Efficacy and Safety of Anifrolumab Compared to Placebo in Adult Subjects With Active Systemic Lupus Erythematosus: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Efficacy and Safety of Anifrolumab Compared to Placebo in Adult Subjects With Active Systemic Lupus Erythematosus (NCT02446899), a randomized phase 3 study. The primary outcome is Number of Participants Who Achieved the British Isles Lupus Assessment Group Based Composite Lupus Assessment (BICLA) Response at Week 52.

2. Clinical Question

BICLA is a composite measure of changes in lupus disease activity across several domains. A participant had to meet all specified response conditions, including improvement in BILAG-2004 assessments without worsening in other organ systems, no worsening in SLEDAI-2K or Physician’s Global Assessment, and no discontinuation of study product or use of restricted medications beyond the permitted threshold. This makes the endpoint a response under a set of joint conditions, rather than a measure of change in just one symptom or organ system.

3. Trial Design

The registry record describes a completed, randomized, double-blind, parallel-group Phase 3 study sponsored by AstraZeneca. It enrolled 373 participants and evaluated intravenous anifrolumab against placebo in adults with active SLE. The intervention descriptions specify anifrolumab 300 mg or placebo every 4 weeks for a total of 13 doses. The primary completion date was 2018-09-27, and the study start date was 2015-07-09.

Random allocation is intended to create groups that are comparable on average at the point of assignment, while double masking is intended to reduce the influence of expectations on participants’ experiences and outcome assessment. These are general features of randomized, masked trials; the exact procedures used to generate and conceal the allocation are not reported in the registry record. The record identifies a parallel design, meaning the two groups received their assigned regimens rather than crossing over between them.

The primary analysis population is defined as the Full Analysis Set: participants who were randomized and received at least one dose of study drug. Its posted endpoint denominators are 180 for anifrolumab and 182 for placebo. Those denominators do not equal the registry’s total enrollment figure; how the enrollment count relates to the endpoint analysis population is not explained in the registry record. The record reports the trial as completed, but does not provide the original sample-size justification, the detailed randomization or masking procedures, or the planned handling of every possible analysis population. Those design details should not be inferred from the posted outcome counts.

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: 8 participants were assigned study drug, but due to non-compliance were analyzed separately (not included in the participant flow). 3 additional participants were assigned study drug \& included in the participant flow, but did not receive study drug due to an adverse event (1 Anifrolumab 300mg) \& failure to meet randomization criteria (2 Placebo).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full Analysis Set: All participants who were randomized and 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: Full analysis set: All participants who were randomized and received at least one dose of the study drug. 3 participants discontinued the study prior to receiving the study drug and are not included in the full analysis set.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Who Achieved the British Isles Lupus Assessment Group Based Composite Lupus Assessment (BICLA) Response at Week 52, measured over Baseline; Week 52.

The primary endpoint is the number of participants who achieved a BICLA response at Week 52. BICLA is a composite disease-activity endpoint. The registry definition requires improvement in all baseline BILAG-2004 A and B findings according to specified transitions, with no defined worsening in other organ systems. It also requires no worsening from baseline in SLEDAI-2K and no worsening in the Physician’s Global Assessment, as well as no discontinuation of investigational product and no restricted-medication use beyond the permitted threshold.

In general, a composite endpoint can capture response across more than one aspect of a condition and can make a trial outcome more clinically encompassing than a single measurement. Its interpretation depends on the exact components and response rules: a participant must satisfy all the stated conditions to count as a responder. A count of responders therefore does not reveal which individual component was most influential, nor does it describe every change experienced by participants.

The endpoint is assessed at Week 52, with baseline measurements part of the stated timeframe. The registry record defines the response criteria but does not report how missing component measurements were treated, or whether additional sensitivity analyses were used to test alternative handling of missing data. Those details matter because a composite response can depend on information from multiple assessments and on the rules used when an assessment is unavailable.

