AFQ056 for Language Learning in Children With FXS: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for AFQ056 for Language Learning in Children With FXS (NCT02920892), a randomized phase 2 study. The primary outcome is Change in Weighted Child Intentional Communication Score.

2. Clinical Question

The question is about the added effect associated with AFQ056 when both randomized groups received language intervention. It is not a comparison of language intervention with no language intervention, because language intervention was part of both groups. The primary outcome concerns child-initiated communication during a structured play session, with greater weight assigned to communication using more symbols. The registry reports a least-squares mean measure and a mixed models comparison, rather than a simple count of children who did or did not improve.

The study was registered as a Phase 2 trial and is marked completed. Its primary comparison ends at Month 8, at the end of the double-blind period. Participants were then offered an open-label extension in which all participants received AFQ056. The primary question therefore concerns the placebo-controlled period, not the later extension.

3. Trial Design

The registry record describes a randomized, double-blind, parallel-group Phase 2 trial of AFQ056 for language learning in children with FXS. It reports enrollment of 110 participants and a completed status. The stated age range in the trial summary is 3-6 years. Both randomized groups received language intervention; one group received AFQ056 and the other matching placebo.

Before randomization, participants underwent a 4-month single-blind placebo lead-in. The double-blind treatment period lasted 8 months. AFQ056 and placebo were administered as oral suspensions twice per day, according to the arm descriptions. The AFQ056 description says treatment began at 25 mg, but the remainder of the dose instructions is not reported in the registry record. The complete titration schedule and the corresponding detailed placebo instructions are also not reported in the registry record.

At the end of the placebo-controlled period, assessments were completed and participants were offered an open-label extension. In that extension, all participants received AFQ056. The record describes 2 months of flexible dose titration at the beginning of the extension, but the remaining details of the extension regimen are not reported in the registry record. Since treatment was no longer blinded or controlled with placebo in that period, it is distinct from the primary randomized comparison.

The registry gives a primary completion date of 2022-05-17 and a study start date of 2017-08-17. The primary endpoint was assessed from baseline through Month 8. The record does not report the randomization ratio, the full eligibility criteria, or the specific methods used to preserve masking. It also does not report a detailed analysis-population definition. Those omissions limit how fully the design and analysis can be reconstructed from the registry record alone.

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: Prior to randomization, there will be a 4-month single-blind placebo lead in after screening assessments during which subjects will have therapy treatments as usual to control for the effects of the language intervention. The placebo lead-in will also control for placebo effects which are prominent in clinical trials in FXS.

Placebo Lead-In Period

Reasons for not completing this period:

Placebo-Controlled Period

Reasons for not completing this period:

Open-Label Extension (OLE) Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Comparison of WCS from baseline and from month 8 visit (end of double-blind period) in both AFQ056 \& Language Intervention and Placebo \& Language Intervention groups.

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 Change in Weighted Child Intentional Communication Score, measured over Baseline through Month 8.

The primary endpoint is change in Weighted Child Intentional Communication Score (WCS) from baseline through Month 8. The score comes from a 22 minute semi-structured play session between an examiner and a child. It captures intentional communication in both structured and unstructured parts of the session.

Each intentional communication act receives a weight according to its form: nonverbal communication is weighted 1, a single symbol is weighted 2, and multiple symbols are weighted 3. The weighted components are added to form the total score. The scale has no maximum score, and higher scores indicate more child-initiated communication. The scale was administered at Baseline and Months 2, 4, 6, and 8.

The recorded endpoint is a composite representing the average of estimated change scores at Months 2, 4, 6, and 8. Its unit is log10(scores on a scale), and the reported arm summaries are least-squares means. This approach is intended to represent change across the double-blind period rather than relying only on a single communication observation. The registry does not report further details about how the estimates at the individual visits were combined beyond describing the composite average.

Intentional communication is relevant to the stated language-learning question because the score emphasizes communication initiated by the child and distinguishes communication by its symbolic complexity. It is nevertheless a measure from a particular examiner-child play session, not a direct measure of every context in which a child communicates.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary analysis is reported as a mixed models analysis of the slope, with a 90% confidence interval and a p-value. The registry names the method but does not report the full statistical model, its covariates, its assumptions, or its treatment of missing observations. The interpretation should therefore stay close to the reported estimate and interval rather than attributing unreported details to the analysis.

