← Clinical Trials
Ischemic Cardiomyopathy Phase 1 Crossover Design NCT00203203

FOCUS HF: Complete Statistical Analysis of Autologous Stem Cells in Ischemic Cardiomyopathy

An independent statistical review of the randomized phase 1 FOCUS HF trial evaluating intramyocardial injection of stem cells via NOGA mapping in patients with ischemic cardiomyopathy, with particular attention to its crossover design and safety endpoint.

Trial period: 2004-04 to 2008-08  ·  Lead sponsor: Texas Heart Institute  ·  Enrollment: 30
Registry context

This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.

1. Trial at a Glance

FOCUS HF was a randomized, single-masked, phase 1 crossover trial evaluating intramyocardial injection of stem cells via NOGA mapping in patients with ischemic cardiomyopathy. The registry reports an enrollment of 30 and identifies safety of autologous-bone-marrow injections as the primary endpoint.

30
Enrollment
Randomized trial
2
Arms
Crossover design
1
Phase
Phase 1
6 mo
Safety window
Final listed time point
FeatureFOCUS HF
PhasePhase 1
ConditionIschemic Cardiomyopathy
AllocationRandomized
Design modelCrossover
MaskingSingle
Primary purposeTreatment
Enrollment30
Number of arms2
Lead sponsorTexas Heart Institute
StatusCompleted
Study start2004-04
Primary completion2008-08

2. Clinical Question

The clinical question was whether intramyocardial injection of autologous bone-marrow-derived stem cells, administered via NOGA mapping, could be evaluated as a treatment in patients with ischemic cardiomyopathy while monitoring safety at prespecified post-procedure time points.

Population

Patients with ischemic cardiomyopathy enrolled in the FOCUS HF trial.

Intervention

Intramyocardial injection of stem cells via NOGA mapping.

Comparator

Control, followed by stem cell therapy in the crossover design.

Primary question

What safety experience was observed following autologous bone-marrow injections at up to 2 weeks, 3 months, and 6 months after the procedure?

3. Trial Design

01
Randomize30 participants
02
CrossoverTwo treatment sequences
03
Stem cellsIntramyocardial injection
04
Safety reviewAdverse events
05
Follow-upUp to 6 months
ARM 1 · CONTROL THEN THERAPY

Control, Then Stem Cell Therapy

  • Control was followed by stem cell therapy.
  • Registry-reported serious adverse events: 7 affected out of 10 at risk.
ARM 2 · STEM CELL THERAPY

Stem Cell Therapy

  • Stem cell therapy was included in the treatment sequence.
  • Registry-reported serious adverse events: 11 affected out of 20 at risk.
The crossover structure is central to interpretation. A crossover trial does not compare two permanently separate groups in the same way as a conventional parallel-group trial. Participants can contribute observations under different treatment conditions, so treatment sequence, period, and carryover effects can become important statistical considerations. The registry identifies the design as crossover but does not report a formal crossover-model analysis.

4. Endpoints

EndpointTime frameRegistry definition
Safety of Autologous-bone-marrow Injections up to 2 weeks post-procedure, 3 months and 6 months Safety of cell injections was assessed by reviewing adverse events at 3 time points: up to 2 weeks post-procedure, 3 months post-procedure, and 6 months. Major adverse events were adjudicated, including hospitalization, arrhythmia, exacerbation of congestive HF (CHF), acute coronary syndrome, myocardial infarction, stroke, or death.

The primary endpoint is therefore a longitudinal safety assessment rather than a conventional efficacy endpoint such as survival, response, or change from baseline. The registry defines several clinically important adverse-event categories and specifies three follow-up points.

5. Statistical Methodology

Safety as a repeated assessment

The primary endpoint evaluates adverse events over three specified post-procedure periods. For a safety endpoint of this type, the first statistical task is descriptive: identify how many participants experience each relevant adverse event and summarize the number or proportion affected within the appropriate treatment or sequence group.

Because the endpoint is assessed at multiple time points, the timing of an event matters. A participant experiencing an event during the first post-procedure period and remaining event-free thereafter should not be treated as though the same observation independently occurred at every later visit. Repeated safety assessments therefore require careful definitions of whether the analysis concerns first occurrence, events during each interval, cumulative incidence, or participants ever affected.

