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 Epanova® for Lowering Very High Triglycerides II (EVOLVE II) (NCT02009865), a randomized phase 3 study. The primary outcome is Percent Change in Triglyceride for All Subjects.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02009865 |
| Phase | Phase 3 |
| Condition | Hypertriglyceridemia |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Enrollment | 379 |
| Arms in the primary outcome | 2 |
| Lead sponsor | AstraZeneca |
| Results posted | Yes |
2. Clinical Question
Population
Subjects with severe hypertriglyceridemia were enrolled. The primary analysis population was the Full Analysis Set (FAS). The registry record does not report a fuller definition of that population or the criteria used to enter the primary analysis.
Intervention
Participants assigned to the Epanova arm received Epanova 2 g/day. The study investigated it as an adjunct to diet; the record describes a screening period that included diet and lifestyle stabilization and a washout period.
Comparator
The comparison arm received Olive Oil 2 g/day. The study was randomized, double-blind, and olive oil-controlled. The registry record does not describe further details of how allocation was concealed or how masking was maintained.
Primary question
The primary endpoint was percent change in triglycerides for all subjects from baseline to the Week 12 endpoint. The statistical comparison asks whether the distributions of change differed between the Epanova and olive oil arms under the reported analysis. The registry record does not give a threshold for a clinically important difference.
The question is about a measured lipid outcome over the treatment period, not about longer-term outcomes. The primary endpoint is distinct from the listed secondary outcomes, which include triglyceride change in subjects with at least one qualifying triglyceride >885 mg/dL, non-high-density lipoprotein cholesterol, high-density lipoprotein cholesterol, and triglyceride change among subjects with biochemically defined Fredrickson Type V. These outcomes address additional features of the lipid profile and participant subsets, but the main comparison is the all-subject triglyceride endpoint.
The study record describes Epanova as an adjunct to diet. It therefore frames the comparison as Epanova plus the study diet context versus olive oil plus that context, rather than as a comparison of diet with no diet. Details of other care or restrictions are not reported in the registry record.
3. Trial Design
EVOLVE II was a Phase 3 randomized, double-blind, parallel-group study with an olive oil control. The registry record lists 379 participants enrolled. The study description gives an approximately 8-week screening period, including diet and lifestyle stabilization and a washout period, followed by a 12-week treatment period. The primary endpoint was evaluated from baseline to the Week 12 endpoint.
The two listed treatment arms were Epanova 2 g/day and Olive Oil 2 g/day. For the primary endpoint, the posted results give 81 participants in each arm. These arm counts are the numbers reported for this endpoint and should not be treated as the total enrollment. The record does not explain the difference between overall enrollment and the endpoint analysis counts.
Randomization and double masking can reduce some sources of bias in a treatment comparison when they are implemented successfully. The registry record does not report the randomization sequence method, allocation concealment procedures, who was masked, or whether masking was assessed. It also does not report protocol details about adherence, treatment discontinuation, or how participants were followed after discontinuation.
The registry lists triglyceride change as the primary outcome and four secondary outcomes. The primary outcome was tested in parallel with the first secondary outcome, each at a 0.025 Type I error rate. The record does not report a broader testing hierarchy or how the remaining secondary outcomes were treated for multiplicity.
Enrollment was 379 and the primary endpoint analysis lists 81 participants in each arm. The registry record does not explain how many randomized participants contributed to each analysis, how the analysis population was constructed, or why participants may not appear in the primary endpoint counts. It does not report a sample-size justification for the original trial. Those details matter when judging how well the analysis population reflects the randomized groups and how precisely the trial was designed to estimate an effect.
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.
Before assignment: In total, 379 subjects were screened from the 43 sites.
