Long Term Safety of Anifrolumab in Adult Subjects With Active Systemic Lupus Erythematosus: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Long Term Safety of Anifrolumab in Adult Subjects With Active Systemic Lupus Erythematosus (NCT02794285), a randomized phase 3 study. The primary outcome is Exposure-adjusted Incidence Rates (EAIRs) of Adverse Events of Special Interest (AESIs).

2. Clinical Question

3. Trial Design

This was a phase 3, randomized, quadruple-masked, parallel-group, placebo-controlled extension study. Its official title describes a multicentre study intended to characterize the long-term safety and tolerability of anifrolumab in adults with active systemic lupus erythematosus. The registry lists AstraZeneca as sponsor and the study status as completed. The recorded start date is 2016-06-30, and the primary completion date is 2021-12-21.

The registry reports total enrollment of 559. The primary safety outcome uses the FAS for the LTE study, defined as participants who were randomized and received at least one dose of investigational product in the LTE study. The listed analysis groups are Randomised Anifrolumab 300 mg, Placebo Feeder + Placebo LTE, Placebo Feeder + Anifrolumab 300 mg LTE, and Anifrolumab 150 mg Feeder + Anifrolumab 300 mg LTE. The feeder-treatment wording matters: the groups are not described solely by LTE treatment, and the comparison labels retain that treatment history.

The study purpose in the registry is to characterize long-term safety and tolerability of intravenous anifrolumab. The primary outcome timeframe is up to a maximum of 1114 days. This is an exposure-based safety assessment, rather than a report of a single outcome measured at one common follow-up visit. The registry record does not report further details of how participants entered the extension or the duration of exposure for each participant. It also does not report the trial's original sample-size justification or design assumptions.

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: The LTE population was comprised of participants who had completed the 52-week double-blind treatment period in one of the Phase III feeder studies (D3461C00004 \[NCT02446899\] or D3461C00005 \[NCT02446912\]), met all LTE eligibility criteria, and were willing to continue into the extension study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: FAS - LTE Study: consisted of all participants who were randomised and received at least 1 dose of investigational product in the LTE Study.

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 (FAS) - LTE Study: consisted of all participants who were randomised and received at least 1 dose of investigational product in the LTE Study.

6. Primary Endpoint

The registry lists the primary outcome as Exposure-adjusted Incidence Rates (EAIRs) of Adverse Events of Special Interest (AESIs), measured over Up to a maximum of 1114 days.

The primary endpoint is titled Exposure-adjusted Incidence Rates (EAIRs) of Adverse Events of Special Interest. Rates are expressed as events per 100 participant years. The registry defines the rate as the number of participants with an event divided by the total exposure time during the LTE study, including follow-up, and scaled to 100 participant years. Exposure for each participant in a period is calculated from the end date minus the start date plus one day. The outcome timeframe is up to a maximum of 1114 days.

The predefined adverse events of special interest were non-opportunistic serious infections, opportunistic infections, anaphylaxis, malignancy, herpes zoster, tuberculosis including latent tuberculosis, influenza, vasculitis other than systemic lupus erythematosus, and major cardiovascular events according to the Cardiovascular Event Adjudication Committee.

General statistical context: exposure-adjusted rates are useful when participants contribute different amounts of follow-up, because the denominator reflects time at risk rather than treating each participant's observation period as identical. The rate remains a summary of event occurrence over exposure; it does not by itself describe when events occurred, whether an individual had repeated events, or the cause of an event. The registry record does not report separate numerical rates for each predefined event type in the primary outcome results.

7. Results Posted in the Registry

• Randomised Anifrolumab 300 mg: 11.00 versus 9.60 (Events per 100 participant years) on control; rate ratio 1.15 (95% CI 0.77–1.71, approximate, Poisson model); p ≈ 0.507.
• Placebo Feeder + Anifrolumab 300 mg LTE: 12.90 versus 9.60 (Events per 100 participant years) on control; rate ratio 1.34 (95% CI 0.85–2.11, approximate, Poisson model); p ≈ 0.201.
• Anifrolumab 150 mg Feeder + Anifrolumab 300 mg LTE: 12.50 versus 9.60 (Events per 100 participant years) on control; rate ratio 1.30 (95% CI 0.78–2.18, approximate, Poisson model); p ≈ 0.317.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is an exposure-adjusted incidence rate. The registry describes its numerator as the number of participants with an event and its denominator as the sum of exposure time, including follow-up, for participants in the analysis set. That quantity is scaled to events per 100 participant years. Exposure in a period is calculated as end of period minus start of period plus one. The method expresses event occurrence in relation to time observed, which is relevant when participants have differing follow-up durations.

