Randomized Placebo-controlled Trial of FCM as Treatment for Heart Failure With Iron Deficiency / and Sub-Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Randomized Placebo-controlled Trial of FCM as Treatment for Heart Failure With Iron Deficiency / and Sub-Study (NCT03037931), a randomized phase 3 study. The primary outcome is Number of Deaths.

2. Clinical Question

The registry describes a treatment comparison focused on efficacy and safety in heart failure with reduced ejection fraction and iron deficiency. The posted primary result addresses mortality as a binary outcome: whether a participant died during the stated timeframe. That question is narrower than the full clinical objective, which also refers to safety and treatment efficacy more broadly. Although the registry lists 3 outcome measures, the other outcome details are not included in the record summarized by these results, and no secondary posted results are listed.

The comparison is between assignment to the ferric carboxymaltose regimen and assignment to placebo. Because the trial is randomized and double-blind, its design is intended to support a controlled comparison while limiting some sources of bias. The registry does not report enough implementation detail to assess the success of masking or other conduct features. The primary endpoint should therefore be interpreted on its own terms: it records deaths in the intent-to-treat population over 1 year, rather than a complete account of all benefits, harms, symptoms or changes in heart failure status.

3. Trial Design

This was a randomized, double-blind, parallel-group phase 3 study sponsored by American Regent, Inc. The registry title describes it as a placebo-controlled trial of FCM for heart failure with iron deficiency, with a sub-study. The brief summary identifies the central objective as assessing the efficacy and safety of intravenous ferric carboxymaltose relative to placebo in participants with heart failure with reduced ejection fraction and iron deficiency.

The record lists enrollment of 3065 participants. The posted primary outcome group sizes are 1532 for ferric carboxymaltose and 1533 for placebo. The intervention arm received ferric carboxymaltose intravenously in 2 doses 7 days apart, with dosing and repeat treatment described in relation to body weight, maximum doses, and iron indices. The placebo arm received normal saline in 2 doses 7 days apart, with repeat dosing every 6 months. These are registry descriptions; additional treatment-delivery details are not reported in the registry record.

The study start date is 2017-03-15, the primary completion date is 2023-02-02, and the status is completed. Those dates indicate the recorded study timeline but do not establish the follow-up experience of each participant. The registry does not report the original sample-size justification, the trial's own statistical analysis plan, or details of any interim analyses. It also does not describe the methods used to generate or conceal the random allocation sequence, or specify the roles covered by double-blinding.

The primary outcome is the number of deaths among participants over 1 year in the intent-to-treat population. The record reports 131 deaths among 1532 participants in the ferric carboxymaltose group and 158 among 1533 in the placebo group. Those are the posted primary endpoint counts. They should not be confused with the death counts listed in the adverse-event information, which are recorded in a separate safety context and have different values.

The registry lists 3 outcome measures and no posted analyses. The comparison statistics available for the primary result are therefore the relevant numerical basis for interpretation. Details about the sub-study, additional outcome definitions, and their results 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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent-to-treat Population

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 Number of Deaths, measured over 1 year.

The primary endpoint counts participants who died during the 1-year timeframe. It is defined as the number of participants who died out of the total treated group and is reported for the intent-to-treat population. In the posted results, the ferric carboxymaltose group had 131 deaths among 1532 participants, and the placebo group had 158 deaths among 1533 participants.

A binary endpoint records whether an event occurred for each participant, rather than measuring the timing or duration of the event. For this endpoint, a participant is counted as having the outcome if they died during the stated period. The endpoint therefore provides a direct comparison of the recorded counts and the associated risk estimates. It does not show when within the year a death occurred, how long participants were followed individually, or the causes of death.

Mortality is used as an outcome because it is an unambiguous event that can be counted, and it is clinically important in a study of heart failure. That general rationale does not mean that this endpoint captures every aspect of treatment effect or safety. Other outcomes may address different aspects of health or adverse events, but the available registry result does not provide their definitions or findings. The primary mortality endpoint should consequently be read as one specific measure, with its stated population and timeframe, rather than as a complete evaluation of all outcomes relevant to the trial.

