Red Cell Storage Duration and Outcomes in Cardiac Surgery: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Red Cell Storage Duration and Outcomes in Cardiac Surgery (NCT00458783), a randomized na study. The primary outcome is Primary Aim is to Determine Whether Length of Storage of Red Blood Cells is Related to Postoperative Morbid Outcomes in Patients Undergoing Cardiac Surgery..

2. Clinical Question

The comparison addresses the storage age of transfused red blood cells in cardiac surgery, rather than comparing transfusion with no transfusion. The groups are described by the age of the available matching blood: oldest available matching units for prolonged storage and youngest available matching units for short storage. This makes the result specific to the two allocation strategies recorded in this trial.

The primary question combines several postoperative outcomes into a single endpoint. A participant counts as having an event if mortality or any of the listed complications occurs during the defined follow-up. The composite includes cardiac, pulmonary, neurologic, renal, infectious, gastrointestinal, reoperation, and vascular events. The registry record does not report separate primary-outcome estimates for each component, so the posted primary result should be understood as an assessment of the composite rather than of any individual complication.

3. Trial Design

This was a randomized, double-masked, parallel-group trial sponsored by The Cleveland Clinic. The registry record gives an enrollment of 3942 and a status of completed. The recorded start is 2007-04 and primary completion is 2018-12. The phase is listed as NA. The registry record does not report the method used to generate the random allocation sequence, the concealment procedure, or the operational details of masking.

The two intervention groups were prolonged RBC storage and short RBC storage. The former involved transfusion with the oldest available matching red blood cells; the latter involved transfusion with the youngest available matching red blood cells. This comparison concerns a policy for selecting available blood units. The registry record does not state the actual storage duration of units received by participants or how much those durations differed between groups.

The primary endpoint was assessed 30 days post surgery. Its event count is posted for 686 participants in the prolonged-storage group and 701 in the short-storage group. The record does not explain why these outcome-group counts differ from the enrollment figure, nor does it provide a participant-flow account that would identify exclusions, withdrawals, or missing outcome data. Those details are not reported in the registry record.

The posted analysis includes a generalized estimating equation analysis of the primary outcome and a Cox regression analysis for postoperative hospital length of stay. The registry record does not describe the full statistical analysis plan, the analysis populations, or whether outcome assessors and treating staff remained masked throughout follow-up. Randomized allocation supports a comparative trial design, but interpretation still depends on the details of implementation, follow-up, and analysis that 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: From July 2007 to May 2016, 3942 adults undergoing coronary artery bypass grafting, cardiac valve procedures, or ascending aorta repair, either alone or in combination were enrolled in the study. 3835 adults were randomized to transfusion of RBCs stored for ≤ 14 days (younger units) or for ≥ 20 days (older units) intraoperatively and throughout the postoperative hospitalization

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.

6. Primary Endpoint

The registry lists the primary outcome as Primary Aim is to Determine Whether Length of Storage of Red Blood Cells is Related to Postoperative Morbid Outcomes in Patients Undergoing Cardiac Surgery., measured over 30 days post surgery..

The primary endpoint is a composite measure of mortality or postoperative morbidity within 30 days post surgery. The registry definition includes mortality; multisystem organ failure; cardiac events such as ventricular tachycardia, fibrillation, asystole, and atrial fibrillation; pulmonary events such as pneumonia, prolonged postoperative ventilation, and pulmonary embolus; neurologic events such as stroke and coma; renal failure; infection, including deep sternal wound infection and sepsis; gastrointestinal ischemia or infarction; reoperation for bleeding, tamponade, or cardiac dysfunction; and vascular events such as dissection or limb ischemia. The definition is described as aligned with the STS database.

A composite endpoint is used to capture whether any event from a defined set occurred. In general, this can provide a broad summary when outcomes span multiple kinds of complications, but it also means the overall result does not reveal whether effects differ among the components. A frequent or less severe component can contribute substantially to the composite, while a rare or serious component may contribute less often. The registry record does not give component-specific primary-outcome estimates or explain the rationale for selecting these components.