What a mixed model does

General statistical explanation: A mixed model can account for repeated measurements from the same participant by modeling the relationship among that participant’s observations, while also estimating average patterns across participants. This can be useful when an outcome is measured at multiple visits, as WCS was at Baseline and Months 2, 4, 6, and 8. The model’s estimated treatment contrast depends on its specification, including which visits and covariates are included and how within-participant correlation is handled. Those details are not reported in the registry record.

The reported slope is -0.1, with a 90% confidence interval from -0.22 to 0.02 and p = 0.159. A confidence interval describes a range of estimates compatible with the observed data and the model under its assumptions. It is not a range containing a known fraction of individual participants’ outcomes. Here, the interval crosses zero, so the result does not pin down a single direction for the comparative change with high precision. A p-value describes how unusual data at least as inconsistent with a null hypothesis would be if that null hypothesis and the model assumptions were true. It is not the probability that the null hypothesis is true.

Why use a confidence interval?

General statistical explanation: A point estimate alone can conceal substantial uncertainty. An interval communicates both the estimated direction and the range of values that remain compatible with the data. Wider intervals generally indicate less precision, while narrower intervals generally indicate greater precision. The registry reports a 90% interval for the primary comparison. The percentage describes the interval procedure’s long-run coverage under its assumptions; it does not mean there is a 90% probability that this particular fixed interval contains the true effect.

The registry’s primary outcome is based on repeated assessment over time, and it reports arm-level least-squares means alongside a model-based comparison. Least-squares means are model-based summaries, not necessarily raw averages. The record does not state the covariate structure or missing-data assumptions underlying those summaries, so they should not be treated as simple unadjusted means.

Recalculation and the sponsor analysis

The posted materials do not provide a recalculated primary contrast from summary statistics. The available primary result is the sponsor-reported mixed models analysis. A straightforward comparison of the two arm summaries would not reproduce a repeated-measures model result, and the registry does not provide enough model detail to independently reconstruct that analysis. The comparison should therefore be interpreted as reported, with the stated estimate, interval, and p-value.

For the secondary CGI-I responder outcome, a simple unadjusted risk-difference recalculation is reported separately from the sponsor’s mixed models odds ratio. Such recalculations use the observed counts and are not substitutes for the trial’s own analysis. The recalculated risk difference is -0.044 with a 95% confidence interval from -0.196 to 0.108 and an unadjusted p-value of 0.57. The sponsor-reported odds ratio is 0.95 with a 90% confidence interval from 0.4 to 2.27 and p = 0.92. The estimates use different effect measures and different analytical approaches, and their intervals use different confidence levels, so they should not be compared as though they were the same calculation.

Planning illustrations

General statistical explanation: A sample-size planning illustration asks how many participants might be needed to detect a specified effect under chosen assumptions. It does not describe the number who were actually enrolled or analyzed, and it is not observed power. The registry’s planning assumptions specify a two-sided significance level of 0.05, equal allocation, and a normal approximation. For the binary secondary outcome, the illustration uses fractions of the observed effect. It gives participants per arm for 80% power and participants per arm for 90% power for each stated fraction. These are planning scenarios based on the observed effect, not the trial’s original design justification. The registry record does not report the trial’s own sample-size justification.

The registry does not report an interim-analysis plan, multiplicity adjustment, or a detailed prespecified statistical analysis plan. It also does not report whether the primary model used an intention-to-treat population, a per-protocol population, or another analysis population. In the absence of those details, the reported numerical result can be described, but not fully audited for all analytic choices.

10. Analyses Reported by the Sponsor

For the primary endpoint, the registry reports the sponsor’s mixed models analysis: slope -0.1, 90% confidence interval -0.22 to 0.02, and p = 0.159. It does not give a separate simple recalculation of the primary contrast from the arm summaries. Because the reported method uses a model and repeated observations, the sponsor’s estimate is the available primary comparison; the arm-level least-squares means should not be subtracted to create a different result.