Binary safety proportions

When a safety outcome is summarized as affected versus not affected, the basic descriptive quantity is a proportion:

Basic safety proportion
p = x / n

where x is the number of participants affected and n is the relevant number at risk. For a small study, confidence intervals for a proportion should generally use methods appropriate for small binomial samples rather than relying automatically on large-sample normal approximations.

Why the crossover design changes the analysis

In a parallel-group trial, treatment assignment remains fixed throughout the comparison. In a crossover trial, treatment condition can change within a participant. This creates the possibility of within-participant comparisons, but it also means that the statistical analysis must distinguish treatment effects from period effects and sequence effects.

For a continuous endpoint, a crossover analysis might use a model containing treatment, period, and sequence terms, with participant-level variation accounted for through the repeated observations. For a binary or event-based safety endpoint, the appropriate model depends on the exact event definition, timing, and whether the analysis is based on first events, interval-specific events, or cumulative events.

Adverse-event adjudication

The registry states that major adverse events were adjudicated. Adjudication is important because clinically important events may require consistent classification before they are counted in a statistical safety analysis. The listed categories include hospitalization, arrhythmia, exacerbation of congestive HF, acute coronary syndrome, myocardial infarction, stroke, and death.

Small-sample inference

The registry explicitly characterizes the study as preliminary and states that the small sample size limits the statistical rigor of the analysis. With an enrollment of 30, statistical estimates can be unstable, confidence intervals can be wide, and formal hypothesis tests can have limited power to distinguish modest differences from random variation.

Small samples require restraint. A numerical difference in adverse-event proportions is not, by itself, evidence of a treatment effect. In a small crossover study, the denominator, timing of exposure, treatment sequence, and definition of the event all influence how a safety comparison should be interpreted.

6. Statistical Methods Explained

Why does the crossover design matter?

A crossover design allows participants to receive different treatment conditions during the study rather than remaining in only one randomized group. This can provide within-participant information, but it introduces additional statistical structure. Treatment sequence and study period can influence the observed outcome, and treatment effects may persist into a later period.

Why is a safety endpoint usually summarized with proportions?

For many adverse-event outcomes, the clinically direct quantity is the proportion of participants affected. For example, the registry reports serious adverse events as affected participants divided by participants at risk in each arm. This is immediately interpretable, but the denominator must be defined correctly and must correspond to the relevant exposure or analysis population.

Why are three follow-up points important?

The registry specifies safety assessment up to 2 weeks post-procedure, at 3 months, and at 6 months. These time points distinguish early post-procedure safety from later follow-up. Combining all observations into a single number can obscure when events occurred.

What does "affected / at risk" mean?

The registry reports serious adverse events as 7/10 for the Control, Then Stem Cell Therapy group and 11/20 for the Stem Cell Therapy group. The numerator identifies participants affected, while the denominator identifies the participants at risk for the reported summary. The ratio should not automatically be interpreted as an incidence rate or as a causal treatment effect.

Why is statistical inference difficult with 30 participants?

Small samples provide relatively little information for estimating event probabilities precisely. A small change in the number of participants experiencing an event can materially change the observed proportion. Exact or small-sample methods can improve the validity of interval estimates, but they cannot create information that the study size does not contain.

What would a formal crossover analysis need to consider?

A formal analysis would generally distinguish treatment from period and sequence effects and would account for the repeated contribution of participants across treatment conditions. For safety events, the precise model would depend on whether the endpoint is defined as an event during a particular period, first occurrence, or cumulative occurrence. The registry does not report a formal statistical model for the primary safety endpoint.

7. Planned Analysis

The registry reports results for the primary safety endpoint but does not post formal statistical analyses for that endpoint. The available safety results are therefore descriptive rather than accompanied by a reported treatment-effect estimate, confidence interval, or p-value.

