A Study of the Efficacy and Safety of Upadacitinib (ABT-494) in Participants With Moderately to Severely Active Crohn's Disease Who Have Inadequately Responded to or Are Intolerant to Biologic Therapy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of the Efficacy and Safety of Upadacitinib (ABT-494) in Participants With Moderately to Severely Active Crohn's Disease Who Have Inadequately Responded to or Are Intolerant to Biologic Therapy (NCT03345836), a randomized phase 3 study. The primary outcome is Percentage of Participants With Clinical Remission Per Crohn's Disease Activity Index (CDAI) at Week 12.

2. Clinical Question

3. Trial Design

This was a Phase 3, randomized, parallel-group, placebo-controlled induction study sponsored by AbbVie. The official title describes a multicenter, randomized, double-blind study; the registry lists quadruple masking. The record identifies Crohn’s disease as the condition and records 624 participants enrolled. The study is marked completed. Its recorded start date is 2017-11-29, and its primary completion date is 2021-08-11.

The main comparison relevant to the primary endpoint took place in Part 1, the double-blind induction period. Participants received either matching placebo or upadacitinib 45 mg orally once daily for 12 weeks. The primary outcome was assessed at Week 12. The ITT1 population included randomized participants who received at least one dose of double-blind study drug during Part 1. The outcome counts for that analysis were 171 in the placebo group and 324 in the upadacitinib group.

The registry also describes later study parts. Part 2 was an open-label induction period with upadacitinib 45 mg. Part 3 was an extended-treatment period that included double-blind or open-label groups receiving upadacitinib 45 mg or 30 mg, depending on earlier assignment and clinical response at Week 12. These later groups should not be conflated with the original placebo-controlled comparison: the primary result concerns the two Part 1 groups.

The registry record does not report the original sample-size justification, detailed randomization procedures, or the full statistical analysis plan. The study’s masking is listed as quadruple, while the Part 1 description and official title call the induction study double-blind. The record does not explain the meaning of the masking label in further detail.

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: This study had 3 Parts: Part 1:randomized,double-blind,placebo-controlled InductionPeriod(IP); Part 2:Once enrollment for Part1 completed,participants were further enrolled in open-label,single-arm active IP to receive upadacitinib 45mg.Clinical non-responders from Parts1 and 2 entered Part3; Part 3:ExtendedTreatmentPeriod for non-responders from Part1 or 2 had 3 cohorts:Cohort 1=placebo participants from Part 1,Cohort2=upadacitinib participants from Part 1.Cohort3=participants from Part2.

Period 1: DB and OL Induction (12 Weeks)

Reasons for not completing this period:

Period 2: 12-Week Extended Treatment

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: ITT1 Population included all randomized participants who received at least one dose of DB study drug during Part 1.

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: Intent-to-Treat (ITT) Population for the 12-week DB Induction Period (Part 1) i.e., ITT1 included all randomized participants who received at least one dose of DB study drug during Part 1. ITT Population for the 12-week OL Induction Period (Part 2) i.e., ITT2 included all participants who received at least one dose of study drug in Part 2.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants With Clinical Remission Per Crohn's Disease Activity Index (CDAI) at Week 12, measured over Week 12.

The primary endpoint is the percentage of participants with clinical remission at Week 12 according to the Crohn’s Disease Activity Index. The registry defines remission as CDAI <150. CDAI measures disease activity using a one-week assessment of eight items: liquid or very soft stool frequency, abdominal pain, Crohn’s disease complications, abdominal mass, hematocrit, body weight, antidiarrheal use, and general condition. The total score ranges from 0 to about 600; higher scores indicate more severe disease. Scores below 150 represent remission, while scores over 450 represent very severe Crohn’s disease.

A remission endpoint turns a multi-item clinical activity score into a participant-level outcome: each person either meets the defined threshold or does not. The percentage therefore communicates how common the defined state was in each group, while the between-group comparison describes the difference in those percentages. The threshold gives the endpoint a stated clinical interpretation, but it does not capture every aspect of Crohn’s disease or every change in a participant’s symptoms.