7. Results Posted in the Registry

• Anifrolumab 300 mg: 47.8% (86/180) versus 31.3% (57/182) on control; risk difference 16.5 percentage points (95% CI 6.5 to 26.4), risk ratio 1.53 (1.17–1.99); unadjusted two-proportion z-test p = 0.001.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Number of Participants Who Achieved the British Isles Lupus Assessment Group Based Composite Lupus Assessment (BICLA) Response at Week 52 in the IFN Test-High Sub-group

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Anifrolumab 300 mg: 48.0% (72/150) versus 30.5% (46/151) on control; risk difference 17.5 percentage points (95% CI 6.7 to 28.4), risk ratio 1.58 (1.18–2.11); unadjusted two-proportion z-test p = 0.002.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Anifrolumab 300 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Number of Participants Who Achieved and Maintained an Oral Corticosteroids (OCS) Dose of ≤7.5 mg/Day at Week 52 in the Sub-Group of Participants With Baseline OCS ≥10 mg/Day

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Anifrolumab 300 mg: 51.7% (45/87) versus 30.1% (25/83) on control; risk difference 21.6 percentage points (95% CI 7.2 to 36.0), risk ratio 1.72 (1.17–2.52); unadjusted two-proportion z-test p = 0.004.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Anifrolumab 300 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Number of Participants With a ≥50% Reduction in Cutaneous Lupus Erythematosus Disease Area and Severity Index (CLASI) Activity Score at Week 12 in The Sub-Group of Participants With Baseline CLASI Activity Score of ≥10

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Anifrolumab 300 mg: 49.0% (24/49) versus 25.0% (10/40) on control; risk difference 24.0 percentage points (95% CI 4.6 to 43.4), risk ratio 1.96 (1.07–3.60); unadjusted two-proportion z-test p = 0.021.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Anifrolumab 300 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The detailed secondary results are directionally aligned with the primary comparison: in each of the reported outcomes, the proportion meeting the stated criterion was higher in the anifrolumab group than in the placebo group. These analyses address a mixture of questions, including BICLA response in the IFN test-high subgroup, corticosteroid-dose reduction in a subgroup with higher baseline use, and skin response among participants with a specified baseline CLASI score. They therefore add context about selected subgroups and outcomes, rather than serving as direct replications of the full-population primary endpoint.

The secondary analyses have narrower populations and different assessment times or criteria. Their denominators are smaller than those for the primary endpoint, and subgroup estimates may be less precise. The registry also names other outcomes, including joint-count response, annualised flare rate, adverse events, adverse events of special interest, and vital-sign changes. Detailed comparative estimates for all listed outcomes are not reported in the registry record.

Unadjusted secondary results carry less weight because the registry record does not report a multiplicity strategy or a hierarchy governing these tests. In general, when many outcomes are tested, the chance of at least one apparently persuasive finding can increase unless the testing plan controls that risk. A small unadjusted p-value for an individual secondary outcome should therefore be viewed as supportive or exploratory evidence, not as independent confirmation on the same footing as the designated primary endpoint. The secondary pattern may be consistent with the primary finding, but it cannot establish that every endpoint or subgroup has the same effect.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Number of Participants Who Achieved the British Isles Lupus Assessment Group Based Composite Lupus Assessment (BICLA) Response at Week 52 in the IFN Test-High Sub-group
• Number of Participants Who Achieved and Maintained an Oral Corticosteroids (OCS) Dose of ≤7.5 mg/Day at Week 52 in the Sub-Group of Participants With Baseline OCS ≥10 mg/Day
• Number of Participants With a ≥50% Reduction in Cutaneous Lupus Erythematosus Disease Area and Severity Index (CLASI) Activity Score at Week 12 in The Sub-Group of Participants With Baseline CLASI Activity Score of ≥10
• Number of Participants With ≥50% Reduction in Joint Counts at Week 52 in The Sub-group of Participants With ≥6 Swollen and ≥6 Tender Joints at Baseline
• Annualised Flare Rate Through 52 Weeks
• Number of Participants With One or More Adverse Events (AEs)
• Number of Participants With One or More Adverse Events of Special Interest (AESIs)
• Number of Participants With a Potentially Clinically Important Change From Baseline in Vital Sign Measurements

9. Statistical Interpretation

10. Statistical Methods Explained

For a binary endpoint such as whether a participant achieved a response, a natural summary is the proportion responding in each group. The primary endpoint is posted as counts: 86 responders among 180 participants assigned to anifrolumab and 57 among 182 assigned to placebo. The unadjusted recalculation expresses the contrast as a risk difference, estimated at 0.165. In general, a risk difference subtracts the comparator-group response proportion from the intervention-group response proportion. It is measured on the probability scale and keeps the contrast tied to an absolute difference in response proportions.

