Protege Encore Study- Clinical Trial of Teplizumab (MGA031) in Children and Adults With Recent-Onset Type 1 Diabetes Mellitus: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Protege Encore Study- Clinical Trial of Teplizumab (MGA031) in Children and Adults With Recent-Onset Type 1 Diabetes Mellitus (NCT00920582), a randomized phase 3 study. The primary outcome is Proportion of Subjects With Both a Total Daily Insulin Dose of Less Than 0.5 U/kg/Day and Hemoglobin A1c (HbA1c) Level of Less Than 6.5%..

2. Clinical Question

The endpoint combines two conditions: participants had to meet the insulin-dose threshold and the HbA1c threshold together to count. It is therefore not a result for either measure considered on its own. The question is framed as a comparison of each teplizumab regimen with placebo, rather than as a direct comparison among the three active regimens. The registry record identifies the study as phase 3, randomized, double-blind, multinational, and placebo-controlled. The listed outcome is assessed at a specified follow-up time, 52 weeks after randomization.

3. Trial Design

This was a phase 3, randomized, parallel-group study with triple masking, according to the registry record. Its official title describes a double-blind, multinational, placebo-controlled study. The study evaluated three teplizumab regimens and a placebo arm in children and adults with recent-onset type 1 diabetes mellitus. The record lists enrollment of 254 participants and identifies MacroGenics as the sponsor. The status is terminated; a reason for termination is not reported in the registry record.

The Herold Regimen was a 14-day course of intravenous teplizumab, with daily doses of 51 µg/m2, 103 µg/m2, 207 µg/m2, and 413 µg/m2 on Study Days 1-4, followed by one dose of 826 µg/m2 on each of Study Days 5-14. It was scheduled to repeat at Week 26. The 33.3% Herold Regimen used daily doses of 17 µg/m2, 34 µg/m2, 68 µg/m2, and 136 µg/m2 on Study Days 1-4, followed by 273 µg/m2 on Study Days 5-14, also repeated at Week 26. The Curtailed Herold Regimen used the first four escalating doses and 826 µg/m2 on Study Days 5-6, followed by intravenous placebo on Study Days 7-14; it too was scheduled to repeat at Week 26. The placebo arm received daily intravenous placebo for 14 days, repeated at Week 26.

The primary outcome was assessed 52 weeks after randomization. The primary outcome denominators were 63 participants for the Herold Regimen, 66 for the 33.3% Herold Regimen, 63 for the Curtailed Herold Regimen, and 62 for placebo. How the analysis population was defined, and how participants were handled if outcome data were unavailable, are not reported in the registry record. The original design assumptions and the reason for the recorded termination are also not reported in the registry record.

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.

Overall Study

Reasons for not completing this period:

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Proportion of Subjects With Both a Total Daily Insulin Dose of Less Than 0.5 U/kg/Day and Hemoglobin A1c (HbA1c) Level of Less Than 6.5%., measured over 52 weeks after randomization.

The primary endpoint counts participants who met both of two criteria at 52 weeks after randomization: a total daily insulin dose of less than 0.5 U/kg/day and an HbA1c level of less than 6.5%. A participant who met only one criterion would not meet the combined endpoint as defined. The registry lists the measure as a count of participants, so the analysis compares the number meeting both conditions in each treatment arm with the number in that arm’s outcome denominator.

A combined endpoint can represent a clinically relevant state that depends on more than one aspect of disease management. Here, the insulin-dose condition concerns the amount of daily insulin, while HbA1c is a measure of blood sugar control over time. Requiring both criteria makes the result more specific to participants who reached the joint target, but it does not reveal how either component behaved on its own. The registry lists other outcomes, including HbA1c and insulin-related measures, but the primary endpoint should be interpreted as the joint count rather than as a substitute for those separate measures.

The endpoint is binary at the participant level: each participant either meets both thresholds or does not. The observed counts were 8 of 63 for the Herold Regimen, 8 of 66 for the 33.3% Herold Regimen, 9 of 63 for the Curtailed Herold Regimen, and 6 of 62 for placebo. These counts describe the posted results at the stated follow-up; they do not by themselves establish a treatment effect. The comparative estimates and their uncertainty are necessary to assess how compatible the data are with different possible between-arm effects.