The registry reports the primary result using non-responder imputation, incorporating multiple imputation to handle missing data due to COVID-19 (NRI-C). The ITT1 population included randomized participants who received at least one dose of double-blind study drug during Part 1. The endpoint is therefore interpreted in the context of that population and the specified approach to missing outcome data.

7. Results Posted in the Registry

• Part 1 (Double Blind): Upadacitinib 45 mg: 38.9% (126/324) versus 21.1% (36/171) on control; risk difference 17.8 percentage points (95% CI 9.7 to 25.9), risk ratio 1.85 (1.34–2.55); 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.

Percentage of Participants With Clinical Remission Per Patient-Reported Outcomes (PROs) at Week 12

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Part 1 (Double Blind): Upadacitinib 45 mg: 39.8% (129/324) versus 14.0% (24/171) on control; risk difference 25.8 percentage points (95% CI 18.3 to 33.2), risk ratio 2.84 (1.91–4.21); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Part 1 (Double Blind): Upadacitinib 45 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Participants With Endoscopic Remission at Week 12

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Part 1 (Double Blind): Upadacitinib 45 mg: 19.1% (62/324) versus 2.3% (4/171) on control; risk difference 16.8 percentage points (95% CI 12.0 to 21.6), risk ratio 8.18 (3.03–22.10); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Part 1 (Double Blind): Upadacitinib 45 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline in Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F) Total Score at Week 12

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • Part 1 (Double Blind): Upadacitinib 45 mg: mean 11.40 versus 3.90 on control; difference 7.50 (95% CI 5.17 to 9.83), standardized effect d = 0.66; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Part 1 (Double Blind): Upadacitinib 45 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes extend beyond the primary CDAI-defined remission endpoint. They include patient-reported clinical remission, endoscopic remission, corticosteroid discontinuation combined with CDAI remission among participants taking corticosteroids at baseline, fatigue, inflammatory bowel disease questionnaire score, and clinical response measures. The registry lists ten posted secondary outcomes, spanning symptom reports, endoscopic assessment, treatment use, and patient-reported functioning.

Among the secondary endpoints with detailed results described in the record, the displayed comparisons generally point in the same direction as the primary result: the upadacitinib group has a higher reported remission percentage or a larger reported improvement than placebo. The findings provide context across distinct outcome domains, but they do not measure exactly the same construct as CDAI remission. Patient-reported remission, endoscopic remission, and fatigue-score change each have their own definition and interpretation.

Secondary results carry less confirmatory weight when the registry record does not report how multiplicity was handled. Testing many outcomes can increase the chance of at least one apparently notable result arising by chance. The record lists the number of outcome measures and analyses, but the multiplicity procedure is not reported in the registry record. This means the secondary results are useful as supportive or exploratory evidence, but isolated p-values should not be interpreted as if each were the only planned test.

The full pattern across all posted secondary endpoints cannot be characterized from the detailed results. The secondary findings should therefore be considered alongside their own estimates and intervals, endpoint definitions, analysis populations, and any multiplicity plan if available. They complement the primary result rather than replace it.

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

• Percentage of Participants With Clinical Remission Per Patient-Reported Outcomes (PROs) at Week 12
• Percentage of Participants With Endoscopic Remission at Week 12
• Percentage of Participants Who Discontinued Corticosteroid Use for Crohn's Disease (CD) and Achieved Clinical Remission Per CDAI at Week 12, in Participants Taking Corticosteroids at Baseline
• Change From Baseline in Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F) Total Score at Week 12
• Change From Baseline in Inflammatory Bowel Disease Questionnaire (IBDQ) Total Score at Week 12
• Percentage of Participants Achieving Clinical Response 100 (CR-100) at Week 2
• Percentage of Participants Achieving Clinical Response 100 (CR-100) at Week 12
• Percentage of Participants With Clinical Remission Per Crohn's Disease Activity Index (CDAI) at Week 4

9. Statistical Interpretation

10. Statistical Methods Explained

What does the primary comparison estimate?