The registry-reported result uses a Cochran-Mantel-Haenszel analysis and labels its parameter as a mean difference in final values. It reports 16.3 with a 95% confidence interval from 6.3 to 26.3 and p = 0.001. The registry does not describe the strata, covariates, or other implementation details for that analysis. The unadjusted recalculation and the registry-reported analysis are therefore different summaries: one is a straightforward contrast recalculated from posted group counts, while the other is the sponsor’s reported method and estimate. They should not be presented as though they were the same calculation.

What does a confidence interval describe?

In general, a confidence interval communicates the uncertainty around an estimated effect under a specified statistical procedure and its assumptions. A narrower interval indicates greater precision than a wider interval on the same measurement scale. The interval is not a range of individual participant responses, and a 95% confidence interval does not mean that there is a 95% probability that the fixed true effect lies inside this particular calculated interval. Its practical interpretation also depends on the study design, data quality, analysis population, and whether the analysis method was specified in advance.

What does the p-value contribute?

In general, a p-value compares the observed data with what would be expected under a null hypothesis, often a hypothesis of no group difference. A small p-value indicates that the observed data would be relatively unusual under that null model, assuming the model and analysis are appropriate. It does not measure the size or importance of an effect and does not indicate the probability that a result will replicate. The registry-reported p-value and the unadjusted recalculated p-value are close, but the registry-reported analysis remains the sponsor’s own analysis for interpreting the trial’s reported result.

Why distinguish the analysis populations?

The primary outcome description identifies the Full Analysis Set as participants randomized and given at least one dose. This definition sets out who contributes to the reported endpoint analysis. In general, decisions about which participants to include can affect an estimate, especially when some participants do not complete follow-up or do not have an evaluable endpoint. The registry record does not report the detailed missing-data rules, participant flow by reason for exclusion, or sensitivity analyses for the primary outcome. Consequently, the effect of those choices cannot be assessed from the posted counts alone.

What is the role of secondary analyses?

Secondary endpoints can examine related questions, such as response within a defined subgroup, corticosteroid dose, skin activity, joint counts, flares, or safety. Their statistical interpretation depends on their design, denominators, and whether the analysis plan accounted for testing multiple outcomes. In general, testing many outcomes increases the chance that at least one result appears statistically persuasive by chance if no adjustment or hierarchy is used. The registry record does not report a multiplicity strategy. Unadjusted secondary p-values should therefore be read as exploratory evidence, not as though each one carried the same confirmatory status as the primary endpoint.

What can be said about over-dispersion?

Over-dispersion is a general concern for some count models when observed variability exceeds what the model assumes. The primary endpoint is a binary responder outcome presented as participant counts, not a repeated-event rate or a continuous measurement with a dispersion estimate. The registry record does not identify an over-dispersion assessment for the primary analysis, and no such assessment should be inferred. For outcomes such as annualised flare rate, model details would be needed to assess how repeated events and variability were handled; those details are not reported in the registry record.

These methods describe different aspects of the evidence: counts show what was observed in each group, an effect estimate summarizes the contrast, a confidence interval conveys precision, and a p-value evaluates the data relative to a null hypothesis. None of these alone resolves questions about missing outcomes, multiplicity, clinical importance, or generalizability.

11. Confidence Intervals and Precision

The primary unadjusted risk-difference estimate is 0.165, with a 95% confidence interval from 0.065 to 0.264. This interval is entirely on the positive side of zero, indicating that the posted estimate is compatible with a positive group difference under the calculation’s assumptions. Its width also shows that the exact size of the difference is uncertain: the data support a range of plausible magnitudes rather than one precisely known value.

