Pharmacokinetic Comparability of Benralizumab Using Accessorized Pre-Filled Syringe or Autoinjector in Healthy Volunteers: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pharmacokinetic Comparability of Benralizumab Using Accessorized Pre-Filled Syringe or Autoinjector in Healthy Volunteers (NCT02968914), a randomized phase 1 study. The primary outcome is Area Under the Concentration-time Curve From Zero to Infinity (AUCinf).

2. Clinical Question

The clinical question is pharmacokinetic, not a test of whether benralizumab improves asthma or chronic obstructive pulmonary disease. The registry summary describes a single-dose study intended to compare drug exposure following use of two delivery devices in healthy volunteers. It does not report a clinical efficacy endpoint.

Exposure is assessed using AUCinf, the area under the concentration-time curve from zero to infinity. This measure represents overall drug exposure across the observed concentration profile and its estimated terminal continuation. The primary endpoint was measured using serum samples collected at pre-dose on Day 1 and at Days 2, 4, 5, 6, 8, 15, 29, 43 and 57.

In practical terms, the comparison asks whether the observed exposure distributions are sufficiently aligned to support the registry’s stated pharmacokinetic comparability objective. That judgment depends not only on the estimated ratio and its precision, but also on a prespecified acceptable range and the analysis rules. Those criteria are not reported in the registry record. The available result therefore describes the estimated exposure contrast and its uncertainty, without establishing the decision rule used for any formal comparability conclusion.

3. Trial Design

This was a phase 1, randomized, open-label, parallel-group study of benralizumab administration using an accessorized pre-filled syringe or an autoinjector. The registry lists an enrollment of 180 and status as completed. The study is described as multicenter, and its stated objective was to compare pharmacokinetic exposure following a single subcutaneous administration.

Randomization assigns participants to treatment groups by chance, which is intended to reduce systematic differences between groups. In a parallel-group design, participants remain in their assigned group for the comparison rather than crossing over between devices. The study was open-label, meaning that masking was not used. The registry record does not report further details about who knew the assigned device or whether outcome assessment was masked.

The arm descriptions refer to stratification by weight group, with ranges of 55 to 69.9 kg, 70 to 84.9 kg and 85 to 100 kg. They also refer to allocation among six treatment combinations, but the descriptions are truncated. The full combination structure and the way those combinations map to the two posted device arms are not reported in the registry record. The posted primary results include 88.0 participants for the AI arm and 87.0 for the AFPS arm in the pharmacokinetic analysis.

The primary endpoint is AUCinf, and six secondary outcomes are listed, including other pharmacokinetic measures, adverse events and antidrug antibody status. The record reports nine outcome measures and three analyses overall. It does not provide a complete statistical analysis plan, detailed enrollment rationale, or full account of all analysis populations. The primary pharmacokinetic analysis population is described as participants who received benralizumab, had at least one quantifiable post-dose serum pharmacokinetic observation, and whose samples were assumed not to be affected by factors such as protocol violations; the description is incomplete.

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: Subjects attended a Screening Visit within 28 days before receiving their first dose of Benralizumab. All subjects underwent inclusion exclusion criteria assessment and all eligible subjects signed the informed consent before undergoing any study-related procedures.

Overall Study

Analysis population for the primary outcome, as defined in the registry: The PK analysis set consisted of all subjects who received benralizumab for whom PK blood samples were assumed not to be affected by factors such as protocol violations (e.g., disallowed medications or incomplete dose administration) and who had at least one post dose quantifiable serum PK observation.

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 Area Under the Concentration-time Curve From Zero to Infinity (AUCinf), measured over At Pre-dose (Day 1) and at Days 2, 4, 5, 6, 8, 15, 29, 43 and 57.

The primary endpoint, AUCinf, is the area under the serum concentration-time curve from zero to infinity, reported in day·ng/mL. It summarizes overall drug exposure after administration: observed concentrations contribute across the sampling period, while the portion extending beyond the final observed period is estimated from the terminal concentration profile. The registry record describes the objective as comparing AUCinf after a single subcutaneous dose delivered by the two devices.

Exposure measures are useful in a device-comparability study because they address whether the delivery method is associated with a difference in the amount of drug reaching the circulation over time. AUCinf is a summary of the concentration-time profile, rather than a direct measure of symptoms, disease control or clinical benefit. It therefore answers a pharmacokinetic question and should not be interpreted as an efficacy outcome.

The registry reports arm-level geometric means and geometric coefficients of variation. The AI arm has a geometric mean of 72210.0 day·ng/mL and a geometric coefficient of variation of 29.7, based on 88.0 participants. The AFPS arm has a geometric mean of 76220.0 day·ng/mL and a geometric coefficient of variation of 32.1, based on 87.0 participants. The reported comparative estimate is a geometric mean ratio with a 90% confidence interval.

