This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Montelukast Therapy on Alzheimer's Disease (NCT03991988), a randomized phase 2 study. The primary outcome is Number of Participants With Any Gastrointestinal (GI) Symptoms.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03991988 |
| Phase | Phase 2 |
| Condition | Alzheimer Disease |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Enrollment | 32 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Emory University |
| Results posted | Yes |
2. Clinical Question
Population
Individuals with mild cognitive impairment and early Alzheimer’s disease dementia, as described in the registry summary. The condition listed in the record is Alzheimer Disease.
Intervention
Montelukast, with escalating doses of 10, 20 to 40 mg as described in the summary and arm description.
Comparator
Matched placebo, described as a placebo pill in the arm information and as a placebo oral tablet in the interventions.
Primary question
How many participants in each group had any reported gastrointestinal symptoms—diarrhea, nausea or vomiting—at baseline or 1 year, and what was the unadjusted between-group risk difference?
The central statistical question concerns a recorded symptom outcome, not the cognitive, cerebrospinal fluid or neuroimaging measures described in the trial summary. Those other measures include cognitive function, CSF biomarkers and neuroimaging, while the posted primary outcome specifically counts participants with any gastrointestinal symptoms. The record gives a result for each randomized group and a risk-difference comparison.
This distinction matters when reading a trial with several kinds of outcomes. The trial’s broader scientific aims include measures relevant to cognition and brain biomarkers, but the primary result is about symptom reporting. It should not be treated as a result for cognitive function, Alzheimer’s disease progression or any of the listed secondary outcomes. The registry record does not report a primary result for those other measures in the information available for this analysis.
The reported comparison addresses the difference between the groups in the proportion of participants counted as having symptoms. The result does not, by itself, describe symptom severity, duration, timing within follow-up or whether a participant experienced more than one of the listed symptoms. Those details are not reported in the registry record.
3. Trial Design
EMERALD, formally titled Effects of Montelukast Therapy on Alzheimer’s Disease, is a completed phase 2 randomized, double-blind, parallel-group clinical trial sponsored by Emory University. The registry lists an enrollment of 32. The summary describes a one-year comparison of montelukast with placebo in individuals with mild cognitive impairment and early Alzheimer’s disease dementia. The recorded start date is 2019-09-25, and the primary completion date is 2022-11-18.
The two study arms are the Montelukast Group and the Placebo Group. Each arm has 16 participants in the posted primary-outcome results. The montelukast arm is described as receiving 10, 20, or 40 mg, with the trial summary specifying escalating doses. The comparator is described as a matched placebo pill. The record identifies the study as randomized and double masked, but details about the randomization procedure, concealment, masking assessment, and dose-escalation schedule are not reported in the registry record.
The design is parallel: participants are assigned to a group rather than described as crossing from one treatment to another. The primary outcome is the number of participants with any gastrointestinal symptoms, defined as diarrhea, nausea or vomiting. Its listed timeframe is baseline and 1 year. The registry also lists 11 outcome measures and four posted secondary outcomes: CSF Amyloid, CSF Tau Levels, Clinical Dementia Rating Score and NIH Toolbox Cognition Battery.
The completion status is recorded as completed. The registry record does not report the detailed protocol, treatment adherence, participant flow, reasons for withdrawal, or how many participants contributed data at each measurement time. Although the summary describes a one-year trial, the available primary outcome information does not explain exactly how symptoms were collected or combined across baseline and follow-up.
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
| Stage | Montelukast Group | Placebo Group |
|---|---|---|
| STARTED | 16 | 16 |
| COMPLETED | 14 | 14 |
| NOT COMPLETED | 2 | 2 |
Reasons for not completing this period:
| Reason | Montelukast Group | Placebo Group |
|---|---|---|
| Withdrawal by Subject | 2 | 2 |
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.
| Characteristic | Unit | Montelukast Group | Placebo Group | Total |
|---|---|---|---|---|
| Participants analysed | 16 | 16 | 32 | |
| Age, Continuous | years | 71.3 (SD 8.8) | 73.9 (SD 7.7) | 72.6 (SD 8.2) |
| Sex: Female, Male: Female | Participants | 12 | 9 | 21 |
| Sex: Female, Male: Male | Participants | 4 | 7 | 11 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 0 | 0 | 0 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 16 | 16 | 32 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Asian | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Black or African American | Participants | 4 | 0 | 4 |
| Race (NIH/OMB): White | Participants | 12 | 16 | 28 |
| Race (NIH/OMB): More than one race | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 0 | 0 | 0 |
| Region of Enrollment: United States | participants | 16 | 16 | 32 |
| Mild Cognitive Impairment (MCI) | Participants | 10 | 9 | 19 |
| Dementia | Participants | 6 | 7 | 13 |
6. Primary Endpoint
The registry lists the primary outcome as Number of Participants With Any Gastrointestinal (GI) Symptoms, measured over Baseline, 1 year.
