reports serious adverse events and other adverse events separately by treatment group. For serious adverse events, 5 participants were recorded as affected among 21 at risk in the dexmedetomidine group, and 3 were affected among 27 at risk in the placebo group. For other adverse events, 12 were recorded as affected among 21 at risk in the dexmedetomidine group, and 17 among 27 at risk in the placebo group. The listed deaths were 0 in each group.

These counts are descriptive safety summaries. They do not, by themselves, establish that one treatment caused more or fewer events. The registry record does not give a statistical comparison, confidence intervals, event timing, severity grading details, or enough information to assess differences in exposure and follow-up. It also does not clarify whether the affected-participant counts represent people with at least one event or how recurrent events were handled.

The listed serious-event terms include respiratory failure, atrial fibrillation, cardiac arrest, alcoholic hepatitis, enterococcal infection, and septic shock, with percentages attached in the record. Those term-level figures should be interpreted in the context of the registry’s event definitions and denominators, which are not fully described in the record. Absence of a reported death does not mean absence of serious illness or establish comparative safety.

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

The posted primary data describe ICU duration in each randomized group, with lower observed median time in the dexmedetomidine arm than in the placebo arm. The sponsor’s rank-based test reports a p-value of 0.245, and no confidence interval or effect estimate for that primary comparison is reported. Together, these registry entries do not provide strong statistical evidence of a group difference under the reported test, but they do not establish that the groups are equivalent or that any possible difference is clinically unimportant.

The finding is specific to time from randomization to ICU transfer orders, within a timeframe of up to 28 days. It should not be recast as evidence about time to withdrawal resolution, hospital discharge, ventilator duration, or long-term recovery. Those are separate outcomes. The registry’s hypothesis included several such outcomes, but detailed results are not available for all of them in the record.

Some secondary discharge-scale comparisons have unadjusted recalculated estimates. These results are exploratory, especially because multiple outcomes were posted and no adjustment strategy is reported. A nominally notable result among several outcomes does not provide the same strength of evidence as a single prespecified primary analysis. The scale scoring and clinical interpretation are not described in enough detail to infer the practical meaning of the numerical differences.

In general, a randomized design can support causal interpretation when assignment is preserved, follow-up is sufficiently complete, and analyses are appropriate. The registry record does not report the analysis population, missing-data procedures, or detailed protocol adherence. It also records that the study was terminated without giving a reason. These limitations make the posted results less informative than a fully documented trial report with prespecified methods and complete outcome reporting.

No treatment recommendation follows from these results. They describe the posted data and the limits of the available statistical information. The record does not establish benefit, harm, or equivalence for dexmedetomidine compared with placebo on the primary outcome, and it does not resolve the broader hypothesis about recovery and discharge outcomes.

14. Why This Trial Matters Statistically

This trial illustrates why an outcome’s definition is central to statistical interpretation. ICU length of stay is defined through the timing of transfer orders, not a direct measure of symptom resolution. A reader who treats those endpoints as interchangeable could overstate what the primary result addresses. The registry’s stated hypothesis spans several dimensions of critical illness and recovery, while the primary endpoint captures one specific transition in care.

It also shows the difference between a descriptive group summary and an inferential comparison. Medians and interquartile ranges communicate the observed distributions in each arm, while the reported rank-based p-value addresses a group comparison using ranks. Neither alone gives a precise treatment-effect estimate with uncertainty. The absence of a primary confidence interval makes it harder to judge whether the data exclude effects that might matter clinically.

The available secondary results emphasize another statistical issue: multiple outcomes can produce a scattered collection of estimates and p-values. Without a reported hierarchy or adjustment, individual secondary findings should be treated cautiously. Their relevance is not eliminated, but their role is more naturally exploratory than confirmatory. Keeping primary and secondary evidence distinct helps avoid selecting whichever result appears most persuasive after the fact.

Finally, registry reporting provides important design and results details, but omissions affect appraisal. Here, the reason for termination, sample-size justification, missing-data procedures, analysis population, and multiplicity approach are not reported. Recognizing those gaps is part of statistical interpretation, not a claim about what investigators did. This example demonstrates how transparent reporting of estimands, analysis decisions, and uncertainty would help readers assess what a small, terminated randomized study can and cannot establish.

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 registry record leaves important aspects of this trial’s conduct and analysis unclear. It does not report the reason for termination, the original sample-size justification, or whether the study reached its planned enrollment. The recorded enrollment is 49, but the record does not explain the relationship between that figure and the arm-specific analysis populations. It also does not report recruitment flow, losses to follow-up, or the number contributing to each outcome.

The primary result lacks a between-group effect estimate and confidence interval. The arm-level medians and interquartile ranges are useful descriptive summaries, but they cannot show the precision of the primary contrast. The reported rank-test p-value does not answer how large a difference is compatible with the data or what range of effects could be clinically relevant. The record does not state whether the test’s assumptions were assessed.

There is limited information about treatment delivery. The active regimen description is incomplete, and details of lorazepam rescue, usual care, adherence, and protocol deviations are not reported in the registry record. The record also does not explain the masking procedures in detail or identify the analysis population. These omissions constrain assessment of whether the intended randomized comparison was maintained throughout follow-up.

Outcome reporting is incomplete for evaluating the full stated hypothesis. The registry lists multiple secondary outcomes, but the numerical summaries available for this analysis cover only selected discharge measures. No multiplicity strategy, missing-data approach, or secondary-outcome analysis plan is reported. Recalculations from summary statistics are unadjusted and cannot reproduce participant-level or sponsor-planned analyses.

Finally, ICU transfer orders may be influenced by clinical judgment and local operational circumstances in addition to a participant’s condition. This is a general limitation of the endpoint, not a documented problem specific to this trial. The registry record does not describe transfer practices or provide enough information to assess how these circumstances affected the results.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Dexmedetomidine (Precedex®) for Severe Alcohol Withdrawal Syndrome (AWS) and Alcohol Withdrawal Delirium (AWD) — NCT01362205.
• Registry record last updated 2017-11-06; status last verified 2017-10.
• Results first posted 2017-07-17.
• 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

NCT01362205 is a randomized, triple-masked, placebo-controlled, parallel-group phase 4 study of dexmedetomidine with lorazepam rescue for critically ill adults with severe AWS or AWD. The registry lists 49 enrolled participants and a terminated status. The primary outcome was time from randomization to ICU transfer orders, measured in hours up to 28 days.

The reported primary summaries are medians of 79.2 hours for dexmedetomidine and 104.9 hours for placebo, with their respective interquartile ranges. The sponsor reports a Wilcoxon (Mann-Whitney) comparison with p = 0.245, but no primary effect estimate or confidence interval. Selected secondary discharge comparisons are available as unadjusted recalculations; they are exploratory in the absence of a reported multiplicity plan. The reason for termination, analysis population, missing-data methods, and sample-size justification are not reported in the registry record.