The posted outcome is a count of participants with an event. It does not report the total number of events per participant, timing of individual events, or a separate estimate for mortality alone. Interpretation therefore concerns occurrence of the composite during the stated follow-up, not the severity, recurrence, or individual type of event.

7. Results Posted in the Registry

• Short RBC Storage: 51.6% (362/701) versus 52.5% (360/686) on control; risk difference -0.8 percentage points (95% CI -6.1 to 4.4), risk ratio 0.98 (0.89–1.09); unadjusted two-proportion z-test p = 0.755.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Effect of RBC Age on Postoperative Hospital Lengths of Stay.

• Endpoint type: Time to event: compared as a hazard ratio.
• Test: Hazard ratio as reported by the sponsor (Regression, Cox); it was not recalculated.
• Result:
  • Prolonged RBC Storage vs Short RBC Storage: hazard ratio 1.00 (95% CI 0.85–1.18), reported p = 0.070 (as reported in the registry).
• Interpretation:
  • The 95% confidence interval includes no difference (a ratio of 1), so these data do not show a statistically significant difference between Prolonged RBC Storage vs Short RBC Storage and the control. The interval gives the range of effects that remain compatible with the data.

The registry record lists intensive care unit length of stay and postoperative hospital length of stay as secondary outcomes. The detailed posted result concerns postoperative hospital length of stay: both groups have a median of 9.0 days, and the reported Cox hazard ratio is 1.0 with a 95% confidence interval from 0.85 to 1.18 and p = 0.07. A detailed ICU length-of-stay result is not reported in the registry record.

Overall, the available secondary information does not provide a broad set of independent estimates that could confirm or challenge the primary composite result. The posted hospital-stay summaries are similar between groups, and the hazard-ratio interval includes its null value. The primary endpoint covers mortality and multiple postoperative complications, while length of stay is a different type of outcome. Similarity in a length-of-stay summary does not determine whether the composite’s component events differed, and a composite result cannot by itself determine length of stay.

Secondary outcomes are often exploratory or supportive, especially when many outcomes are assessed. If multiple comparisons are made, some may appear notable by chance even when there is no underlying difference. Unless the analysis plan specifies a hierarchy or a multiplicity adjustment, secondary results generally carry less confirmatory weight than a prespecified primary analysis. The registry record lists 13 outcome measures and 13 analyses but does not report whether multiplicity was controlled, how outcomes were prioritized, or whether the secondary analyses were prespecified. The available secondary results should therefore be interpreted as context rather than independent confirmation of the primary finding.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Effect of RBC Age on Intensive Care Unit (ICU) Length of Stay.
• Effect of RBC Age on Postoperative Hospital Lengths of Stay.

9. Statistical Interpretation

10. Statistical Methods Explained

The registry record presents a binary primary outcome: whether a participant experienced at least one event in the composite by the stated follow-up. For such an outcome, a risk is the proportion of participants in a group who experienced the event. A risk difference subtracts one group’s risk from the other’s, so zero represents no difference. A risk ratio divides one group’s risk by the other’s, so 1 represents no relative difference. An odds ratio compares the odds of an event rather than the risks; its null value is also 1. These are distinct measures and can look different in magnitude even when calculated from the same underlying groups.

What does the unadjusted comparison represent?

The recalculated risk difference, risk ratio, odds ratio, and p-value use the posted event counts and group sizes without adjustment for other variables. “Unadjusted” describes the calculation, not a judgment about whether the analysis is appropriate. It provides a direct comparison of the observed groups. It does not account for any covariates, clustering, or other structure that a model might include. The registry record does not report enough detail to determine whether the posted GEE model addresses particular sources of dependence or adjusts for specified covariates.

The risk difference is expressed on an absolute scale. It indicates how far apart the group event risks are, with a negative estimate in this comparison pointing toward a lower observed risk in the short-storage group under the stated subtraction order. The relative measures express the contrast as a ratio. Neither scale is universally preferable: absolute differences communicate how much the observed risks separate, while ratios express a relative contrast. Both should be read with their confidence intervals and with the outcome definition in mind.