The primary outcome is binary at the participant level: remission is defined by whether the CDAI is below the stated threshold. The group percentages estimate the proportion meeting that definition in each arm. A risk difference expresses the contrast between the two proportions on an absolute scale. The sponsor reports an adjusted risk difference of 17.9, with a 95% confidence interval from 10.0 to 25.8. This is the registry’s reported adjusted comparison, not an independently reconstructed result.

The registry identifies the primary analysis method as Cochran-Mantel-Haenszel. In general, this method can estimate an association or treatment contrast while accounting for stratification factors. The registry record does not report the stratification factors or enough detail to reconstruct the adjustment. Accordingly, the adjusted estimate should be described as the sponsor’s reported analysis rather than as a calculation that can be reproduced from the arm totals alone.

How should the p-value be understood?

A p-value measures how incompatible the observed data are with a specified null hypothesis, under the statistical model and analysis assumptions. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The primary sponsor analysis reports p < 0.001 and is labeled a superiority analysis. The unadjusted recalculation reports p <0.001. These are differently reported analyses; the sponsor’s adjusted analysis is the appropriate source for the registry’s formal conclusion.

What does the confidence interval add?

A confidence interval gives a range of estimates compatible with the observed data under the method used. For the primary adjusted risk difference, the reported 95% interval extends from 10.0 to 25.8. It describes sampling uncertainty for the adjusted contrast, not the range of individual participant outcomes and not a guarantee that the true value lies in that interval. It also does not incorporate every potential source of bias or uncertainty.

How were missing outcomes handled?

The primary description states that results used non-responder imputation incorporating multiple imputation to handle missing data due to COVID-19, abbreviated NRI-C. Non-responder imputation generally classifies missing participants as not having achieved the response or remission outcome; the record additionally identifies multiple imputation for the COVID-19-related missing data. Multiple imputation is a general statistical approach that represents uncertainty about missing values by creating and combining analyses of imputed datasets. The registry record does not report the detailed imputation model, the amount or pattern of missing data, or sensitivity analyses that would show how results change under alternative assumptions.

What is the analysis population?

The primary endpoint population is ITT1, defined in the record as all randomized participants who received at least one dose of double-blind study drug during Part 1. This definition is narrower than a population that includes every randomized participant regardless of treatment receipt. The arm counts for the primary outcome are 171 for placebo and 324 for upadacitinib 45 mg. The enrollment total is 624, but the record does not explain in the registry record how participants not included in these outcome counts were distributed across study parts or analysis populations.

For the reported secondary outcomes, the methods vary. The registry identifies Cochran-Mantel-Haenszel for the binary secondary comparisons and MMRM for the FACIT-F change-from-baseline analysis. In general, a mixed model for repeated measures (MMRM) can analyze outcomes measured repeatedly over time while modeling within-participant correlation and using available observations under its assumptions. The posted FACIT-F result is an adjusted treatment difference. The registry record does not give the full model specification, covariance structure, or covariates.

Several details needed for a complete statistical audit are not reported in the registry record, including the full protocol-based analysis plan, the original sample-size justification, the multiplicity procedure across endpoints, and whether interim analyses were planned or conducted. These omissions limit conclusions about how all endpoints and analyses were prioritized.

11. Confidence Intervals and Precision

The primary adjusted risk difference is 17.9 percentage points, and its 95% confidence interval runs from 10.0 to 25.8. The interval is entirely above zero, so the reported adjusted comparison is not compatible with no difference under the stated interval method. Its width also matters: the data support a range of possible effect sizes rather than fixing the treatment contrast at 17.9. The interval should be read as uncertainty around the estimate, not as a range containing most individual participants’ outcomes.

The arm-specific estimates are 21.1% for placebo, with a 95% confidence interval from 14.9 to 27.2, and 38.9% for upadacitinib, with a 95% confidence interval from 33.6 to 44.2. These intervals overlap. Overlap or non-overlap of separate arm intervals is not the formal test of the between-group difference; the registry reports a separate adjusted comparison with its own confidence interval and p-value.