7. Results Posted in the Registry

• Herold Regimen: 12.7% (8/63) versus 9.7% (6/62) on control; risk difference 3.0 percentage points (95% CI -8.0 to 14.1), risk ratio 1.31 (0.48–3.56); unadjusted two-proportion z-test p = 0.592.
• 33.3% Herold Regimen: 12.1% (8/66) versus 9.7% (6/62) on control; risk difference 2.4 percentage points (95% CI -8.3 to 13.2), risk ratio 1.25 (0.46–3.41); unadjusted two-proportion z-test p = 0.658.
• Curtailed Herold Regimen: 14.3% (9/63) versus 9.7% (6/62) on control; risk difference 4.6 percentage points (95% CI -6.7 to 16.0), risk ratio 1.48 (0.56–3.90); unadjusted two-proportion z-test p = 0.428.

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.

The Change in Beta-cell Function as Measured by C-peptide Secretory Response Following a Mixed Meal

• 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 deviations.
• Result:
  • Herold Regimen: mean 0.04 versus -0.04 on control; difference 0.08 (95% CI -0.11 to 0.27), standardized effect d = 0.23; z-test p = 0.409.
  • 33.3% Herold Regimen: mean -0.05 versus -0.04 on control; difference -0.01 (95% CI -0.24 to 0.22), standardized effect d = -0.02; z-test p = 0.932.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Herold Regimen and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between 33.3% Herold Regimen and the control. The interval gives the range of effects that remain compatible with the data.

Number of Subjects With Both a Total Daily Insulin Dose of Less Than 0.5 U/kg/Day and Hemoglobin A1c (HbA1c) Level of Less Than 6.5%.

• 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:
  • Herold Regimen: 7.1% (1/14) versus 5.9% (1/17) on control; risk difference 1.3 percentage points (95% CI -16.3 to 18.8), risk ratio 1.21 (0.08–17.71); unadjusted two-proportion z-test p = 0.887.
  • 33.3% Herold Regimen: 15.4% (2/13) versus 5.9% (1/17) on control; risk difference 9.5 percentage points (95% CI -13.1 to 32.1), risk ratio 2.62 (0.27–25.81); unadjusted two-proportion z-test p = 0.390.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Herold Regimen and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between 33.3% Herold Regimen and the control. The interval gives the range of effects that remain compatible with the data.

The Proportion of Subjects Who Have Both a Total Daily Insulin Dose < 0.5 U/Kg/Day and Hemoglobin A1c (HbA1c) Level < 7.0%

• 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:
  • Herold Regimen: 20.6% (13/63) versus 11.3% (7/62) on control; risk difference 9.3 percentage points (95% CI -3.4 to 22.1), risk ratio 1.83 (0.78–4.27); unadjusted two-proportion z-test p = 0.154.
  • 33.3% Herold Regimen: 15.2% (10/66) versus 11.3% (7/62) on control; risk difference 3.9 percentage points (95% CI -7.8 to 15.6), risk ratio 1.34 (0.54–3.31); unadjusted two-proportion z-test p = 0.520.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Herold Regimen and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between 33.3% Herold Regimen and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes broaden the picture beyond the primary joint threshold. They include change in C-peptide secretory response after a mixed meal, combined insulin-dose and HbA1c thresholds at other follow-up times, and other listed measures such as daily insulin injections, hypoglycemia events, HbA1c change, and adverse events. The registry record identifies nine posted secondary outcomes, though the available results do not cover every listed outcome in the same detail. These measures address related but distinct aspects of the study; a finding on one cannot stand in for all the others.

Across the available secondary results, the evidence is not a uniform pattern of clearly separated active and placebo groups. The reported C-peptide analyses and the posted combined-threshold comparisons have uncertainty, and the listed estimates do not establish consistent superiority across outcomes. This broad pattern aligns with the primary result in the limited sense that neither set of reported comparisons provides a precise, decisive separation. Secondary results should not be treated as confirmation or refutation of the primary endpoint without considering their different definitions, follow-up times, denominators, and methods.

In general, multiple outcomes create more opportunities for apparently noteworthy results to arise by chance when each is examined separately. An unadjusted secondary p-value does not account for the full set of comparisons. The registry record does not report whether a multiplicity strategy was prespecified or applied to the secondary analyses. For that reason, unadjusted secondary findings carry less weight than a clearly designated primary analysis and should be viewed as supportive or exploratory rather than as stand-alone evidence. The registry does not report enough information to determine how the secondary outcome results were prioritized in the original analysis plan.