Why is the reported GEE result different in form?

A generalized estimating equation is a regression approach that can estimate population-average associations while accounting for correlated observations under a specified working correlation structure. In general, correlation can arise when observations are not statistically independent. The registry labels the primary analysis as GEE and reports an odds ratio, but does not report the model specification, correlation structure, adjustment variables, or details of how standard errors were obtained. It is therefore possible to describe the reported estimate and interval, but not to reconstruct the model from the registry record.

An odds ratio is not the same as a risk difference or risk ratio. The reported odds ratio of 0.77 should not be read as a reduction of 0.23 in the event risk. Its meaning is on the odds scale, and a direct translation to an absolute risk contrast would require additional model information. The reported confidence interval extends from 0.50 to 1.2 and includes 1, the null value for an odds ratio. The p-value is 0.08. These inferential summaries do not supply evidence of a precise, definitive separation between the groups.

What does the Cox method estimate?

Cox regression is a time-to-event method. In general, it compares event rates over follow-up while allowing the underlying event rate to vary over time; the hazard ratio is the model-based relative contrast. The registry reports a hazard ratio of 1.0 for postoperative hospital length of stay, with a 95% confidence interval from 0.85 to 1.18. It also reports a median of 9.0 days in each group and interquartile ranges from 7.0 to 13.0 days. A hazard ratio and a median summarize different features of time-to-event data. The hazard ratio is not a difference in days, and it cannot be interpreted as the ratio of the two medians.

The registry does not specify the Cox event definition, censoring rules, or whether proportional hazards were assessed. In general, a Cox model relies on assumptions about how the relative hazards behave over time. Without those details, the reported estimate can be quoted, but the assumptions and diagnostic checks cannot be evaluated from the registry record.

How should p-values and intervals be read?

A p-value measures how incompatible the observed data are with a stated null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. A confidence interval describes a range of estimates that are reasonably compatible with the data and model under the interval procedure. It is not a guarantee that the true value lies in the displayed range. Precision depends on the width of the interval as well as the scale and context of the estimate.

For the primary outcome, the unadjusted interval crosses zero for the risk difference and crosses 1 for the ratio measures. The GEE odds-ratio interval also crosses 1. This pattern indicates that the data do not sharply distinguish the null from some non-null possibilities. It should not be converted into a claim that the groups are identical. A formal equivalence or non-inferiority conclusion would require a prespecified margin and suitable analysis; those are not reported in the registry record.

The registry lists 13 outcome measures and 13 analyses. When many outcomes or comparisons are examined, the chance of at least one apparently notable result can increase. In general, multiplicity procedures or a clearly specified hierarchy can address this issue, but the registry record does not report whether such procedures were planned or applied. It also does not report the complete analysis plan, missing-data methods, or the population used for each analysis. These omissions limit how fully the inferential results can be assessed.

11. Confidence Intervals and Precision

The primary unadjusted risk difference is -0.008, with a confidence interval from -0.061 to 0.044. In general, a confidence interval expresses the uncertainty around an estimate under the method used to construct it. For a risk difference, zero represents no difference. Because the interval includes zero, the posted data are compatible with a small difference in either direction as well as with no difference. The interval is relatively broad compared with the point estimate, so the point estimate alone should not be treated as a precise description of the underlying contrast.

The recalculated risk ratio is 0.984, with an interval from 0.889 to 1.089, and the recalculated odds ratio is 0.967, with an interval from 0.783 to 1.194. For these ratio measures, 1 is the null value. Their intervals include 1. The sponsor-reported GEE odds ratio is 0.77, with a 95% confidence interval from 0.50 to 1.2, which also includes 1. The different point estimates reflect differences in analysis and estimand; they should not be treated as interchangeable versions of one measure.

For hospital length of stay, the reported Cox hazard ratio is 1.0, with a 95% confidence interval from 0.85 to 1.18. The posted group medians are both 9.0 days, with interquartile ranges from 7.0 to 13.0 days. These summaries give different views of the time outcome, and neither removes uncertainty about the full distribution or time-to-event model.