For the Month 8 CGI-I positive-response outcome, the sponsor reports an odds ratio of 0.95, a 90% confidence interval from 0.4 to 2.27, and p = 0.92. The unadjusted recalculation uses a risk difference instead: -0.044, with a 95% confidence interval from -0.196 to 0.108 and p = 0.57. The calculations differ in effect measure, confidence level, and analytical method. The recalculated result is a simple summary-statistic comparison, not a replacement for the sponsor’s model-based analysis. The registry record does not provide enough detail to reproduce the mixed models analysis independently.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary outcome and several secondary outcome domains, including language, development, adaptive behavior, and a CGI-I response. The record reports eight outcome measures and eight analyses, but it does not describe a hierarchical testing procedure, family-wise error control, or another multiplicity strategy. It is therefore not reported in the registry record whether conclusions across the outcomes were adjusted to account for multiple testing.

Multiplicity matters because testing multiple outcomes can increase the chance that at least one result appears statistically notable even when none reflects a reliable underlying difference. This is a general statistical concern, not a claim that any particular result in this trial is false. The primary outcome is the clearest focal endpoint because it is explicitly identified as primary. Secondary outcomes should be interpreted with more restraint when their relationship to the primary endpoint and the overall testing strategy are not stated.

The registry record does not report a detailed missing-data approach for the primary or secondary analyses. It does not state how missing visits were handled, whether any imputation was used, or what assumptions were made about participants with incomplete follow-up. The difference between the enrollment total of 110 and the primary arm counts of 49 and 50 is not explained in the registry record. It should not be interpreted as a dropout count or attributed to any particular reason.

The record also does not report the analysis populations used for the primary or secondary outcomes. In particular, it does not specify whether participants were analyzed according to randomized assignment, whether any exclusions were made, or whether a separate per-protocol analysis was conducted. Those definitions affect how results are interpreted, especially when follow-up is incomplete. Without them, the reported estimates can be described but cannot be fully assessed for the influence of missing observations or exclusions.

No interim-analysis plan or stopping rule is reported in the registry record. These omissions do not establish that such procedures were absent; they mean that the public record does not describe them. The same distinction applies to the statistical plan: the reported methods are identifiable, but a complete account of prespecified analyses is not available in the registry record.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: atypical pneumonia (2.0%), croup infectious (1.1%), seizure (0.9%).

The registry reports adverse events in four periods: the placebo lead-in, the double-blind placebo period, the double-blind AFQ056 period, and the open-label AFQ056 period. In the lead-in, 1 serious adverse event was reported among 110 at risk. During the double-blind period, 1 serious adverse event was reported among 49 at risk in the placebo group, and none among 50 at risk in the AFQ056 group. During the open-label period, 1 serious adverse event was reported among 89 at risk. No deaths were reported in any of these periods.

The registry lists Atypical Pneumonia at 2.0%, Croup Infectious at 1.1%, and Seizure at 0.9% among the top serious adverse-event terms. These are adverse-event summaries, not evidence that a treatment caused or prevented an event. The record does not provide enough detail to attribute individual events to treatment or to establish comparative risk.

Serious adverse events are distinct from other adverse events. The record also reports other adverse events, but those counts should not be treated as counts of serious events. Periods differ in their populations at risk and treatment conditions, particularly because the open-label extension gave AFQ056 to all participants. The figures are descriptive and should be read within those design differences.

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

The primary result describes uncertainty in the comparison of change in WCS between the AFQ056-plus-language-intervention and placebo-plus-language-intervention groups through Month 8. The reported slope is -0.1, with a 90% confidence interval from -0.22 to 0.02 and p = 0.159. The interval includes zero, and the reported estimate does not precisely establish a direction for the comparative change.

This result does not show that the groups are identical. A non-significant p-value does not prove that there is no difference, and the confidence interval leaves a range of possible comparative values. At the same time, the result does not establish a difference in favor of either randomized group. The registry does not report a threshold for a clinically important change in WCS, so statistical uncertainty cannot be translated into a conclusion about the practical importance of a particular effect size.

The comparison concerns AFQ056 added to language intervention, relative to placebo added to language intervention. It does not estimate the effect of language intervention itself, because there was no randomized group without language intervention in the described double-blind comparison. The open-label extension is also not a continuation of the placebo-controlled comparison: all participants received AFQ056 during that period.