Overall Study
| Stage | Epanova | Olive Oil |
|---|---|---|
| STARTED | 81 | 81 |
| COMPLETED | 80 | 76 |
| NOT COMPLETED | 1 | 5 |
Reasons for not completing this period:
| Reason | Epanova | Olive Oil |
|---|---|---|
| Adverse Event | 0 | 2 |
| Protocol Violation | 0 | 2 |
| Withdrawal by Subject | 1 | 1 |
Analysis population for the primary outcome, as defined in the registry: Full Analysis Set (FAS)
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 | Epanova | Olive Oil | Total |
|---|---|---|---|---|
| Participants analysed | 81 | 81 | 162 | |
| Age, Continuous | Years | 50.3 (SD 10.60) | 50.0 (SD 10.88) | 50.2 (SD 10.71) |
| Sex: Female, Male: Female | Participants | 16 | 19 | 35 |
| Sex: Female, Male: Male | Participants | 65 | 62 | 127 |
| Lipid-altering drug use: Yes | Participants | 36 | 37 | 73 |
| Lipid-altering drug use: No | Participants | 45 | 44 | 89 |
| At least 1 qualifying TG category > 885 mg/dL: Yes | Participants | 53 | 52 | 105 |
| At least 1 qualifying TG category > 885 mg/dL: No | Participants | 28 | 29 | 57 |
6. Primary Endpoint
The registry lists the primary outcome as Percent Change in Triglyceride for All Subjects, measured over From Baseline to Week 12 Endpoint.
| Element | Registry entry |
|---|---|
| Result type | Median |
| Measure of spread | Inter-Quartile Range |
| Unit | Percentage of change (%) |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is percent change in triglycerides from baseline to the Week 12 endpoint for all subjects. It describes the relative change in a laboratory measurement over the treatment period, rather than the triglyceride level at a single time alone. The registry identifies the endpoint unit as percentage of change and reports the arm summaries as medians with interquartile ranges.
A change endpoint can help describe how measurements shift from each participant’s starting value. Expressing change as a percentage is one way to put changes in context of baseline levels. The exact calculation used to derive the percentage change, including any handling of unusual or unavailable measurements, is not reported in the registry record.
Triglycerides are the target measure named in the study objective, which describes investigating Epanova as an adjunct to diet for reduction of triglyceride levels in subjects with severe hypertriglyceridemia. The endpoint therefore directly reflects the measured outcome that motivated the trial. The record does not report a threshold that would distinguish a clinically important change from a smaller change, nor does it connect this laboratory endpoint to longer-term outcomes.
The primary outcome is designated for all subjects in the FAS. The registry record does not define the FAS in detail. Consequently, the endpoint description identifies the intended measurement and time point, but does not by itself explain which participants or measurements were included in the analysis.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
For the primary endpoint, the reported median percent change was -28.1 in the Epanova arm and -10.2 in the olive oil arm. The reported interquartile ranges were -42.1 to -5.5 and -36.4 to 23.0, respectively. These summaries describe the observed distributions in the analyzed participants; they do not state how every participant changed.
The reported comparison was a Wilcoxon (Mann-Whitney) analysis. The registry gives a median difference (final values) of -14.2, with a 95% confidence interval from -26.2 to -2.8, and p = 0.017. The reported groups are Epanova and Olive Oil, in that order. The estimate is negative, and the interval is also entirely negative. The reported p-value is below the stated 0.025 Type I error rate for this primary outcome.
These are statistical summaries of the posted primary endpoint. The confidence interval communicates uncertainty around the reported difference under the analysis used; it does not show the range of individual participant responses. The p-value describes how compatible the observed data are with a null comparison under the test’s assumptions. Neither quantity, by itself, establishes the clinical importance of the difference. The registry record does not report a threshold for clinical importance.
The arm summaries are medians, while the comparison is identified as a Wilcoxon (Mann-Whitney) test. A rank-based comparison evaluates the relative ordering of observations between groups. Its result is not automatically a test of the difference between the two reported medians alone. The reported difference is labeled “Median Difference (Final Values),” but the registry record does not provide further detail on how that estimate or its confidence interval was constructed.