For an endpoint of this type, a typical statistical analysis would begin with participant-level counts and proportions of adverse events at each prespecified follow-up period. Major adverse events would be classified according to the adjudicated definitions, and cumulative as well as time-specific summaries could be considered if they were prespecified. In a crossover design, any formal comparative analysis would also need to account for treatment sequence and period, with particular attention to the possibility that an earlier treatment could influence observations in a later period.

Registry reporting status: ClinicalTrials.gov does not post formal statistical analyses for the primary safety endpoint. The record does report serious adverse events by arm, allowing a descriptive safety summary but not a reported inferential comparison with a confidence interval or p-value.

8. Safety Results

The registry reports serious adverse events by arm using an affected-versus-at-risk format. These are the principal numerical safety results reported in the record.

ArmSerious adverse eventsInterpretation of reported format
Control, Then Stem Cell Therapy 7 / 10 7 participants affected among 10 at risk.
Stem Cell Therapy 11 / 20 11 participants affected among 20 at risk.
Serious adverse events · affected participants
Control, Then Stem Cell Therapy
7 / 10
Stem Cell Therapy
11 / 20

The denominators differ between the two reported arm summaries, so the raw event counts alone should not be compared. The reported affected-to-at-risk ratios are descriptive summaries of serious adverse events in the registry record, not adjusted estimates of treatment effect.

Reading the safety numbers carefully

The reported figures describe serious adverse events, but the registry's primary endpoint is broader: safety of autologous-bone-marrow injections assessed through adverse events at three specified time points, with major adverse events adjudicated. A serious-adverse-event summary should therefore not be treated as a complete description of every component of the primary endpoint.

Statistical interpretation

The registry reports 7/10 serious adverse-event participants in the Control, Then Stem Cell Therapy group and 11/20 in the Stem Cell Therapy group. These are descriptive affected-to-at-risk summaries. They do not establish that one treatment sequence caused more or fewer serious adverse events.

The confidence interval around a safety proportion would describe statistical uncertainty in the estimated proportion under a specified sampling model. No such confidence intervals are posted for these results in the registry.

A p-value, had one been reported, would address compatibility with a specified null hypothesis under the chosen statistical model. It would not measure the size or clinical importance of a safety difference.

9. Crossover Analysis: What Would Need to Be Considered?

The crossover structure is one of the most important statistical features of FOCUS HF. Unlike a simple parallel-group study, the treatment condition can change within a participant. That creates opportunities for paired information but also creates additional sources of bias and dependence.

Treatment effect
The analysis needs to distinguish the effect associated with stem cell therapy from differences that arise simply because observations occur under different treatment conditions.
Period effect
Outcomes can change over calendar or study periods for reasons unrelated to treatment, particularly when the participant is observed at different stages of follow-up.
Sequence effect
Participants receiving different treatment sequences may have systematically different outcome patterns if sequence interacts with the treatment or study period.
Carryover
If an earlier treatment has effects that persist into a later period, the later observation may not represent an untreated or independently exposed state.

These issues are especially important for safety outcomes because adverse events can occur at specific times relative to a procedure. The timing of exposure and follow-up must therefore be aligned with the definition of the endpoint.

Conceptual crossover model
Outcome = treatment + period + sequence + participant-level variation

This is a conceptual representation of the factors a formal crossover analysis may need to distinguish. The registry identifies the crossover design but does not report that this particular model was used.

10. Safety Analysis and Serious Adverse Events

Safety analysis has a different statistical objective from efficacy analysis. The primary question is generally not whether treatment produces a statistically significant improvement, but whether adverse outcomes occur, how often they occur, when they occur, and whether their frequency or severity differs meaningfully between treatment conditions.

The FOCUS HF record specifically identifies major adverse events that were adjudicated: hospitalization, arrhythmia, exacerbation of congestive HF (CHF), acute coronary syndrome, myocardial infarction, stroke, or death.

Major adverse-event category named in the registryRole in endpoint
HospitalizationIncluded among major adverse events subject to adjudication.
ArrhythmiaIncluded among major adverse events subject to adjudication.
Exacerbation of congestive HF (CHF)Included among major adverse events subject to adjudication.
Acute coronary syndromeIncluded among major adverse events subject to adjudication.
Myocardial infarctionIncluded among major adverse events subject to adjudication.
StrokeIncluded among major adverse events subject to adjudication.
DeathIncluded among major adverse events subject to adjudication.