The registry-reported estimate is 16.3, with a 95% confidence interval from 6.3 to 26.3. The lower and upper limits are positive, consistent with the superiority result reported by the sponsor. The registry’s reported analysis and the unadjusted recalculation use different methods and scales of presentation, so their interval limits should not be compared as if they were produced by an identical procedure. The sponsor’s reported analysis takes precedence as the trial’s own analysis.

In general, interval precision depends on the amount of information in the data, the variability of the outcome, and the statistical method. The primary endpoint denominators are 180 and 182, and the response counts are 86 and 57. These data provide an estimate but do not remove sampling uncertainty. A confidence interval does not account for every source of uncertainty: it cannot correct for unreported analysis choices, missing-data assumptions, biases in conduct, or limited applicability beyond the studied population.

The secondary intervals are wider for some subgroup outcomes, where the posted denominators are smaller. In general, estimates from smaller subgroups tend to be less precise, although the interval also depends on the event pattern and calculation method. The confidence intervals should be read alongside the outcome definitions and analysis populations rather than as standalone evidence of a treatment effect.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are illustrations based on the observed primary effect, not a record of how the trial was originally designed. They assume a two-sided significance level of 0.05, equal allocation between the two groups, and a normal approximation. For each scenario, the effect to detect is set to the stated fraction of the observed effect. The original trial’s sample-size justification and design assumptions are not reported in the registry record.

The figures are expressed as participants per arm for 80% power and participants per arm for 90% power. In plain language, 80% power means that, if the assumed effect and planning assumptions held across repeated studies, the design would have an 80% chance of detecting an effect at the stated significance level. The 90% column represents a higher target chance of detection under the same kind of assumptions. Neither value is a probability that this particular trial’s conclusion is correct.

The planning scenarios use 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. The figures rise as the assumed effect becomes smaller: smaller differences are harder to distinguish from random variation, so more participants are generally needed to maintain the same target power. For example, at the 100% scenario, the primary-outcome planning illustration lists 138 participants per arm for 80% power and 184 participants per arm for 90% power. At the 50% scenario, it lists 530 participants per arm for 80% power and 708 participants per arm for 90% power.

These figures are planning illustrations, not observed power. They do not show whether the actual trial was adequately designed under its original assumptions, because those assumptions are not reported. They also should not be mistaken for observed participant counts or a post hoc measure of the strength of the result.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.313; treatment proportion p₁ = 0.478; average p̄ = 0.395.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6915 = 1.3553.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6816 = 0.5737.
4. n = (1.3553 + 0.5737)² / (0.1646)² = 137.4, rounded up to 138 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.313186813186813                          # control proportion
p1 <- 0.477777777777778                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(2446899)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 138 (80% power), 184 (90% power)
 75% of observed effect: n per arm = 241 (80% power), 322 (90% power)
 50% of observed effect: n per arm = 530 (80% power), 708 (90% power)
power.prop.test: 137.4 
Simulated power at n = 138 per arm: 0.803 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry-reported primary analysis uses the Cochran-Mantel-Haenszel method and gives a mean difference of 16.3, with a 95% confidence interval from 6.3 to 26.3 and p = 0.001. The unadjusted recalculation from the posted counts gives a risk difference of 0.165, with a 95% confidence interval from 0.065 to 0.264 and p = 0.001. Both summaries point in the same direction, but they are not identical estimates or methods.

The recalculation is a simple, unadjusted contrast based on the displayed group counts. The registry-reported analysis is the sponsor’s own analysis and should take precedence when describing the trial’s formal result. The registry record does not specify the strata or all implementation details underlying the Cochran-Mantel-Haenszel analysis, so the difference between the reported and recalculated estimates cannot be explained further from this record.

