Comparison of Gabapentin and Metoclopramide for Treating Hyperemesis Gravidarum: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparison of Gabapentin and Metoclopramide for Treating Hyperemesis Gravidarum (NCT02163434), a randomized phase 2 study. The primary outcome is Baseline Adjusted Mean Daily Motherisk-PUQE Total Scores (Pregnancy-unique Quantification of Emesis and Nausea Scale) for Days 5-7.

2. Clinical Question

The comparison therefore concerns symptom scores during a short, controlled treatment period in the enrolled trial population. It does not, from the posted primary result alone, establish what happened during the optional open-label period or how outcomes varied across participant subgroups.

3. Trial Design

The study, titled Comparison of Gabapentin and Metoclopramide for Treating Hyperemesis Gravidarum, is registered as NCT02163434 and is listed as completed. The registry identifies it as a phase 2, randomized, parallel-group trial. It lists triple masking, while the brief summary calls the initial phase double-blinded. The record does not explain this difference, so the exact masking arrangement cannot be clarified from the registry information alone.

The brief summary describes a one-week randomized comparison, followed by an offer of open-label gabapentin with rescue metoclopramide until symptoms no longer required treatment. The primary endpoint timeframe is also listed as one week. Findings for the later open-label period are not reported in the registry record. The summary says enrollment was to occur at the University at Buffalo and the Universities of Rochester and Wisconsin; the completed record lists 31 participants, while the summary describes a planned enrollment of 60.

The registry lists gabapentin and metoclopramide as the two drug interventions. The reported primary outcome is a baseline-adjusted daily Motherisk-PUQE score averaged for days 5-7. Five secondary outcomes are listed, addressing nausea, oral nutrition, repeat intravenous hydration or hospital admission, treatment satisfaction, and desire to continue therapy. The record lists six outcome measures and five analyses, but does not provide further details about the analysis sets or how the secondary outcomes were tested.

The reason enrollment was 31 rather than the 60 described in the brief summary is not reported in the registry record. Neither the sample-size justification nor information about recruitment progress is reported. These gaps limit what can be inferred about the trial's planning and conduct beyond the design features explicitly recorded.

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 Baseline Adjusted Mean Daily Motherisk-PUQE Total Scores (Pregnancy-unique Quantification of Emesis and Nausea Scale) for Days 5-7, measured over 1 week.

The primary endpoint is the baseline-adjusted mean daily total score on the Motherisk-PUQE, the Pregnancy-unique Quantification of Emesis and Nausea Scale, for days 5-7. The registry defines the scale range as 6-30 and states that higher scores indicate a worse outcome. It is a continuous symptom measure, intended to capture the reported burden of nausea and emesis rather than a single event or a yes-or-no clinical outcome.

Using a daily score and summarizing days 5-7 provides a measure of symptoms near the end of the one-week randomized treatment period. Baseline adjustment is intended, in general, to account for starting symptom levels when comparing follow-up outcomes. It can make the comparison more informative than looking only at follow-up scores, especially if participants began with different scores. The registry does not describe how the adjustment was implemented, which baseline variables entered the model, or how repeated daily observations were handled.

The posted mean summaries are 6.35 for gabapentin and 13.22 for metoclopramide. Their reported spread is a standard error, not a standard deviation. These group summaries describe the endpoint as posted, but they are not themselves the model-based treatment contrast. The registry does not report that contrast or a confidence interval for it, so the size and precision of the adjusted difference cannot be read directly from the arm means alone.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a mean summarize?

A mean is the arithmetic average of the measured values in a group. For a continuous outcome such as the Motherisk-PUQE score, it gives a compact description of the group's average symptom score. The posted means are 6.35 for gabapentin and 13.22 for metoclopramide. Because higher scores indicate worse outcomes, these summaries place the gabapentin group at a lower average score. A difference between two arm means is descriptive unless it is paired with an appropriate inferential analysis that accounts for the study design and data structure.

The registry labels the endpoint as baseline-adjusted, while the displayed arm summaries are presented as means with standard errors. It is not clear from the record whether those summaries are adjusted estimates from the fitted model or another summary of the endpoint data. The model-based treatment contrast is not reported. For that reason, the arm means should not be treated as a substitute for an adjusted between-group estimate.