The displayed arm counts also differ, with 171 in the placebo group and 324 in the upadacitinib group. Precision depends on the amount of information available and on the analysis method, among other factors. The reported intervals account for sampling uncertainty as handled by their respective methods, but they do not resolve issues such as missing-data assumptions, generalizability, or unreported design details. The 95% level describes the long-run coverage property of the interval procedure under its assumptions; it is not a 95% probability statement about this particular interval containing the true effect.

Precision should therefore be judged from the estimate and its interval together. The primary estimate is positive in the reported adjusted analysis, while the interval indicates that the exact size remains uncertain. The statistical evidence does not, by itself, determine the clinical importance of any particular value within that interval.

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 sample-size figures are planning illustrations based on the observed effect, not the trial’s original design calculations. For these binary outcomes, the entries refer to participants per arm, not total events. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is the stated fraction of the observed effect. The original trial’s own sample-size assumptions and justification are not reported in the registry record.

The 80% power column describes the number of participants per arm in a hypothetical study designed to have an 80% chance of detecting the chosen effect under the stated assumptions. The 90% power column uses a higher target chance of detecting that effect, so it generally requires more participants per arm. These are design-planning concepts, not descriptions of what happened in this completed trial.

The rows for 100%, 75%, and 50% of the observed effect illustrate how the planning requirement changes if the effect a future study aims to detect is smaller than the observed estimate. Smaller effects are harder to distinguish from random variation and therefore require more participants per arm under otherwise similar assumptions. The counts are illustrative: they rely on the observed estimate and the listed simplified assumptions, and should not be treated as the trial’s actual enrollment target or design rationale.

These calculations are not observed power. Observed power is not a useful substitute for interpreting the estimated effect and its confidence interval after a study has been completed. The estimates and intervals describe the posted results; the planning figures address a different question about hypothetical future study designs.

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 = 0.211; treatment proportion p₁ = 0.389; average p̄ = 0.300.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6479 = 1.2698.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6355 = 0.5348.
4. n = (1.2698 + 0.5348)² / (0.1784)² = 102.4, rounded up to 103 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 <- 0.210526315789474                          # control proportion
p1 <- 0.388888888888889                          # 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(3345836)
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 = 103 (80% power), 137 (90% power)
 75% of observed effect: n per arm = 175 (80% power), 234 (90% power)
 50% of observed effect: n per arm = 374 (80% power), 501 (90% power)
power.prop.test: 102.4 
Simulated power at n = 103 per arm: 0.807 (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 sponsor’s primary analysis uses the Cochran-Mantel-Haenszel method and reports an adjusted risk difference of 17.9, with a 95% confidence interval from 10.0 to 25.8 and p < 0.001. The simple recalculation from the posted summary counts is unadjusted and reports a risk difference of 0.178, with an interval from 0.097 to 0.259 and p <0.001. The recalculation is expressed on a proportion scale; the sponsor’s estimate is expressed in percentage points.

The estimates are closely aligned in direction and magnitude, but they come from different analyses. The sponsor’s reported estimate is adjusted, while the recalculation uses only the posted group-level counts and does not reproduce the adjustment. The registry record does not provide enough detail to reconstruct all aspects of the sponsor’s analysis. Because there is no reported discrepancy, there is no conflict that requires a recalculated result to displace the sponsor’s result. The sponsor’s own analysis takes precedence for describing the formal registry finding.

The p-values are also reported at different thresholds. Both indicate strong statistical evidence against the relevant null under their respective analyses, but the adjusted sponsor result should be cited when discussing the trial’s reported primary analysis. The unadjusted calculation is best understood as a transparent check based on summary data, not as a correction to or replacement for the sponsor’s analysis.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates multiple outcomes, time points, treatment comparisons, or analysis populations. Each additional statistical test creates another opportunity for a small p-value to appear by chance, even when no effect exists for any individual test. Common approaches include prespecifying a hierarchy of endpoints or adjusting the threshold for inference across a family of tests. These are general methods, not procedures that can be attributed to this study without documentation.