In general, an interval that includes the null does not prove that there is no effect. It means that the data and analysis do not exclude the null at the interval’s stated confidence level. The registry record does not provide a clinically meaningful threshold for the primary composite, so statistical precision cannot be translated into a judgment about whether a particular difference would matter clinically. The interval endpoints should be considered alongside the composite definition, group counts, and limitations in the analysis details.

12. 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 sample-size display is a planning illustration based on the observed primary effect, not the trial’s original design calculation. The registry record does not report the trial’s original sample-size assumptions or justification. For this binary-outcome illustration, the figures are participants per arm, not total events. The assumptions stated for the illustration are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning scenario uses the stated fraction of the observed effect as the effect to detect.

The column for 80% power gives the number of participants per arm that the planning calculation indicates would be needed for an 80% chance of detecting the specified effect under those assumptions, if that effect were the true effect. The column for 90% power gives the corresponding planning requirement for a 90% chance. Higher target power generally requires more participants when other assumptions remain the same. The rows compare planning for 100% of the observed effect, 75% of the observed effect, and 50% of the observed effect. Smaller effects need more participants to distinguish them from random variation at the stated significance level and power target.

The listed planning values are 55,839 and 74,752 participants per arm for 100% of the observed effect; 99,253 and 132,871 for 75% of the observed effect; and 223,278 and 298,906 for 50% of the observed effect, for the respective 80% and 90% targets. These are hypothetical planning outputs based on the observed estimate and stated assumptions, not counts enrolled or analyzed in this trial. This is not observed power.

The display should be read as a sensitivity illustration: it shows how a planning requirement changes when the target effect is smaller or the desired power is higher. Because the observed effect is itself uncertain, planning from it does not guarantee that a future study will achieve the displayed power. The registry record does not report the assumptions used to design the actual trial.

13. 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 = 0.525; treatment proportion p₁ = 0.516; average p̄ = 0.521.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7065 = 1.3847.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.7065 = 0.5946.
4. n = (1.3847 + 0.5946)² / (-0.0084)² = 55,838.9, rounded up to 55,839 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 <- 0.524781341107872                          # control proportion
p1 <- 0.516405135520685                          # 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(458783)
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 = 55,839 (80% power), 74,752 (90% power)
 75% of observed effect: n per arm = 99,253 (80% power), 132,871 (90% power)
 50% of observed effect: n per arm = 223,278 (80% power), 298,906 (90% power)
power.prop.test: 55838.9 
Simulated power at n = 55,839 per arm: 0.799 (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.

14. Analyses Reported by the Sponsor

The sponsor-reported primary analysis uses generalized estimating equations and reports an odds ratio of 0.77, with a 95% confidence interval from 0.50 to 1.2 and p = 0.08. The recalculated summary uses the posted event counts and group sizes to give an unadjusted risk difference of -0.008, with an interval from -0.061 to 0.044 and p = 0.755. It also reports unadjusted ratio measures. The two approaches differ in both method and effect scale, so their point estimates and p-values are not direct substitutes for one another.

The GEE result is the sponsor’s reported model-based analysis. The recalculation is a simple comparison of the posted group summaries and does not reproduce the GEE model. The registry record does not state the covariates or correlation structure used in that model, so those details cannot be assessed or matched by the unadjusted calculations. The registry record does not flag a discrepancy between the reported and recalculated primary results.

Both analyses have uncertainty that includes their respective null values: 1 for the odds-ratio measures and zero for the risk difference. The sponsor’s result therefore does not establish a clear difference, while the unadjusted result likewise does not establish one. The sponsor analysis should be read as the trial’s reported inferential model; the recalculation is supplementary context from the posted counts, not a replacement for that analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists 13 outcome measures and 13 analyses. Assessing multiple outcomes creates a multiplicity issue: the more hypotheses examined, the greater the opportunity for a seemingly notable result to arise by chance. In general, researchers may address this with an ordered testing strategy, adjusted significance thresholds, or other prespecified procedures. The registry record does not report whether multiplicity was considered, whether any adjustment was used, or how the primary and secondary outcomes were prioritized. It is not possible to determine from the record whether the reported p-values are adjusted or unadjusted, except that the recalculated primary p-value is an unadjusted calculation.

