A Randomized, Controlled, Double-Masked, Clinical Trial of Autologous Serum Eye Drops for Severe Ocular Chronic Graft-versus-Host Disease (GVHD) in Hematopoietic Stem Cell Transplant (HSCT) Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Randomized, Controlled, Double-Masked, Clinical Trial of Autologous Serum Eye Drops for Severe Ocular Chronic Graft-versus-Host Disease (GVHD) in Hematopoietic Stem Cell Transplant (HSCT) Patients (NCT01972438), a randomized phase 2 study. The primary outcome is Proportion of Participants Who Experienced a ≥ 50% Reduction in the Combined Score of the Modified Oxford Punctate Keratopathy Grading and the NIH/National Eye Institute (NEI) Visual Analogue Scale in the Study Eye From Baseline to Month 3..

2. Clinical Question

The trial asked whether autologous serum eye drops (ASEDs), prepared from each participant’s own blood, were safe and more effective than normal saline for people with severe chronic ocular graft-versus-host disease (GVHD) after hematopoietic stem cell transplant. Its primary question focused on whether participants had a substantial reduction in a combined score of corneal findings and reported eye symptoms after three months.

The primary result counts one treatment success among six participants in the ASEDs-then-saline sequence and zero among six in the saline-then-ASEDs sequence. Because the endpoint was evaluated at month three, these counts concern the first treatment period, before the planned crossover. They are not a full-period estimate of the effect of receiving ASEDs compared with saline.

The registry record describes the intended clinical question and the posted outcome counts, but does not report an estimated treatment effect with a confidence interval or a significance test. The result therefore gives a limited description of the observed first-period outcomes rather than a precise estimate of how the treatments compare.

3. Trial Design

This was a randomized, controlled, double-masked phase 2 trial with a crossover model. The registry lists triple masking. The planned enrollment was 18, and the reported status is terminated. The study began in 2013-11, with primary completion in 2015-05. The sponsor was the National Eye Institute (NEI).

Participants were assigned to one of two sequences. In the ASEDs-then-saline sequence, participants used ASEDs daily during the first three months and then crossed over to daily saline from month three through month six. In the saline-then-ASEDs sequence, they used saline first and then crossed over to ASEDs. The registry describes participants as unaware of which drops they were receiving. Each participant was to have blood drawn to prepare their own ASEDs.

Although treatment was scheduled across two periods, the primary endpoint was assessed at month three, after the first period and before the crossover. This timing matters: the posted primary counts compare the two randomized sequences at the end of their initial assigned treatments. They do not use the later period to estimate the primary outcome.

The primary outcome analysis included six participants in each sequence, for 12 participants in the reported analysis. The registry says the analysis was per protocol, defined as participants who adhered to the protocol before drug unavailability. Three participants did not start drug, and three participants lacked visit information beyond baseline. Further detail on exclusions and the disposition of all participants is not reported in the registry record.

The registry lists 23 outcome measures and 22 posted secondary outcomes. It does not report a statistical analysis plan, the reason for termination, or details of any interim analyses. These aspects of the trial design are 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.

Baseline-Month 3 Crossover Period

Reasons for not completing this period:

Months 3-6 Crossover Period

Reasons for not completing this period:

Months 6-12 Study Period

Reasons for not completing this period:

Post Month 12 Study Period

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: This analysis is conducted on the per-protocol population, defined as those who adhered to the protocol prior to drug unavailability, not on the intent-to-treat population due to lack of visit information beyond baseline for 3 participants. Three never started drug: one was unable to donate blood, one was ineligible at screening and one died.

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 Participants Who Experienced a ≥ 50% Reduction in the Combined Score of the Modified Oxford Punctate Keratopathy Grading and the NIH/National Eye Institute (NEI) Visual Analogue Scale in the Study Eye From Baseline to Month 3., measured over Baseline and 3 months.

