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
IRONMAN was a randomized, parallel, single-masked, phase 4 treatment trial enrolling 1160 participants with chronic heart failure, iron deficiency, and left ventricular systolic dysfunction. The registered primary endpoint was cardiovascular mortality or hospitalization for worsening heart failure, with the analysis specified to include both first and recurrent hospitalizations.
| Feature | IRONMAN |
|---|---|
| Clinical trial | IRONMAN |
| NCT identifier | NCT02642562 |
| Phase | Phase 4 |
| Status | Completed |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Single |
| Primary purpose | Treatment |
| Enrollment | 1160 |
| Number of arms | 2 |
| Intervention | Ferric derisomaltose |
| Lead sponsor | University of Glasgow |
| Start date | 2016-08 |
| Primary completion date | 2022-08-26 |
2. Clinical Question
The central clinical question is whether treatment with ferric derisomaltose affects the occurrence of cardiovascular mortality or hospitalization for worsening heart failure in patients with chronic heart failure, iron deficiency, and left ventricular systolic dysfunction.
Population
Patients with chronic heart failure, iron deficiency, and left ventricular systolic dysfunction.
Intervention
Ferric derisomaltose, administered intravenously.
Comparator
The other randomized trial arm. The registry information available here does not specify its intervention.
Primary question
How does randomized assignment to ferric derisomaltose compare with the other randomized arm for cardiovascular mortality or hospitalization for worsening heart failure?
3. Trial Design
IRONMAN used a randomized, parallel-group, single-masked design. The trial was classified as phase 4 and had a primary purpose of treatment. The registry records 1160 enrolled participants and two study arms.
Random assignment is intended to balance prognostic factors between treatment groups in expectation, allowing differences in outcomes to be interpreted in relation to randomized treatment assignment.
Participants remain in their randomized treatment groups rather than sequentially receiving both interventions as part of a crossover design.
The registry classifies the study as single-masked. The available record does not specify here which trial participants or personnel were masked.
The registered primary purpose is treatment, consistent with evaluating an intervention in a clinical population.
4. Trial Population and Clinical Setting
The registry identifies three conditions for the study: chronic heart failure, iron deficiency, and left ventricular systolic dysfunction. These conditions define the clinical population to which the registered trial question applies.
| Population feature | Registry description |
|---|---|
| Heart failure | Chronic heart failure |
| Iron status | Iron deficiency |
| Cardiac function | Left ventricular systolic dysfunction |
| Enrollment | 1160 |
| Study status | Completed |
The available registry information does not provide a baseline demographic table, treatment-group baseline counts, or subgroup characteristics. Those details therefore cannot be used to characterize balance between the two randomized arms here.
5. Primary Endpoint
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Primary endpoint | CV mortality or hospitalisation for worsening heart failure (analysis will include first and recurrent hospitalisations) | Minimum of 3 months follow-up from last patient recruited |
This endpoint combines two clinically important types of event: cardiovascular death and hospitalization for worsening heart failure. The registry specifically states that the analysis will include both first and recurrent hospitalizations, which is an important statistical distinction from an analysis that counts only the first hospitalization for each participant.
6. Why the Endpoint Requires More Than a Simple Event Proportion
A conventional binary analysis asks whether each participant ever experienced an event. That approach does not naturally distinguish between a participant who has one hospitalization and another who has several hospitalizations. IRONMAN's registered endpoint explicitly includes recurrent hospitalizations, so the timing and multiplicity of hospitalization events matter.
The endpoint also contains a terminal event: cardiovascular mortality. Once a participant dies, additional hospitalization events cannot occur afterward. This creates a statistical structure in which recurrent nonfatal events and a terminal event are related.
First-event analysis
A first-event approach records the time until the first qualifying event and treats later hospitalizations as outside the primary event definition.
Recurrent-event analysis
A recurrent-event approach allows multiple hospitalizations from the same participant to contribute information rather than discarding later events.
Terminal event
Cardiovascular death prevents further hospitalization observations after death, so death has a different statistical role from a repeat hospitalization.
Composite endpoint
The primary endpoint combines cardiovascular mortality with hospitalization for worsening heart failure, so interpretation depends on the contributions of both event types.
7. Planned Analysis
The registry defines the primary outcome as cardiovascular mortality or hospitalization for worsening heart failure and explicitly states that the analysis will include first and recurrent hospitalisations. The follow-up requirement is a minimum of 3 months from the last patient recruited.
The registry record available here does not post a formal statistical analysis for this endpoint. For an endpoint with recurrent hospitalizations, an analysis would typically account for repeated events within the same participant rather than treating every hospitalization as statistically independent.
This structure illustrates why ordinary methods for a single binary outcome may discard information contained in recurrent hospitalizations. Appropriate recurrent-event methods can account for within-participant dependence and for the terminal role of death.