7. Results Posted in the Registry

• Ferric Carboxymaltose: 8.6% (131/1532) versus 10.3% (158/1533) on control; risk difference -1.8 percentage points (95% CI -3.8 to 0.3), risk ratio 0.83 (0.67–1.03); unadjusted two-proportion z-test p = 0.096.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary: each participant either had the recorded event during the stated timeframe or did not. The registry gives counts and denominators for each randomized group, then reports comparative measures. The primary measure is the risk difference. It expresses the difference between the event risks in the two groups on an absolute scale. A negative estimate indicates a lower observed risk in the ferric carboxymaltose group relative to placebo under the stated comparison direction. Its value is -0.018.

What does the risk difference describe?

The risk difference compares the proportion of participants with the event in one group with the proportion in the other. The posted group risks are 0.086 for ferric carboxymaltose and 0.103 for placebo. The registry reports the risk difference directly, with an interval from -0.038 to 0.003. The interval reflects uncertainty around the estimated contrast. Because it includes zero, the range includes a result with no risk difference. It also includes values on either side of zero, so the direction and size of a possible underlying difference are not precisely determined by this interval.

How do other effect measures differ?

The registry also reports a risk ratio of 0.830 and an odds ratio of 0.814. A risk ratio compares the risks as a ratio; a value representing no difference is 1. An odds ratio compares event odds rather than risks, and its no-difference value is also 1. Their reported intervals are 0.665 to 1.034 for the risk ratio and 0.638 to 1.038 for the odds ratio. Both intervals include 1. These measures present the same group comparison on relative scales, rather than independent confirmations of an effect. For a binary outcome that is not extremely rare, an odds ratio can differ in interpretation from a risk ratio, so the measures should not be treated as interchangeable.

What does the p-value add?

The unadjusted p-value is 0.096. A p-value describes how compatible the observed data are with a statistical model in which there is no underlying group difference, given the assumptions of that model. It is not the probability that the null hypothesis is true, nor does it give the probability that the result is due to chance. A p-value also does not measure clinical importance. Its interpretation depends on the analysis and assumptions, including how the comparison was specified.

The planning illustration uses a two-sided significance level of 0.05, but the original trial's own design assumptions are not reported in the registry record. The posted p-value is unadjusted. No multiplicity adjustment is reported, and the record lists no posted analyses. It is therefore not possible to determine from the registry whether this comparison was part of a larger prespecified testing sequence or how the trial's investigators handled multiple endpoints or analyses.

Why does the interval matter?

An interval estimate gives a range of values compatible with the data under the statistical method used. It communicates uncertainty about the point estimate, rather than only indicating whether a test crosses a decision threshold. Here, the risk-difference interval crosses zero. The relative-measure intervals cross their no-difference value of 1. That pattern indicates that the posted estimates do not tightly distinguish a reduction from no difference, and the risk-difference interval also permits a small result in the opposite direction.

The confidence level and exact interval method are not reported in the registry record. Accordingly, the interval can be described as the posted interval, but its calibration and construction cannot be assessed from the available information. The same limitation applies to the p-value: the registry does not identify the statistical test or provide a full analysis specification.

What is the role of randomization and the intent-to-treat population?

Randomization aims to make treatment groups comparable at the start of a study, on average, including for characteristics that may not be measured. An intent-to-treat analysis generally compares participants according to their randomized assignment, regardless of later treatment changes. The registry identifies the primary outcome population as intent-to-treat, but detailed rules for defining that population, handling participants with incomplete outcomes, and managing treatment discontinuation are not reported. The posted result should therefore be understood using the population label, without assuming unreported analytic procedures.

The record describes the trial as double-blind and parallel-group. Masking can reduce some forms of bias in treatment administration and outcome assessment, while a parallel design compares groups followed under separate assigned regimens. These are general design concepts; registry descriptions alone do not establish how completely masking was maintained or how outcomes were adjudicated. Further details on allocation concealment, outcome ascertainment, model choice, covariate adjustment and missing-data methods are not reported.