The registry lists 13 outcome measures and 12 analyses, including a primary endpoint and multiple secondary outcomes. A multiplicity adjustment or testing hierarchy is not reported in the registry record. Without that information, the primary finding can be described using its reported analysis, but the degree of family-wise protection for secondary claims cannot be assessed. Secondary p-values should consequently be read cautiously rather than treated as independent confirmation.

Missing data are addressed in the primary endpoint description through non-responder imputation incorporating multiple imputation for COVID-19-related missing data. In general, missing-data methods rely on assumptions about why observations are missing and what outcomes might have been observed. The registry record does not report the amount of missing data, the detailed imputation model, the assumptions used for the COVID-19 component, or sensitivity analyses under alternative assumptions. Those details matter because results can be affected when missingness is related to outcomes or treatment experience.

The primary population is ITT1: randomized participants who received at least one dose of double-blind study drug during Part 1. It is not defined as all randomized participants regardless of whether they received study drug. The record does not provide a complete accounting in the registry record of participants excluded from this population or their outcomes. Further details about protocol deviations, analysis sets for every secondary endpoint, interim analyses, and the full statistical plan are not reported in the registry record.

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: crohn's disease (21.4%), abdominal abscess (7.1%), gastroenteritis (7.1%), hypokalaemia (7.1%), covid-19 pneumonia (1.6%).

In Part 1, serious adverse events were reported for 17 of 171 participants in the placebo group and 30 of 324 participants in the upadacitinib 45 mg group. Other adverse events were reported for 69 of 171 participants in the placebo group and 126 of 324 participants in the upadacitinib group. The registry records zero deaths in the placebo group and one death in the upadacitinib group.

These are counts of participants with reported events, not a causal comparison. The record’s serious-adverse-event groups also include later open-label and extended-treatment parts, with different treatment histories and group sizes. Those groups should not be combined with the randomized Part 1 comparison as though they were equivalent randomized arms. Serious-event terms listed in the record include Crohn’s disease, abdominal abscess, gastroenteritis, hypokalaemia, COVID-19 pneumonia, and febrile neutropenia; the listed percentages are registry descriptions, not evidence that treatment caused those events.

Safety counts are descriptive and should be interpreted with the study’s exposure periods, group sizes, event definitions, and attribution information. A comprehensive causal safety assessment is not possible from the counts alone.

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

The registry results show that, under the study’s CDAI definition and primary analysis, a higher percentage of participants in the upadacitinib 45 mg group met the Week 12 remission threshold than in the placebo group. The adjusted treatment difference was reported as 17.9 percentage points, with a 95% confidence interval from 10.0 to 25.8 and p < 0.001. This is a description of the randomized Part 1 comparison and its reported statistical analysis.

The result does not mean every participant in the upadacitinib group achieved remission, nor that no participant in the placebo group did. It does not predict an individual participant’s outcome. The endpoint classifies remission using a defined CDAI threshold and cannot represent every symptom, complication, or longer-term disease outcome.

The result also does not by itself establish the clinical importance of the estimated difference. Statistical evidence and clinical importance are related but distinct questions. The confidence interval conveys uncertainty in the estimated group contrast; judgments about importance require clinical context that is not established by the p-value alone.

The registry describes a 12-week induction comparison, alongside open-label and extended-treatment parts. The primary result should not be read as evidence about every later treatment period, every dose, or every patient population. The study population was limited to people with moderately to severely active Crohn’s disease who had inadequately responded to or were intolerant to biologic therapy, and the primary analysis used the defined ITT1 population.

Finally, a superiority result for the primary endpoint is not a complete benefit-risk conclusion. Safety counts are descriptive, and the record does not provide all details needed to evaluate causality or longer-term outcomes. The data support interpretation of the posted comparison, not individualized medical advice.