Why use an exposure-adjusted rate?

General statistical context: a simple count can be hard to compare when the groups have different amounts of observation time. A rate places event occurrence over a time-at-risk denominator, allowing the summary to reflect differences in accumulated exposure. It does not remove every source of bias, establish that the compared groups are exchangeable, or explain whether a rate reflects many participants with an event or a smaller number with repeated events. The registry's stated numerator is participants with an event, while the endpoint is presented as an event rate; details about repeat-event handling are not reported in the registry record.

How to read the rate ratios

The posted comparisons express each anifrolumab group rate relative to the placebo feeder plus placebo LTE rate. A rate ratio of 1 represents equal rates in the two groups; a value above 1 represents a higher estimated rate in the named group, and a value below 1 represents a lower estimated rate. These are general interpretations of a rate ratio. The reported estimates are 1.146, 1.344, and 1.302 for the three comparisons. They are accompanied by confidence intervals and unadjusted p-values. The results list a Poisson approach for the comparisons. The registry record does not report additional details of the model specification or its assumptions.

A Poisson model is commonly used for counts or rates. General statistical context: such models relate event counts to exposure, often using exposure time as an offset. A basic Poisson model assumes that the count variability is tied to its mean. Over-dispersion occurs when variability is greater than that assumption allows; if present and not addressed, uncertainty estimates can be too small. The registry record does not report a dispersion estimate or describe an over-dispersion adjustment for these results. No independent recalculation is needed to interpret the posted comparisons.

Intervals and tests

The confidence intervals describe uncertainty around the estimated rate ratios under the analysis method. Their exact confidence level is not reported in the registry record, so no level should be assumed. All three intervals include 1, and each is wide enough to cover both lower and higher relative rates than the reference value. The listed p-values are unadjusted: 0.507, 0.201, and 0.317. A p-value is a measure of how compatible the data are with a specified null hypothesis under the statistical model; it is not the probability that the null hypothesis is true, nor does it measure clinical importance. The registry record does not describe a multiplicity adjustment for the three comparisons or the predefined event categories.

The primary population is the FAS for the LTE study: randomized participants who received at least one dose of investigational product in the LTE study. The record does not provide a fuller account of exclusions, missing observations, censoring rules, or how incomplete follow-up was handled. Exposure-adjustment uses recorded time in the denominator, but the details of how that time was determined for participants with incomplete observation are not reported in the registry record.

The posted groups have different participant counts: 257, 112, 111, and 67. These are the counts listed for the four primary outcome groups. The record also gives the corresponding rates and comparisons, but does not report further baseline characteristics or covariate-adjusted estimates. Consequently, the rate ratios should be treated as the posted comparisons, not as adjusted effects accounting for other participant characteristics.

There are no posted analysis entries listed in the registry record, although the outcome results include rate estimates, intervals, and p-values. The record does not report a full statistical analysis plan, the prespecified confidence level, an interim analysis plan, or the original sample-size justification. The planning illustrations are separate from the trial's own design: they use the observed effect and stated planning assumptions, and they do not reconstruct the original protocol.

10. Confidence Intervals and Precision

Precision is reflected in the width of an estimate's confidence interval. A narrower interval indicates less statistical uncertainty under the method used; a wider interval leaves a broader range of plausible values. General statistical context: an interval does not assign a probability to the true value after the data have been observed. Its interpretation depends on the method and confidence level, and the confidence level is not reported in the registry record.