The result’s meaning depends on the prespecified acceptable range for comparability, as well as the analysis population and model. The acceptable range and detailed model specification are not reported in the registry record. The endpoint can therefore be described, and the estimate and interval can be interpreted statistically, but the available information does not establish a formal pass-or-fail determination against a protocol criterion.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary analysis reports geometric mean AUCinf by device and a geometric mean ratio with a 90% confidence interval. The method is identified as ANOVA. These elements describe a comparison of exposure on a relative scale, rather than a difference expressed in the original concentration-time units. The registry record does not report the complete model specification, transformations, covariates or other implementation details.

What does a geometric mean summarize?

A geometric mean is a way to summarize positive measurements when relative changes are more meaningful than equal-sized absolute changes. It is based on the logarithmic scale and is commonly used for pharmacokinetic measurements, which can be right-skewed. On this scale, a ratio gives a relative comparison between group summaries. The posted arm-level geometric means are 72210.0 day·ng/mL for AI and 76220.0 day·ng/mL for AFPS. The reported ratio is 94.46. The registry record does not state the exact ratio calculation formula or provide further model details, so interpretation should rely on the reported ratio and its stated group comparison.

What does the geometric coefficient of variation describe?

The geometric coefficient of variation describes relative variability around a geometric mean. It can help readers understand the dispersion of pharmacokinetic measurements in each analyzed arm, but it is not an interval estimate for the group mean and should not be confused with the confidence interval for the ratio. The posted values are 29.7 for AI and 32.1 for AFPS. A coefficient of variation does not show the full distribution, identify influential observations, or explain why participants’ concentrations differ. The registry record does not report plots or participant-level values in the narrative description.

How should the confidence interval be read?

The 90% confidence interval for the reported ratio extends from 88.16 to 101.21. In general, a confidence interval communicates the range of parameter values compatible with the data and the analysis procedure, subject to its assumptions. It reflects sampling uncertainty, not the range of individual responses. The interval includes 100, the ratio value associated with equal group geometric means. That inclusion is not proof of equality. Conversely, an interval that includes equality is not automatically evidence against comparability, because comparability is assessed against an acceptable margin, not merely against equality.

For a formal comparability assessment, a margin or range is normally defined before examining the results. The estimate and its confidence interval are then judged relative to that range. The registry record does not report a margin, and does not state the confidence-interval criterion used to reach any formal decision. Thus the ratio and interval can be described, but the criterion for declaring comparability cannot be reconstructed from this record.

What does ANOVA contribute?

ANOVA is the reported analysis method for the group comparison. In general, analysis of variance is a framework for comparing means while accounting for variation within groups. In pharmacokinetic analyses, a model may be applied on a transformed scale to support a ratio-based comparison, but the registry record does not specify whether or how a transformation was used in this analysis. It also does not list covariates, treatment terms, assumptions checks or the exact method used to construct the confidence interval. These details should not be inferred from the method label alone.

Model-based estimates depend on assumptions about the data and model errors. General concerns include whether the model represents the observed distribution adequately and whether observations are sufficiently independent for the analysis. Over-dispersion, in general, occurs when variability is greater than a model assumes; if not accommodated, uncertainty can be understated. The registry record does not report a specific over-dispersion assessment or indicate that over-dispersion occurred in this study.

What is the role of the analysis population?

The primary endpoint population is described as participants who received benralizumab, had at least one post-dose quantifiable serum pharmacokinetic observation, and whose samples were assumed not to be affected by factors such as protocol violations, including disallowed medications or incomplete dose administration. This is an analysis-set definition, not a description of all randomized participants. The registry wording is truncated, so the complete eligibility criteria and final participant flow are not reported in the registry record.

Analyzing a defined pharmacokinetic set can make the endpoint more directly reflect interpretable exposure profiles, but exclusions after randomization can affect how well the randomized groups remain comparable. The record does not report the reasons for the difference between the pharmacokinetic counts and the safety denominators, or describe how missing or non-quantifiable samples were handled. It does not provide enough information to evaluate the consequences of exclusions for the ratio estimate.

How much can the reported statistics establish?

The posted ratio and interval quantify the relative exposure estimate and its uncertainty under the sponsor’s reported analysis. They do not establish a clinical effect, demonstrate an acceptable device-use experience, or answer safety questions by themselves. A statistical result should be read in the context of its endpoint, study population, model, analysis set and prespecified decision rule. The population and endpoint are described, and ANOVA is named, but important details of the model, missing-data handling and comparability criterion are not reported in the registry record.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s comparative analysis as an ANOVA, with a geometric mean ratio of 94.46 and a 90% confidence interval from 88.16 to 101.21. Those are the posted comparative results. The registry record also gives arm-level geometric means and geometric coefficients of variation, but it does not report a separate unadjusted re-calculation or a second confidence interval for comparison.