Possible statistical frameworks for this type of endpoint include recurrent-event survival models, marginal rate-based approaches, joint models for recurrent and terminal events, and other prespecified composite-event methods. The choice among these approaches depends on the protocol and statistical analysis plan, including how recurrent hospitalizations, death, censoring, and the estimand are defined.
8. Statistical Methodology
Randomization
Randomization assigns participants to treatment groups using a mechanism intended to prevent systematic allocation based on prognosis. Its statistical value comes from creating comparability in expectation across randomized groups, particularly for characteristics that may be difficult to measure or are not known in advance.
For IRONMAN, the registered allocation is randomized and the design is parallel. The principal treatment comparison therefore concerns outcomes according to randomized assignment.
Time-to-event analysis
The primary endpoint contains events that occur over follow-up rather than at a single fixed assessment time. Time-to-event methods preserve information about when an event occurs and can accommodate participants who are followed for different lengths of time.
The survival function is the probability that the event time T exceeds time t. In recurrent-event settings, additional structure is needed because a participant can experience more than one nonfatal event.
Recurrent events
When hospitalizations can occur repeatedly, observations from the same participant are correlated. Counting every hospitalization as though it came from a different independent person can underestimate uncertainty and produce inappropriate inference.
Recurrent-event methods address this by explicitly modeling repeated events, using either participant-level risk structures, event-specific processes, marginal event rates, or other approaches that account for within-person dependence.
Terminal events and competing processes
Cardiovascular mortality is different from a recurrent hospitalization because death ends the possibility of observing subsequent hospitalizations. Statistical methods therefore need to distinguish recurrent nonfatal events from a terminal event when the analysis uses information beyond the first event.
Censoring
In time-to-event analysis, a participant may reach the end of available follow-up without experiencing the event of interest. Such an observation is censored rather than treated as if the event never could occur. Valid inference depends on appropriate definitions of the censoring mechanism and the analysis population.
9. Statistical Methods Explained
Why does recurrent hospitalization change the analysis?
If only the first hospitalization were relevant, a standard time-to-first-event framework could summarize the time until the first qualifying event. Once recurrent hospitalizations are included, later events contain additional information. A suitable recurrent-event analysis must therefore account for multiple observations arising from the same participant.
Why is cardiovascular death different from a repeat hospitalization?
A participant can experience multiple hospitalizations while alive, but death prevents further hospitalization events from occurring afterward. Cardiovascular death is therefore a terminal event within the recurrent-event process. Treating death as simply another repeat hospitalization would ignore this structural difference.
Why might a simple proportion be insufficient?
A proportion such as "the percentage hospitalized" collapses timing and recurrence into one number. Two treatment groups could have the same percentage of participants hospitalized while differing substantially in how quickly hospitalization occurred or how often participants were hospitalized.
Why are time-to-event methods useful?
Time-to-event methods incorporate both event occurrence and follow-up time. They can also handle right-censoring, allowing participants who have not experienced the event by the end of observation to contribute information without pretending their eventual outcome is known.
Why is within-participant correlation important?
Repeated hospitalizations from one participant are not independent observations in the same way that outcomes from unrelated participants are. A participant with repeated admissions contributes a cluster of related observations. Statistical inference therefore needs to account for that dependence.
What would a hazard ratio mean if one were reported?
A hazard ratio would be a model-based relative comparison of event rates over time under the assumptions of the selected survival model. It would not directly represent a percentage of participants who benefit, a percentage reduction in the probability of hospitalization, or a difference in the number of hospitalizations.
What would a recurrent-event rate ratio mean?
A rate ratio would compare the estimated occurrence rate of qualifying recurrent events between randomized groups under the specified model and follow-up framework. Its interpretation would depend on exactly how exposure time, recurrent hospitalizations, cardiovascular death, and dependence between repeated events were handled.
10. Analysis Population and Estimand Considerations
Randomization makes the treatment assignment the natural anchor for the primary causal comparison. In a randomized parallel trial, an intention-to-treat analysis generally compares participants according to the groups to which they were randomized, regardless of subsequent adherence or treatment changes.
For IRONMAN, the registry information available here does not specify an analysis population definition, treatment-discontinuation rules, handling of protocol deviations, or a formal estimand. Those details are important because the estimated treatment effect can depend on whether the analysis targets treatment assignment, treatment received, or another treatment strategy.
| Estimand consideration | Why it matters for this endpoint |
|---|---|
| Randomized assignment | Preserves the principal advantage of randomization for the treatment comparison. |
| Hospitalization recurrence | Determines whether repeated hospitalizations contribute to the estimand rather than only the first event. |
| Cardiovascular death | Acts as a terminal event that prevents further hospitalization observations. |
| Follow-up | The registry specifies a minimum of 3 months of follow-up from the last patient recruited. |
| Censoring | Determines how participants without an observed qualifying event contribute information. |
11. Missing Data, Censoring, and Follow-Up
The registered primary endpoint is assessed over time, making follow-up and censoring central to its analysis. The registry specifies a minimum of 3 months of follow-up from the last patient recruited.