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

• The Change in Beta-cell Function as Measured by C-peptide Secretory Response Following a Mixed Meal
• The Change in Beta-cell Function as Measured by C-peptide Secretory Response Following a Mixed Meal
• Mean Number of Total, Major, Minor and Nocturnal Hypoglycemia Events
• Mean Number of Daily Insulin Injections
• Number of Subjects With Both a Total Daily Insulin Dose of Less Than 0.5 U/kg/Day and Hemoglobin A1c (HbA1c) Level of Less Than 6.5%.
• The Proportion of Subjects Who Have Both a Total Daily Insulin Dose < 0.5 U/Kg/Day and Hemoglobin A1c (HbA1c) Level < 7.0%
• The Mean HbA1c Change From Baseline
• Number of Participants With Adverse Events

9. Statistical Interpretation

10. Statistical Methods Explained

The primary endpoint is binary because each participant either met both stated criteria or did not. The arm-level results are counts and denominators. The re-calculations compare the proportion meeting the endpoint in each teplizumab arm with the proportion in the placebo arm, using an unadjusted risk difference. A risk difference is the active-arm proportion minus the comparator proportion. A value above zero means the observed proportion was higher in the active arm; a value below zero means it was lower. The estimate is descriptive of these trial data and does not by itself account for uncertainty.

What does a confidence interval describe?

A confidence interval gives a range of values compatible with the data under the method used. For a risk difference, zero represents no difference between the proportions. When an interval includes zero, the data are compatible with no difference and with some effects on either side of zero. Interval width matters as well: a wide interval indicates less precision and leaves more possible effect sizes consistent with the observations. The intervals for the primary risk differences include zero and extend on both sides, so they do not identify a precise difference.

How does an odds ratio differ from a risk difference?

An odds ratio compares the odds of meeting the endpoint in two arms. An odds ratio of 1 represents equal odds; values above or below 1 express different relative odds. Odds ratios and risk differences describe the same binary outcome using different scales, so their numerical values are not directly interchangeable. The registry’s reported primary analyses use odds ratios, while the additional unadjusted re-calculations use risk differences. A confidence interval for an odds ratio is interpreted relative to 1, rather than zero.

What method did the registry report?

For the three primary comparisons, the registry reports Mantel Haenszel odds ratios. This method combines information across strata when estimating an association. The registry record does not describe the stratification variables, so their role cannot be specified further. The registry reports the odds ratio, a 95% confidence interval, and a p-value for each comparison. These are the sponsor’s reported analyses and should be distinguished from the simple unadjusted re-calculations, which use posted summary counts and denominators.

Why might analysis methods give different estimates?

Methods can differ because they use different effect scales, adjustment choices, or analysis assumptions. A risk difference is an absolute contrast in proportions, whereas an odds ratio is a relative contrast in odds. Mantel Haenszel estimation can also account for stratification in a way that a simple unadjusted comparison does not. Therefore, small differences between estimates from different methods are not necessarily contradictions. The registry record does not report enough detail to reconstruct all modeling or adjustment decisions.

How should p-values be read?

A p-value measures how surprising the observed data, or data more extreme, would be under a specified null hypothesis and analysis method. It is not the probability that the null hypothesis is true, nor does it measure the clinical importance of a finding. The registry’s primary p-values are 0.609, 0.601, and 0.400. The re-calculation p-values are 0.592, 0.658, and 0.428. These values should be read together with the effect estimates and confidence intervals, not as stand-alone evidence. The record does not report whether or how the study’s statistical plan adjusted for multiple comparisons.

For the posted continuous secondary outcome of change in C-peptide secretory response, the registry reports ANCOVA estimates of mean difference in final values for two comparisons. ANCOVA is a model-based method that can compare groups while adjusting for covariates; the covariates used are not reported in the registry record. The posted re-calculations for this outcome instead compare arm summaries as unadjusted mean differences. A mean difference is expressed in the outcome’s units, nmol/L. Its confidence interval describes uncertainty around the estimated difference, with zero corresponding to no difference.

These methods address different questions and rely on their stated analysis approaches. Registry analyses should not be treated as if they were identical to unadjusted re-calculations from summary statistics. The methods and findings must also be interpreted alongside outcome denominators and the unreported details of the analysis population and missing-data handling.