The arm summaries are descriptive statistics; the sponsor’s reported ratio and interval are the inferential comparison. A ratio based on a model need not be reproduced exactly by taking a simple quotient of the displayed arm summaries, because model specifications, transformations and analysis populations can matter. The registry record does not provide the details needed to reproduce the sponsor’s estimate independently. No discrepancy is identified in the registry record, but the full model cannot be audited from the information posted.

The reported interval includes 100. Whether the sponsor regarded the result as meeting the study’s comparability objective cannot be determined from the estimate alone: the acceptable range and decision rule are not reported in the registry record. Accordingly, the sponsor’s own analysis should be treated as the reported inferential result, while avoiding a formal conclusion that requires unreported protocol criteria.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists nine outcome measures and three analyses overall, including the primary AUCinf measure and six secondary outcomes. The secondary outcomes include time when maximum concentration is observed, terminal half-life, apparent extravascular clearance, apparent volume of distribution based on the terminal phase, participants with adverse events, and antidrug antibody status. The record does not identify which outcomes were included in each of the three analyses or provide a complete analysis hierarchy.

Multiplicity arises when several outcomes, groups or statistical tests are considered. In general, conducting multiple tests can increase the chance that at least one result appears unusual by chance. Common approaches include designating a primary endpoint, adjusting significance thresholds, or treating secondary analyses as descriptive. Which approach was prespecified or used for this study is not reported in the registry record. The primary endpoint is identified, but that alone does not reveal whether multiplicity adjustments were planned for secondary outcomes or other analyses.

The record does not report interim analyses, stopping rules, or a complete statistical analysis plan. Their absence from the registry record should not be taken to mean they did or did not occur; they are not reported in the registry record. No p-value is posted for the primary ratio analysis, and the confidence interval is the available inferential summary.

Missing-data handling is also not reported in sufficient detail. The primary pharmacokinetic analysis description requires at least one quantifiable post-dose observation and says samples were assumed not to be affected by specified factors, but the description is truncated. The record does not state how missing samples, values below quantification limits, incomplete profiles or participants excluded from analysis were handled. It does not provide a participant flow accounting for the pharmacokinetic counts of 88.0 and 87.0 relative to 90 participants at risk in each safety group.

The analysis population is therefore important to interpretation. Randomization applies to assignment, while the primary result is reported for a defined pharmacokinetic set. Without complete participant flow and missing-data rules, readers cannot assess whether exclusions were balanced or how they may affect the estimate. The registry record also does not report a sensitivity analysis testing how alternative assumptions about missing or excluded observations might alter the result.

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 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events for 0 of 90 participants at risk in the autoinjector group and 0 of 90 participants at risk in the accessorized pre-filled syringe group. It also reports 50 participants with other adverse events in the autoinjector group and 46 in the accessorized pre-filled syringe group. Deaths are listed as 0 in each group.

These are counts within the safety groups, not estimates of comparative risk adjusted for follow-up or other factors. The record does not provide event-level detail for the reported serious adverse events, and its list of top serious terms is empty. It does not specify in the posted summary how “other” adverse events are defined, whether participants could contribute more than one event, or the full duration of safety observation. The study is described as single-dose, but the posted safety summary does not explain the event collection window.

Zero reported serious adverse events in each group means that none were recorded in these safety summaries; it does not establish that the risk is absent. The safety counts should be considered separately from the pharmacokinetic endpoint and should not be used to infer a safety advantage for either device. The registry record does not report a comparative statistical analysis of adverse events.

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

The result addresses whether benralizumab exposure differed between two delivery devices after a single subcutaneous administration in healthy volunteers. The posted geometric mean ratio is 94.46, and its 90% confidence interval is 88.16 to 101.21. The interval includes the ratio value associated with equal geometric means. This is a description of the estimated exposure contrast and the uncertainty around it under the reported ANOVA.

The result does not, on its own, prove that the devices produce identical exposure. Nor does it establish formal pharmacokinetic comparability, because the acceptable margin and the decision rule are not reported in the registry record. A formal conclusion depends on those criteria, in addition to the estimate and confidence interval. The posted numbers should therefore not be converted into a pass-or-fail conclusion that relies on unreported protocol details.

The study is not an efficacy trial in people with asthma or chronic obstructive pulmonary disease. Its population is described as healthy volunteers, and its primary endpoint is a pharmacokinetic measure. AUCinf describes drug exposure, not symptom change, disease control, or clinical benefit. The registry’s condition labels do not alter the stated population for this study.