The available registry information does not specify an imputation method for missing outcome information, a detailed censoring algorithm, or rules for handling incomplete hospitalization records. These are methodological decisions that would ordinarily be documented in a statistical analysis plan for a recurrent-event endpoint.
Observed event
A qualifying cardiovascular death or hospitalization can contribute directly to the event process.
Censored observation
A participant who reaches the end of observable follow-up without a qualifying event may contribute information up to that point.
Repeated hospitalization
A later hospitalization can contribute additional information because the registered endpoint explicitly includes recurrent hospitalizations.
Death
Cardiovascular death terminates the possibility of observing subsequent hospitalization events.
12. Multiplicity, Interim Analysis, and Other Design Features
The registry information available here identifies a single registered primary endpoint. It does not report an alpha-spending strategy, multiplicity adjustment, interim efficacy analysis, non-inferiority margin, Bayesian analysis, factorial structure, or crossover design.
| Design topic | Registry information |
|---|---|
| Primary endpoint | One registered composite primary endpoint |
| Multiplicity strategy | Not reported in the available registry information |
| Interim analysis | Not reported in the available registry information |
| Non-inferiority margin | Not reported in the available registry information |
| Bayesian methods | Not reported in the available registry information |
| Factorial design | Not reported; the registered design model is parallel |
| Crossover | Not reported |
| Stratification factors | Not reported in the available registry information |
13. Results Status
The ClinicalTrials.gov record identifies the study as completed, with a primary completion date of 2022-08-26. However, the record considered here does not contain posted statistical analyses for the registered primary endpoint.
Primary endpoint
The registry record does not provide a treatment-effect estimate, confidence interval, or p-value for the primary endpoint.
Registered endpoint: CV mortality or hospitalisation for worsening heart failure, including first and recurrent hospitalisations.
The absence of a posted statistical analysis is distinct from the trial being incomplete. The registry reports a completed study, while the available statistical-analysis fields do not contain a formal result for the primary endpoint.
14. Safety and Secondary Outcomes
The available registry information identifies the intervention and the primary endpoint but does not provide posted secondary endpoint results or serious adverse event counts by randomized arm.
| Outcome category | Available registry information |
|---|---|
| Primary endpoint | CV mortality or hospitalisation for worsening heart failure, including first and recurrent hospitalisations |
| Secondary endpoint results | No posted statistical analyses in the available record |
| Serious adverse events by arm | Not reported in the available information |
| Treatment-emergent adverse events by arm | Not reported in the available information |
Safety and efficacy answer different questions. Even when safety events are summarized alongside a primary efficacy endpoint, they should remain separate components of the statistical interpretation because their denominators, observation periods, event definitions, and clinical meanings can differ.
15. Limitations
- No posted primary endpoint analysis: the available ClinicalTrials.gov record does not provide an effect estimate, confidence interval, or p-value for the primary endpoint.
- Recurrent-event complexity: the registered endpoint includes recurrent hospitalizations, requiring attention to within-participant dependence and event recurrence.
- Terminal event: cardiovascular death prevents subsequent hospitalization events, creating an additional modeling consideration.
- Incomplete statistical specification: the available registry information does not specify the formal recurrent-event model, censoring rules, missing-data strategy, or analysis population.
- Limited comparator information: the available record identifies ferric derisomaltose as the intervention but does not specify the intervention used in the other randomized arm in the information considered here.
- Limited baseline information: the available record does not provide treatment-group baseline characteristics or subgroup distributions.
- Limited safety information: serious adverse event counts by randomized arm are not provided in the available record.
- Follow-up definition: the registered primary endpoint specifies a minimum of 3 months of follow-up from the last patient recruited; longer-term follow-up information is not specified here.
16. Why This Trial Matters Statistically
IRONMAN provides a useful teaching example because its primary endpoint goes beyond a simple first-event comparison. The registry explicitly requires analysis of first and recurrent hospitalizations together with cardiovascular mortality. That combination illustrates why the statistical definition of an endpoint matters as much as the clinical wording.
| Concept | How it appears in IRONMAN |
|---|---|
| Randomization | The study uses randomized allocation. |
| Parallel design | The registry classifies the design model as parallel. |
| Single masking | The registry classifies the masking as single. |
| Time-to-event analysis | The primary endpoint is defined over follow-up rather than as a single fixed-time measurement. |
| Composite endpoint | Cardiovascular mortality and hospitalization for worsening heart failure are combined into one primary endpoint. |
| Recurrent events | The analysis explicitly includes first and recurrent hospitalizations. |
| Terminal event | Cardiovascular mortality ends the opportunity for subsequent hospitalization events. |
| Censoring | Follow-up duration determines how long participants can contribute observed event information. |
| Estimand | The definition of the treatment effect depends on how recurrent events, death, censoring, and randomized assignment are incorporated. |
17. Interpreting the Primary Endpoint Without a Posted Estimate
The registered primary endpoint asks whether participants experience cardiovascular mortality or hospitalization for worsening heart failure during follow-up, with recurrent hospitalizations included in the analysis.