For Randomised Anifrolumab 300 mg versus Placebo Feeder + Placebo LTE, the rate ratio is 1.146, with a reported interval from 0.766 to 1.713. For Placebo Feeder + Anifrolumab 300 mg LTE, the rate ratio is 1.344, with an interval from 0.854 to 2.114. For Anifrolumab 150 mg Feeder + Anifrolumab 300 mg LTE, the rate ratio is 1.302, with an interval from 0.776 to 2.184. All three intervals include 1, the value representing equal rates. Each interval also includes values below and above 1, so the posted estimates do not tightly pin down the direction or size of a relative difference.

The group counts are 257, 112, 111, and 67. Differences in group size and exposure contribute to the information available for estimating rates, but participant count alone does not describe total exposure or the distribution of event times. The record does not report confidence intervals for the individual EAIRs, nor does it report the confidence level for the rate-ratio intervals. The unadjusted p-values are 0.507, 0.201, and 0.317; these do not replace the intervals as a description of the range of uncertainty.

11. 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 participants per arm for 80% power and participants per arm for 90% power. They are planning illustrations based on the observed effect, not the trial's original design calculations. The stated assumptions are two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation; each planning scenario uses the stated fraction of the observed effect. The original trial sample-size assumptions and justification are not reported in the registry record. These figures are not observed power.

For the Randomised Anifrolumab 300 mg comparison, planning for 100% of the observed effect gives 2,709 participants per arm for 80% power and 3,627 for 90% power. Planning for 75% of the observed effect gives 4,894 and 6,552; planning for 50% gives 11,192 and 14,982. For Placebo Feeder + Anifrolumab 300 mg LTE, the corresponding pairs are 536 and 717, 984 and 1,317, and 2,288 and 3,063. For Anifrolumab 150 mg Feeder + Anifrolumab 300 mg LTE, they are 681 and 911, 1,246 and 1,667, and 2,888 and 3,866.

The 80% and 90% columns describe the planned probability of detecting the specified effect under the stated assumptions if that effect were present. The 90% target requires more participants than the 80% target in each displayed scenario. Smaller effects require more participants because they are harder to distinguish from random variation. These are hypothetical planning quantities; they are not evidence that the completed trial had a particular probability of detecting an effect.

12. 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, λ₀ and λ₁ are event rates per patient-year, and T is follow-up per participant in years. This assumes Poisson counts and equal follow-up.

Worked calculation for this trial

1. Control rate λ₀ = 0.0960; treatment rate λ₁ = 0.1100 events per patient-year; follow-up T = 3.05 years.
2. n = (1.96 + 0.842)² × (1/0.0960 + 1/0.1100) / (3.05 × [ln(1.146)]²) = 2,708.9, rounded up to 2,709 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
lambda0 <- 0.096                      # control rate, events per patient-year
lambda1 <- 0.11                      # treatment rate
tt      <- 3.04996577686516                       # follow-up per participant, years
n_per_arm <- function(power, frac = 1) {
  l1 <- lambda0 * (lambda1 / lambda0)^frac   # smaller effect = fraction of the observed log rate ratio
  ceiling(round((z_a + qnorm(power))^2 * (1 / lambda0 + 1 / l1) / (tt * log(l1 / lambda0)^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 = ",")))

# Simulation 1: Poisson counts. Simulation 2: over-dispersed counts (variance = 1.5 x mean), analysed with the matching variance.
set.seed(2794285)
n    <- n_per_arm(0.80)
reps <- 20000
m0 <- lambda0 * n * tt; m1 <- lambda1 * n * tt
x0 <- rpois(reps, m0); x1 <- rpois(reps, m1)
z  <- log(x1 / x0) / sqrt(1 / x0 + 1 / x1)
cat(sprintf("Simulated power, Poisson counts, n = %s per arm: %.3f\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE)))
phi <- 1.5
y0 <- rnbinom(reps, mu = m0, size = m0 / (phi - 1)); y1 <- rnbinom(reps, mu = m1, size = m1 / (phi - 1))
z2 <- log(y1 / y0) / sqrt(phi * (1 / y0 + 1 / y1))
cat(sprintf("Simulated power, over-dispersed counts (phi = 1.5), same n: %.3f\n", mean(abs(z2) > z_a, na.rm = TRUE)))