The primary score is based on intentional communication observed during a semi-structured play session. Its weighting scheme emphasizes communication using symbols, but the score is not a complete measure of language use across daily settings. The registry’s reported outcome should be interpreted as the result for this specific measure and assessment period.

Finally, the registry summary does not report enough detail about missing data, analysis populations, or the full model specification to establish how sensitive the result is to those choices. The appropriate reading is limited to the posted estimate and its uncertainty, alongside the design and measurement information described in the record.

14. Why This Trial Matters Statistically

This trial illustrates why a clinical-trial result cannot be reduced to whether a p-value crosses a conventional threshold. The primary estimate is accompanied by a confidence interval that includes values on either side of zero. Looking at the estimate together with its interval makes the uncertainty and the range of plausible comparative effects more visible than a binary significant-or-not label would.

It also illustrates the importance of distinguishing the outcome summary from the treatment comparison. The arm-level least-squares means describe estimated changes within each randomized group, whereas the mixed models slope is the reported comparison across groups. Those quantities answer related but different questions, and simple subtraction of the arm summaries would not recreate the reported repeated-measures analysis.

The study also separates the effect of AFQ056 from the shared language intervention. Since both randomized groups received language intervention, the primary contrast concerns the addition of AFQ056 relative to placebo in that context. Keeping the comparator clear prevents the result from being misread as a test of language intervention against no intervention.

Finally, the record demonstrates the limits of public registry summaries. It reports a primary model result and some outcome counts, but does not report the full analysis plan, missing-data methods, or analysis populations. These omissions matter when readers assess precision and reproducibility. The reported estimates can still be interpreted, but the extent of independent statistical auditing is limited by the methodological detail available in the registry record.

15. Limitations

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

The primary comparison is reported with an estimate, confidence interval, and p-value, but the full mixed-model specification is not reported in the registry record. The covariates, covariance structure, and handling of incomplete repeated measurements are not described. These omissions limit independent reproduction and make it difficult to assess sensitivity to modeling choices.

The record does not explain why the primary arm summaries include 49 and 50 participants while overall enrollment is 110. It does not provide a participant flow accounting or identify the analysis population. This prevents readers from determining from the registry alone how many participants contributed to each visit or why the displayed endpoint counts differ from enrollment.

The WCS is a specific observed measure from an examiner-child play session. The registry describes its scoring and assessment schedule, but does not report an established threshold for a clinically important difference. It also does not provide the visit-specific primary estimates, which would help readers understand how the composite change developed during the double-blind period.

The secondary outcomes cover several constructs, but the available registry results do not support a complete numerical synthesis across them. The registry also does not report a multiplicity strategy, making secondary findings harder to distinguish as confirmatory or exploratory. The single posted binary secondary comparison has a different outcome definition and analysis method from the primary endpoint.

The arm descriptions in the registry are incomplete: they do not report the full AFQ056 titration instructions, and the placebo regimen description is also truncated. The record does not provide the original sample-size justification, an interim-analysis plan, or a detailed statistical analysis plan. These are reporting limitations; they do not establish that such procedures or documents did not exist.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: AFQ056 for Language Learning in Children With FXS — NCT02920892.
• Registry record last updated 2023-10-10; status last verified 2023-09.
• Results first posted 2023-10-10.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

The registry record describes a completed, randomized, double-blind, parallel-group Phase 2 trial of AFQ056 for language learning in children with Fragile X syndrome. Enrollment was 110. After a 4-month single-blind placebo lead-in, participants were randomized to AFQ056 or placebo, with language intervention in both groups, during an 8 month double-blind period.

The primary endpoint was change in Weighted Child Intentional Communication Score through Month 8. The reported mixed models result was a slope of -0.1, with a 90% confidence interval from -0.22 to 0.02 and p = 0.159. The record also describes secondary developmental, language, adaptive-behavior, and CGI-I outcomes, with limited usable comparative results available in the registry.

After the controlled period, participants were offered an open-label extension in which all received AFQ056. The record reports no deaths and includes serious adverse-event counts by study period. Important details about the full statistical plan, missing-data handling, and analysis populations are not reported in the registry record.