A Study of Tirzepatide (LY3298176) in Healthy Participants and Participants With Type 2 Diabetes (T2DM): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Tirzepatide (LY3298176) in Healthy Participants and Participants With Type 2 Diabetes (T2DM) (NCT02759107), a randomized phase 1 study. The primary outcome is Number of Participants With One or More Serious Adverse Event(s) (SAEs).

2. Clinical Question

This was an early-phase study with stated purposes that included assessing safety, tolerability, pharmacokinetics, and pharmacodynamics. The primary endpoint is specifically a count of participants with investigator-attributed serious adverse events, rather than a measure of all side effects. Other serious and non-serious adverse events, regardless of causality, were reported separately in the adverse-events module. These different definitions matter: the primary endpoint does not describe every event experienced during the trial, and the broader adverse-event counts do not by themselves establish that study drug caused an event.

3. Trial Design

The registry identifies NCT02759107 as a completed phase 1, randomized, double-masked, parallel study sponsored by Eli Lilly and Company. It enrolled 142 participants and included people described as healthy as well as participants with type 2 diabetes mellitus. The study had three parts, and each participant enrolled in only one part. The registry reports a start date of 2016-05-11 and a primary completion date of 2017-06-26.

Part A examined single subcutaneous doses. Tirzepatide arms used escalating dose levels, while a separate arm received a single placebo dose. Parts B and C examined four weekly subcutaneous doses. Their tirzepatide arms included fixed-dose and titrated-dose regimens, and each part included placebo; Part B also included dulaglutide. As a result, the trial is not a single comparison of one tirzepatide regimen against one common control. The treatment regimens and group sizes varied across parts.

Randomization and double masking are important design features, but the registry record does not report the randomization procedure, masking details, allocation ratio, or whether masking applied identically across every treatment group. It also does not report a sample-size justification. The brief summary describes approximate study durations for the parts, but the primary endpoint’s stated observation periods are baseline through Day 43 for Part A and Day 57 for Parts B and C.

Results for the primary endpoint are counts of participants with at least one qualifying serious adverse event, accompanied by group sizes. The record also lists five secondary outcomes, covering pharmacokinetic area under the concentration-versus-time curve and glucose area-under-the-curve ratios in Part C. No inferential analyses are posted in the registry record. Therefore, the posted primary results are descriptive group counts and calculated risk-difference comparisons, not a reported full analysis of the trial’s objectives.

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.

Only the first 6 of 18 groups are shown.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants who received at least one dose of study drug.

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.

Population: All randomized participants.

Only the first 6 of 19 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With One or More Serious Adverse Event(s) (SAEs), measured over Baseline through Day 43 (Part A) and Day 57 (Part B and C).

The primary endpoint is the number of participants with one or more serious adverse events that the investigator considered related to study-drug administration. It was assessed from baseline through Day 43 in Part A and Day 57 in Parts B and C. The endpoint is expressed as a participant count, so a person with multiple qualifying events is still counted once for this measure.

In general, serious adverse events are used in clinical research to capture events considered serious under the study’s reporting definitions, while attribution asks whether the investigator judged the event related to study drug. These are distinct aspects of the endpoint: an event’s seriousness does not alone establish a causal relationship. The registry also reports serious and other non-serious adverse events regardless of causality in a separate module, which gives a broader account than this primary measure.

A participant-level count is useful for describing how many people experienced at least one qualifying event during the specified follow-up. It is not a measure of event severity beyond the serious-event classification, the number of recurrent events, or the full pattern of tolerability. Nor does it answer the pharmacokinetic, pharmacodynamic, or efficacy questions described in the study summary. The registry does not report further endpoint adjudication details or a statistical analysis plan for this outcome.