Output from running the code

100% of observed effect: n per arm = 2,709 (80% power), 3,627 (90% power)
 75% of observed effect: n per arm = 4,894 (80% power), 6,552 (90% power)
 50% of observed effect: n per arm = 11,192 (80% power), 14,982 (90% power)
Simulated power, Poisson counts, n = 2,709 per arm: 0.798
Simulated power, over-dispersed counts (phi = 1.5), same n: 0.625

The first simulation draws Poisson counts and should land close to the target power. The second allows extra variability (variance 1.5 times the mean, a purely illustrative value), which lowers the power at the same sample size, so rate trials usually need more participants than the Poisson formula suggests.

13. Multiplicity, Missing Data and Analysis Populations

The primary outcome covers multiple predefined categories of adverse events of special interest, and the posted results include three rate comparisons against the placebo feeder plus placebo LTE group. The p-values are identified as unadjusted. The registry record does not report whether a multiplicity procedure was planned or used, how the family of comparisons was defined, or whether any interpretation was adjusted for testing multiple event categories or treatment comparisons. The unadjusted p-values should therefore be read as individual comparison results, not as an account of family-wise error control.

The analysis population is defined as participants randomized and receiving at least one dose of investigational product in the LTE study. This definition identifies who is included in the stated FAS, but does not by itself explain how missing outcome information or incomplete exposure was handled. The registry record does not report the missing-data method, rules for censoring or follow-up termination, or sensitivity analyses evaluating missingness. The endpoint definition includes follow-up in exposure, but the handling of specific patterns of missing data is not described.

The record reports no posted analysis entries, and it does not include a full statistical analysis plan, an interim analysis plan, or the original sample-size justification. The registry results do provide the outcome rates, rate ratios, intervals, and unadjusted p-values. Details beyond those reported, including whether adjustments for baseline characteristics were performed, are not reported in the registry record. No assumption about such procedures should be inferred from the presence of the posted estimates.

14. 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: systemic lupus erythematosus (5.4%), pneumonia (4.5%), covid-19 pneumonia (3.7%), herpes zoster (2.7%), covid-19 (2.4%).

The safety summaries list the number of participants at risk and the number affected by serious adverse events, other adverse events, and deaths for each group. For serious adverse events, the figures are 56 of 257 in Randomised Anifrolumab 300 mg, 27 of 112 in Placebo Feeder + Placebo LTE, 26 of 111 in Placebo Feeder + Anifrolumab 300 mg LTE, and 14 of 67 in Anifrolumab 150 mg Feeder + Anifrolumab 300 mg LTE. The corresponding counts affected by other adverse events are 195, 79, 81, and 53. The reported death counts are 3, 1, 3, and 2, respectively.

The listed serious adverse-event terms include Systemic lupus erythematosus at 5.4%, Pneumonia at 4.5%, COVID-19 pneumonia at 3.7%, Herpes zoster at 2.7%, COVID-19 at 2.4%, and Gastroenteritis at 1.8%. These summaries should not be added to or substituted for the exposure-adjusted primary endpoint. The registry record does not report a causal assessment in these figures, and the counts alone do not establish that treatment caused an event.

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

The posted results describe exposure-adjusted rates of predefined adverse events of special interest during an extension study and compare three anifrolumab 300 mg groups with the placebo feeder plus placebo LTE group. The reported rates were 11.0, 9.6, 12.9, and 12.5 events per 100 participant years across the four listed groups. The rate ratios for the three comparisons were 1.146, 1.344, and 1.302, and their reported intervals all included 1. The estimates and intervals show uncertainty about the relative rates; they do not establish a clear difference between the groups.

The results do not establish that anifrolumab caused or prevented a particular event. The endpoint is a safety measure, and it does not answer whether treatment changes lupus disease activity or other clinical outcomes. The event categories are grouped into the primary outcome, and separate numerical rate estimates for each category are not reported in the results. Serious adverse-event counts and the listed event terms provide additional safety summaries, but they are not the same measure as the exposure-adjusted endpoint.

