Risk of Oxygen During Cardiac Surgery Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Risk of Oxygen During Cardiac Surgery Trial (NCT02361944), a randomized phase 2, phase 3 study. The primary outcome is Intraoperative Systemic Oxidative Damage.

2. Clinical Question

The trial also registered postoperative outcomes, including arrhythmia, stroke, respiratory failure, myocardial injury or infarction, cognitive dysfunction, chronic kidney disease, mitochondrial function, and vascular reactivity. These outcomes cover distinct biological and clinical questions. They should not be treated as interchangeable measures of the primary oxidative-damage question.

3. Trial Design

ROCS is a randomized, double-masked, parallel-group trial listed as phase 2 and phase 3. The registry records an enrollment of 213 and a completed status. The trial began on 2016-04-05, with primary completion on 2020-10-08. The brief summary describes a plan to recruit and randomize 200 elective cardiac surgery patients; the registry record does not explain the difference between that planned figure and the listed enrollment.

Participants were assigned to normoxia or hyperoxia during surgery. The normoxia strategy targeted hemoglobin oxygen saturation of 95-97% or arterial PaO2 80-110 mmHg. The hyperoxia strategy specified a fraction of inspired oxygen of 1.0 during mechanical ventilation and 0.8 during cardiopulmonary bypass. These are materially different oxygenation targets, so the comparison concerns the assigned intraoperative strategies rather than a small variation in a shared target.

The primary endpoint was systemic oxidative damage measured using F2-isoprostanes and isofurans at a specified intraoperative time. The registry gives median values and inter-quartile ranges for 100 participants in each arm. It does not report a primary-endpoint test, confidence interval, or adjusted estimate. Therefore the arm summaries describe the observed distributions, but do not by themselves establish the precision of a between-arm contrast.

The registry lists 33 outcome measures and 31 secondary outcomes posted. It records no reported statistical analyses. Details such as allocation concealment, the exact masking procedures, analysis populations, handling of protocol deviations, and a prespecified statistical analysis plan are not reported in the registry record. The listed double masking and randomization describe the registered design, but do not fill in those procedural details.

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: 213 patients signed consent. 12 participants were withdrawn prior to randomization. 201 were randomized.

Overall Study

Reasons for not completing this period:

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.

Population: 101 participant were randomly assigned hyperoxia. 1 of these participant's surgery was cancelled, and this participant was therefore withdrawn from the study. We do not have baseline characteristics for this participant.

6. Primary Endpoint

The registry lists the primary outcome as Intraoperative Systemic Oxidative Damage, measured over separation from cardiopulmonary bypass or completion of off-pump coronary artery bypass grafting (approximately 3-5 hours into surgery and intervention).

The primary endpoint is intraoperative systemic oxidative damage, quantified by measuring F2-isoprostanes and isofurans. The endpoint is reported in pg/ml as a median with an inter-quartile range. Its timing is separation from cardiopulmonary bypass or completion of off-pump coronary artery bypass grafting, approximately 3-5 hours into surgery and intervention.

These biomarkers were selected to address the trial’s biological hypothesis: the investigators proposed that physiologic oxygenation might reduce generation of reactive oxygen species and oxidative damage compared with hyper-oxygenation. In general, a biomarker endpoint can assess a biological process more directly or promptly than a later clinical outcome. It does not, by itself, demonstrate that a patient-important outcome changes, and interpretation depends on the measurement’s validity, timing, and relationship to clinical outcomes.

The registry reports a median of 69 pg/ml with an inter-quartile range of 54 to 89 for normoxia, and a median of 83 pg/ml with an inter-quartile range of 62 to 108 for hyperoxia. The inter-quartile range describes the middle half of observed values, not the uncertainty around the median or the treatment contrast. A confidence interval for the between-arm primary comparison and a statistical test are not reported in the registry record. The posted summaries therefore allow descriptive comparison, but do not establish how precisely the difference between strategies was estimated.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies the trial as randomized, double-masked, and parallel-group. Randomization is intended to make assigned groups comparable on average, including with respect to factors that are not measured. Masking is intended to reduce the influence of knowledge about assignment on care or outcome assessment. Those are general purposes of the design features; the registry does not report the allocation-concealment procedures, who was masked, or how masking was maintained.