The primary endpoint was the number of participants who experienced at least a 50% reduction from baseline to month three in the combined score for the study eye. The score combined modified Oxford punctate keratopathy grading with the NIH/NEI visual analogue scale. A reduction meeting that threshold was defined as a treatment success.

The Oxford grading measures punctate keratopathy in the cornea on a scale from 0 to 5. The visual analogue scale captures participants’ assessments of ocular dryness, redness, and irritation. Each symptom is scored from 0 to 3, where 0 means none and 3 means severe. Together, the two components produce a combined score from 0 to 14; higher scores represent a worse outcome.

Combining an objective corneal assessment with participant-reported symptoms is intended to reflect more than one aspect of ocular GVHD. The endpoint is also expressed as a threshold-based success count: participants either met the defined reduction criterion or did not. This can be straightforward to interpret clinically, but it discards information about how much the score changed for each participant and how close other participants came to the threshold.

The outcome was measured at month three, before the crossover treatment period. The posted results report one success among six participants in the ASEDs-first sequence and no successes among six in the saline-first sequence. Individual baseline and follow-up scores, and the distribution of score changes, are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The posted primary endpoint is binary: each analyzed participant either met the success definition or did not. A binary endpoint can be summarized by counting participants in each category. Here, the registry reports the success counts and the number analyzed in each randomized sequence. These counts are the direct basis for describing the observed primary result.

What does a confidence interval add?

In general, a confidence interval expresses uncertainty around an estimated quantity, such as a difference between groups. It reflects both the observed data and the assumptions of the statistical method. With small groups and few events, estimates can be especially uncertain. The registry record does not report a confidence interval for the primary treatment comparison, so the range of plausible values under a specified analysis cannot be read from the posted primary result.

What does a significance test answer?

In general, a significance test evaluates how compatible the observed data are with a stated null hypothesis under a statistical model. A p-value is not the probability that a treatment works, and it does not measure the size or clinical importance of an effect. No p-value or other formal test result for the primary comparison is reported in the registry record. The raw counts should therefore not be treated as a reported significance finding.

Why does the crossover timing matter?

In a crossover design, participants are scheduled to receive more than one treatment in separate periods. The primary outcome in this trial was assessed at month three, after the initial treatment period and before the planned switch. Consequently, the primary comparison is based on the randomized first-period assignments. A crossover analysis might need to consider period, sequence, and possible carryover effects, but the registry record does not describe such an analysis for the primary endpoint. It would be inappropriate to infer that later-period data were incorporated into the posted primary counts.

What does the per-protocol population mean?

The registry defines the primary analysis population as participants who adhered to the protocol before drug unavailability. It states that the analysis was not intent to treat because three participants lacked visit information beyond baseline. In general, an intent-to-treat analysis retains participants in the groups to which they were randomized, while a per-protocol analysis restricts the population according to adherence or eligibility criteria. The latter can be useful for describing outcomes among participants meeting those criteria, but it may introduce bias if the reasons for inclusion or exclusion relate to outcomes. The registry does not report enough detail to assess that possibility.

What is over-dispersion?

In general, over-dispersion occurs when the variability in count data is greater than a statistical model expects. It can make standard errors too small if the model does not account for the extra variability. The primary endpoint is a binary participant-level outcome rather than a reported count-rate model, and the registry does not report an over-dispersion adjustment. This concept should not be taken as a description of an analysis performed in this trial.

How should recalculations and planning illustrations be read?

In general, a recalculated estimate based on summary counts is an unadjusted calculation, not a substitute for the sponsor’s prespecified analysis. No recalculated primary comparison is reported in the registry record. The registry also describes sample-size illustrations based on the observed effect and specified planning assumptions. Such illustrations concern hypothetical future study sizes; they are not observed power and do not show that the trial was adequately powered. The registry record does not report the trial’s own sample-size justification or its planned target effect.