| Element | Registry entry |
|---|---|
| Result type | Count Of Participants |
| Measure of spread | not applicable |
| Unit | Participants |
| Endpoint classification | Binary Endpoint |
The primary endpoint is the number of participants reported to have any gastrointestinal symptoms. The listed symptoms are diarrhea, nausea and vomiting, and the stated timeframe is baseline and 1 year. A participant who had any of these symptoms is counted for this endpoint; the record does not provide separate primary results for each symptom or a measure of symptom severity.
This type of endpoint is useful for describing how often a defined event was reported in each randomized group. Expressing the result as a participant count makes the outcome relatively direct to interpret: it concerns the number counted as having symptoms, rather than a change in a continuous score. The comparison can then be expressed as a risk difference, which contrasts the groups’ event proportions.
The registry reports one participant with symptoms in the Montelukast Group and one in the Placebo Group, with 16 participants in each group for the posted result. The primary comparison reports a risk difference of 0.0. This is a descriptive finding about the recorded endpoint in these groups; it does not establish that the true difference between the treatments is exactly zero.
The record does not specify whether the endpoint means symptoms at either measurement, symptoms newly occurring after baseline, or symptoms present at a particular visit. It also does not report the collection instrument, the rules for handling recurrent symptoms, or whether the endpoint was assessed by self-report, clinical interview or another method. These omissions limit interpretation of precisely what the count represents over the stated timeframe.
7. Results Posted in the Registry
- Montelukast Group: 6.2% (1/16) versus 6.2% (1/16) on control; risk difference 0.0 percentage points (95% CI -16.8 to 16.8), risk ratio 1.00 (0.07–14.64); unadjusted two-proportion z-test p = 1.000.
Risk difference: Montelukast Group
95% CI: -16.8 to 16.8 percentage points
Recalculated from the posted summary statistics
8. Statistical Interpretation
In the posted primary-outcome results, one participant in the Montelukast Group and one participant in the Placebo Group are counted as having any gastrointestinal symptoms. The group sizes contributing to these results are 16 and 16. The estimated risk difference is 0.0, and the reported unadjusted p-value is 1.0. These figures describe the result for the specified symptom endpoint; they are not results for cognition, CSF biomarkers or neuroimaging.
The point estimate of 0.0 indicates that the observed event proportions were the same in the two groups for this endpoint. That statement is limited to the posted counts and does not imply that the underlying risks are known to be identical. The confidence interval spans values on both sides of the null value, so the data are compatible with differences in either direction. Its breadth is important: the estimate is based on few recorded events, and the result does not tightly locate the possible difference.
The unadjusted p-value is a measure of how unusual results at least as incompatible with a specified null model would be if that model and the test’s assumptions held. It is not the probability that the treatments are equivalent, nor the probability that the null hypothesis is true. A p-value of 1.0 does not prove no difference. With sparse binary outcomes, a non-significant result can reflect limited information as well as a small estimated contrast.
The registry lists risk ratio and odds ratio estimates in addition to the risk difference. With one event in each group, those relative comparisons are also based on very sparse data; the reported intervals are wide. For this outcome, the absolute contrast is often easier to interpret because it expresses the comparison in terms of the difference between group event proportions. None of these estimates should be interpreted as a precise description of the underlying treatment effect.
The outcome is reported at baseline and 1 year, but the registry does not explain how these time points were used to form the count. Without that definition, it is not possible to distinguish an event occurring at baseline from an event arising during follow-up, or to determine how repeated reports were handled. The record also does not report participant-level outcome data or an analysis population definition.
The appropriate reading is therefore narrow: the posted results show equal observed counts for the defined symptom outcome and an estimated risk difference of 0.0, with substantial uncertainty around the comparison. They do not establish equivalence, and they do not answer whether montelukast changes cognitive or Alzheimer’s disease outcomes. No treatment recommendation follows from this symptom result.
9. Statistical Methods Explained
What does a risk difference describe?
A risk difference compares the proportion of participants with an event in one group with the corresponding proportion in another. It is an absolute measure: it describes the gap between event proportions rather than expressing one group’s event rate as a multiple of the other. In this trial, the event is having any reported gastrointestinal symptom as defined in the registry. The reported estimate is 0.0, reflecting equal observed proportions in the two arms.