Because these outcomes differ substantially in clinical meaning and frequency, a single composite safety summary can conceal important differences among its components. A detailed safety analysis would normally examine the individual components as well as any prespecified composite definition.

11. Analysis Populations and Denominators

The registry reports overall enrollment of 30, while the serious-adverse-event summaries use denominators of 10 and 20 for the two reported arm descriptions. Those denominators should be preserved rather than substituted with the overall enrollment count.

Why denominators matter

Affected participants cannot be interpreted without knowing the corresponding population at risk. The same event count can represent very different frequencies in different denominators.

Why crossover complicates denominators

When participants receive more than one treatment condition, the population contributing to an exposure-specific safety summary may not correspond directly to the original randomized group size.

Participant-level analysis

Adverse-event analyses should distinguish participants from events. One participant can potentially experience more than one event, so event counts and affected-participant counts are not interchangeable.

Time-specific analysis

The registry specifies three follow-up points. The analysis therefore needs to preserve the temporal relationship between the procedure and the adverse event.

12. What a Confidence Interval Would Add

The registry does not report confidence intervals for the serious-adverse-event summaries. Nevertheless, confidence intervals are useful for understanding why a small study's observed proportion should not be treated as a precise estimate of the underlying event probability.

Precision versus magnitude

An observed safety proportion is an estimate based on a finite number of participants. A confidence interval expresses uncertainty around that estimate under a specified statistical framework. With small denominators, the interval can be substantially wider than the point estimate might suggest.

What the interval would not mean

A confidence interval would not describe the range of adverse-event frequencies that individual patients could experience. It would quantify uncertainty about a population-level parameter under the statistical model used to construct it.

Why no p-value is needed to describe the result

For a preliminary safety study, descriptive reporting can be informative without converting the observed data into a binary significant-versus-nonsignificant conclusion. The clinical importance of a serious adverse event does not depend on crossing an arbitrary statistical threshold.

13. Limitations

The registry itself states that the study is a preliminary study based on a small sample size, which limits the statistical rigor of the analysis. That limitation is particularly important when interpreting safety results from a crossover study.

14. Why This Trial Matters Statistically

FOCUS HF is a useful teaching case because its statistical challenge is different from that of a conventional large parallel-group efficacy trial. The central issues are the design and interpretation of safety data in a small randomized crossover study.

ConceptHow it appears in FOCUS HF
RandomizationThe registry identifies the allocation as randomized.
Crossover designThe design model is explicitly listed as crossover.
Single maskingThe registry identifies the trial as single-masked.
Safety analysisThe primary endpoint evaluates safety of autologous-bone-marrow injections.
Repeated follow-upSafety is assessed up to 2 weeks, at 3 months, and at 6 months.
Event adjudicationMajor adverse events are adjudicated.
Small-sample inferenceThe registry explicitly notes that the small sample size limits statistical rigor.
Denominator disciplineSerious adverse events are reported as affected participants divided by participants at risk.
Within-participant dependenceCrossover means observations can be related within participants.
Period and sequence effectsThese are important considerations whenever treatment conditions change during a crossover study.

15. Clinical Interpretation vs Statistical Interpretation

Statistical interpretation

The available registry results are descriptive serious-adverse-event counts by arm. No formal confidence intervals, p-values, or treatment-effect estimates are posted for the primary safety endpoint.

Clinical interpretation

The study focuses on safety following autologous bone-marrow injections in ischemic cardiomyopathy, with adverse events reviewed at specified post-procedure time points and major events adjudicated.

16. The Difference Between an Event Count and a Treatment Effect

One of the most important statistical distinctions in this study is the difference between describing what happened and estimating whether treatment caused a difference.

Descriptive versus comparative question
Descriptive: How many participants were affected?
Comparative: How does the probability or rate differ between treatment conditions?

The first question can be answered directly from the reported affected-to-at-risk counts. The second requires a prespecified comparative statistical framework that accounts for the crossover design and the timing of exposure and events.

