To Study the Effect of Early Cooling in Acute Subdural Hematoma Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for To Study the Effect of Early Cooling in Acute Subdural Hematoma Patients (NCT02064959), a randomized na study. The primary outcome is Number of Participants With Favorable Glasgow Outcome Score-Extended (GOS-E) at 6 Months Post Injury.

2. Clinical Question

The registry identifies this as a randomized, prospective trial and describes its purpose as assessing whether very early cooling affects outcome after traumatic subdural hematoma evacuation. The posted primary outcome compares the number of participants in each group classified as having a favorable score. That binary classification gives a straightforward account of how many participants met the favorable-outcome definition, but it does not preserve the full range of the Glasgow Outcome Scale-Extended. The registry record does not report further details about the score assessment process or the clinical characteristics of participants.

3. Trial Design

The registry describes a randomized, prospective, parallel-group trial with single masking. It lists the study as terminated and reports enrollment of 32. The record names The University of Texas Health Science Center, Houston as sponsor. The official title describes a multicenter randomized clinical trial, but the number of participating centers is not reported in the registry record.

The two posted groups are Normothermia and Hypothermia, with 16 participants listed in each. The normothermia group is described as standard care at 37°C. The hypothermia group is described as cooling to 33°C. The trial summary gives a more detailed intervention sequence: cooling to 35°C prior to opening the dura, followed by maintenance at 33°C for a minimum of 48h. It identifies an intravascular temperature-management device for inducing hypothermia or maintaining normothermia. The record does not report actual temperature exposure, treatment adherence, or whether participants reached the intended temperatures.

The primary outcome is favorable Glasgow Outcome Scale-Extended status at 6 months post-injury. The registry also lists safety as assessed by the number of adverse events reported per participant, intensive care unit length of stay, hospital length of stay, and incidence of cortical spreading depolarization. Numerical results for these secondary outcomes are not reported in the registry record. The phase is listed as NA. The trial's original design assumptions, including its planned sample-size rationale, are not reported in the registry record.

The start date is 2014-03-22, and the primary completion date is 2019-02-07. These dates describe the registry record's study timeline; they do not establish when individual participants were enrolled or assessed. Further details about random-sequence generation, allocation concealment, which role was masked, and the reason for termination 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

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 Participants With Favorable Glasgow Outcome Score-Extended (GOS-E) at 6 Months Post Injury, measured over 6 months post injury.

The primary endpoint is the number of participants with a favorable Glasgow Outcome Scale-Extended score at 6 months post-injury. The registry defines favorable status as moderate disability to good recovery. Severe disability, vegetative state, or death is classified as unfavorable. The outcome is therefore represented as a count in each group, with the group size providing the context for interpreting that count.

The Glasgow Outcome Scale-Extended is used in the registry's primary question to summarize functional outcome after injury. Converting the score into favorable and unfavorable categories makes the primary comparison a binary outcome: each participant falls into one of the two stated categories. This simplifies interpretation of the proportion meeting the favorable definition, but it also means that differences within either category are not reflected in the reported count. A participant with moderate disability and one with good recovery are both included as favorable, despite the distinctions in the scale's wording.

The specified assessment time is 6 months post-injury. This gives the primary outcome a defined follow-up point rather than an unspecified measurement during hospitalization. The record does not report details such as the assessment instrument procedure, assessor training, or how missing follow-up scores were handled. The posted results give the favorable counts by group; they do not provide the full distribution across individual Glasgow Outcome Scale-Extended levels.

7. Results Posted in the Registry

• Hypothermia: 37.5% (6/16) versus 25.0% (4/16) on control; risk difference 12.5 percentage points (95% CI -19.3 to 44.3), risk ratio 1.50 (0.52–4.32); unadjusted two-proportion z-test p = 0.446.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the risk difference describe?

The risk difference compares the proportion with a favorable outcome in the hypothermia group with the proportion in the normothermia group. The posted estimate is 0.125. A positive value means the observed favorable-outcome proportion was higher in the hypothermia group for this comparison. It is an absolute measure: unlike a ratio, it describes the separation between the group probabilities on the probability scale. The estimate is based on the posted counts, without adjustment for other variables.

How should the confidence interval be read?

