CytoSorb Reduction of Free Hemoglobin During Cardiac Surgery: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for CytoSorb Reduction of Free Hemoglobin During Cardiac Surgery (NCT02566525), a randomized na study. The primary outcome is Assessment of Serious Device Related Adverse Events.

2. Clinical Question

The study asked whether adding the CytoSorb device to standard care during elective cardiac surgery was associated with serious device-related adverse events through discharge from the intensive care unit (ICU). The registry describes a prospective, multicenter, randomized feasibility study. Participants were assigned to standard care plus the device or to standard care alone; the brief summary states that assignment was planned in a 1:1 ratio.

The population was people undergoing elective cardiac surgery. The intervention was the CytoSorb device installed on the cardiopulmonary bypass (CPB) machine alongside standard care. The comparator was standard care. The primary safety question concerned the serious device-related adverse-event profile from enrollment through ICU discharge. The stated timeframe gives an example ICU stay of approximately 4 days, not a fixed follow-up duration for every participant.

The registry also lists time in the ICU, 30 Day All Cause Mortality, and adverse-event rates as secondary outcomes. The posted serious-event counts describe affected participants in each group, but they do not provide a comparative estimate, confidence interval, or test for the primary endpoint. The record therefore supports a descriptive reading of the reported safety counts, not a precise estimate of the difference between groups.

The clinical question is safety-focused. It is not a test of whether the device improves clinical outcomes, and the registry’s listed outcome information does not establish that it does or does not. The available record also does not specify enough detail to determine how the primary endpoint’s device-related events were selected from all serious events.

3. Trial Design

This was a prospective, multicenter, randomized feasibility clinical study. The registry identifies the allocation as randomized and the design as parallel. Its brief summary says participants would be assigned in a 1:1 ratio to standard care alone or standard care with the CytoSorb device. The registry lists no masking. The phase is recorded as NA.

The intervention group received standard care plus the CytoSorb device installed on the CPB machine. The control group received standard care. The condition listed is elective cardiac surgery. Enrollment is recorded as 52. The posted serious-event data give a risk group of 23 for the CytoSorb group and 23 for the control group. The record does not explain the relationship between the enrollment figure and the group-specific risk counts, so they should not be reconciled by assumption.

The study status is completed. The recorded start is 2015-11 and primary completion is 2016-10. The sponsor is CytoSorbents, Inc. These registry entries describe the study’s administrative and design context; they do not, by themselves, establish how participants were recruited, how allocation was generated or concealed, or whether the analysis followed the original assignment.

The primary endpoint was assessment of serious device-related adverse events through ICU discharge. Listed secondary outcomes were time in the ICU, 30 Day All Cause Mortality, and adverse-event rates. Seven outcome measures are recorded, while the registry record lists zero analyses. No detailed statistical analysis plan, allocation procedure, masking procedure, or rationale for the enrollment target is reported in the registry record.

Feasibility studies can characterize whether a study approach is workable and describe preliminary outcome data. That general purpose does not make a safety count a definitive comparison. Interpretation depends on the endpoint definition, completeness of follow-up, and an analysis that accounts for uncertainty; those details 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.

Before assignment: Potential patients had to have a planned 3+ hour bypass time. Isolated coronary bypass graft or single valve procedures were excluded. Cases included were aortic reconstructions, valve repairs, valve replacements and combinations thereof, and permitted combination with coronary bypass graft. Emergency and endocarditis patients were excluded. 52 patients enrolled in the study, of which 6 withdrew from the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All patients who had at least on Adverse Event

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 Assessment of Serious Device Related Adverse Events, measured over From time of enrollment through ICU discharge (e.g. ICU stay of approximately 4 days).

The primary endpoint is titled “Assessment of Serious Device Related Adverse Events.” Its description says the safety profile was assessed through serious device-related adverse events from enrollment through ICU discharge. The timeframe is from enrollment through ICU discharge, with an ICU stay of approximately 4 days given as an example. The stated unit is participants, and the outcome is recorded as a number.

A serious adverse-event endpoint is used to describe important safety events during a defined observation period. Restricting the endpoint to serious events focuses attention on events with greater clinical significance than minor or transient events. Following participants until ICU discharge sets a practical boundary tied to the perioperative and intensive-care period, but it does not capture events occurring after that boundary unless a separate follow-up outcome does so.