Absolute measures are often useful for clinical interpretation because the unit is a difference in event probability. But a point estimate is only one summary of the data. Its precision depends on the amount of information in the outcome, including the numbers of participants and events. Sparse event data can leave considerable uncertainty even when the observed proportions match.
What does the confidence interval add?
A confidence interval gives a range of values compatible with the data under the method used to calculate it. Its endpoints represent uncertainty around the estimated risk difference; they are not a range of participant-level results. The reported interval includes the null difference and extends to both negative and positive values. Thus it does not rule out a difference favoring either group. A confidence interval is not a guarantee that the true value lies within its endpoints, and it should be read together with the estimate, design and data quality.
The registry reports the interval and an unadjusted p-value, but it does not report the named statistical test, the confidence-interval method, or any continuity correction or exact method. Those details matter particularly with sparse binary outcomes, where different reasonable methods can yield different interval limits or p-values. The precise computational procedure is not reported in the registry record.
How should the p-value be understood?
The p-value evaluates the observed data against a null hypothesis within a specified statistical model and test. It is not a measure of effect size and does not quantify the probability that a treatment is effective, ineffective or equivalent to placebo. The posted unadjusted p-value is 1.0. This should be read as a test result for this comparison, not as proof that the groups have identical underlying risks.
“Unadjusted” indicates that the reported p-value is not presented as adjusted for other variables or for testing multiple outcomes. The record does not report a covariate-adjusted model for this primary comparison, and it does not provide a detailed statistical analysis plan. It is therefore not possible to determine whether the posted analysis exactly follows a prespecified analysis or whether additional analyses were conducted outside the registry entry.
Why does sparse data matter?
When event counts are small, each counted event can have a substantial influence on an estimated proportion. Statistical procedures developed for larger counts may rely on approximations that perform differently in sparse settings. Exact methods can be considered in general, but the registry record does not identify which method was used. Readers should focus on the uncertainty rather than treating the point estimate as a stable estimate of the underlying risk difference.
Relative measures such as risk ratios and odds ratios compare the groups on a multiplicative scale. They can be useful, but with sparse events their estimates may be highly uncertain. The record posts these measures and their intervals, but the primary interpretation is clearer on the absolute scale: equal observed proportions produce a risk-difference estimate of 0.0, while the uncertainty interval remains broad. No new relative or absolute effect calculation is needed to make that distinction.
What do the planning illustrations mean?
The accompanying sample-size illustrations are planning calculations based on the observed effect, not the trial’s own design. Their stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation; the effect to detect is defined as a stated fraction of the observed effect. Such calculations describe hypothetical sample-size requirements under chosen assumptions. They are not observed power, do not show the probability that this completed trial detected an effect, and do not establish that the original enrollment was adequate.
The registry record does not report the original sample-size justification, target power, assumed effect used in planning, or whether the trial’s design was changed. It also does not report interim analyses or a complete statistical plan. Accordingly, the illustrations should not be mistaken for the trial’s original protocol calculations or for evidence about what the trial was designed to detect.
10. Confidence Intervals and Precision
The risk-difference estimate is 0.0, but the estimate alone does not convey how precisely the underlying group difference has been determined. The reported confidence interval ranges across negative and positive values and includes the null difference. That means the observed data are consistent with a range of possible group contrasts, not only with an exact absence of difference.
The interval is broad in relation to the point estimate. In practical terms, the small number of recorded symptom events leaves limited information for estimating the difference between groups. Although each group has 16 participants in the posted result, only one participant in each group is counted as having symptoms. Sparse events make estimates sensitive to the observed counts and limit precision. Equal observed counts therefore should not be mistaken for strong evidence that the true event proportions are equal.
A confidence interval is not a prediction interval for future participants and does not describe the range of symptom experiences among individuals. It quantifies uncertainty in the estimated group contrast according to the method used. The registry record does not identify the interval method, so the exact inferential procedure cannot be assessed from the posted information alone.
The reported unadjusted p-value of 1.0 is consistent with the point estimate being at the null value, but it does not replace the interval as a description of uncertainty. Nor does it show that a clinically important difference is impossible. Whether a difference would be important depends on a prespecified clinical threshold, which is not reported in the registry record.
Precision also depends on what the outcome definition captures. The record specifies symptoms and a baseline-to-1-year timeframe, but it does not explain how reports at the two time points were combined. Uncertainty about this operational definition is distinct from sampling uncertainty and cannot be resolved by the confidence interval. The result should therefore be read as imprecise for the posted endpoint and limited to the way that endpoint was recorded.