10. Confidence Intervals and Precision

Precision concerns how much uncertainty surrounds an estimated effect. For the primary risk difference, the posted estimate is -0.018 and its interval extends from -0.038 to 0.003. The interval crosses zero, the value corresponding to equal risks. Thus, the data are compatible with a lower risk in the ferric carboxymaltose group, little or no difference, and a small difference in the opposite direction. The point estimate alone does not capture this uncertainty.

The alternative relative measures tell a similar story about precision. The risk ratio is 0.830, with an interval from 0.665 to 1.034; the odds ratio is 0.814, with an interval from 0.638 to 1.038. Their intervals cross 1, the no-difference value for those measures. The ranges are broad enough to include materially different estimates, so the magnitude of any underlying effect is not tightly pinned down by the posted results.

A confidence interval is generally a range generated by a statistical procedure to express uncertainty around an estimate. Its interpretation depends on the method and confidence level. The registry record does not report the confidence level or the method used to construct these intervals. No more detailed assessment of interval calibration is possible from the posted result alone.

The unadjusted p-value is 0.096. It should be read alongside the interval and effect estimate, not as a substitute for them. Neither an interval crossing its no-difference value nor a p-value above a chosen threshold proves that the groups are identical. Rather, these results leave uncertainty about whether an underlying difference exists and, if so, its size and direction. The registry does not report precision estimates for other outcomes in the available results.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total events. They are illustrative calculations based on the observed effect, not the trial's original design justification. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is specified as a fraction of the observed effect. The trial's own original sample-size assumptions are not reported in the registry record.

For planning for 100% of the observed effect, the illustration gives 4,348 participants per arm for 80% power and 5,821 participants per arm for 90% power. For planning for 75% of the observed effect, it gives 7,892 participants per arm for 80% power and 10,564 participants per arm for 90% power. For planning for 50% of the observed effect, it gives 18,117 participants per arm for 80% power and 24,253 participants per arm for 90% power.

Power is the probability that a planned study will detect an effect of the specified size under the planning assumptions, if that effect is present. The 80% power column corresponds to a design with an 80% chance of detecting the assumed effect; the 90% column corresponds to a higher chance under those same assumptions. Smaller effects are harder to distinguish from random variation, so planning for a smaller fraction of the observed effect requires more participants per arm. These are planning illustrations, not observed power for the completed trial.

Because the figures use the observed effect as their reference, they should not be read as the sample-size rationale for this trial. They also depend on the assumptions listed for the illustration and do not account for unreported features of the original protocol or analysis plan. The registry does not report whether its planned enrollment was based on this effect, a different endpoint, or other considerations.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.103; treatment proportion p₁ = 0.086; average p̄ = 0.094.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4133 = 0.8100.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4131 = 0.3477.
4. n = (0.8100 + 0.3477)² / (-0.0176)² = 4,347.9, rounded up to 4,348 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.103065883887802                          # control proportion
p1 <- 0.085509138381201                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(3037931)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 4,348 (80% power), 5,821 (90% power)
 75% of observed effect: n per arm = 7,892 (80% power), 10,564 (90% power)
 50% of observed effect: n per arm = 18,117 (80% power), 24,253 (90% power)
power.prop.test: 4347.9 
Simulated power at n = 4,348 per arm: 0.805 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when several outcomes, comparisons or analyses are examined. When many statistical tests are conducted, the chance of finding a nominally unusual result can increase unless the analysis plan accounts for the testing structure. This is a general statistical concern, not evidence that a particular error occurred in this trial. The registry lists 3 outcome measures, but no secondary results are posted, and the number of posted analyses is 0. The record does not state whether the primary endpoint was tested within a prespecified hierarchy or whether any multiplicity adjustment was planned or used.

The primary result is reported with an unadjusted p-value of 0.096. No adjusted p-value or multiplicity procedure is reported. Without the full analysis plan and the complete set of analyses, it is not possible to judge how the primary result relates to other planned tests or whether the reported p-value was intended to be confirmatory on its own. The sample-size illustration's two-sided significance level of 0.05 describes that planning exercise; it does not establish the original trial's testing plan.

The primary outcome population is identified as intent-to-treat. This label indicates that the analysis is intended to reflect randomized assignment, but the precise analysis population rules are not reported. The registry record also does not report how missing primary outcomes were handled, whether there were losses to follow-up, how participants with incomplete observation were treated, or whether sensitivity analyses assessed alternative assumptions about missing data.