General statistical interpretation: an endpoint needs a clear definition of what qualifies as an event, whether the count refers to people or event episodes, how attribution to the device is decided, and how the observation window is applied. Those details affect which observations enter the endpoint and how groups can be compared. The registry record identifies the endpoint and its broad timeframe, but does not report the detailed adjudication rules or the criteria for judging an event to be device-related.

The outcome description also gives a population phrase of “All patients who had at least on Adverse Event.” That wording does not clarify the analysis population for the primary endpoint. The posted group risk counts are useful context, but the record does not explain whether participants without an adverse event were included in the primary endpoint analysis denominator. This limits interpretation of the endpoint as a risk comparison.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Interpreting a safety comparison begins with identifying the unit of analysis. The primary endpoint gives the unit as participants and records a number. The group data distinguish people at risk from people affected. For the serious-event data, the CytoSorb group has 23 at risk and 16 affected; the control group has 23 at risk and 11 affected. These counts describe participants rather than the number of separate event episodes, according to the outcome’s stated unit. If a participant could experience more than one event, a participant count would not convey the total burden or timing of all events.

What does a binary outcome represent?

General explanation: a binary endpoint classifies each participant according to whether the defined event occurred during the specified observation period. A comparison can then be expressed as a difference or ratio between group event risks, provided the denominators and analysis population are clear. The registry gives group risk counts and affected counts, but the primary endpoint’s population wording is unclear and no comparative estimate is reported. No new comparison should be inferred from the raw counts alone.

Endpoint definition matters as much as the arithmetic. The primary title specifies serious device-related adverse events, whereas the posted data label the affected counts as serious adverse events. A valid analysis needs a consistent rule for seriousness, attribution to the device, and whether an event occurred within the follow-up window. The registry does not report detailed rules or adjudication procedures. If these rules differ across groups or are applied inconsistently, the comparison can be difficult to interpret even when the counts are correct.

Why is randomization relevant?

General explanation: random assignment is intended to make treatment groups comparable, on average, with respect to both measured and unmeasured characteristics at the start of a study. In an individual study, randomization does not guarantee that every characteristic is balanced, and chance imbalances can occur. The registry identifies the study as randomized and reports a planned 1:1 assignment, but does not describe the sequence-generation method, concealment of assignment, or whether analysis used the assigned groups. Those features are important for judging the protection randomization provides against bias.

The design is parallel, meaning participants were assigned to one of the described study groups rather than receiving both regimens in sequence. The registry reports no masking. General explanation: lack of masking can affect how events are recognized, reported, evaluated, or attributed, particularly when outcome assessment involves judgment. It does not prove that bias occurred; it identifies a possible source of influence that should be assessed alongside event definitions and assessment procedures. The registry does not describe independent outcome adjudication.

How should uncertainty be assessed?

General explanation: a point estimate gives a single summary of a group contrast, while a confidence interval indicates the precision of that estimate under the chosen statistical model. A wide interval can be compatible with a range of materially different effects; a narrow interval indicates greater precision, though it does not correct biased measurement or an unsuitable endpoint. The registry record does not report a primary-endpoint effect estimate or confidence interval. It also lists zero analyses, so there is no reported inferential analysis to interpret for this endpoint.

A significance test, if reported, would address compatibility with a specified null hypothesis under its assumptions. It would not establish that the endpoint was measured without bias, that the study had adequate precision, or that the result applies beyond the study population. No primary-endpoint p-value is reported in the registry record. It would therefore be inappropriate to infer statistical significance or non-significance from the counts.

What do the planning assumptions mean?

The registry’s planning note describes illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation; the effect to detect is a stated fraction of the observed effect. It specifies that non-time-to-event illustrations are expressed as participants per arm. Such calculations are planning illustrations, not observed power and not evidence that the study was originally designed around those assumptions. The registry does not report the study’s own sample-size justification. The primary outcome is not reduced to one comparison in the posted analysis information, so no single primary-endpoint planning comparison should be inferred from the note.

General explanation: planning calculations depend on the effect size, variability or event rate, significance threshold, allocation, and analysis method. A calculation based on an observed effect is especially dependent on the observed estimate, which can be imprecise in a small feasibility study. It should not be treated as a reliable forecast of future results or a retrospective measure of whether the trial was adequately powered.

Why do analysis populations and missingness matter?