11. Multiplicity, Missing Data and Analysis Populations
The registry lists 11 outcome measures and identifies four secondary outcomes: CSF Amyloid, CSF Tau Levels, Clinical Dementia Rating Score and NIH Toolbox Cognition Battery. Testing multiple outcomes can increase the chance of at least one apparently notable result arising by chance if each is assessed independently without a multiplicity strategy. Statistical approaches can address that issue, but the registry record does not report a multiplicity adjustment, a hierarchy for outcomes, or a comprehensive testing plan. The primary comparison’s p-value is explicitly unadjusted.
The record reports zero analyses in its analysis listing. That does not establish that no analyses were performed; it means the registry information available for this page does not list analysis records. The statistical plan, model specification, test selection, covariate handling, and whether outcomes were transformed or categorized are not reported in the registry record. It is also not reported whether the primary outcome definition or analysis was changed after trial initiation.
Missing-data handling is not described. The posted primary results give group sizes and symptom counts, but do not state whether every randomized participant contributed information at baseline and 1 year, how missing symptom reports were handled, or whether any imputation or sensitivity analysis was used. Participant flow, discontinuations, and reasons for missing observations are not reported. Without these details, the possibility and consequences of missing outcome data cannot be evaluated.
The analysis population is likewise not defined in the registry record. It does not say whether the result follows an intention-to-treat approach, includes only participants who completed treatment, or uses another population. The posted group counts should not be used to infer adherence, completion or the handling of participants with incomplete follow-up.
These reporting gaps do not change the arithmetic of the posted comparison, but they limit assessment of how robustly it represents the randomized groups and how it fits within the broader set of trial outcomes. They also prevent a full evaluation of selective reporting and of whether the primary analysis was specified before results were known.
12. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Montelukast Group | 0 | 16 | 0.0% |
| Placebo Group | 0 | 16 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Montelukast Group | 0 of 16 |
| Placebo Group | 0 of 16 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Montelukast Group | 3 of 16 |
| Placebo Group | 2 of 16 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Montelukast Group | Placebo Group |
|---|---|---|---|
| Nausea | Gastrointestinal disorders | 1 of 16 | 0 of 16 |
| Vomiting | Gastrointestinal disorders | 0 of 16 | 1 of 16 |
| Rash | General disorders | 1 of 16 | 0 of 16 |
| Dizziness | General disorders | 1 of 16 | 0 of 16 |
| Headache | General disorders | 0 of 16 | 1 of 16 |
| Fatigue | General disorders | 0 of 16 | 1 of 16 |
| Low Mood | General disorders | 0 of 16 | 1 of 16 |
The registry’s serious-adverse-event figures list zero affected participants in the Montelukast Group and zero in the Placebo Group. The listed deaths are also zero in both groups. These are the posted counts for the safety fields; they should not be read as proof that no adverse experiences occurred or that the groups have identical safety profiles.
The record also gives other adverse-event figures of 3 for the Montelukast Group and 2 for the Placebo Group. These are distinct from the serious-adverse-event figures. The registry does not provide the specific terms in a serious-event list, and the top serious terms list is empty. The record’s empty list should be interpreted as no serious terms listed in that field, not as a complete description of all safety monitoring or reporting procedures.
Safety counts depend on the definitions, observation period and reporting process used. The registry information does not explain those details, give event-level descriptions for the other adverse events, or state how many participants had multiple events. It also does not report a statistical comparison for these safety figures. No inference about comparative safety should be drawn from these counts alone.
13. What the Result Does, and Does Not, Mean
The posted result answers a narrow descriptive question: how many participants were recorded as having any of the specified gastrointestinal symptoms in each group, with the outcome timeframe listed as baseline and 1 year. One participant is counted in each group. The estimated risk difference is 0.0, and the reported unadjusted p-value is 1.0. The confidence interval nevertheless permits differences in either direction, reflecting limited precision.
This result does not establish equivalence between montelukast and placebo. An equivalence conclusion requires a prespecified margin and an analysis designed to assess whether the uncertainty falls within that margin; neither is reported in the registry record. A non-significant p-value is not a substitute for such evidence. The result also does not prove that there is no difference in the true risk of gastrointestinal symptoms.
The primary outcome is not a measure of cognitive change, disease progression or brain biomarkers. The trial summary describes those broader areas of interest, and the registry lists related secondary outcomes, but the result cannot be generalized to them. Results for those outcomes are not reported in the registry record used for this analysis.