11. Confidence Intervals and Precision

Precision is best judged from the confidence intervals, not from the point estimates alone. For the primary risk-difference re-calculations, the Herold Regimen estimate is 0.030 with a confidence interval from -0.080 to 0.141. The 33.3% Herold Regimen estimate is 0.024, with an interval from -0.083 to 0.132. The Curtailed Herold Regimen estimate is 0.046, with an interval from -0.067 to 0.160. All three intervals include zero and extend to both negative and positive values. This leaves uncertainty about both the direction and the size of the underlying difference.

The registry’s own primary analyses also have broad 95% confidence intervals: 0.431 to 4.219 for the Herold Regimen, 0.453 to 4.230 for the 33.3% Herold Regimen, and 0.521 to 5.066 for the Curtailed Herold Regimen. These are odds-ratio intervals, for which 1 represents no difference. Each interval includes 1 and spans values on both sides. Thus, the registry’s analyses likewise do not estimate the comparisons precisely.

In general, confidence-interval width reflects the amount of information available and the variability of the outcome under the analysis method. Here, the primary outcome counts are modest in each arm, and the confidence intervals are broad relative to their point estimates. The intervals are not a guarantee that the true effect lies within their limits; they express uncertainty under the assumptions of the method used. Precision should not be confused with clinical importance: a precise estimate could still be small, while an imprecise estimate may leave a wide range of effects plausible.

The reported comparisons do not establish equivalence or rule out potentially meaningful differences. The registry record does not report an equivalence margin or a non-inferiority margin. The uncertainty is therefore best described directly: the point estimates favor a higher observed endpoint proportion in each active arm, but the intervals also allow no difference and effects in the opposite direction.

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 own original design assumptions. For these binary outcomes, the figures are participants per arm. They are not the total number of participants across all arms and are not counts of 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 assumed effect to detect is the stated fraction of the observed effect. The original trial’s own sample-size justification and assumptions are not reported in the registry record.

The 80% power column gives the planned participants per arm under those assumptions for a design with an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column gives the corresponding planning figure for a 90% chance. These are design targets, not descriptions of the probability that the trial’s observed result is correct. For example, for planning around 100% of the observed primary risk difference for the Herold Regimen comparison, the illustration gives 1,708 participants per arm for 80% power and 2,286 participants per arm for 90% power. For planning around 50% of that observed effect, the corresponding figures are 6,428 and 8,605 participants per arm. Smaller effects require more participants because they are harder to distinguish from random variation.

The sample-size curves express the same planning relationship across effects: as the target effect becomes smaller, the required participants per arm increase. The observed trial arm sizes were much smaller than many of the illustrated planning figures, but these illustrations use the trial’s observed effects and stated assumptions rather than the original design rationale. This is not observed power. It is a forward-looking illustration of sample sizes under specified hypothetical planning assumptions, and it should not be used to reinterpret the trial’s result as if it had a known probability of success.

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 = 0.097; treatment proportion p₁ = 0.127; average p̄ = 0.112.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4458 = 0.8737.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4453 = 0.3748.
4. n = (0.8737 + 0.3748)² / (0.0302)² = 1,707.9, rounded up to 1,708 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 <- 0.0967741935483871                          # control proportion
p1 <- 0.126984126984127                          # 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(920582)
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 = 1,708 (80% power), 2,286 (90% power)
 75% of observed effect: n per arm = 2,948 (80% power), 3,945 (90% power)
 50% of observed effect: n per arm = 6,428 (80% power), 8,605 (90% power)
power.prop.test: 1707.9 
Simulated power at n = 1,708 per arm: 0.809 (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 reports the sponsor’s primary comparisons as Mantel Haenszel odds ratios. For the Herold Regimen versus placebo, the reported odds ratio is 1.348, with a 95% confidence interval of 0.431 to 4.219 and p = 0.609. For the 33.3% Herold Regimen, the reported odds ratio is 1.356, with a 95% confidence interval of 0.453 to 4.230 and p = 0.601. For the Curtailed Herold Regimen, it is 1.624, with a 95% confidence interval of 0.521 to 5.066 and p = 0.400.

The additional re-calculations use unadjusted risk differences from the posted counts and denominators, rather than the sponsor’s odds-ratio method. Their estimates and p-values need not match the sponsor’s figures because the scales and methods differ. The sponsor’s reported analysis is the principal registry analysis; the re-calculation is a supplementary, unadjusted way to express the posted summary data. The registry record does not provide all details needed to reproduce the sponsor’s analysis, including the stratification details. Both sets of results show substantial uncertainty, and the sponsor’s confidence intervals include 1, the no-difference value for an odds ratio.

