Effects of Sildenafil on Choroidal Thickness in AMD: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effects of Sildenafil on Choroidal Thickness in AMD (NCT01830790), a non-randomized phase 2 study. The primary outcome is Change in Choroidal Thickness as Assessed on Enhanced-Depth Imaging Optical Coherence Tomography (EDI-OCT).

2. Clinical Question

The study description frames the rationale as a hypothesis about choroidal thickness in AMD. It notes that choroidal thinning has been hypothesized to participate in AMD pathogenesis, and asks whether sildenafil might increase thickness in eyes with AMD. It also proposes that a small thickness response could be consistent with reduced choroidal vascular compliance or increased stiffness in this condition. These are proposed interpretations, not findings from this study.

The registry summary also mentions visual acuity, central macular thickness (CMT), and macular volume as measures that investigators hoped to relate to changes in choroidal thickness. The listed secondary outcome is change in central foveal thickness assessed by OCT. The record does not report results for these measures or correlations among them.

3. Trial Design

The registry identifies this as a Phase 2 study sponsored by Duke University, with a non-randomized allocation and no masking. Its model is listed as single group. At the same time, the study description names AMD patients and healthy controls as groups, and describes comparing post-treatment choroidal-thickness changes between them. The record cautions that the arms are not a like-for-like control comparison. These labels and descriptions should be read together rather than treating the healthy-control group as a randomized control arm.

The intervention was a planned single 100mg oral dose of sildenafil citrate. Participants were to undergo EDI-OCT imaging before treatment and after treatment. The primary outcome timeframe lists baseline, 1 hour, and 3 hours post-treatment. The protocol description says that the images would be used to assess choroidal thickness and also mentions CMT and macular volume. A secondary outcome is listed for change in central foveal thickness assessed by OCT.

The registry lists enrollment as 10 and gives a study start of 2013-03 and primary completion of 2014-05. The status is terminated. The primary outcome description states that the study was terminated early because of inadequate support to complete recruitment and data analysis, and that no outcome measure data were collected. The enrollment figure does not establish how many participants received the planned dose or contributed to any particular measurement.

Non-randomized studies can describe patterns within treated participants or compare groups, but group differences may reflect characteristics other than the intervention. Without random allocation, masking, or reported outcome data, this record cannot establish whether any observed change occurred or whether a difference between groups would be attributable to sildenafil. In this case, there are no outcome estimates to interpret in the first place.

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

Analysis population for the primary outcome, as defined in the registry: Study was terminated early due to inadequate support to complete recruitment/data analysis. No outcome measure data was collected.

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 Choroidal Thickness as Assessed on Enhanced-Depth Imaging Optical Coherence Tomography (EDI-OCT), measured over Baseline, 1 hour, and 3 hours post-treatment.

The primary endpoint was change in choroidal thickness assessed on enhanced-depth imaging optical coherence tomography. The planned assessment points were baseline, 1 hour, and 3 hours post-treatment. EDI-OCT is an imaging approach intended to visualize the choroid so that its thickness can be measured. In general, a repeated measurement can describe how an anatomical quantity changes over time within a participant after an intervention.

Choroidal thickness was selected because the study rationale concerns a possible relationship between the choroid and AMD. The investigators proposed that sildenafil might increase thickness in AMD eyes, and that the size or absence of a change could inform hypotheses about choroidal vascular properties. Such mechanisms remain hypotheses in the registry description; an imaging endpoint alone would not establish a change in disease course.

The record does not give a numerical definition of how thickness was measured, the unit, the summary statistic, or the planned statistical contrast. Those details are not reported in the registry record. More importantly, the endpoint description explicitly says that no outcome measure data were collected. There is therefore no observed change, estimate, uncertainty interval, or group comparison to interpret. The planned endpoint explains what the study aimed to measure, not what it found.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods help define what a study will estimate, how observations will be compared, and how uncertainty will be described. The registry record identifies the intended outcome and several design features, but it does not report a statistical analysis plan. It lists zero analyses and states that no outcome measure data were collected. The general explanations below clarify the methods that would ordinarily be relevant; they should not be mistaken for methods confirmed as having been used in this study.