General explanation: a randomized comparison is commonly interpreted according to the assigned groups, but other analysis populations can answer different questions. Excluding participants after randomization may compromise comparability if exclusions relate to prognosis or events. For a safety endpoint, it is also important to know who was observed, whether follow-up ended at ICU discharge as planned, and how missing outcome status was handled. The registry does not report these analysis-population details or missing-data methods. The unusual primary population phrase referring to patients who had at least one adverse event makes the intended denominator particularly important to clarify.

Finally, adverse-event outcomes may involve multiple terms, severity categories, or time windows. General explanation: examining many outcomes raises multiplicity concerns because the chance of at least one apparently unusual result can increase as more comparisons are made. The registry lists seven outcome measures and several secondary outcomes, but does not report a multiplicity strategy, prespecified hierarchy, or interim-analysis plan. Those omissions mean the record cannot establish how the results were selected or how multiple comparisons were handled.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists seven outcome measures, including the primary safety endpoint and secondary outcomes of time in the ICU, 30 Day All Cause Mortality, and adverse-event rates. It lists zero analyses. No multiplicity adjustment, outcome hierarchy, or prespecified strategy for interpreting multiple endpoints is reported in the registry record. General explanation: when a study evaluates multiple outcomes or event categories, chance alone can produce apparently notable findings; a prespecified approach helps readers distinguish primary conclusions from exploratory observations.

The record does not report an interim-analysis plan or state whether any interim monitoring was conducted. It also does not provide a statistical analysis plan, primary-endpoint p-value, confidence interval, or adjusted analysis. These details should be treated as not reported in the registry record, rather than assumed to have been absent from the study itself.

Missing-data handling and analysis populations are also not reported in the registry record. The primary endpoint’s population wording refers to patients who had at least one adverse event, but does not clarify whether the primary comparison included all randomized participants, only participants with an event, or another defined population. The record does not describe withdrawals, incomplete ICU follow-up, or how unknown endpoint status was treated.

These gaps affect interpretation because event counts depend on both the event definition and who was included in the denominator. They also prevent assessment of whether the reported groups were analyzed consistently with the randomized assignment. The available counts can be read descriptively, but the record does not support a full appraisal of missingness, multiplicity, or analysis-population choices.

11. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: cardiogenic shock (47.8%), pleural effusion (13.0%), atrial fibrilation (8.7%), atrial flutter (8.7%), cardiac arrest (4.3%).

The registry reports 16 participants affected by serious adverse events among 23 at risk in the CytoSorb group and 11 affected among 23 at risk in the control group. For other adverse events, it reports 23 affected among 23 at risk in the CytoSorb group and 21 among 23 at risk in the control group. These are group-specific descriptive counts; no comparative estimate or uncertainty interval is reported.

The listed serious-event terms include Cardiogenic shock (47.8%), Pleural effusion (13.0%), Atrial Fibrilation (8.7%), Atrial flutter (8.7%), Cardiac arrest (4.3%), and Right ventricular dysfunction (4.3%). The registry does not specify the denominator for these term percentages in the record. They should not be read as group-specific risks or as percentages of all randomized participants.

The primary endpoint refers to serious device-related events, but the data do not explain which serious events were attributed to the device. The record also does not define whether participants could contribute to multiple terms or how events were adjudicated. The reported death counts are 2 in the CytoSorb group and 1 in the control group; their timing and analysis are not detailed. These descriptive figures do not establish a causal safety difference or show that the groups are equivalent.

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

The registry documents a randomized, parallel feasibility study comparing standard care plus CytoSorb with standard care in elective cardiac surgery. It identifies a safety-focused primary endpoint covering serious device-related adverse events through ICU discharge. The posted group data describe serious and other adverse-event counts, and the registry lists terms associated with serious events.

Those data do not provide a complete comparative analysis of the primary endpoint. The registry record reports no effect estimate, confidence interval, or hypothesis test, and it does not explain how serious events were judged to be device-related. The counts therefore do not show the precision of any difference between groups. Nor do they establish that the groups have the same safety profile: lack of a reported statistical difference is not evidence of equivalence, and no equivalence analysis is reported.

The trial’s feasibility context is important. A feasibility study can generate descriptive safety information and inform later study planning, but it is not automatically capable of resolving relatively uncommon or uncertain safety outcomes. General statistical explanation: a small amount of observed data can be consistent with a broad range of underlying event rates. Confidence intervals and clearly defined analysis populations help express that uncertainty; they are not reported for the primary endpoint in this record.