15. Multiplicity, Missing Data and Analysis Populations

The study has one listed primary endpoint and multiple listed secondary outcomes, including repeated or related measures of insulin use, HbA1c, C-peptide response, hypoglycemia, and adverse events. When several outcomes and comparisons are examined, the chance of at least one apparently unusual result can rise if each is tested separately without adjustment. In general, a prespecified multiplicity plan can define which comparisons are confirmatory and how error rates are controlled. Whether the original protocol used such a plan, and whether any adjustment was applied across the primary or secondary comparisons, are not reported in the registry record.

The sponsor’s reported primary analyses give Mantel Haenszel odds ratios for each teplizumab regimen versus placebo. The record does not report whether the three comparisons were part of a hierarchical testing strategy or whether the p-values were adjusted. The additional risk-difference estimates are unadjusted re-calculations and should not be interpreted as multiplicity-adjusted results. The registry also does not report interim analyses or stopping rules, so no conclusion about those procedures can be drawn from this record.

The primary outcome lists denominators of 63, 66, 63, and 62 participants across the four arms. Other posted outcome denominators differ, including for the C-peptide outcome and the 104-week combined endpoint. Such differences indicate that the analysis denominators vary by outcome, but the reason for each difference is not reported in the registry record. It is not possible to infer whether this reflects missing measurements, eligibility for a particular analysis, or another rule.

The analysis populations are not reported in the registry record. In particular, the record does not state whether the primary analysis followed an intention-to-treat, per-protocol, or other population definition. Missing-data methods, imputation rules, and sensitivity analyses are also not reported. These omissions matter because results can depend on who is included and how unavailable outcome data are handled. The posted estimates remain interpretable as reported comparisons, but the record does not support a full assessment of the robustness of those estimates to analysis-population or missing-data choices.

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: diabetic ketoacidosis (4.8%), hyperglycaemia (3.2%), hypoglycaemic seizure (3.2%), gastroenteritis (1.6%), lymphopenia (1.6%).

The serious adverse event summary lists 7 affected participants among 63 at risk in the Herold Regimen arm, 6 among 66 in the 33.3% Herold Regimen arm, 8 among 63 in the Curtailed Herold Regimen arm, and 3 among 62 in the placebo arm. The registry also reports zero deaths in each arm. These are arm-level counts from the posted safety summary; they do not show the timing, duration, severity grading, or causal attribution of each event.

The most frequently listed serious adverse event terms include Diabetic ketoacidosis at 4.8%, Hyperglycaemia at 3.2%, and Hypoglycaemic seizure at 3.2%, followed by Gastroenteritis, Lymphopenia, and Abdominal pain at 1.6% each. The summary also records other adverse events affecting all participants at risk in each arm. These overall figures do not establish that every participant experienced the same type or number of events, and the registry record does not provide a comparative statistical analysis of safety outcomes in the information described.

Safety counts should be read as descriptive summaries rather than as proof that an event was caused by treatment. The record does not report a formal causal assessment for the listed events. Differences in counts across arms can also be difficult to interpret without event-level detail and an analysis that accounts for exposure and follow-up. The posted information does not establish a safety comparison or support a clinical recommendation.

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

The result addresses a specific question: how many participants met both the insulin-dose threshold of less than 0.5 U/kg/day and the HbA1c threshold of less than 6.5% at 52 weeks after randomization. The observed counts were 8 of 63, 8 of 66, and 9 of 63 in the three teplizumab arms, compared with 6 of 62 in the placebo arm. The estimated risk differences are positive, but their confidence intervals include zero and extend in both directions. The sponsor’s odds-ratio confidence intervals also include 1. Together, these results describe uncertain comparisons rather than a precise separation between treatment and placebo.

The result does not show that the treatment arms are equivalent to placebo. A non-significant or imprecise comparison does not prove there is no difference; the intervals allow a range of possible effects. Nor does the result establish that any regimen improves or worsens the combined outcome. It is not a direct comparison among the three teplizumab regimens, because the reported primary comparisons are each against placebo. The registry record does not provide a prespecified equivalence or non-inferiority framework.

The combined endpoint also does not say how many participants met the insulin criterion alone or the HbA1c criterion alone. Those separate results would be needed to understand the components. Other secondary outcomes provide related information, but they have different definitions, time points, denominators, or analysis methods. They should not be substituted for the primary endpoint.