What does a change endpoint represent?

For an imaging outcome measured at baseline and after treatment, a change endpoint compares measurements taken at different times. In general, an analysis may summarize changes within participants, compare changes between groups, or do both. A within-participant comparison uses each participant's earlier measurement as a reference. A between-group comparison asks whether the pattern differs across groups. These are different questions and need distinct definitions of the target estimate.

The record specifies imaging at baseline, 1 hour, and 3 hours post-treatment, but it does not say whether the intended primary contrast was change at each follow-up time, a combined summary, or another contrast. It does not report a measurement unit, a method for combining repeated readings, or how measurements from both eyes, if relevant, would have been handled. These details are not reported in the registry record. Without a defined contrast and observed measurements, no numerical estimate can be reconstructed.

How should uncertainty be understood?

In general, a confidence interval describes the range of estimates reasonably compatible with the observed data under a specified statistical model. Its width depends on factors such as variability, sample size, and the analysis method. A narrow interval generally indicates greater precision than a wide interval, but an interval cannot correct for bias in the study design or measurement process. This study has no reported primary outcome estimate or confidence interval because no outcome measure data were collected.

A p-value, in general, measures how unusual data at least as incompatible with a specified null hypothesis would be if that hypothesis and the analysis assumptions held. It is not the probability that a hypothesis is true, and it does not measure clinical importance. No p-value for a study outcome is reported. Statistical significance therefore cannot be assessed, and it would be misleading to describe the trial as having a positive or negative statistical result.

Why does allocation matter?

Random allocation is a design method intended to make comparison groups more similar, on average, with respect to measured and unmeasured characteristics. This study is listed as non-randomized and unmasked, and the record warns that the arms are not a like-for-like control comparison. In general, differences between non-randomized groups can be associated with group membership as well as with treatment exposure. Adjustment methods can address measured factors under assumptions, but they cannot guarantee removal of unmeasured confounding. The record does not report an adjustment plan or an analysis that could be evaluated.

Repeated imaging also raises general measurement questions. Instrument settings, image quality, reader decisions, and timing can affect thickness estimates. Masking can reduce the chance that knowledge of group or treatment influences measurement or interpretation. The registry lists no masking, but does not give details about image grading, reader training, repeatability, or quality-control procedures. Those aspects are not reported in the registry record.

What do the planning assumptions mean?

The registry describes a sample-size planning illustration using a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It says that the effect to detect is a stated fraction of the observed effect, and that the planning counts are participants per arm for 80% power except for time-to-event tables, which use total events. These assumptions describe hypothetical planning calculations. They do not establish the actual design rationale, and the registry does not report a sample-size justification for the trial itself.

Power is a property of a planned statistical test under specified assumptions, including an effect size, variability, allocation, and significance level. In general, a power calculation is not a measure of how likely a completed trial's result is to be correct. The planning values should not be called observed power. In this record, there is a further interpretive issue: the planning description refers to an observed effect, while the primary outcome record states that no outcome data were collected. The illustration cannot be treated as an effect observed in this trial.

Over-dispersion is a general concern when data vary more than a statistical model expects. If present, it can make standard errors too small and confidence intervals too narrow when the model does not account for it. The registry does not describe the distribution of the imaging outcomes or an analysis model, so there is no basis for assessing over-dispersion in these data.

Finally, no result can be recalculated from summary statistics because the registry record contains no outcome summaries. Recalculation is useful in general as a check on a reported estimate when the necessary group summaries are available, but it cannot create missing measurements or replace the sponsor's analysis. Here, no sponsor outcome analysis is reported, and the record lists zero analyses.

10. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates several outcomes, time points, groups, or statistical contrasts. In general, testing many hypotheses can increase the chance that at least one result appears noteworthy by chance. A statistical plan may address this through a hierarchy of outcomes, adjusted testing, or clearly identified exploratory analyses. The registry identifies choroidal-thickness change as the primary outcome and central foveal thickness as a secondary outcome. Its study summary also mentions CMT, macular volume, and visual acuity. It does not report a multiplicity strategy, a testing hierarchy, or which additional measures would be analyzed inferentially. These details are not reported in the registry record.