What does the standard error describe?

A standard error describes the uncertainty in an estimated quantity, such as a group mean, under a statistical model. It is different from a standard deviation, which describes dispersion among individual observations. The reported spreads of 2.44 and 2.39 are identified as standard errors. They therefore should not be interpreted as the typical distance of an individual participant's score from the group mean, and they do not reveal the full shape of the score distribution.

Standard errors can help readers judge how precisely group means have been estimated, but they do not by themselves provide the uncertainty in the contrast between groups. That uncertainty depends on the comparison, including how the observations are related and what model is used. The registry does not report a confidence interval for the treatment contrast. A confidence interval is a range of values compatible with the data and the stated model; it helps show both the estimated direction and the precision of a result.

What does a mixed model do?

In general, a mixed model can represent both average patterns shared across participants and participant-specific variation. It is often useful when an outcome is measured repeatedly or when observations are grouped, because measurements from the same participant may be related rather than independent. A model can account for that structure while estimating a comparison between randomized groups. The registry names a mixed-model analysis, but the details needed to assess its exact form are not reported in the registry record.

In particular, the record does not state the model terms, covariance structure, handling of repeated measurements, covariates beyond the baseline adjustment named in the endpoint, or method for dealing with missing observations. It also does not report whether the p-value was adjusted for any other analyses. These are important details because different reasonable model specifications can produce different estimates or uncertainty, especially when the available sample is limited.

Over-dispersion is a general concern when observed variation is greater than a model assumes. If a model understates variation, its uncertainty estimates can be too small and inferential results can appear more precise than they are. Whether over-dispersion is relevant to this continuous endpoint, how the outcome distribution was checked, and whether any corrective approach was used are not reported in the registry record. This is a general statistical consideration, not a claim that over-dispersion occurred in this trial.

How should the p-value be read?

The registry reports p = 0.01 for the mixed-model superiority analysis. In general, a p-value measures how unusual data at least as inconsistent with a specified null hypothesis would be if that null hypothesis and the model assumptions held. It is not the probability that the null hypothesis is true, and it does not say how large an effect is. A small p-value can coexist with an imprecise effect estimate when the estimate and its interval are not reported; it cannot replace those quantities.

The registry does not identify the exact null hypothesis formulation or a prespecified decision threshold. It also does not provide the model-based contrast, confidence interval, or test statistic. Consequently, readers can report the posted p-value and the direction of the arm summaries, but should avoid reconstructing an inferential result from the displayed means and standard errors. The code-generated recalculation is unavailable because the two-arm continuous comparison is listed as having values missing or not estimable.

What do the planning illustrations mean?

In general, sample-size planning asks how many participants would be needed to detect a chosen effect with a specified statistical procedure and level of power. A planning illustration based on an observed effect is a conditional calculation: its result depends on treating the observed effect, or a fraction of it, as the effect worth detecting and on the assumptions stated for the illustration. It is not evidence that the completed trial had that level of power, and it is not an observed study result.

The registry record does not report the trial's own sample-size justification. The available planning assumptions describe a two-sided significance level of 0.05, equal allocation, a normal approximation, and effect sizes expressed as fractions of the observed effect. They do not establish that these were the design assumptions used by the investigators. The resulting participant counts are not reported in the registry record, so no such counts can be described. A planning illustration should not be used to infer whether the trial was adequately powered or to reinterpret its observed p-value as observed power.

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("Gabapentin", "Metoclopramide")
n   <- c(12, 9)
m   <- c(6.35, 13.22)
sp  <- c(2.44, 2.39)
sd  <- sp * sqrt(n)      # standard errors converted to SDs
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  Gabapentin  12  6.350  8.452  0.980  11.720
CARM  Metoclopramide  9  13.220  7.170  7.709  18.731

11. Analyses Reported by the Sponsor

The registry's reported analysis is a mixed-model superiority comparison with p = 0.01. It does not include a model-based effect estimate or confidence interval. A separate unadjusted recalculation is not available for this continuous two-arm endpoint: the registry analysis data are identified as missing or not estimable for that comparison. The arm means and standard errors alone are not enough to recreate the mixed-model result.