The comparisons also retain feeder-study treatment history in their labels. They should not be interpreted as clean comparisons of LTE treatment alone, independent of prior treatment. Moreover, an exposure-adjusted rate accounts for time in the denominator but cannot, on its own, account for differences in other participant characteristics or study experience. The record does not report the adjustments or detailed missing-data procedures needed to assess those issues further.

These findings are best understood as registry-reported safety summaries for the studied groups and follow-up period. They are not a complete assessment of every possible safety outcome, and they do not support conclusions beyond the outcomes and methods reported in the registry record.

16. Why This Trial Matters Statistically

Long-term safety studies raise a statistical challenge: participants may contribute different amounts of time under observation, and adverse events are interpreted in relation to that exposure. The EAIR expresses occurrence per 100 participant years, making the denominator explicit rather than presenting only a count of participants with events. This helps readers understand the scale of observed event rates in an extension setting.

The result also illustrates why point estimates need to be read together with uncertainty. The three rate-ratio estimates are above 1, but their intervals include 1 and extend across values on either side. The unadjusted p-values likewise do not provide a precise account of the range of plausible relative rates. The statistical question is therefore not answered by looking only at the direction of a point estimate.

Group labels preserve feeder-study treatment history, which is important for interpreting comparisons in an extension study. A rate difference between groups could reflect more than the LTE treatment assignment alone, and the posted comparisons should not be detached from those labels. The record also distinguishes the primary exposure-adjusted endpoint from participant counts for serious and other adverse events.

Finally, the planning illustrations show how strongly hypothetical sample-size requirements can depend on the effect size targeted. They are useful for understanding future study planning, but they do not describe the design rationale or power of this completed trial. The registry record's missing analysis details, including multiplicity and missing-data handling, are themselves relevant to a careful statistical reading.

17. 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.
• Rate comparisons use a Poisson approximation unless the sponsor posted confidence limits; real event counts are often over-dispersed.

The registry record gives a defined primary endpoint and reported rate comparisons, but does not provide the full statistical analysis plan, the confidence level for the reported intervals, or detailed model specifications. It does not report an over-dispersion assessment or adjustment. These omissions limit how fully the assumptions behind the uncertainty estimates can be evaluated.

The record does not describe the handling of missing data, incomplete follow-up, censoring, or repeat events. Although the endpoint definition specifies exposure time and includes follow-up, the operational rules for participants with incomplete observation are not reported. Separate rates for the individual predefined adverse events of special interest are not included in the results.

The three comparisons are reported with unadjusted p-values, and no multiplicity procedure is described. The groups also have different listed participant counts, and the comparison labels include different feeder-study treatment histories. Further participant characteristics or adjusted comparisons are not reported, so the posted estimates do not resolve potential differences beyond the treatment labels.

The registry reports enrollment of 559, while the primary endpoint results are given for the four listed groups. The record does not explain the relationship between total enrollment and the group counts used for this endpoint. The original sample-size rationale, interim analysis plan, and further details of extension entry are also not reported in the registry record. These gaps constrain interpretation but do not alter the values the registry reports.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Long Term Safety of Anifrolumab in Adult Subjects With Active Systemic Lupus Erythematosus — NCT02794285.
• Registry record last updated 2023-01-13; status last verified 2022-12.
• Results first posted 2023-01-13.
• 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.

21. Record Summary

NCT02794285 is a completed phase 3 randomized, quadruple-masked, parallel-group, placebo-controlled extension study sponsored by AstraZeneca. It enrolled 559 adults with active systemic lupus erythematosus, with a recorded start date of 2016-06-30 and primary completion date of 2021-12-21.

The study's stated purpose was to characterize long-term safety and tolerability of intravenous anifrolumab. Its primary outcome was exposure-adjusted incidence rates of adverse events of special interest, reported as events per 100 participant years for up to a maximum of 1114 days. The posted comparison rates were 11.0, 12.9, and 12.5 in the three anifrolumab 300 mg LTE groups, against 9.6 in the placebo feeder plus placebo LTE group. The three rate-ratio intervals included 1. The registry record does not report a full statistical analysis plan, multiplicity procedure, or detailed missing-data method.