Multiplicity cannot be evaluated from outcome estimates because no outcome measure data were collected. The record lists zero analyses. There are no reported estimates or tests whose multiplicity adjustment could be checked, and it would be wrong to infer that unadjusted or adjusted analyses were performed.

The registry record also does not report how missing measurements would have been handled, whether any participants had incomplete imaging, or whether any imputation or sensitivity analysis was planned. The stated reason for termination was inadequate support to complete recruitment and data analysis; the endpoint description says no outcome measure data were collected. This is not the same as a reported analysis of partially missing outcomes. The number and pattern of missing observations, if any, are not reported in the registry record.

Analysis populations are also not described. The record does not specify whether planned analyses would include all enrolled participants, only those receiving the dose, or only those with usable imaging. The enrollment entry does not establish any of these populations or their sizes. Without a prespecified analysis population and observed outcome data, neither the handling of missing data nor the population to which an estimate might apply can be assessed.

11. 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 0% reporting threshold.

The registry adverse-event entries should be read as recorded counts in the categories shown, not as a complete safety assessment. For healthy controls, the record lists zero at risk for serious events and zero at risk for other events, with zero affected in both categories. For AMD patients, it lists 10 at risk for serious events and zero affected, and 10 at risk for other events and zero affected. Death counts are not reported in the registry record.

The healthy-control entries with zero at risk do not provide a meaningful denominator for assessing event frequency in that group. The AMD entries record zero affected among those listed at risk, but the record does not give an adverse-event observation period, details of event collection, or a fuller account of participant exposure. It also does not say whether the listed at-risk counts correspond to everyone enrolled, everyone treated, or another safety population. The enrollment entry should not be substituted for these group-specific records.

Zero recorded events does not demonstrate that an intervention has no risk. Event counts can be informative only in relation to who was monitored, for how long, and how events were sought and classified. The registry record's status is terminated, and it says no outcome measure data were collected; it does not provide enough safety detail to make broader claims about tolerability. These entries are descriptive registry data and should not be read as an estimate of the intervention's safety profile.

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

This record establishes that a study was planned to examine choroidal-thickness change after a single 100mg oral dose of sildenafil citrate in people with AMD and healthy controls. It also establishes that the study was non-randomized, unmasked, listed as a single-group model, and ultimately terminated. The primary outcome description says that termination occurred early because of inadequate support to complete recruitment and data analysis, and that no outcome measure data were collected.

Accordingly, the record does not establish whether choroidal thickness changed after treatment. It does not establish a difference between the AMD and healthy-control groups, or a relationship between thickness and visual acuity, CMT, macular volume, or central foveal thickness. It cannot show whether the hypothesized vascular explanation is correct, and it cannot show whether any anatomical change would alter AMD progression. The proposed interpretations in the study summary are hypotheses, not findings.

The design information also limits what a completed comparison could establish. Non-randomized allocation does not ensure that groups are comparable, and the record warns against treating the arms as a like-for-like control comparison. In general, a change measured after treatment is not by itself proof that treatment caused the change. But these design considerations remain secondary to the central limitation: there are no posted outcome measurements to analyze.

The adverse-event entries report zero affected in the categories recorded for the AMD group, while the healthy-control group is listed with zero at risk. These entries do not establish absence of risk, and death counts are not reported in the registry record. Likewise, the sample-size planning assumptions do not turn into evidence about the outcome or the statistical power achieved by this terminated study.

The most defensible conclusion is therefore limited: the registry documents the intended question, planned measurements, and early termination, but supplies no trial outcome estimate. Neither a beneficial nor an adverse effect on choroidal thickness can be inferred from the absence of results.

13. Why This Trial Matters Statistically

This study illustrates why a clearly defined clinical question and a recorded outcome estimate are both necessary for statistical interpretation. The planned endpoint, choroidal-thickness change assessed with EDI-OCT, is identifiable, as are the intended measurement times. Yet a defined endpoint does not itself constitute evidence. Without collected outcome data, there is no estimate to compare with a clinically meaningful threshold, no uncertainty to quantify, and no basis for evaluating whether the observed pattern is consistent with the study hypothesis.