The result also does not establish that the devices are interchangeable in every respect. Device usability, successful administration, preference, or performance in other populations are not the primary endpoint. The posted adverse-event counts provide limited safety information, but do not establish equivalence of safety or absence of risk. The registry does not report a complete safety analysis or comparative inference for those events.

What can be stated is narrower: among the participants included in the posted pharmacokinetic analysis, the registry reports the arm-level geometric means and a ratio estimate with its interval. The precision and formal interpretation depend on the analysis model and decision criteria, some of which are not reported. The result is informative about the stated exposure comparison, but should not be extended to outcomes or settings the study did not measure.

14. Why This Trial Matters Statistically

This study illustrates the distinction between estimating a relative pharmacokinetic effect and making a formal comparability decision. The geometric mean ratio and its confidence interval summarize the estimated relationship between exposure in the device groups. A comparability conclusion requires an acceptable range set out in advance; without that range, the estimate cannot be judged against the study’s intended threshold from the registry record alone.

It also shows why the analysis population matters. The registry reports 88.0 and 87.0 participants in the pharmacokinetic groups, while the safety summaries list 90 participants at risk in each group. These summaries answer different questions and use different participant counts. Readers need the participant flow and analysis rules to understand which randomized participants contributed to each endpoint and why.

Geometric summaries are commonly useful for pharmacokinetic endpoints because relative comparisons are central to exposure questions. Yet a geometric mean and coefficient of variation do not reveal the entire distribution or the influence of individual observations. The model, missing-data rules and sensitivity checks are part of interpreting the estimate, and those details are not reported in the registry record.

Finally, the study keeps the question appropriately narrow: it compares exposure after administration by two devices in healthy volunteers. Statistical evidence about drug exposure is not evidence about clinical efficacy or a complete assessment of device experience. Keeping the endpoint, population and decision criterion distinct helps prevent a pharmacokinetic comparison from being overinterpreted as a broader clinical conclusion.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record does not report a prespecified comparability margin or a complete decision rule. Without those criteria, the ratio and 90% confidence interval cannot be used to determine from the record alone whether the intended formal comparability standard was met. The reported analysis method is ANOVA, but its full specification, including transformations, covariates and interval construction, is not reported.

The description of the primary pharmacokinetic analysis population is truncated. It refers to participants who received benralizumab, had at least one quantifiable post-dose serum pharmacokinetic observation, and whose samples were assumed not to be affected by factors such as protocol violations. The complete criteria, participant flow and reasons for the difference between pharmacokinetic and safety counts are not reported in the registry record.

Missing-data and below-quantification handling are not reported in the registry record. Neither are sensitivity analyses, model diagnostics, detailed sampling completeness, or any assessment of whether exclusions could have affected the comparison. The registry gives sampling time points through Day 57, but does not describe the estimation details used for the terminal portion of AUCinf.

The record describes stratification by weight ranges of 55 to 69.9 kg, 70 to 84.9 kg and 85 to 100 kg, but the arm descriptions are incomplete. The allocation text refers to six treatment combinations, yet the full combinations are not reported. Device naming is also inconsistent: the title and study description use APFS, whereas a posted arm title uses AFPS.

The study is described as open-label and conducted in healthy volunteers. Its primary endpoint is pharmacokinetic exposure, so it does not directly address clinical outcomes in people with asthma or chronic obstructive pulmonary disease. Safety reporting is limited to counts, with no serious-event terms listed and no comparative safety analysis reported in the registry record. These limitations bound what can be concluded from the posted results.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Pharmacokinetic Comparability of Benralizumab Using Accessorized Pre-Filled Syringe or Autoinjector in Healthy Volunteers — NCT02968914.
• Registry record last updated 2019-07-05; status last verified 2019-06.
• Results first posted 2019-07-05.
• 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

NCT02968914 is a completed phase 1, randomized, open-label, parallel-group study sponsored by AstraZeneca. The registry lists enrollment of 180 and describes a single-dose pharmacokinetic comparison of benralizumab delivered subcutaneously by an accessorized pre-filled syringe or an autoinjector in healthy volunteers.

The primary endpoint was AUCinf in day·ng/mL, measured using serum samples from pre-dose on Day 1 through Day 57. The posted comparative result is a geometric mean ratio of 94.46 with a 90% confidence interval from 88.16 to 101.21. The pharmacokinetic analysis includes 88.0 participants in the AI group and 87.0 in the AFPS group. Safety summaries report 0 serious adverse events in each group, with 90 participants at risk per group.

The estimate describes relative exposure, but the registry record does not report the comparability margin or full statistical analysis plan. The result should therefore be interpreted as a pharmacokinetic comparison, not as proof of clinical efficacy or a definitive formal comparability conclusion.