7. Results Posted in the Registry

• Tirzepatide (Part A): 0.0% (0/42) versus 0.0% (0/14) on control; risk difference 0.0 percentage points (95% CI 0.0 to 0.0); unadjusted two-proportion z-test p = n/a.
• Tirzepatide (Part B): 0.0% (0/25) versus 0.0% (0/14) on control; risk difference 0.0 percentage points (95% CI 0.0 to 0.0); unadjusted two-proportion z-test p = n/a.
• Tirzepatide (Part C): 0.0% (0/42) versus 0.0% (0/14) on control; risk difference 0.0 percentage points (95% CI 0.0 to 0.0); unadjusted two-proportion z-test p = n/a.
• Placebo (Part C): 0.0% (0/11) versus 0.0% (0/14) on control; risk difference 0.0 percentage points (95% CI 0.0 to 0.0); unadjusted two-proportion z-test p = n/a.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a binary participant-level event summarized as a count: whether a participant had one or more qualifying serious adverse events. The arm results pair that count with the number of participants in the arm. This is a natural descriptive presentation for an outcome defined by the presence or absence of at least one event. It does not retain information about the number of events per participant once the participant has met the endpoint definition.

What does the risk difference describe?

A risk difference compares the proportion of participants with the endpoint in one group with the corresponding proportion in another group. In general, it is expressed as the first group’s event proportion minus the comparator group’s event proportion. A value of 0.0 means the observed proportions are the same for that particular comparison. It is not a relative measure and does not describe how many events occurred overall. In this study, the posted risk-difference estimates are 0.0 for the reported tirzepatide-versus-placebo comparisons and the Part C placebo comparison. The associated reported limits are 0.0 to 0.0.

Those comparisons must be read as part-specific. Part A used single doses, while Parts B and C used weekly dosing for four weeks, with different fixed or titrated regimens. The comparator group sizes also differ. A risk difference from one part is not automatically a summary of the other parts, and the results do not represent one pooled contrast across all participants. The Part B dulaglutide arm has one primary-endpoint event, but the posted comparison list does not include a risk-difference contrast for that arm.

How should confidence intervals be understood?

In general, a confidence interval describes uncertainty around an estimate under the assumptions of the method used to construct it. It does not guarantee that the true value lies within its limits, and it should be considered alongside the estimate, outcome frequency, and study design. Here, the reported risk-difference intervals have identical lower and upper limits of 0.0. With no events in either group in these comparisons, the displayed intervals do not express the uncertainty that sparse data ordinarily leave about the underlying event risk. The registry record does not report the interval-construction method, so the method behind these values cannot be assessed from the posted record.

What is not reported about the analysis?

The registry lists no posted analyses for the outcome. It does not report a statistical analysis plan, a test statistic, p-values, adjustment variables, or an inferential model. It also does not report whether any planned analyses were conducted outside the registry record. Accordingly, the risk differences and their limits can be described as posted results, but their calculation and inferential interpretation cannot be verified from the registry entry.

In general, randomized allocation can help reduce systematic differences between groups, but it does not remove chance variation, especially when groups are small and the endpoint is uncommon. Masking can reduce some forms of bias, but the registry’s label does not provide the operational details needed to assess how masking was maintained. The parallel design indicates that participants were assigned to treatment groups rather than crossing through treatments, but the record does not describe the allocation sequence or concealment process.

The study’s sample-size material is explicitly a planning illustration computed from the observed effect, using a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It is not the study’s own design justification, and the calculated sample sizes are not observed power. More generally, planning calculations based on an observed effect should not be read as evidence that the original trial had a particular probability of detecting a true effect. The registry record does not report the trial’s own sample-size rationale.

Because the endpoint is sparse and the registry posts no inferential analysis, the most defensible reading is descriptive: identify the number of participants meeting the endpoint in each arm, note the groups and follow-up period, and avoid treating zero observed events as proof of zero risk. A broader statistical assessment would require methods and details not reported in the registry record.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data identify the value of interest. For a rare or sparse endpoint, precision depends not only on the observed count but also on the group sizes and the method used to quantify uncertainty. In this trial, the primary-endpoint counts are zero in the tirzepatide and placebo arms and one in the Part B dulaglutide arm. Some groups, especially in Part B, include few participants. With so few observed events, the posted counts cannot characterize the frequency of uncommon outcomes with much assurance.

The registry’s risk-difference comparisons report estimates of 0.0, with lower and upper limits both 0.0. These displayed limits are therefore numerically narrow, but they should not be interpreted as evidence that the underlying risk difference is known without uncertainty. The registry record does not describe how the intervals were constructed. When both compared groups have zero observed events, the absence of observed events does not rule out a nonzero underlying risk in either group. The displayed interval also does not resolve uncertainty about an event that may occur rarely.