The record also highlights the distinction between planned comparisons and credible controls. It describes AMD patients and healthy controls, but allocation was non-randomized and the record explicitly cautions that the arms are not a like-for-like control comparison. Statistical comparisons depend on how groups were formed, not only on the test used. A sophisticated model cannot automatically resolve differences that the design leaves uncontrolled, particularly if relevant factors were not measured or reported.

Early termination is statistically important because it can prevent completion of recruitment and analysis, as the registry says occurred in this case. The enrollment figure cannot substitute for analyzable endpoint data. It is possible to have a registered enrollment count without a reported estimate for the outcome of interest. Reporting the reason for termination and explicitly stating that no outcome data were collected helps readers distinguish an unreported estimate from an estimate showing little or no change.

The planning illustration also demonstrates why sample-size calculations must be interpreted in context. Planning calculations depend on assumptions about the effect to detect and the design. They are not findings, and they cannot supply an empirical effect when outcomes were not collected. The tension between a planning description that refers to an observed effect and an outcome record that reports no data makes it especially important not to treat those calculations as evidence from the trial.

Statistically, the most useful message is about evidential limits: a registry can document a research question and design while still providing no answer to that question. Distinguishing protocol rationale, planning assumptions, safety entries, and actual outcome estimates prevents the absence of data from being mistaken for a null result.

14. 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 principal limitation is explicit in the registry record: the study was terminated early because of inadequate support to complete recruitment and data analysis, and no outcome measure data were collected. There is therefore no primary endpoint result, secondary endpoint result, or analysis of the proposed relationships among anatomical and functional measures. The record cannot support an estimate of effect or its precision.

The intended group comparison is constrained by the non-randomized allocation and the registry's warning that the arms are not a like-for-like control comparison. Masking is listed as none. These features would require careful consideration even if outcome data were available, because group differences and measurement decisions may affect interpretation. The registry does not report adjustment methods or safeguards to address such concerns.

Important design and analysis details are not reported in the registry record. These include the numerical measurement unit, the precise primary contrast, the statistical model, handling of repeated measurements, analysis populations, missing-data procedures, multiplicity strategy, and measurement quality-control procedures. The record also does not describe the final group composition or the number of participants who received the planned intervention and completed imaging. The enrollment entry alone does not answer those questions.

The safety entries are limited as well. The healthy-control group is listed with zero at risk in the adverse-event categories, the AMD group has zero affected in the categories recorded, and death counts are not reported. The record does not specify the safety observation period or enough detail to establish the scope of event ascertainment.

Finally, the sample-size planning description cannot resolve the absence of outcome data. It is a planning illustration and is described in relation to an observed effect, yet the primary outcome record states that no measurements were collected. No trial-specific outcome effect or achieved precision can be inferred from that illustration.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Effects of Sildenafil on Choroidal Thickness in AMD — NCT01830790.
• Registry record last updated 2015-06-23; status last verified 2015-04.
• Results first posted 2015-05-14.
• 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.

18. Record Summary

The registry record describes a Duke University Phase 2 study of sildenafil citrate and choroidal thickness in age-related macular degeneration. The planned intervention was a single 100mg oral dose, with EDI-OCT imaging before treatment and at baseline, 1 hour, and 3 hours post-treatment. The study description includes AMD patients and healthy controls older than 65, while the record lists non-randomized allocation, no masking, and a single-group model.

The record lists enrollment as 10, a study start of 2013-03, primary completion of 2014-05, and status as terminated. It attributes early termination to inadequate support to complete recruitment and data analysis. No outcome measure data were collected, and zero analyses are listed. There is no reported estimate of change in choroidal thickness or result for the secondary outcome.

The adverse-event record lists zero affected in the recorded categories for the AMD group; healthy controls are listed with zero at risk, and death counts are not reported. These entries do not establish a safety conclusion. Overall, the registry documents the planned question and why the study ended, but provides no outcome evidence with which to answer the clinical question.