Details about missing data and analysis populations are also limited. The registry record does not report how missing primary outcomes were handled, whether any imputation was used, or whether participants were analyzed according to their randomized assignment. It does not describe exclusions after randomization, withdrawal, loss to follow-up, or the treatment of participants who did not receive the assigned storage strategy. Although an enrollment figure and primary outcome group counts are posted, the record does not provide a flow account that reconciles them. The posted counts alone cannot establish why those figures differ or whether any outcomes were unavailable.

The record identifies GEE for the primary sponsor-reported analysis and Cox regression for hospital length of stay, but it does not provide the full analysis plan or detailed model specifications. For the GEE analysis, adjustment variables and correlation structure are not reported. For the Cox analysis, the event definition, censoring approach, and assessment of model assumptions are not reported. The registry record also does not report interim analyses or stopping rules. These are reporting gaps, not evidence that a particular method was or was not used. They limit the ability to evaluate the full set of analytic decisions and their effects on uncertainty.

16. 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

Serious events with the highest arm-level rate: atrial fibrillation (36.8%), prolonged ventilation (15.6%), graft occlusion/valve dysfunction, noncardiac reasons (6.9%), bleeding/tamponade (6.4%), renal morbidity (5.2%).

The registry’s serious adverse event entries list 360 affected participants among 686 at risk in the prolonged-storage group and 362 among 701 at risk in the short-storage group. The entries also list 23 deaths for prolonged storage and 15 for short storage. Other adverse events are recorded as 0 in each group. The registry lists atrial fibrillation at 36.8%, prolonged ventilation at 15.6%, graft occlusion or valve dysfunction for noncardiac reasons at 6.9%, bleeding or tamponade at 6.4%, renal morbidity at 5.2%, and gastrointestinal morbidity at 4.8% among the top serious adverse terms.

These are descriptive registry figures. The record does not explain the denominator used for each listed percentage, the event-collection window, whether terms can overlap within a participant, or how seriousness was defined and adjudicated. The affected counts should not be assumed to be counts of unique event types or compared as if they were adjusted risks. The record does not provide a statistical comparison for these safety entries. Differences in raw counts alone do not establish a causal safety difference, and the entries should be interpreted with the stated limitations in reporting.

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

The primary result describes the occurrence of a broad composite of mortality or postoperative complications among participants undergoing cardiac surgery. The observed event counts and unadjusted estimates are close between the posted groups, while the confidence intervals include the null and allow effects in either direction. The sponsor-reported GEE estimate also has a confidence interval that includes its null value. Together, these results do not provide a statistically clear separation between the storage strategies on the recorded primary composite.

That interpretation is narrower than saying that the strategies are equivalent. Equivalence is a specific statistical claim that requires a prespecified margin defining how much difference is acceptable, along with an analysis designed to test that margin. The registry record does not report an equivalence margin or an equivalence analysis. A non-significant result is not proof of no effect, and the observed point estimates should not be read as proof that every clinically relevant difference has been ruled out.

The composite result also does not show whether mortality, cardiac events, pulmonary events, infection, or other individual components differed. The registry does not provide separate primary estimates for those components. The posted hospital length-of-stay result concerns a separate outcome and does not answer the primary composite question. A detailed ICU length-of-stay result is not reported in the registry record.

The evidence applies to the specific randomized comparison recorded: transfusion with the oldest available matching red blood cells versus the youngest available matching red blood cells in cardiac surgery. The record does not report the actual storage ages received or the range of storage durations in each arm. It also lacks details about missing data, analysis populations, model specifications, and multiplicity procedures. These gaps limit precision about how broadly the result can be interpreted. The findings should be described as the posted estimates and their uncertainty, rather than as a universal conclusion about all transfusion settings or all postoperative outcomes.