In general, an adjusted or stratified analysis may differ from a crude comparison because it accounts for features of the analysis design or data structure. That is a general possibility, not a claim about which features produced the reported estimate in this trial. The close direction and similar scale of the two estimates provide context, while inference about the trial should remain anchored to the registry-reported method and its confidence interval.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies BICLA response at Week 52 as the primary endpoint and lists multiple secondary outcomes. The record also includes comparative analyses for three secondary outcomes. It does not report whether the secondary tests were arranged in a prespecified hierarchy, whether p-values were adjusted for multiplicity, or whether a family-wise error or false-discovery strategy was used. The original statistical analysis plan and full details of the testing sequence are not reported in the registry record.

In general, multiplicity arises when several hypotheses are tested. If each is judged using an unadjusted threshold, the chance of at least one false-positive result across the set can be greater than the chance for a single test. That concern is especially relevant when secondary results are interpreted as confirmatory. The primary result remains the clearest designated test in the posted record; the secondary p-values should be read with caution because their role in the overall testing plan is unknown.

The endpoint population is described as participants randomized and receiving at least one dose. The record does not explain how missing Week 52 assessments, missing components of the composite, withdrawal, or protocol deviations were handled in the primary analysis. It also does not report detailed participant flow or sensitivity analyses comparing alternative missing-data assumptions. In general, missing-data rules can affect binary composite outcomes because classification as a responder depends on satisfying all required components.

The registry record does not provide enough information to assess whether missingness differed between groups or whether the chosen analysis could be sensitive to missing outcomes. It would be inappropriate to assume a particular imputation method, a particular treatment of participants who discontinued, or a specific set of sensitivity analyses. These absences limit assessment of robustness but do not by themselves show that the posted analysis is biased.

Randomization and masking are important design features, but they do not remove every uncertainty about analysis conduct. The detailed randomization procedures, allocation concealment, prespecified analysis plan, and completeness of outcome ascertainment are not reported in the registry record. These missing details should be kept distinct from the observed counts and reported estimates.

16. 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: pneumonia (3.8%), systemic lupus erythematosus (3.3%), gastroenteritis viral (1.1%), radius fracture (1.1%), acute coronary syndrome (0.6%).

The posted serious adverse event counts are 16 among 180 participants at risk in the anifrolumab group and 34 among 182 participants at risk in the placebo group. The record also reports 1 death in the anifrolumab group and 0 in the placebo group. These are participant counts and should be read with their group denominators; they are not, on their own, evidence that a particular event was caused by the assigned treatment.

The registry lists serious-event terms including pneumonia at 3.8%, systemic lupus erythematosus at 3.3%, gastroenteritis viral at 1.1%, radius fracture at 1.1%, acute coronary syndrome at 0.6%, and Chilaiditi’s syndrome at 0.6%. The listed terms are not identified by treatment group in the record, so they should not be interpreted as group-specific comparisons. The record also gives other adverse event counts of 125 among 180 participants and 99 among 182 participants. The definition and breakdown of that category are not further described in the available results.

In general, safety counts require context about event definitions, duration of exposure, severity, and attribution. The registry record does not provide enough detail to determine causation or compare the full safety profile. These counts describe what was posted, not a recommendation about treatment.

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

The posted primary result indicates that more participants in the anifrolumab group met the registry’s complete BICLA response definition at Week 52 than in the placebo group. The estimated group difference is positive, and the registry-reported 95% confidence interval is also positive. The result is specific to the endpoint definition, study population, treatment regimen, and follow-up time described in the record.

It does not mean that every participant in the anifrolumab group responded, that every aspect of SLE improved, or that all secondary outcomes had the same result. BICLA requires several conditions to be met together; the overall response count does not identify which components drove the group contrast. The endpoint also does not quantify the full course of disease activity or establish how outcomes changed after Week 52.

A statistically significant result is not, by itself, a judgment about clinical importance. That judgment depends on the value placed on the response definition, the size and durability of the difference, other outcomes, and the balance of benefits and harms. The registry record provides a primary response estimate and selected safety counts, but it does not report every detail needed for a comprehensive assessment of those considerations.