What does the primary endpoint summary describe?

The primary outcome is reported as a median and inter-quartile range rather than a mean and standard deviation. In general, a median marks the central position of a distribution, while the inter-quartile range describes its middle half. These summaries can be useful when values are skewed or include extreme observations. They do not quantify uncertainty about the difference between groups. For that, an inferential comparison would ordinarily require a method suited to the endpoint distribution and a confidence interval or other uncertainty measure. Such a comparison is not reported in the registry record.

The primary values are 69 pg/ml with an inter-quartile range of 54 to 89 for normoxia, and 83 pg/ml with an inter-quartile range of 62 to 108 for hyperoxia. It would be inappropriate to infer a p-value, confidence interval, or adjusted contrast from these summaries without additional information and a stated method. The registry record does not identify the primary analysis method or report a recalculated primary comparison.

How should the binary outcomes be read?

The selected secondary clinical outcomes are counts of participants with events. Their comparisons are expressed as risk differences, with hyperoxia compared with normoxia. In general, a risk difference is the event risk in one group minus the event risk in the reference group, expressed on the risk scale. Positive estimates indicate a higher observed risk in the compared group; negative estimates indicate a lower observed risk. The interpretation is tied to the specified group order and the estimation method.

A confidence interval describes a range of values reasonably compatible with the data under the model and assumptions used. It conveys uncertainty around an estimate; it is not a range containing a fixed proportion of individual outcomes. When an interval includes zero for a risk difference, the data do not rule out no difference at the interval’s stated confidence level. An interval that also spans meaningful values in both directions indicates that the direction and magnitude remain uncertain. The registry comparisons are unadjusted, and the analysis method for the reported intervals is not described.

The secondary comparisons report unadjusted p-values. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a specified null model and assumptions. It is not the probability that the null hypothesis is true, nor a direct measure of clinical importance. The record does not report a multiple-testing procedure, so the p-values should not be interpreted as if a family of secondary outcomes had been controlled for false-positive findings.

Why does over-dispersion matter?

For count outcomes, a model that assumes variability is exactly what its simplest probability model predicts can be too optimistic when observed variability is greater. This is called over-dispersion. In general, failing to account for it can make standard errors too small and confidence intervals too narrow. The posted secondary outcomes are participant counts for binary events, not a reported count-rate analysis. The registry record does not describe a model or provide enough information to assess over-dispersion; no such adjustment should be presumed.

What do the planning estimates mean?

The planning illustrations are based on observed effects and state assumptions of two-sided significance level 0.05, equal allocation, and a normal approximation. They give participants per arm for 80% power and 90% power under the listed fractions of the observed effect. In general, such calculations show how sample-size requirements change with the effect assumed for planning and the target operating characteristics. They are not observed power, do not demonstrate that the trial had a particular probability of detecting an effect, and do not reveal the trial’s original design rationale. The registry record does not report the trial’s own sample-size assumptions or justification.

These planning estimates also inherit uncertainty from the observed effect. An effect estimated imprecisely can produce an unstable planning value, while using a fraction of that estimate imposes a hypothetical target rather than a known treatment effect. The calculations therefore serve as illustrations for future planning, not as evidence that an outcome is important or that a particular sample size would be adequate in another setting.

What is missing from the analysis description?

The registry lists no reported analyses. The record does not report an analysis population, missing-data strategy, covariate adjustment, handling of protocol deviations, or a prespecified statistical analysis plan. It also does not report whether the primary endpoint was transformed, how multiple biomarker measures were combined, or how measurements outside the stated timing were handled. These omissions prevent a full appraisal of the analytic choices. A statistical method should not be inferred solely from the endpoint format or from the presence of summary statistics.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 33 outcome measures and 31 secondary outcomes posted, spanning biomarkers, physiological measures, and postoperative events. When many outcomes are tested, the chance of at least one apparently unusual result can increase. In general, multiplicity procedures address this issue by controlling an error rate or by organizing outcome testing in a prespecified hierarchy. No such procedure or hierarchy is reported in the registry record. The posted secondary p-values are unadjusted and should not be treated as confirmatory evidence in isolation.