Overall, the available method details are limited. The record identifies the endpoint, its assessment time, the analyzed sequence counts, and the per-protocol population, but a prespecified analysis plan, inferential results, handling of missing data beyond the stated rationale, and details of crossover modeling are not reported in the registry record.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("ASEDs - Saline", "Saline - ASEDs")
n      <- c(6, 6)
events <- c(1, 0)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  ASEDs - Saline  6  1  16.7  0.4  64.1  3.0  56.4  3.7  57.9
ARM  Saline - ASEDs  6  0  0.0  0.0  45.9  0.0  39.0  0.4  41.0

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 23 outcome measures and 22 posted secondary outcomes. The primary endpoint is one outcome among a broader set that includes adverse events and several score changes at different time points and in different eyes. When many outcomes are examined, the chance of finding an apparently notable result by chance can increase. Statistical methods can address this issue, but the multiplicity strategy is not reported in the registry record. No adjustment for multiple testing should be assumed.

The primary analysis is described as per protocol and not intent to treat. The stated reason is that three participants lacked visit information beyond baseline. The record also says that three participants never started drug, including one who was unable to donate blood; the remaining detail in that description is incomplete. The analyzed primary groups each contain six participants. How all enrolled participants were assigned to the analysis population, and how missing outcomes were handled beyond the stated rationale, is not reported in the registry record.

In general, missing outcome data can reduce precision and can bias results if the likelihood of missingness is related to treatment or outcome. A per-protocol analysis may also differ from the randomized population if adherence or availability affects inclusion. These are interpretive concerns, not claims that a particular bias occurred. The registry does not report sensitivity analyses, imputation methods, or an analysis comparing the primary result across different populations.

The limited analysis detail means the primary counts are best read as the result for the reported per-protocol population. They do not provide a complete account of how outcomes would look under an intent-to-treat analysis or under alternative assumptions about missing data.

12. Safety Results

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

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: upper respiratory tract infection (5.6%), lymphadenopathy (5.6%), aspergillus infection (5.6%), lung infection (5.6%).

The registry reports serious adverse event counts by intervention category. It lists 18 participants at risk in each category: ASEDs, saline, ASEDs or saline, and no intervention. The recorded numbers affected by serious adverse events are, respectively, one, zero, two, and one. Death counts are not reported in the registry record.

These categories should not be added together or treated as four independent groups. The ASEDs-or-saline category and the no-intervention category are distinct registry groupings, and the record does not explain whether participants or events overlap across categories. The figures are counts, not adjusted comparisons of risk. The registry does not report a formal statistical comparison of serious adverse events between ASEDs and saline.

For other adverse events, the corresponding recorded counts are zero for ASEDs, two for saline, zero for ASEDs or saline, and one for no intervention, with 18 at risk in each category. The record also lists upper respiratory tract infection, lymphadenopathy, Aspergillus infection, and lung infection among serious adverse event terms, each with a reported value of 5.6%. The registry does not provide enough detail in these entries to establish whether terms represent unique participants or events.

These safety entries describe what was recorded in a small trial. They do not establish comparative safety or rule out uncommon events. The trial’s reported status is terminated, but the reason for termination is not reported in the registry record.

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

The primary outcome is a threshold-based measure of improvement in a combined corneal and symptom score. The posted result shows one participant in the ASEDs-first sequence meeting the success definition by month three, compared with no participants in the saline-first sequence. This is a description of the analyzed participants, not a precise estimate of a treatment effect.

The result does not establish that ASEDs are more effective than saline, nor does it establish that the treatments are equivalent. The registry does not report a confidence interval, p-value, or adjusted between-sequence estimate for the primary endpoint. With only six participants reported in each sequence, the available counts leave substantial uncertainty about how the result would generalize beyond these participants.

The result also does not describe the full six-month crossover experience. The primary endpoint was assessed at month three, before the treatment switch. Outcomes after the switch may address different questions, but the posted primary counts do not combine both treatment periods.