This distinction prevents a common error in clinical-trial interpretation: treating a numerical difference between two observed proportions as though it were automatically an estimated causal treatment effect.

17. Longitudinal Safety Assessment

The three listed follow-up windows provide a natural framework for thinking about safety longitudinally. Events occurring soon after a procedure may have a different relationship to treatment exposure than events identified months later.

Up to 2 weeks post-procedure

Early safety

The first registry-specified assessment window captures adverse events during the early post-procedure period.

3 months

Intermediate follow-up

The second assessment extends safety evaluation beyond the immediate post-procedure period.

6 months

Later listed safety assessment

The final registry-specified time point extends the safety assessment to 6 months after the procedure.

A longitudinal analysis should preserve these distinctions rather than treating every adverse event as if it occurred at the same time. If participants can change treatment condition during the study, exposure timing becomes even more important.

18. What the Registry Does and Does Not Establish Statistically

QuestionWhat the record supports
Was the study randomized?Yes. Allocation is listed as randomized.
Was the design crossover?Yes. The design model is listed as crossover.
Was the trial phase 1?Yes.
Was safety a primary endpoint?Yes. Safety of Autologous-bone-marrow Injections is the registered primary endpoint.
Were three safety time points specified?Yes: up to 2 weeks post-procedure, 3 months, and 6 months.
Were major adverse events adjudicated?Yes.
Are serious adverse events reported by arm?Yes: 7/10 and 11/20.
Are formal primary-endpoint confidence intervals posted?No formal statistical analyses are posted for the primary endpoint.
Are primary-endpoint p-values posted?No formal statistical analyses are posted for the primary endpoint.

19. Important Statistical Interpretation Issues

20. Why Crossover Design Is Statistically Interesting

Crossover trials are often attractive because each participant can provide information under more than one treatment condition. In principle, this can reduce some between-person variability. But that advantage is only realized when the endpoint and treatment effects are compatible with crossover analysis.

Potential advantage

Within-participant comparisons can reduce variability attributable to stable differences between participants.

Period effects

The outcome can change over time independently of treatment, making treatment and time difficult to separate without an appropriate model.

Carryover

A treatment effect that persists into a subsequent period can contaminate the later comparison.

Safety-specific concern

For adverse events, exposure timing and procedure timing can be central to determining which treatment condition is associated with an event.

These considerations make FOCUS HF particularly useful for understanding why the choice of statistical method should follow the design and endpoint rather than simply the fact that the trial is randomized.

21. Trial Timeline

2004-04

Study start

The registry lists April 2004 as the study start date.

2008-08

Primary completion

The registry lists August 2008 as the primary completion date.

Completed

Registry status

The trial is listed as completed, with results posted in the registry.

22. Statistical Concepts in This Trial

Learn more about the methods used in this trial:

23. Related Statistical Calculators

24. Sources

Continue through the Clinical Biostats knowledge graph

Explore clinical-trial designs and statistical methods through related tutorials and analytical tools.

25. Record Summary

FOCUS HF provides a compact example of the statistical issues that arise when a randomized clinical study uses a crossover design and focuses on safety rather than a conventional efficacy endpoint. The registry identifies 30 participants, 2 arms, randomized allocation, single masking, and a crossover design. Its primary endpoint evaluates safety of autologous-bone-marrow injections at up to 2 weeks post-procedure, 3 months, and 6 months, with major adverse events adjudicated. The registry reports serious adverse events as 7/10 for the Control, Then Stem Cell Therapy group and 11/20 for the Stem Cell Therapy group.

The most important statistical lesson is that these counts are descriptive safety results. A formal comparative analysis would need to respect the crossover structure, exposure timing, treatment sequence, period effects, and the small sample size. The registry itself emphasizes that the study is preliminary and that its small sample size limits the statistical rigor of the analysis. That context is essential when interpreting the numerical safety findings.

Statistical perspective: FOCUS HF illustrates why clinical-trial interpretation begins with the relationship between design, endpoint, timing, and analysis population. In a small crossover safety study, careful attention to denominators and exposure periods can be more informative than applying a generic hypothesis test without regard to the trial structure.