The registry record lists no reported statistical analyses. For the primary endpoint, it gives group medians and inter-quartile ranges but no inferential comparison. For the selected secondary outcomes, it gives unadjusted comparisons, but does not identify the statistical model or report sponsor analyses. A prespecified statistical analysis plan, including endpoint priorities and analysis methods, is not reported in the registry record.

Missing-data handling and analysis populations are also not reported in the registry record. It is therefore unclear from the registry whether the reported summaries represent all randomized participants, a modified analysis set, or those with available outcome measurements. The record does not explain reasons for missing measurements, exclusions after randomization, protocol deviations, or whether any imputation was used. These details matter because missingness can affect both the observed group summaries and the validity of a between-group comparison.

Interim analyses, stopping rules, and any sample-size justification are not reported in the registry record. Their absence from the record should not be interpreted as proof that such procedures did or did not occur. The planning illustrations based on observed effects do not supply this missing design information: they are explicitly planning calculations under stated assumptions, not an account of the trial’s original design. Without further analysis documentation, readers cannot assess whether the reported outcome selection and inferential approach followed a prespecified plan.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: category: arrythmia, pacemaker or cardiac arrest (5.0%), seizure (2.0%), surgical complications requiring valve replacement revision (2.0%), hypoxia (1.0%), right ventricular failure (1.0%).

The registry reports serious adverse events among 100 participants at risk in each arm: 8 affected participants in the normoxia arm and 11 in the hyperoxia arm. It reports other adverse events among 100 participants at risk in each arm, with 80 affected in each arm. Deaths are reported for 2 participants in the normoxia arm and 0 in the hyperoxia arm. These are descriptive safety counts; the record does not provide an adjusted safety comparison, confidence intervals, or details about event ascertainment.

The listed serious-event categories include “Arrythmia, Pacemaker or Cardiac Arrest” at 5.0%, seizure at 2.0%, surgical complications requiring valve replacement revision at 2.0%, hypoxia at 1.0%, right ventricular failure at 1.0%, and heparin-induced thrombocytopenia at 1.0%. The registry listing does not establish that these category percentages belong to a particular arm, nor does it provide a causal attribution to the oxygen strategy.

Adverse-event counts need context, including definitions, timing, severity criteria, and whether people could contribute more than one event. Those details are not reported in the registry record. The figures therefore describe what was posted but do not, on their own, establish comparative safety or explain the clinical significance of the events.

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

The primary result is a descriptive comparison of oxidative-damage biomarker distributions at a defined point during surgery. The normoxia arm’s reported median is 69 pg/ml, with an inter-quartile range of 54 to 89; the hyperoxia arm’s median is 83 pg/ml, with an inter-quartile range of 62 to 108. The registry does not report a primary-endpoint confidence interval, p-value, or adjusted comparison. The observed medians alone do not show how precisely the between-arm contrast was estimated.

The result does not establish whether the biomarker distributions differ beyond sampling variability, because the registry record does not provide the corresponding inferential analysis. It also does not establish whether any biomarker difference would translate into a difference in organ injury, recovery, or another patient-important outcome. The trial’s biological rationale motivates the measurement, but a proposed mechanism is not itself proof that the mechanism changed or that clinical outcomes followed.

The secondary outcomes provide additional descriptive information, but their estimates are unadjusted and uncertain. Their intervals include no risk difference, and the registry does not report a strategy for accounting for multiple comparisons. Neither a non-significant p-value nor a confidence interval that includes zero demonstrates equivalence. Equivalence or non-inferiority requires a suitable design and prespecified margins, which are not reported in the registry record.

Randomization is a strength for comparing assigned strategies in principle, but interpretation also depends on complete outcome data, adherence, analysis populations, and valid outcome assessment. The registry does not report enough detail on these matters to evaluate their effects. The safest reading is therefore limited: the registry records the assigned oxygenation targets, the primary arm summaries, selected secondary event counts, and safety counts. It does not provide a complete inferential account of the primary endpoint or a basis for broad claims about clinical benefit, harm, or equivalence.

The findings should be understood within the trial’s setting of elective cardiac surgery and the specified intraoperative oxygen strategies. Generalization to other populations, procedures, oxygen targets, or care settings is not established by the registry record. This interpretation concerns the posted registry results and is not medical advice.