18. Why This Trial Matters Statistically

This trial illustrates why statistical interpretation should not rest on a point estimate or a p-value alone. For the primary composite, the point estimates are close to the null, but the confidence intervals still span values on either side of the null. That combination is informative: the observed comparison is not strongly separated, and the data do not exclude every possible difference. Reporting the estimate, interval, outcome definition, and analysis scale together gives a more complete account than a binary significant-or-not label.

The trial also shows why estimates from different methods must be distinguished. The unadjusted risk difference summarizes an absolute contrast from the posted counts, while the GEE analysis reports a model-based odds ratio. They use different scales and do not answer precisely the same statistical question. The same principle applies to the hospital-stay result: a median in days and a Cox hazard ratio summarize different features of the outcome.

Composite endpoints can be useful for summarizing a broad set of outcomes, but they can obscure component-specific patterns. A single overall result does not indicate whether the components contribute evenly or whether a particular event drives the composite. The registry record gives the composite definition but does not provide separate primary estimates for its components, making careful wording especially important.

Finally, the record illustrates the value of transparent methods reporting. Details on the analysis population, missing data, multiplicity, and model specifications would help readers evaluate how robustly the estimates answer the trial question. The posted results still support a clear account of the observed comparison and its uncertainty, while leaving important questions about analysis implementation and outcome detail unresolved.

19. 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 does not report actual red blood cell storage durations in either group or the degree of separation achieved between the oldest-available and youngest-available matching units. The intervention descriptions therefore identify the selection strategies but do not quantify the exposure contrast. That limits interpretation of what storage-age difference the estimates represent.

The primary endpoint combines mortality and multiple types of postoperative complications. Separate primary estimates for the components are not reported, so it is not possible to determine whether the overall composite reflects a consistent pattern across events. The record also does not describe the rationale for selecting the composite components or provide details about event adjudication beyond the stated STS definition.

The relationship between the enrollment figure of 3942 and the posted primary outcome groups is not explained. A participant-flow account, reasons for exclusions, and details of missing outcome data are not reported in the registry record. Analysis populations and handling of participants who did not follow their assigned strategy are also not reported.

The primary GEE model’s adjustment variables and correlation structure are not reported. The record does not give the complete statistical analysis plan, multiplicity procedures, interim analyses, or original sample-size assumptions. For the Cox analysis, the event definition, censoring rules, and assumption checks are not reported. These limitations do not invalidate the posted estimates, but they constrain independent evaluation of model assumptions, selection effects, and the full inferential context.

Safety reporting also lacks detail on the percentage denominators, event timing, overlap among adverse-event terms, and statistical comparisons. A detailed ICU length-of-stay result is not reported in the registry record. Consequently, the available secondary and safety information does not support broad conclusions beyond the specific posted summaries.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Red Cell Storage Duration and Outcomes in Cardiac Surgery — NCT00458783.
• Registry record last updated 2021-06-03; status last verified 2021-05.
• Results first posted 2021-06-03.
• 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.

23. Record Summary

The registry record describes a completed, randomized, double-masked, parallel-group trial sponsored by The Cleveland Clinic, comparing transfusion with the oldest available matching red blood cells against transfusion with the youngest available matching red blood cells in cardiac surgery. Enrollment is listed as 3942; the recorded start is 2007-04 and primary completion is 2018-12.

The primary endpoint is a 30-day composite of mortality or specified postoperative complications. The posted counts are 360 events among 686 participants in the prolonged-storage group and 362 among 701 in the short-storage group. The unadjusted risk difference is -0.008, with a confidence interval from -0.061 to 0.044. The sponsor-reported GEE odds ratio is 0.77, with a 95% confidence interval from 0.50 to 1.2 and p = 0.08.

The registry also reports postoperative hospital length-of-stay summaries and a Cox regression estimate. ICU length-of-stay results, detailed analysis-population information, missing-data methods, and multiplicity procedures are not reported in the registry record.