The finding also should not be generalized automatically to populations or regimens not described in the record. The study concerned adults with active SLE, and the brief summary specifies moderately to severely active, autoantibody-positive disease. The posted data do not establish effects in other populations, with other dosing schedules, or on outcomes that were not reported. They also do not replace consideration of the trial’s analysis limitations, including unreported missing-data and multiplicity procedures.

In short, the registry reports a group difference on a defined composite endpoint, with an interval that expresses statistical uncertainty. The data do not establish individual outcomes, the mechanism of any difference, or a complete account of long-term efficacy and safety.

18. Why This Trial Matters Statistically

This trial illustrates how a binary clinical endpoint can be analyzed without reducing its meaning to a single p-value. The responder counts describe the observed data, the risk difference summarizes the absolute group contrast, and the confidence interval shows the uncertainty around that estimate. Keeping these quantities together helps distinguish the size of an estimated difference from the strength and precision of the evidence.

It also shows why the sponsor’s reported analysis and a simple recalculation should be distinguished. The registry-reported Cochran-Mantel-Haenszel analysis is the trial’s own result, while the unadjusted recalculation provides a transparent comparison from the posted counts. Their estimates are similar in direction, but method and scale matter when interpreting either one.

The endpoint itself highlights a central issue in clinical trial statistics: a composite outcome can represent multiple dimensions of disease activity, but its response definition must be understood to interpret the result. A participant counts as a responder only when all specified conditions are satisfied. The overall response contrast cannot identify which condition mattered most or establish effects on outcomes not included in the composite.

Finally, the registry record makes visible the distinction between statistical evidence and completeness of reporting. It provides a primary estimate and selected secondary and safety results, while some important details—such as multiplicity procedures, missing-data handling, and the original sample-size rationale—are not reported. Recognizing both the evidence and its reporting limits is essential to a careful statistical reading.

19. 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.

Several limitations are specific to what the registry record reports. The original sample-size justification and trial design assumptions are not reported, so the planning illustrations based on the observed effect cannot establish whether the original design targeted a particular effect size or power. The procedures used for randomization and allocation concealment are also not reported in detail.

The primary endpoint is composite, and the record does not provide component-level response counts. It is therefore not possible to determine from the overall BICLA count which components contributed most to the group difference. The record describes a Full Analysis Set of randomized participants who received at least one dose, but it does not give detailed participant flow or explain how missing Week 52 data and missing composite components were classified.

The registry does not report a multiplicity plan for secondary outcomes. Three detailed secondary comparisons are available, while further outcomes are listed without detailed comparative estimates in the record. The smaller subgroup denominators limit precision for those outcomes, and the absence of a reported testing hierarchy makes confirmatory interpretation difficult.

Safety reporting is also limited in the posted information. Serious adverse event participant counts and selected event terms are given, but the listed terms are not attributed to treatment groups in the record. The available counts do not provide a complete assessment of event timing, duration, exposure, attribution, or severity for all adverse events.

These limitations concern reporting and interpretation; they do not demonstrate that the trial was improperly conducted. They identify questions that cannot be resolved from the registry record alone.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Efficacy and Safety of Anifrolumab Compared to Placebo in Adult Subjects With Active Systemic Lupus Erythematosus — NCT02446899.
• Registry record last updated 2023-01-10; status last verified 2022-12.
• Results first posted 2020-03-18.
• 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.

23. Record Summary

NCT02446899 is a completed, randomized, double-blind, parallel-group Phase 3 study comparing anifrolumab 300 mg intravenous infusion with placebo, each administered every 4 weeks for 13 doses, in adults with active SLE. The registry reports 373 enrolled participants. The primary endpoint was BICLA response at Week 52 in participants randomized and given at least one dose.

The posted primary counts are 86 of 180 for anifrolumab and 57 of 182 for placebo. The registry-reported analysis gives a mean difference of 16.3, with a 95% confidence interval from 6.3 to 26.3 and p = 0.001. The unadjusted recalculation gives a risk difference of 0.165, with a 95% confidence interval from 0.065 to 0.264 and p = 0.001.

The record includes selected secondary comparisons and safety counts. Detailed information about the original sample-size rationale, multiplicity, and primary missing-data handling is not reported in the registry record.