The endpoint analysis lists 81 participants in each arm, whereas enrollment was 379. Without an account of participant flow, missing measurements, or the exact FAS definition, it is not possible to determine from the registry record why the reported analysis counts differ from enrollment. This limits assessment of whether missing or excluded observations could affect the comparison.
The registry states that the primary endpoint and the first secondary endpoint were tested in parallel, each at a 0.025 Type I error rate. That context matters when interpreting the primary p-value: the stated criterion was not simply 0.05. The record does not describe the broader multiplicity approach for all secondary outcomes, so the primary finding should not be treated as a complete account of every outcome tested.
Overall, the posted data describe a more negative median percent change in the Epanova arm than in the olive oil arm, alongside a negative reported comparison estimate and a p-value below the stated primary testing threshold. Interpretation should remain limited to this endpoint, treatment period, analysis population as recorded, and statistical method. The data do not establish effects on outcomes not measured or reported in this trial record.
9. Statistical Methods Explained
The primary endpoint is a continuous measurement expressed as percent change in triglycerides from baseline to the Week 12 endpoint. The registry reports the arm summaries as medians and interquartile ranges. A median is the center of an ordered set of observations: half lie at or below it and half at or above it. An interquartile range describes the middle portion of the observations, rather than the uncertainty around the median or the full spread of individual values.
Why use a rank-based comparison?
The reported method is the Wilcoxon (Mann-Whitney) test. In general, this test compares the ranks of observations across groups instead of relying directly on a comparison of their arithmetic means. Rank-based methods can be useful when outcomes are skewed or when extreme observations make a mean-based model less suitable. This is a general explanation of the method, not evidence that a particular distributional problem was found in this trial.
A Wilcoxon (Mann-Whitney) result is not necessarily equivalent to a test that compares medians. Under certain distributional patterns it can be interpreted in terms of relative location, but it can also respond to differences in distributional shape. The registry reports medians and a median difference, but does not provide details of the model assumptions or the exact procedure used to derive the reported difference and interval. Those aspects are not reported in the registry record.
How to read the estimate and interval
The comparison is labeled “Median Difference (Final Values)” and is reported as -14.2, with a 95% confidence interval from -26.2 to -2.8. In general, a confidence interval gives a range of values compatible with the data and the interval method, while reflecting sampling uncertainty. It is not a range that contains a specified proportion of individual participants’ outcomes. The precise coverage properties depend on how the interval was calculated; the registry record does not describe that calculation.
The confidence interval is a 95% interval as reported. The primary endpoint’s stated Type I error rate is 0.025, and the record says it was tested in parallel with the first secondary endpoint at that rate. The registry does not explain whether the reported confidence interval was adjusted for this testing arrangement. Thus the interval and the testing threshold should be read as distinct reported elements, not assumed to be outputs of an identical multiplicity adjustment.
What the p-value says
The reported p-value is 0.017. In general, a p-value describes the probability, under a specified null hypothesis and statistical model, of obtaining data at least as inconsistent with that null as the observed data. It is not the probability that the null hypothesis is true, nor does it measure the size or importance of an effect. The stated threshold for the primary endpoint is 0.025; the record does not give the null hypothesis in formal detail.
The arm medians and interquartile ranges provide descriptive context, while the rank-based test provides the reported inferential comparison. Both are useful, but they answer different questions. Descriptive summaries show the observed distribution in each arm; the test and estimate address a between-arm comparison under the chosen method.
What cannot be recovered from the record
The registry identifies the primary population as the Full Analysis Set, but its inclusion rules are not reported in the registry record. The record also does not describe handling of missing baseline or endpoint measurements, transformations, covariate adjustment, outlier rules, or sensitivity analyses. It does not state whether the test or interval was adjusted for baseline values. These details could affect interpretation and reproducibility, so they should not be inferred from the reported method name alone.