No discrepancy is identified between a recalculation and the reported result, because a recalculated comparison is not available. This is not evidence that the two approaches agree; there is no independent recalculated estimate to compare with the sponsor's analysis. The reported p-value should therefore be described as the registry's mixed-model result, without implying that it has been independently reproduced from the posted summaries.

The arm summaries remain useful as descriptive information: the means are 6.35 for gabapentin and 13.22 for metoclopramide, with lower scores indicating less severe symptoms on the scale. But the adjusted comparison's magnitude and precision cannot be established from those summaries. Details of the model specification and analysis population are also not reported in the registry record, limiting independent assessment of the analysis.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists five secondary outcomes in addition to the primary endpoint and reports five analyses. The listed secondary outcomes cover daily nausea scores, oral nutrition scores, repeat intravenous hydration or hospital admission, treatment satisfaction, and desire to continue therapy. The registry does not provide the results for those secondary outcomes in the information recorded here, so their direction, magnitude, and precision cannot be assessed.

When a study tests several outcomes, the chance of at least one apparently unusual result can increase if each test is interpreted without regard to the others. General approaches to this issue include specifying a primary outcome in advance, using a testing hierarchy, or adjusting inference across a family of tests. The registry record does not report a multiplicity strategy, a testing hierarchy, or whether the reported p-value accounts for other outcomes. It is therefore not possible to determine from the record how multiplicity was handled.

Missing-data methods are also not reported in the registry record. The displayed primary outcome summaries include 12 participants in the gabapentin arm and 9 in the metoclopramide arm, compared with total enrollment of 31. The registry does not explain whether this reflects missing endpoint data, analysis exclusions, or another reason. It does not state how missing scores were handled or whether any sensitivity analyses were conducted. Those omissions matter because an analysis can be affected if participants with unavailable outcomes differ systematically from those with observed outcomes.

The analysis population is not reported in the registry record. It is not clear whether the primary analysis included all randomized participants, only participants with observed endpoint data, or another defined group. The registry also does not describe adherence, protocol deviations, interim analyses, or whether any analyses were planned after examining the data. These details cannot be inferred from the randomized design or the reported p-value.

13. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent non-serious adverse event terms

The serious adverse event entries list 0 affected participants in the gabapentin group, among 12 at risk, and 0 in the metoclopramide group, among 9 at risk. The death counts are also 0 in both groups. These are the serious-event and death figures recorded for the trial; zero reported events does not establish that no risk exists, nor does it show that the groups have identical safety profiles.

For other adverse events, the record lists 4 affected participants among 15 at risk in the gabapentin group and 3 among 16 at risk in the metoclopramide group. These counts use risk populations that differ from the primary endpoint summary counts. They should not be silently substituted for the numbers analyzed for the symptom endpoint, and the record does not explain the different denominators. No serious adverse event terms are listed.

Adverse-event counts depend on how events were defined, collected, and attributed, as well as the duration and completeness of follow-up. The registry record does not provide additional detail about event severity, timing, relatedness, or the methods used for safety monitoring. With a completed enrollment of 31 and a short primary treatment period, these entries are a limited description of events recorded, not a comprehensive estimate of comparative safety.

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

The posted primary outcome summaries show lower average Motherisk-PUQE scores in the gabapentin arm than in the metoclopramide arm. Because higher values on the scale indicate worse symptoms, that is the direction of the descriptive group means. The registry also reports a mixed-model superiority analysis with p = 0.01. Taken together, these are the main posted statistical findings for the one-week primary comparison.

Those findings do not reveal the size of the adjusted difference or its precision, because the registry does not report the model-based estimate or confidence interval. A p-value is not a measure of clinical importance. The registry also does not report enough model detail to evaluate all assumptions behind the comparison, and the displayed endpoint counts are smaller than total enrollment without an explanation. Those limits should remain part of any reading of the reported result.

The evidence in the record concerns the enrolled participants and the primary symptom endpoint over one week. It does not establish what happened during the optional open-label phase, how long any difference persisted, or whether results would be similar in other populations or settings. The summaries also do not describe individual participant responses; group averages can conceal variation among people.