Precision is further limited by the trial’s structure. Parts A, B, and C had different dosing regimens and different follow-up periods, and each participant enrolled in one part. The group sizes vary, and the comparisons are not a single pooled estimate. The posted results therefore offer limited precision for each specific arm and period and do not establish the precision of a general treatment effect across the study.

The record does not provide inferential analyses, p-values, or a stated method for confidence-interval construction. Consequently, the interval values should be read as the registry’s posted comparison output rather than as fully documented uncertainty estimates. The data are most precise about what was actually recorded in these groups during the specified follow-up; they are much less informative about the frequency of rare events outside the observed participants and periods.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists six outcome measures, including the primary serious-adverse-event endpoint and five secondary outcomes. The secondary outcomes concern pharmacokinetic area under the concentration-versus-time curve in Parts A, B, and C, and glucose area-under-the-curve ratios in Part C. The record does not report a multiplicity strategy, such as a hierarchy for testing outcomes or an adjustment for evaluating several outcomes. It also lists no posted analyses. Without a reported testing plan, it is not possible to determine whether or how multiplicity was handled in analyses outside the registry record.

The primary endpoint results are descriptive counts and risk differences. No p-values or formal hypothesis tests are posted for these results. Therefore, there is no reported family of inferential tests from which to assess the consequences of multiple comparisons. The different parts and regimens also create distinct comparisons; interpreting them as one broad set of confirmatory tests would go beyond what the registry reports.

The primary endpoint population is defined as all randomized participants who received at least one dose of study drug. This definition is not identical to every possible randomized analysis population, because it excludes randomized participants who did not receive a dose. The registry record does not report the number of randomized participants who did not receive treatment, nor does it describe an intention-to-treat analysis or other analysis populations.

Missing-data methods are not reported in the registry record. The entry does not state whether endpoint data were missing, how missing follow-up was handled, or whether any participants withdrew before the endpoint period ended. It also does not describe imputation, sensitivity analyses, or assumptions about missing outcomes. Thus, the posted counts can be read as the reported results for the stated population, but the completeness of endpoint ascertainment cannot be evaluated from the registry record alone.

In general, multiplicity and missing-data procedures are important because they affect how confidently results can be interpreted across multiple outcomes and incomplete observations. For this trial, the relevant plans and details are not reported in the registry record, so claims about adjustment or robustness are not supported by the posted information.

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: atrioventricular block second degree (25.0%).

The primary safety count includes participants with one or more serious adverse events that investigators considered related to study-drug administration. The registry reports zero such participants in the Part A tirzepatide and placebo arms, zero in the Part B tirzepatide and placebo arms, and zero in the Part C tirzepatide and placebo arms. The Part B dulaglutide arm reports one participant. No deaths are reported in the listed adverse-event groups.

These are counts during the stated endpoint periods, not estimates of long-term risk. The number of participants at risk differs by group, and some groups are small. A zero count means that no qualifying event was recorded in that group during the reported period; it does not show that such an event is impossible or that the underlying risk is zero.

The adverse-events module also reports serious and other non-serious events regardless of causality. Those broader counts answer a different question from the primary endpoint, which requires investigator attribution to study drug. The registry lists atrioventricular block second degree as a top serious-event term at 25.0%. That term listing should be considered alongside the record’s definitions and should not be assumed to represent the primary endpoint comparison or to establish causation.

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

The results establish what the registry reports for the defined endpoint and follow-up: no investigator-attributed serious adverse events were recorded in the tirzepatide or placebo arms, while one participant in the Part B dulaglutide arm met the endpoint. They also show that the broader adverse-event record includes events beyond those counted in the primary outcome. These observations are a description of the participants and periods recorded in this phase 1 study.

They do not establish that tirzepatide has no serious-event risk, that placebo and tirzepatide have equivalent risks, or that a serious event could not occur in a larger or longer study. A zero count is not proof of absence. The group sizes and sparse endpoint provide limited information about uncommon events, and the posted confidence limits do not remove that limitation. Nor does one event in the dulaglutide group establish that dulaglutide has a higher risk: the comparison is not supported by a reported inferential test, and the group is small.