These omissions matter because missing outcomes can affect estimates if the absence of a measurement is related to participants' event status or treatment assignment. This is a general methodological issue; the record does not establish that missing outcomes occurred or that they changed the result. Similarly, the lack of posted secondary analyses means the primary mortality finding cannot be placed alongside a complete set of other outcomes from the trial. Interpretation should remain limited to the primary estimate and the information actually recorded.

14. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (3.7%), acute kidney injury (3.0%), covid-19 (2.5%), chronic obstructive pulmonary disease (1.5%), sepsis (1.4%).

The adverse-event record lists serious adverse events affecting 413 participants in the ferric carboxymaltose group, among 1532 at risk, and 401 in the placebo group, among 1533 at risk. Other adverse events are listed for 424 participants in the ferric carboxymaltose group and 412 in the placebo group, with the same group-specific at-risk counts. These are counts of affected participants, not event totals, unless otherwise specified by the registry. They should be read as separate safety summaries rather than as the primary mortality outcome.

The listed top serious-event terms are Pneumonia at 3.7%, Acute kidney injury at 3.0%, COVID-19 at 2.5%, Chronic obstructive pulmonary disease at 1.5%, Sepsis at 1.4%, and Cellulitis at 1.4%. The registry record does not state in this summary whether these percentages are pooled across groups or provide the detailed denominators and coding rules for each term. No comparative estimates or uncertainty intervals for these specific terms are posted.

Deaths in the adverse-event information are 354 in the ferric carboxymaltose group and 367 in the placebo group. These are distinct from the primary endpoint counts of 131 and 158, respectively. The registry presents them in different outcome contexts; their definitions and observation periods are not reported in enough detail to reconcile them. They should not be substituted for the primary 1-year mortality result or combined with it. Adverse-event counts alone do not establish causation or provide a complete comparison of risk.

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

The posted result answers a defined statistical question: how the 1-year death risk in the intent-to-treat population compared between participants assigned to ferric carboxymaltose and those assigned to placebo. The point estimate for the risk difference is -0.018, while its interval ranges from -0.038 to 0.003. The group risk estimates are 0.086 and 0.103, and the unadjusted p-value is 0.096. These figures indicate a numerically lower observed risk in the ferric carboxymaltose group, but the interval includes no difference and does not rule out a small difference in the opposite direction.

The result does not establish that the groups had exactly the same mortality risk. A result that does not clearly distinguish a difference from no difference is not proof of equivalence. The registry record does not report an equivalence or noninferiority design, a prespecified margin, or the analyses needed to make such a claim. Equally, the numerical direction of the point estimate is not by itself proof of a treatment effect. Interpretation must take account of the uncertainty and the analysis details that are not reported.

The result also does not describe when deaths occurred, their causes, or how mortality varied over time. It is a binary count over a 1-year timeframe rather than a time-to-event analysis. Nor does it answer every aspect of efficacy or safety. The registry lists 3 outcome measures but does not post secondary results in this record, and the adverse-event information is a separate summary that should not be treated as interchangeable with the primary endpoint.

The trial population was described as people with heart failure with reduced ejection fraction and iron deficiency. The posted summary does not provide detailed eligibility criteria or participant characteristics, so it cannot establish how closely the enrolled participants match any particular clinical population. The results are a group-level statistical comparison, not a prediction for an individual participant. They do not provide medical advice or determine what treatment is appropriate for any person.

In short, the registry records a lower point estimate for 1-year mortality with ferric carboxymaltose than with placebo, but the statistical uncertainty is substantial enough that the estimate remains compatible with no difference. The available record does not support a broader conclusion about all clinical outcomes, the mechanisms behind the observed counts, or the consequences of treatment for individuals.

16. Why This Trial Matters Statistically

This trial is statistically informative because it compares a defined binary outcome between randomized groups using a placebo control and a double-blind design. Randomization is intended to limit systematic differences at assignment, while masking can reduce some biases in care and outcome assessment. These design features strengthen the rationale for a comparative interpretation, although the registry's brief description does not provide enough detail to evaluate allocation procedures or how masking was maintained.