A confidence interval communicates uncertainty about an estimated effect under the statistical procedure used to construct it. It is not a range of individual participant outcomes. The risk-difference interval runs from -0.193 to 0.443, including zero. Zero represents no risk difference. Because the interval includes values on both sides of zero, the data are compatible with a lower favorable-outcome probability in the hypothermia group, no difference, or a higher probability. The width of the interval also indicates that the estimate is not precise. The registry record does not report the method used to construct this interval.

What does the p-value add?

The unadjusted p-value is 0.446. A p-value describes how unusual data at least as discrepant as those observed would be under a specified null model and its assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. In this record, the p-value is labeled unadjusted. The registry record does not report the statistical test used, any prespecified significance threshold for the primary analysis, or an adjusted p-value. The p-value should therefore be read alongside the effect estimate and confidence interval, not as a stand-alone verdict.

Why are risk ratios and odds ratios also shown?

A risk ratio compares the outcome probabilities between groups by division; an odds ratio compares the odds. These are relative measures, while the risk difference is absolute. The posted risk ratio is 1.5 and its interval is 0.520 to 4.323. The posted odds ratio is 1.8 and its interval is 0.394 to 8.215. For both measures, the value corresponding to no relative difference lies within the interval. The interval widths show that the relative effects are also uncertain. Odds and probabilities are not interchangeable, so the odds ratio should not be read as a direct percentage-point change.

What does randomization contribute?

Random allocation is intended to make treatment groups comparable, on average, with respect to factors that could affect outcome. It does not guarantee that groups will be alike in every characteristic, particularly in a small trial. The registry identifies the trial as randomized but does not report the sequence-generation or concealment methods. It also does not provide baseline characteristics in the registry record used for this interpretation. Consequently, no assessment of balance on individual prognostic factors can be made from the reported information.

Why can small samples create uncertainty?

With limited observations, each outcome contributes appreciably to the estimated group proportion. This can lead to wide intervals and makes the results sensitive to a small number of outcome classifications. General statistical principle: uncertainty is driven not only by the point estimate but also by the amount and variability of information available. The group counts and confidence intervals in this trial make that uncertainty visible. A nonsignificant p-value does not prove the groups are equivalent; demonstrating equivalence requires a suitable design and prespecified margin, neither of which is reported in the registry record.

What analyses are not described?

The registry lists zero analyses. This does not identify which analytical decisions were made outside the registry record. Details about analysis populations, covariate adjustment, missing-outcome handling, interim analyses, and a prespecified statistical analysis plan are not reported in the registry record. Without those details, the posted unadjusted comparison can be described, but its relationship to any planned primary analysis cannot be established. No assumptions about those unreported choices should be inferred.

10. Confidence Intervals and Precision

Precision concerns how narrowly the data locate the effect, rather than whether the point estimate is positive or negative. The risk-difference estimate for hypothermia compared with normothermia is 0.125, and its 95% confidence interval extends from -0.193 to 0.443. The interval crosses zero and spans effects in both directions. It is therefore consistent with no risk difference, as well as with appreciable differences favoring either group. The estimate should not be treated as a precise measure of the underlying effect.

The relative-effect intervals are also broad: 0.520 to 4.323 for the risk ratio and 0.394 to 8.215 for the odds ratio. Each interval includes the value for no relative difference. The point estimates alone—1.5 for the risk ratio and 1.8 for the odds ratio—do not convey this uncertainty. Intervals show the range of effect values compatible with the data under the interval method and its assumptions; they do not mean that every value in the range is equally plausible.

The posted group counts are 16 in each arm, with favorable outcomes reported for 4 in normothermia and 6 in hypothermia. With this amount of information, the estimate can remain highly uncertain even when a point estimate appears to favor one group. The registry record does not report the confidence-interval construction method, so that aspect of precision cannot be independently characterized from the record. Overall, the posted results are not sufficiently precise to distinguish clearly between no difference and effects in either direction.

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 participants or total events. They are illustrations calculated from the observed effect, not the trial's original sample-size plan. The registry record does not report the original design assumptions or sample-size justification. The listed assumptions for the illustration are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect.

For each effect assumption, the 80% power column gives the planned participants per arm under which the calculation has an 80% chance of detecting an effect of that size if the assumptions hold and that effect is real. The 90% power column uses a higher target chance of detection, and therefore calls for more participants per arm under the same assumptions. These are planning quantities, not outcomes from the trial. They are not observed power.