The findings also do not answer whether tirzepatide changes blood sugar, or establish its pharmacokinetics or pharmacodynamics. Those were among the study’s stated purposes and the registry lists secondary outcomes related to drug exposure and glucose, but the primary results concern serious adverse events. No efficacy conclusion follows from the primary safety count.

Finally, the results apply to the study’s specific parts, dosing regimens, participant groups, and stated observation periods. They should not be collapsed into a single conclusion across all regimens or extended beyond the follow-up recorded. The registry does not report a statistical analysis plan, missing-data handling, or a sample-size justification, which further limits assessment of how the results were intended to support broader conclusions.

14. Why This Trial Matters Statistically

This trial illustrates several statistical features of early-phase safety reporting. First, the endpoint is sparse: most arms report no participant meeting the primary definition, and one arm reports a single participant. Sparse outcomes make descriptive counts easy to state but limit what those counts can tell readers about uncommon events. A zero observed count is a fact about the study record, not a guarantee of zero underlying risk.

Second, the study is divided into three parts with different dosing schemes, comparators, and group sizes. That structure makes the estimand—the precise comparison being described—important. A risk difference for a particular part and regimen should not be interpreted as a universal comparison across every participant or dose. The registry’s separate arm results preserve this context.

Third, the endpoint distinguishes seriousness from attribution. The primary count includes serious adverse events judged related to study drug, whereas the adverse-events module reports serious and other non-serious events regardless of causality. Statistical interpretation depends on respecting those definitions rather than treating all recorded events as the same outcome.

Finally, the posted comparison intervals are 0.0 to 0.0 despite sparse data, and the registry does not report the method used to construct them. This is a useful reminder that numerical precision in a displayed estimate does not necessarily mean practical certainty. Interpretation requires attention to the event count, group size, design, and transparency of the analysis method. The registry’s lack of posted analyses and sample-size justification also distinguishes descriptive results from a documented inferential plan.

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 is based on investigator attribution, and the registry does not report additional details about how relatedness was assessed or whether events underwent independent adjudication. Attribution is a judgment distinct from event occurrence. The primary endpoint also counts participants rather than recurrent events, so it does not describe the total number or pattern of qualifying events per participant.

The outcome is rare in the posted results, with zero participants meeting the endpoint in most arms and one in the Part B dulaglutide arm. Group sizes vary and include small groups. This limits the information available about uncommon harms and makes comparisons sensitive to a small number of observations. The displayed zero-width confidence limits should be read cautiously because the registry record does not report how they were calculated.

The three study parts used different dosing schedules and had different stated endpoint timeframes. Participants enrolled in only one part, so the results do not constitute repeated observations of the same participants across all regimens. Results from one part should not be generalized automatically to another part or combined as though treatment and follow-up were identical.

The record does not report the sample-size justification, detailed randomization or masking procedures, statistical analysis plan, missing-data methods, or analysis populations beyond the primary population definition. It also lists no posted analyses. These gaps prevent assessment of planned inferential methods, the handling of incomplete outcome data, and the rationale for the size of each group. The registry record’s safety findings are consequently best treated as descriptive observations during the specified follow-up, not as a complete characterization of comparative or long-term risk.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Study of Tirzepatide (LY3298176) in Healthy Participants and Participants With Type 2 Diabetes (T2DM) — NCT02759107.
• Registry record last updated 2024-01-16; status last verified 2023-04.
• Results first posted 2024-01-16.
• 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

NCT02759107 was a completed randomized, double-masked, parallel phase 1 study of tirzepatide involving healthy participants and participants with type 2 diabetes mellitus. The study had three parts: Part A used single doses, while Parts B and C used four weekly doses. Placebo was included in all parts, and dulaglutide was included in Part B. Participants enrolled in only one part.

The primary endpoint counted participants with one or more serious adverse events considered related to study drug, from baseline through Day 43 in Part A and Day 57 in Parts B and C. The posted results report zero qualifying participants in the tirzepatide and placebo groups and one in the Part B dulaglutide group. The registry lists five secondary outcomes and no posted analyses. It does not report the trial’s sample-size justification, inferential analysis plan, or missing-data methods, so the primary findings are descriptive and limited to the recorded groups and follow-up.