The primary result also illustrates why a point estimate should not be read in isolation. The observed risk difference is negative, and the ferric carboxymaltose group has a lower posted risk estimate. Yet the interval crosses zero and the unadjusted p-value is 0.096. The statistical message is therefore not simply the direction of the estimate; it is the combination of estimate and uncertainty. That distinction is especially important when comparing a result with a threshold or using it to make a categorical claim.

The alternative risk-ratio and odds-ratio summaries provide relative scales for the same comparison. Their intervals include the no-difference value of 1, reinforcing the uncertainty visible on the absolute risk-difference scale. Looking across scales can help readers understand what is being estimated, but it does not create multiple independent pieces of evidence from a single dataset.

The registry's planning illustration is also a reminder that sample-size requirements depend on the effect a future study aims to detect. Under its stated assumptions, planning for smaller fractions of the observed effect requires more participants per arm. These figures are not the original design rationale and are not evidence of the completed trial's observed power. Distinguishing planning calculations from results prevents a common misunderstanding of power analyses.

Finally, the registry record leaves important statistical details unreported, including the original sample-size justification, missing-data approach, multiplicity strategy and analysis specification. Those gaps limit independent assessment of how the primary comparison was planned and implemented. Statistical significance, precision, design and reporting completeness all contribute to interpretation; none should be replaced by a simple reading of the event counts alone.

17. 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 registry record gives a concise description of the population but does not report detailed inclusion and exclusion criteria or baseline characteristics. As a result, the composition of the enrolled population and its representativeness cannot be assessed from this information alone. Although the study is described as including participants with heart failure with reduced ejection fraction and iron deficiency, finer clinical definitions are not reported.

The primary endpoint is reported as a 1-year participant count, but details about how death status was ascertained, how incomplete follow-up was handled, and whether all participants had the same observation opportunity are not reported in the registry record. The exact statistical test, interval construction method, and confidence level are also not specified. The unadjusted comparison is available, but no adjusted analysis is posted.

The registry lists 3 outcome measures but no secondary posted results and no posted analyses. The record does not provide the original sample-size calculation, interim-analysis information, multiplicity plan, or complete statistical analysis plan. It is therefore not possible to determine how the reported primary comparison relates to all planned endpoints or analyses. These are reporting limitations; they do not establish that the trial lacked such procedures.

Safety information is presented as participant counts and selected serious-event terms. Detailed definitions, event-level follow-up, attribution procedures, and comparative uncertainty estimates are not included in the available summary. The adverse-event death counts differ from the primary endpoint counts, and the record does not provide enough context to reconcile them. The sub-study mentioned in the title is not described in sufficient detail to assess its objectives or findings.

Finally, the posted primary result has an interval that includes no risk difference. The estimate is therefore uncertain, and it should not be generalized to outcomes or populations that the registry does not describe. These limits constrain what can be concluded from the record without invalidating the statistics that are reported.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Randomized Placebo-controlled Trial of FCM as Treatment for Heart Failure With Iron Deficiency / and Sub-Study — NCT03037931.
• Registry record last updated 2024-07-03; status last verified 2024-02.
• Results first posted 2024-07-03.
• 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.

21. Record Summary

NCT03037931 is a completed randomized, double-blind, parallel-group phase 3 study sponsored by American Regent, Inc. It compared intravenous ferric carboxymaltose with placebo in participants described as having heart failure with reduced ejection fraction and iron deficiency. The registry lists enrollment of 3065, a start date of 2017-03-15, and a primary completion date of 2023-02-02.

The primary endpoint was the number of deaths over 1 year in the intent-to-treat population. The record reports 131 deaths among 1532 participants assigned to ferric carboxymaltose and 158 among 1533 assigned to placebo. The risk difference was -0.018, with an interval from -0.038 to 0.003, and the unadjusted p-value was 0.096. The interval includes no difference, so the point estimate should be interpreted with the reported uncertainty.

The registry also includes adverse-event summaries, but those are separate from the primary mortality endpoint. Details of the original statistical plan, missing-data handling, multiplicity procedures, and secondary outcome results are not reported in the registry record.