18. Why This Trial Matters Statistically

This trial illustrates how a clinical endpoint, analysis population, missing-data rule, and treatment comparison jointly shape the interpretation of a result. The primary endpoint is not a vague judgment of improvement: it is a remission classification based on a stated CDAI threshold. That definition makes the target outcome explicit and allows the reported percentages and risk difference to be interpreted against a common criterion.

The contrast between arm percentages and the sponsor’s adjusted risk difference also demonstrates why a trial’s reported analysis should be distinguished from a simple calculation using summary counts. The sponsor used a Cochran-Mantel-Haenszel method, while the recalculation is unadjusted. Their close alignment is informative, but the adjusted analysis remains the registry’s formal result and cannot be reconstructed fully without the analysis details.

The primary analysis also makes the handling of missing outcomes visible. NRI-C is specified, including multiple imputation to address missing data due to COVID-19. That detail matters because the treatment contrast depends not only on observed responses but also on how unavailable outcomes are represented. The record’s lack of further imputation details shows why reporting a method label is useful yet not always sufficient for independent evaluation.

Finally, the registry includes multiple secondary outcomes and several study parts. This makes endpoint prioritization and multiplicity relevant to the strength of claims beyond the primary result. The example underscores a core biostatistical practice: interpret an estimate with its interval, population, analysis method, outcome definition, and reporting limitations together.

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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary comparison is limited to a 12-week induction period. The registry also describes open-label and extended-treatment groups, but the primary randomized placebo-controlled result does not answer questions about every later phase or treatment history. The primary endpoint is a threshold-based CDAI remission measure, so it does not summarize all dimensions of Crohn’s disease or longer-term outcomes.

The analysis population is defined as randomized participants who received at least one dose of double-blind study drug during Part 1. This definition is narrower than all randomized participants, and the registry entries summarized do not provide a full accounting of participants outside the primary outcome population. The registry specifies NRI-C for the primary outcome but does not give the detailed imputation model, missing-data amount, or sensitivity analyses needed to evaluate robustness to alternative assumptions.

The statistical reporting also leaves important design and analysis details unavailable. The original sample-size justification, full analysis plan, multiplicity procedure, and interim-analysis information are not reported in the registry record. The reported adjusted primary analysis is identified as Cochran-Mantel-Haenszel, but the adjustment factors are not specified in the registry record. That limits independent reconstruction of the sponsor’s estimate.

Secondary outcomes assess different domains and are numerous. Without a reported multiplicity approach, secondary p-values warrant caution. Safety counts span groups from different study parts and treatment histories; the counts alone do not establish causation, and they should not be treated as a single randomized comparison. These limitations do not negate the posted primary result, but they define what can and cannot be concluded from the registry information.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study of the Efficacy and Safety of Upadacitinib (ABT-494) in Participants With Moderately to Severely Active Crohn's Disease Who Have Inadequately Responded to or Are Intolerant to Biologic Therapy — NCT03345836.
• Registry record last updated 2022-08-15; status last verified 2022-07.
• Results first posted 2022-08-15.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

23. Record Summary

NCT03345836 is a completed Phase 3 randomized, parallel-group study of upadacitinib 45 mg versus matching placebo as double-blind induction therapy for moderately to severely active Crohn’s disease in participants who inadequately responded to or were intolerant to biologic therapy. Part 1 treatment was oral once daily for 12 weeks. The registry lists 624 participants enrolled.

The primary Week 12 endpoint was CDAI-defined clinical remission, with remission defined as CDAI <150. The sponsor reported an adjusted risk difference of 17.9 percentage points, with a 95% confidence interval from 10.0 to 25.8 and p < 0.001. Results used NRI-C, including multiple imputation for COVID-19-related missing data, in the ITT1 population.

The record includes secondary outcomes, safety counts, and later open-label or extended-treatment parts. The original sample-size justification, multiplicity procedure, and detailed statistical analysis plan are not reported in the registry record.