The one-year description and the baseline, 1-year timeframe do not reveal how symptoms were combined across visits. The available information does not establish whether the count refers to any symptom at either visit, a new symptom after baseline, or a particular follow-up assessment. It also does not characterize severity or duration. These are limits on what can be inferred from the endpoint, separate from the statistical uncertainty around the group comparison.
The most defensible conclusion is that the observed counts were equal for this posted gastrointestinal symptom endpoint, while the data provide limited precision about the underlying group difference. This finding is not a broad judgment about the treatment or its effects on Alzheimer’s disease. It should be interpreted alongside the endpoint definition, incomplete statistical reporting and the other evidence relevant to the trial’s broader aims.
14. Why This Trial Matters Statistically
This trial illustrates why a point estimate and its uncertainty need to be read together. The estimated risk difference is 0.0 because the posted symptom counts match, but the confidence interval includes contrasts in both directions. A simple statement that the groups had the same observed count would be accurate; a statement that they have the same underlying risk would go beyond the evidence.
The endpoint also shows how outcome definition shapes interpretation. The registry names diarrhea, nausea and vomiting and gives a timeframe of baseline and 1 year, but does not explain exactly how reports across those time points were converted into the participant count. This missing detail matters because a count can represent different clinical and statistical quantities depending on its rules for timing, recurrence and baseline status.
The sparse event data make absolute estimates and their uncertainty especially important. Relative measures can be difficult to interpret when event counts are small, and neither a risk ratio nor an odds ratio can remove the underlying lack of information. The absolute risk-difference estimate gives a direct description of the observed contrast, while the interval signals that the estimate remains imprecise.
Finally, the registry record lists several outcome measures but does not document a full analysis plan or multiplicity approach. This is a reminder that interpretation concerns not only a numerical result but also the analysis context: which outcomes were prioritized, how missing observations were handled, and how tests were selected. Those details are not reported in the registry record, so the primary result is informative as a posted comparison but does not support a comprehensive evaluation of the trial’s statistical strategy.
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 main trial-specific limitation is the sparse primary outcome: one participant with a reported gastrointestinal symptom in each arm. With so few events, the risk-difference estimate has limited precision, and the posted confidence interval includes differences in either direction. The matched observed counts cannot establish equivalence or rule out a meaningful difference.
The outcome timeframe is listed as baseline and 1 year, but the registry record does not explain whether the endpoint counts symptoms at either visit, new symptoms during follow-up, or symptoms at a particular assessment. It does not give separate counts by symptom, severity, duration or time point. This makes the clinical meaning of the combined endpoint less specific than its symptom list might suggest.
Important design and analysis details are not reported in the registry record. These include the randomization procedure, allocation concealment, masking assessment, adherence, participant flow, missing-data handling, analysis population, exact statistical test and confidence-interval method. The record also does not provide an original sample-size justification, interim-analysis information, or a complete statistical analysis plan. As a result, it is not possible to assess whether the posted analysis followed a prespecified strategy or how departures and incomplete observations were handled.
The trial summary describes cognitive, CSF and neuroimaging aims, and the registry lists secondary outcomes, but corresponding results are not reported in the information used. The gastrointestinal result therefore cannot stand in for results on those other measures. The listed safety counts are also limited: event details and a comparative safety analysis are not reported. These constraints call for a narrow reading of the primary result rather than a broad conclusion about treatment effects or safety.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: Montelukast Therapy on Alzheimer's Disease — NCT03991988.
- Registry record last updated 2024-03-07; status last verified 2024-02.
- Results first posted 2024-03-07.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
19. Record Summary
EMERALD is a completed, randomized, double-blind, parallel-group phase 2 trial sponsored by Emory University. The registry lists an enrollment of 32 and describes a one-year comparison of escalating-dose montelukast with matched placebo in individuals with mild cognitive impairment and early Alzheimer’s disease dementia. The recorded start date is 2019-09-25, and primary completion is recorded as 2022-11-18.
The primary outcome is the number of participants with any gastrointestinal symptoms—diarrhea, nausea or vomiting—at baseline and 1 year. The posted results count one participant in each group, with 16 participants in each group, and report a risk difference of 0.0 with an unadjusted p-value of 1.0. The confidence interval includes differences in either direction, so the comparison is imprecise and does not establish equivalence.
The registry lists 11 outcome measures and four secondary outcomes. Detailed analysis methods, missing-data procedures, analysis populations and results for the other outcomes are not reported in the registry record.