The listed secondary outcomes include ICU time and 30 Day All Cause Mortality, but the available information does not provide complete comparative results for those outcomes. The death counts alone should not be interpreted as a mortality effect. The record also does not report longer-term safety follow-up beyond the stated primary endpoint period.

Accordingly, the posted information is best read as a record of study design and descriptive safety observations, not as proof that the device caused, prevented, increased, or reduced an adverse outcome. The absence of detail about endpoint attribution, missing data, and analysis methods further limits the strength of conclusions that can be drawn.

13. Why This Trial Matters Statistically

This study is statistically important because safety outcomes require careful separation of event description from causal comparison. A count of participants with an event is an essential starting point, but it does not by itself show the precision of a group contrast or whether the events were related to the intervention. The record’s primary endpoint explicitly concerns serious device-related events, making attribution rules central to interpretation.

The study also illustrates why the denominator and analysis population matter. The registry reports 23 at risk in each group, alongside affected counts, but its primary population wording refers to patients who had at least one adverse event. Without clarification of how that wording was used, readers cannot fully determine the population represented by the primary outcome. This is a reporting issue, not a basis for assuming that the counts are incorrect.

Randomization is valuable for reducing confounding, while lack of masking and unspecified event adjudication raise questions about how safety outcomes were recognized and classified. These design features do not prove bias, but they identify information needed to assess the reliability of a comparison. A completed feasibility study can contribute useful descriptive observations even when the registry does not report inferential statistics.

The record also demonstrates the distinction between a planning illustration and a trial’s original design justification. Calculations based on observed effects can inform future planning but cannot substitute for a prespecified sample-size rationale or an uncertainty analysis of the observed results. No such original justification is reported in the registry record.

For researchers, the key statistical lesson is that transparent definitions, denominators, analysis populations, follow-up, and uncertainty estimates are necessary to interpret safety data. The posted counts are informative as descriptions, but the missing analytical details constrain what can be concluded from them.

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 registry’s serious-event counts do not include a comparative effect estimate or confidence interval, and the record lists zero analyses. The primary endpoint is described as serious device-related adverse events, but the group counts are labeled as serious adverse events. The record does not explain how device relationship was determined or whether the counts correspond exactly to the endpoint definition.

The primary endpoint population wording is unclear, referring to patients who had at least one adverse event. Although group-specific numbers at risk are reported, the registry does not clarify the analysis population or how participants without events were handled. Missing-data methods, withdrawals, completeness of follow-up, and analysis according to randomized assignment are not reported.

The registry identifies no masking and does not describe event adjudication. It does not report how randomization was generated or concealed. These omissions limit assessment of potential bias in allocation and in the recognition or classification of adverse events. The listed percentages for serious-event terms also lack a stated denominator in the record described.

The primary observation period ends at ICU discharge, with an approximate ICU stay given as an example. That does not describe safety beyond ICU discharge. Although 30 Day All Cause Mortality is listed as a secondary outcome, the available record does not supply a comparative analysis for it. The listed death counts do not resolve the primary endpoint or provide longer-term safety information.

The study is described as a feasibility study, and no original sample-size justification or detailed statistical analysis plan is reported in the registry record. These limitations mean the posted information should be interpreted as descriptive rather than as a conclusive estimate of comparative safety.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: CytoSorb Reduction of Free Hemoglobin During Cardiac Surgery — NCT02566525.
• Registry record last updated 2021-08-20; status last verified 2021-07.
• Results first posted 2021-08-20.
• 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 completed, prospective, multicenter, randomized, parallel feasibility study of CytoSorb during elective cardiac surgery. Participants were assigned to standard care plus the device installed on the CPB machine or to standard care alone; the brief summary states a planned 1:1 ratio. The study is listed as unmasked, with phase recorded as NA.

The primary endpoint was serious device-related adverse events from enrollment through ICU discharge. The posted serious-event counts are 16 affected among 23 at risk in the CytoSorb group and 11 among 23 at risk in the control group. No primary-endpoint comparative estimate or inferential analysis is reported. The record lists seven outcome measures and zero analyses.

Secondary outcomes include time in the ICU, 30 Day All Cause Mortality, and adverse-event rates. The registry does not report detailed endpoint adjudication, missing-data methods, analysis populations, or a statistical analysis plan. The findings are therefore limited to the descriptive information recorded and do not establish a causal or precise comparative safety conclusion.