Sample-size calculations based on an observed effect are planning illustrations, not observed power. In general, such illustrations estimate the number of participants that might be needed under stated assumptions to detect a chosen effect. They do not establish the probability that this completed trial detected an effect, and they do not replace the original design justification. The original sample-size justification is not reported in the registry record.
Over-dispersion is a concern in some count-data models when observed variability exceeds what the model assumes. The primary endpoint is reported as a percentage change and analyzed with a rank-based test, not a count-model analysis. The registry record does not report an over-dispersion assessment, and that concept should not be imposed on this analysis without evidence that a relevant count model was used.
The main methodological reading is therefore bounded: the registry provides arm medians, interquartile ranges, a Wilcoxon (Mann-Whitney) comparison, a difference estimate, a confidence interval, and a p-value. It does not provide enough detail to independently reconstruct all analysis decisions or assess every assumption behind the inferential results.
10. Analyses Reported by the Sponsor
| Method | Estimate | 95% CI | P-value |
|---|---|---|---|
| Wilcoxon (Mann-Whitney) | Median Difference (Final Values) -14.2 | -26.2 to -2.8 | 0.017 |
The registry reports a Wilcoxon (Mann-Whitney) analysis comparing Epanova with Olive Oil for the primary endpoint. Its reported median difference (final values) is -14.2, the 95% confidence interval is -26.2 to -2.8, and p = 0.017. These are the posted analysis results and should be distinguished from any simple recalculation based only on summary statistics.
The registry record does not report a separate recalculated estimate for this comparison or give enough detail to reproduce the sponsor’s analysis from the summary values alone. It also does not describe the construction of the confidence interval. The record does not identify a discrepancy between the reported result and a recalculation, but a numerical comparison cannot be described without a separate recalculated value.
The reported test and estimate remain the main inferential evidence available in the registry record. A recalculation from summary statistics, if considered, would be unadjusted and would not necessarily reproduce a rank-based analysis of participant-level data. The distinction matters because the posted result comes from the recorded analysis method, whereas summary-statistic calculations can rely on different information and assumptions.
11. Multiplicity, Missing Data and Analysis Populations
The registry states that the primary endpoint was tested in parallel with the first secondary endpoint, each at a 0.025 Type I error rate. This is the reported multiplicity information for the primary comparison. It does not establish how the remaining secondary outcomes were handled, and a broader testing sequence or adjustment is not reported in the registry record.
Four secondary outcomes are listed: triglyceride percent change for subjects with at least one qualifying triglyceride >885 mg/dL; percent change in non-high-density lipoprotein cholesterol; percent change in high-density lipoprotein cholesterol; and triglyceride change in subjects with biochemically defined Fredrickson Type V, defined in the record by a triglyceride/very-low-density lipoprotein cholesterol ratio at least 6. The record indicates five outcome measures and five analyses, but it does not provide analysis results for every listed outcome. Values are missing or not estimable for a two-arm comparison for another endpoint.
For the primary endpoint, the stated analysis population is the Full Analysis Set. The registry record does not define who qualified for that set, whether participants were analyzed according to randomized assignment, or how post-randomization events were handled. Those choices are important for understanding the population represented by the result.
Missing-data methods are not reported in the registry record. It does not say how missing endpoint values were handled, whether any values were imputed, or whether sensitivity analyses examined alternative assumptions. The endpoint arm counts are 81 in each arm, while the overall enrollment is 379; participant flow and reasons for the difference are not reported.
As a result, interpretation can recognize the stated primary testing threshold and population label, but cannot establish the full family-wise error approach, the extent or pattern of missingness, or the robustness of results to missing-data assumptions. These are limitations of the registry account, not evidence that a particular method was or was not used in the full trial analysis.