The serious adverse event entries report no affected participants in either arm, while other adverse events are recorded in both groups. These observations are not enough to conclude that the treatments have equal safety, particularly because the risk denominators and endpoint denominators differ and additional safety detail is absent. The registry findings are statistical information about this trial record, not individual medical guidance or a definitive account of the overall balance of outcomes.

15. Why This Trial Matters Statistically

This trial is statistically informative because it compares two active drug interventions using random allocation and a continuous symptom score. A randomized design is intended, in general, to support a fair comparison by reducing systematic differences between treatment groups at assignment. The parallel-group structure and named primary endpoint make the central comparison identifiable, while the baseline-adjusted outcome indicates that starting symptom levels were intended to be considered.

The record also illustrates why a p-value alone is not a complete statistical report. The mixed-model analysis gives p = 0.01, but no model-based contrast or confidence interval is recorded. Without those, readers cannot judge the estimated magnitude and precision of the adjusted difference from the inferential result. The arm summaries provide useful descriptive context, but they do not replace an estimate of the treatment contrast.

Enrollment and endpoint reporting deserve attention as well. The brief summary describes a plan to enroll 60 women, whereas the completed record lists 31. The endpoint summaries include 12 and 9 participants in the two arms, and the reasons those counts differ from total enrollment are not reported. This is a reminder that the randomized sample, the enrolled sample, and the analyzed sample may not be interchangeable, and that a clear account of analysis populations and missing data is important.

Finally, the record distinguishes a short blinded comparison from a later open-label phase but reports the primary result for only the one-week timeframe. That distinction prevents readers from extending a short-term endpoint to longer-term outcomes. Together, these details show how design, outcome definition, analysis reporting, and transparent participant accounting shape the interpretation of a clinical trial result.

16. 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 trial-specific details unresolved. It lists triple masking, whereas the brief summary describes a double-blinded initial phase. The record does not explain the difference or identify which participants or study personnel were masked. The actual enrollment of 31 is also lower than the 60 described in the brief summary, but the reason is not reported in the registry record. The sample-size justification and recruitment history are not reported.

The primary outcome is described as baseline-adjusted, but the adjustment method and full mixed-model specification are not recorded. The registry reports p = 0.01 without a model-based effect estimate, confidence interval, test statistic, or detailed account of the assumptions. No independent recalculation is available for the continuous two-arm comparison because values are listed as missing or not estimable. The discrepancy flag is not present, but that cannot substitute for a reproducible comparison.

The endpoint summary counts are 12 and 9, while total enrollment is 31. The registry does not report the analysis population, reasons for unavailable endpoint values, or methods for missing data. It also does not describe protocol deviations, adherence, sensitivity analyses, or interim analyses. Without that information, readers cannot assess how the analyzed participants relate to all those enrolled or whether the result is robust to alternative handling of unavailable observations.

Although five secondary outcomes and five analyses are listed, results and multiplicity procedures are not reported in the registry record. The brief summary describes an open-label period after the randomized comparison, but results for that period are not recorded. Safety reporting is limited to event counts and risk denominators, with no further detail on timing, severity, or attribution. These limitations do not negate the posted findings; they define what can and cannot be concluded from this record.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Comparison of Gabapentin and Metoclopramide for Treating Hyperemesis Gravidarum — NCT02163434.
• Registry record last updated 2023-02-14; status last verified 2023-02.
• Results first posted 2021-09-27.
• 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.

20. Record Summary

NCT02163434 is a completed phase 2 randomized, parallel-group comparison of gabapentin and metoclopramide for hyperemesis gravidarum. The registry lists triple masking, while its brief summary describes a double-blinded initial phase. The primary endpoint is the baseline-adjusted mean daily Motherisk-PUQE total score for days 5-7 during a one-week period; higher scores indicate worse symptoms.

The posted arm means are 6.35 for gabapentin, based on 12 participants, and 13.22 for metoclopramide, based on 9 participants. The reported analysis is a mixed-model superiority comparison with p = 0.01. A model-based effect estimate and confidence interval are not reported, and an unadjusted recalculation is not available. The record lists total enrollment as 31, five secondary outcomes, and 0 serious adverse events in each arm. Details about missing endpoint data, analysis populations, multiplicity, and the later open-label phase are not reported in the registry record.