The endpoint uses a 50% reduction threshold, so it does not show the full distribution of score changes. It cannot indicate how much scores changed among participants who did not meet the threshold or how individual score components behaved. Those individual data are not reported in the registry record.

Finally, the analysis population was per protocol, not intent to treat. This limits interpretation as an estimate for everyone randomized, particularly because the registry notes missing post-baseline visit information and participants who did not start drug. These limitations do not erase the observed counts, but they narrow what can reasonably be concluded from them.

14. Why This Trial Matters Statistically

This trial illustrates how an outcome definition shapes the question a study can answer. The primary endpoint combined a corneal grading measure with participants’ assessments of eye symptoms, then classified a participant as a success only if the combined score fell by at least 50%. That approach offers a clear threshold but does not preserve the detail in the underlying score changes.

It also highlights the importance of timing in a crossover design. Although participants were scheduled to receive both ASEDs and saline, the primary outcome was assessed after the first treatment period. The reported counts therefore concern randomized initial assignments, not a comparison using all periods. Readers need to distinguish the overall design from the specific period used for the primary result.

The gap between planned enrollment and the reported primary analysis is another important statistical feature. The registry lists enrollment of 18, while the primary endpoint is reported for six participants in each sequence. The stated per-protocol analysis and missing post-baseline information make analysis-population definitions central to interpretation. Reporting complete participant flow and sensitivity analyses would help readers assess how those choices influence the result.

Finally, the posted counts are not accompanied by an inferential estimate or a reported multiplicity plan despite the registry’s listing of numerous outcomes. This demonstrates why outcome counts alone are not enough to judge precision or statistical certainty. The record is useful for understanding the trial’s question, endpoint, and observed primary counts, while leaving important analytical details unresolved.

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 analysis includes six participants in each treatment sequence, and the registry identifies it as per protocol. It also states that three participants lacked visit information beyond baseline and that three never started drug. The available record does not fully describe participant flow or provide an intent-to-treat result. These features limit how broadly the primary counts can be interpreted.

The endpoint was measured at month three, before crossover, so the primary result does not itself summarize outcomes across both treatment periods. The registry does not report an analysis of possible period, sequence, or carryover effects. Nor does it provide individual score changes or the distribution of the combined score, which would add context to the threshold-based success counts.

The registry lists many outcomes but does not report a multiplicity strategy, a statistical analysis plan, confidence intervals, or a significance test for the primary comparison. The reason the trial was terminated is not reported in the registry record. The reported primary counts therefore cannot be placed in the context of a documented planned analysis or a complete account of trial conduct.

Safety entries are also limited: the registry reports counts by categories but does not provide a formal comparison, and deaths are not reported. The recorded adverse-event counts should not be interpreted as proof of equal safety or as evidence that uncommon harms were absent.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Randomized, Controlled, Double-Masked, Clinical Trial of Autologous Serum Eye Drops for Severe Ocular Chronic Graft-versus-Host Disease (GVHD) in Hematopoietic Stem Cell Transplant (HSCT) Patients — NCT01972438.
• Registry record last updated 2016-12-08; status last verified 2016-10.
• Results first posted 2016-09-13.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

This randomized, controlled, double-masked phase 2 crossover trial evaluated autologous serum eye drops against normal saline for severe chronic ocular GVHD in hematopoietic stem cell transplant patients. Participants were scheduled to receive each treatment for three months, with the primary endpoint assessed at month three, before crossover.

The primary endpoint was at least a 50% reduction from baseline in the combined study-eye score of modified Oxford punctate keratopathy grading and the NIH/NEI visual analogue scale. The per-protocol result records one success among six participants in the ASEDs-then-saline sequence and none among six in the saline-then-ASEDs sequence.

The record lists enrollment of 18 and status as terminated. It reports 22 secondary outcomes and safety counts, but does not report a primary confidence interval, p-value, statistical analysis plan, multiplicity strategy, or reason for termination. The primary counts are descriptive and do not alone establish a precise comparative treatment effect.