13. Why This Trial Matters Statistically

This trial illustrates why a clinical comparison needs more than arm-level summaries. Medians and inter-quartile ranges show where the observed biomarker distributions lie, but they do not report the uncertainty of the treatment contrast. Without a confidence interval or test for the primary endpoint, a central question remains unanswered statistically: how much information do these data provide about the between-group difference?

It also illustrates the distinction between a mechanistic endpoint and a clinical outcome. F2-isoprostanes and isofurans were selected to assess systemic oxidative damage, while the secondary outcomes include postoperative events and other biological measures. A coherent interpretation requires care not to assume that movement in a biomarker necessarily predicts patient-important benefit or harm.

The secondary results show why sparse outcomes and multiplicity matter. Event estimates can be imprecise, and examining many outcomes without a reported adjustment makes isolated findings difficult to interpret as confirmatory. The registry’s planning illustrations further demonstrate that sample-size calculations depend on assumptions about the effect to be detected. Because these calculations use observed effects, they should be read as planning examples rather than as evidence of the trial’s original power or design.

Finally, this record highlights the value of complete statistical reporting. Analysis populations, missing-data procedures, endpoint hierarchy, and the primary comparison are essential for judging robustness. Their absence does not negate the descriptive information recorded, but it limits what can be concluded from it. For clinical researchers, the distinction between what was measured, what was estimated, and what remains unreported is central to responsible interpretation.

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.

A central limitation of the registry results is that the primary endpoint is reported only through arm-level medians and inter-quartile ranges. The registry record does not report the primary comparison’s uncertainty, statistical test, or model. Consequently, readers cannot assess the precision of the primary contrast from the posted results. The endpoint is also a biomarker measured during surgery; the record does not establish how the measurement relates to subsequent clinical outcomes.

The outcome timing is specific to separation from cardiopulmonary bypass or completion of off-pump coronary artery bypass grafting, approximately 3-5 hours into surgery and intervention. This timing supports interpretation of an intraoperative biological measure, but does not by itself describe the duration of any effect. The registry does not report additional timing details, repeated-measure analysis, or how values from the different surgical pathways were handled.

Important analysis information is absent. The registry does not report the analysis population, missing-data approach, adjustment variables, handling of protocol deviations, or the statistical analysis plan. It also does not report whether outcome assessors were masked or provide detailed masking procedures. These omissions limit assessment of bias and robustness even though the design is listed as randomized and double-masked.

The secondary clinical comparisons include outcomes with few events and are unadjusted. Their confidence intervals are broad enough to leave substantial uncertainty about direction and magnitude. The registry does not report multiplicity control or an outcome hierarchy, so secondary findings should not be elevated to confirmatory conclusions. Safety information is also presented as counts without detailed definitions, attribution, or comparative uncertainty.

Finally, the registry does not report its original sample-size justification. The planning illustrations based on observed effects do not resolve that limitation. The available information is therefore useful for describing the trial’s registered question and posted results, but insufficient for a complete appraisal of the statistical analysis or the strength of evidence for a clinical effect.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Risk of Oxygen During Cardiac Surgery Trial — NCT02361944.
• Registry record last updated 2024-02-01; status last verified 2024-01.
• Results first posted 2023-08-14.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

The Risk of Oxygen During Cardiac Surgery (ROCS) Trial, NCT02361944, is listed as a completed randomized, double-masked, parallel-group phase 2 and phase 3 study sponsored by Vanderbilt University. It enrolled 213 participants, with a start date of 2016-04-05 and primary completion on 2020-10-08.

The trial compared intraoperative normoxia with hyperoxia in cardiac surgery. Its primary endpoint was systemic oxidative damage measured using F2-isoprostanes and isofurans. The registry reports median concentrations and inter-quartile ranges for 100 participants in each arm, but no primary inferential comparison. It also posts secondary outcome and safety information, including selected unadjusted comparisons.

The record does not report the statistical analysis plan, primary-endpoint test, missing-data strategy, analysis populations, or multiplicity procedure. As a result, the posted summaries are descriptive and the precision of the primary contrast cannot be determined from the registry record.