The endpoint definition alone does not establish how large the treatment effect is. An effect estimate requires observed event data and a prespecified statistical analysis.
If two groups have similar numbers of participants with at least one hospitalization but different numbers of total hospitalizations, a first-event analysis could obscure that difference. A recurrent-event analysis can retain information from subsequent admissions.
A participant who dies from cardiovascular causes can no longer experience later hospitalization events. The statistical treatment of death therefore affects the interpretation of any recurrent-hospitalization analysis.
If a formal hypothesis test were reported, its p-value would quantify evidence against a specified null hypothesis under the chosen model and assumptions. It would not by itself describe the magnitude, clinical importance, or absolute impact of the treatment effect.
18. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The registered design supports a randomized comparison, while the primary endpoint requires methods capable of handling a composite outcome with recurrent hospitalization events and cardiovascular death.
Clinical interpretation
The endpoint is directly connected to clinically meaningful outcomes in chronic heart failure: cardiovascular mortality and hospitalization for worsening heart failure.
These perspectives should remain distinct. Statistical analysis determines how the observed event history is summarized and compared, while clinical interpretation determines what those outcomes mean in the context of heart failure care.
19. Timeline
Trial start
The registered study start date was August 2016.
Randomized phase 4 study
The completed study enrolled 1160 participants across two randomized parallel arms.
Cardiovascular mortality or worsening-heart-failure hospitalization
The primary endpoint includes both first and recurrent hospitalizations and specifies a minimum of 3 months of follow-up from the last patient recruited.
Primary completion
The registered primary completion date was August 26, 2022.
20. What Would Make the Primary Analysis Statistically Informative?
A complete statistical result for this endpoint would ideally identify the treatment effect measure, its uncertainty interval, the analysis population, the recurrent-event framework, and the handling of cardiovascular death and censoring. Those elements allow readers to determine exactly what quantity was estimated and how it should be interpreted.
| Result component | Why it matters |
|---|---|
| Effect estimate | Quantifies the relative or absolute difference between randomized groups. |
| Confidence interval | Describes statistical uncertainty around the estimated treatment effect. |
| Event definition | Determines which hospitalizations and deaths contribute to the endpoint. |
| Recurrent-event model | Determines how repeated hospitalizations from the same participant contribute to inference. |
| Terminal-event handling | Determines how cardiovascular death affects the recurrent hospitalization process. |
| Analysis population | Defines which randomized or treated participants contribute to the primary estimate. |
| Censoring rules | Determine when participants leave the observable risk set. |
21. Clinical Biostats Statistical Perspective
The most important statistical feature of IRONMAN is not simply that the trial is randomized. It is that the registered primary endpoint is a composite recurrent-event outcome. A participant can contribute more than one hospitalization, while cardiovascular death permanently ends the possibility of additional hospitalizations.
That structure creates several layers of interpretation. First, the endpoint must be defined precisely. Second, repeated observations within a participant must be handled appropriately. Third, the terminal nature of death must be recognized. Fourth, censoring and follow-up determine which portions of each participant's event history are observed.
This is why a single event proportion can be an incomplete representation of the underlying data. The statistical analysis should correspond to the clinical question: not merely whether hospitalization ever happened, but how the overall burden of qualifying events and cardiovascular mortality differs between randomized groups under the prespecified estimand.
22. Sources
- ClinicalTrials.gov: NCT02642562 — IRONMAN.
Continue through Clinical Biostats
Explore additional clinical trial results, statistical tutorials, and quantitative methods for clinical research.
23. Record Summary
IRONMAN is a completed randomized phase 4 trial of intravenous ferric derisomaltose in 1160 patients with chronic heart failure, iron deficiency, and left ventricular systolic dysfunction. Its registered primary endpoint is cardiovascular mortality or hospitalization for worsening heart failure, with the analysis explicitly including both first and recurrent hospitalizations and requiring a minimum of 3 months of follow-up from the last patient recruited.
Statistically, the endpoint is important because it combines a terminal event with recurrent nonfatal events. That structure requires more careful analysis than a simple binary comparison. The ClinicalTrials.gov record establishes the randomized parallel design and primary endpoint but does not provide a posted statistical analysis or treatment-effect estimate for that endpoint in the available record.