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 |
|---|---|---|---|
| Epanova | 1 | 81 | 1.2% |
| Olive Oil | 2 | 81 | 2.5% |
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 |
|---|---|
| Epanova | 29 of 81 |
| Olive Oil | 28 of 81 |
Most frequent serious adverse event terms
| Event term | Organ system | Epanova | Olive Oil |
|---|---|---|---|
| Rectal haemorrhage | Gastrointestinal disorders | 0 of 81 | 1 of 81 |
| Pancreatitis | Gastrointestinal disorders | 1 of 81 | 0 of 81 |
| Clavicle fracture | Injury, poisoning and procedural complications | 0 of 81 | 1 of 81 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Epanova | Olive Oil |
|---|---|---|---|
| All | Gastrointestinal disorders | 12 of 81 | 7 of 81 |
| ALL | Infections and infestations | 10 of 81 | 6 of 81 |
| Diarrhoea | Gastrointestinal disorders | 5 of 81 | 5 of 81 |
| ALL | Nervous system disorders | 5 of 81 | 3 of 81 |
| ALL | Investigations | 4 of 81 | 3 of 81 |
| ALL | Metabolism and nutrition disorders | 1 of 81 | 5 of 81 |
| ALL | Respiratory, thoracic and mediastinal disorders | 3 of 81 | 2 of 81 |
| ALL | General disorders | 3 of 81 | 2 of 81 |
| ALL | Musculoskeletal and connective tissue disorders | 0 of 81 | 5 of 81 |
| Nasopharyngitis | Infections and infestations | 3 of 81 | 1 of 81 |
Serious events with the highest arm-level rate: rectal haemorrhage (1.2%), pancreatitis (1.2%), clavicle fracture (1.2%).
The registry reports serious adverse events for 1 of 81 participants at risk in the Epanova arm and 2 of 81 at risk in the Olive Oil arm. These are counts during the reported safety observation, not estimates of comparative risk adjusted for follow-up or other factors. The record does not provide event timing, duration of observation for each participant, or further details about attribution.
The listed serious adverse event terms are Rectal haemorrhage, Pancreatitis, and Clavicle fracture, each reported as 1.2%. The registry listing does not provide additional context about severity, recurrence, relationship to treatment, or which treatment arm contributed each term. The figures should not be read as a complete description of all safety outcomes.
Other adverse events are reported for 29 of 81 participants at risk in the Epanova arm and 28 of 81 at risk in the Olive Oil arm. The registry record does not give event-level detail for those counts in the registry record. Death counts are not reported in the registry record. These safety summaries describe what is posted, but do not by themselves establish whether one treatment caused an event or whether the arms differed meaningfully in safety.
13. What the Result Does, and Does Not, Mean
The primary result describes triglyceride percent change over a 12-week treatment period among participants included in the reported FAS analysis. The arm medians and the reported comparison indicate that the observed distributions differed, with the test result meeting the stated 0.025 Type I error threshold. The confidence interval gives an indication of uncertainty for the reported estimate, subject to the interval method and the analysis details that are not available in the registry record.
This result does not establish the clinical importance of the difference. The registry record does not define a clinically meaningful threshold for triglyceride change. It also does not show whether the measured change led to differences in symptoms, complications, quality of life, or longer-term outcomes. Those outcomes cannot be inferred from this laboratory endpoint.
The result does not describe every participant’s response. Medians and interquartile ranges summarize distributions, and the group comparison does not show the degree of individual variation or identify who changed in a particular way. The reported analysis counts are smaller than total enrollment, and the registry does not explain the participant flow or missing-data handling. That uncertainty limits how confidently the analysis can be generalized to all enrolled participants.
The study compared Epanova 2 g/day with Olive Oil 2 g/day in the context of diet, over the specified treatment period. It does not compare Epanova with every possible alternative, nor does it answer questions about treatment duration beyond the study period. Generalization beyond the studied population and conditions requires evidence not reported in this record.
The appropriate statistical conclusion is therefore narrow: the posted primary analysis reports a difference between the arms on the stated triglyceride endpoint, with the estimate, confidence interval, and p-value in the record. It does not, on its own, settle questions of clinical benefit, long-term outcomes, or the complete safety profile.