The findings apply to the participants and outcome definitions represented in the registry results. The study record describes a pediatric and adult population with recent-onset type 1 diabetes, but it does not report enough detail to assess how well the results generalize to other populations or settings. The reason for termination, missing-data handling, and analysis population are not reported in the registry record. These limitations leave important uncertainty about the completeness and robustness of the posted analyses. Statistical results alone do not provide medical advice or determine an individual treatment decision.

18. Why This Trial Matters Statistically

This trial is statistically informative because it puts a clearly defined, combined endpoint into a randomized comparison with placebo and reports results for three dosing regimens. The endpoint requires participants to meet both an insulin-dose threshold and an HbA1c threshold, making clear that the primary result concerns a joint target rather than either component in isolation. The arm-level counts and comparative estimates allow readers to consider both the observed pattern and the uncertainty around it.

The study also illustrates why point estimates should not be read alone. The active-arm counts exceed the placebo count, and the re-calculated risk differences are positive, but all of their confidence intervals include zero. The sponsor’s odds-ratio intervals include 1. The estimates therefore do not give a precise account of the size or direction of the comparison. This is especially important with a binary endpoint and relatively small event counts, where estimates can be uncertain even when arm totals are larger.

The registry’s own odds-ratio analyses and the unadjusted risk-difference re-calculations also show how statistical methods can express a comparison on different scales. Their values are not interchangeable, and the available record does not contain every detail needed to reconstruct the sponsor’s method. Reporting both the effect estimate and its interval gives a more complete account than relying on a p-value alone.

Finally, the record highlights the importance of transparency about analysis populations, multiplicity, missing data, and the original sample-size rationale. Those details are not reported in the registry record. Their absence does not erase the posted results, but it limits how fully readers can assess the design and robustness of the analyses. The trial is a useful example of how a randomized study can provide comparative data while still leaving substantial uncertainty about a specific endpoint.

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.

The registry record gives the trial status as terminated but does not report why it was terminated. It also does not report the original sample-size justification, the planned effect size, or the study’s own design assumptions. Planning illustrations based on the observed results cannot fill that gap. Without the original rationale, it is not possible to determine from the registry record whether enrollment or follow-up matched the intended design.

The primary endpoint has few participants meeting both criteria in each arm. Its confidence intervals are broad, and the posted comparison estimates allow effects in either direction. The endpoint combines insulin dose and HbA1c, so the primary counts alone do not show the separate distributions of either component. The record lists related secondary outcomes, but their varying denominators and definitions make them unsuitable as direct replacements for the primary result.

The analysis populations and handling of missing outcome data are not reported in the registry record. Denominators vary among posted outcomes, but the reasons are not given. This prevents a full assessment of whether missingness or outcome availability could have influenced the estimates. The record also does not report whether the study used interim analyses, what stopping rules applied, or whether multiplicity was addressed across comparisons.

The sponsor’s primary analysis is reported as a Mantel Haenszel odds ratio, but the record does not describe the strata or provide all details needed to reproduce the analysis. The unadjusted risk-difference re-calculations are therefore complementary summaries, not a replacement for the sponsor’s analysis. The record also provides limited safety detail: arm-level serious adverse event counts and selected terms are listed, but event timing, individual-level detail, and comparative safety analyses are not reported in the registry record.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Protege Encore Study- Clinical Trial of Teplizumab (MGA031) in Children and Adults With Recent-Onset Type 1 Diabetes Mellitus — NCT00920582.
• Registry record last updated 2023-12-20; status last verified 2023-10.
• Results first posted 2023-12-20.
• 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

NCT00920582 is a phase 3 randomized, placebo-controlled study of three teplizumab regimens in children and adults with recent-onset type 1 diabetes mellitus. The registry record lists 254 participants, triple masking, parallel groups, MacroGenics as sponsor, and a terminated status. A reason for termination is not reported in the registry record.

The primary endpoint was the number of participants meeting both a total daily insulin dose of less than 0.5 U/kg/day and an HbA1c level of less than 6.5% at 52 weeks after randomization. Counts were 8 of 63, 8 of 66, and 9 of 63 in the three teplizumab arms, compared with 6 of 62 in the placebo arm. The registry’s reported odds-ratio confidence intervals include 1, while the unadjusted risk-difference intervals include zero. Both sets of estimates leave substantial uncertainty about the between-arm effects.

The registry lists secondary outcomes and serious adverse event summaries. Details about the original sample-size assumptions, analysis populations, missing-data methods, multiplicity plan, and reason for termination are not reported in the registry record.