14. Why This Trial Matters Statistically
This trial illustrates how a clinically familiar endpoint can still require careful statistical reading. The primary outcome is percent change in triglycerides, but the registry reports medians and interquartile ranges and uses a Wilcoxon (Mann-Whitney) test. Readers need to distinguish the descriptive summaries from the inferential comparison and avoid treating a rank-based test as a simple test of median values without qualification.
The testing threshold also matters. The registry says the primary endpoint and first secondary endpoint were tested in parallel, each at 0.025. Interpreting the primary p-value against that stated threshold is more appropriate than relying on a conventional threshold without considering the specified testing arrangement. At the same time, the registry does not describe the multiplicity strategy across all secondary outcomes, so the available account has limits.
The record also demonstrates why participant counts and analysis-population definitions are central to interpretation. The primary endpoint counts are 81 per arm, while overall enrollment is 379. The registry identifies the Full Analysis Set but does not define it or describe participant flow. Without that context, readers cannot fully assess how the analyzed group relates to all enrolled participants.
Finally, the data concern a specific laboratory endpoint measured from baseline to the Week 12 endpoint. Statistical evidence about that endpoint should not be mistaken for evidence about outcomes that were not reported. The registry gives enough information to understand the reported comparison at a high level, while leaving important details of missing-data handling, interval construction, and the full analysis plan unreported.
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 primary endpoint analysis lists 81 participants in each arm, but overall enrollment is 379. The registry record does not provide a participant flow account explaining the difference, reasons for exclusion, or numbers with missing baseline or Week 12 values. It is therefore not possible to assess from this record whether the analyzed participants differ from those not represented in the endpoint summaries.
The Full Analysis Set is named but not defined. The record does not state how participants were assigned to that population, how discontinuations or protocol deviations were handled, or whether participants were analyzed according to randomized assignment. Missing-data methods, imputation, and sensitivity analyses are also not reported in the registry record.
The statistical description is limited. The test is identified as Wilcoxon (Mann-Whitney), and the reported estimate is labeled “Median Difference (Final Values),” but the confidence-interval method and detailed analysis specifications are not reported. The record does not describe baseline adjustment, covariates, transformations, or checks of assumptions. The reported 95% confidence interval is not described as adjusted for the parallel primary and first-secondary testing.
The registry gives a short treatment period and does not report longer-term follow-up results in the registry record presented. It also does not provide a clinical-importance threshold for the primary endpoint. The listed safety counts lack details about event timing, attribution, and duration of observation, and death counts are not reported in the registry record.
Finally, the registry outcome list includes secondary measures, but not all have posted two-arm comparison values. This limits a broader account of the trial’s outcomes. These gaps do not show that the full trial lacked such procedures or analyses; they mean those details are not reported in the registry record.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: Epanova® for Lowering Very High Triglycerides II (EVOLVE II) — NCT02009865.
- Registry record last updated 2019-09-10; status last verified 2019-08.
- Results first posted 2016-01-15.
- 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
EVOLVE II (NCT02009865) was a completed Phase 3, randomized, double-blind, parallel-group study of Epanova and Olive Oil in subjects with severe hypertriglyceridemia. The study description includes an approximately 8-week screening period with diet and lifestyle stabilization and washout, followed by a 12-week treatment period.
The primary endpoint was percent change in triglycerides from baseline to the Week 12 endpoint in the Full Analysis Set. The posted summaries report medians of -28.1 for Epanova and -10.2 for Olive Oil. The reported Wilcoxon (Mann-Whitney) comparison gives a median difference of -14.2, a 95% confidence interval from -26.2 to -2.8, and p = 0.017. The registry states that the primary endpoint and first secondary endpoint were tested in parallel at a 0.025 Type I error rate.
Important details—including the FAS definition, participant flow, missing-data handling, and the broader multiplicity approach—are not reported in the registry record. Serious adverse event counts were 1 of 81 and 2 of 81, respectively; death counts are not reported in the registry record.