The rows for 100%, 75%, and 50% of the observed effect show how the planning estimate changes when the assumed effect is smaller. For 100% of the observed effect, the illustration gives 215 participants per arm for 80% power and 287 for 90% power. For 75% of the observed effect, it gives 372 participants per arm for 80% power and 498 for 90% power. For 50% of the observed effect, it gives 812 participants per arm for 80% power and 1,086 for 90% power. Smaller effects need more participants because they are harder to distinguish from random variation. As the starting effect comes from a small trial and has substantial uncertainty, these illustrations should not be treated as definitive sample-size recommendations.

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 = 0.250; treatment proportion p₁ = 0.375; average p̄ = 0.312.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6555 = 1.2848.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6495 = 0.5466.
4. n = (1.2848 + 0.5466)² / (0.1250)² = 214.7, rounded up to 215 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 <- 0.25                          # control proportion
p1 <- 0.375                          # 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(2064959)
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 = 215 (80% power), 287 (90% power)
 75% of observed effect: n per arm = 372 (80% power), 498 (90% power)
 50% of observed effect: n per arm = 812 (80% power), 1,086 (90% power)
power.prop.test: 214.7 
Simulated power at n = 215 per arm: 0.800 (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

The registry lists one primary outcome and four secondary outcomes: adverse events per participant, intensive care unit length of stay, hospital length of stay, and incidence of cortical spreading depolarization. Numerical results for these secondary outcomes are not reported in the registry record. The record also lists five outcome measures and zero analyses. It does not describe which outcomes were included in any inferential testing or whether all listed outcomes were analyzed.

Multiplicity arises when several outcomes, time points, groups, or subgroup comparisons are examined. In general, examining more statistical comparisons can increase the chance of finding an apparently notable result by chance. Methods to manage that issue include prespecifying a primary comparison or adjusting inference across a defined family of tests. Whether such methods were used for this trial is not reported in the registry record. The single posted primary comparison can be described on its own, but the full context of outcome selection and testing cannot be reconstructed from the record.

Missing data can affect an outcome comparison if participants do not have a recorded assessment or if the probability of missingness differs between groups or relates to outcome. The registry record does not report missing primary outcome counts, reasons for missing assessments, imputation methods, or sensitivity analyses. Although the posted group counts and favorable counts are available, no further account of the analysis population or outcome completeness is reported.

The registry identifies the study as randomized, but it does not report whether the primary analysis followed an intention-to-treat population, a per-protocol population, or another defined group. It also does not report a statistical analysis plan, interim analyses, or the prespecified handling of multiplicity and missing data. These omissions limit assessment of how closely the posted comparison follows a planned analysis. No conclusion about those unreported procedures can be drawn from the result alone.

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 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: anemia (37.5%), death (31.2%), stroke (31.2%), pneumonia (31.2%), thromboembolic event (31.2%).

The registry reports serious adverse event figures separately for the two groups. In the normothermia group, 10 of 16 participants were listed as at risk and 10 as affected; in the hypothermia group, 11 of 16 were listed as at risk and 11 as affected. For other adverse events, the corresponding figures are 10 of 16 in normothermia and 12 of 16 in hypothermia. Deaths are listed as 5 in normothermia and 3 in hypothermia. These are descriptive safety counts and should not be interpreted as proof that the intervention caused or prevented any event.

The registry also lists serious-event terms with percentages: anemia, 37.5%; death, 31.2%; stroke, 31.2%; pneumonia, 31.2%; thromboembolic event, 31.2%; and hypermagnesemia, 25.0%. The denominator and the relationship between these term percentages and participants affected are not reported in the registry record. Event terms may not represent mutually exclusive participants, so they should not be added together or treated as a group-level rate without further information.

The record does not report event timing, severity grading beyond the serious-event designation, attribution to treatment, or comparative statistical testing for these safety results. These counts provide a descriptive account of reported events, not a causal safety comparison.

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

The posted primary result describes a small randomized comparison of early hypothermia and normothermia for patients undergoing evacuation of traumatic subdural hematoma. Favorable status at 6 months was reported for 6 participants in the hypothermia group and 4 in the normothermia group. The risk-difference estimate is 0.125, but its confidence interval includes zero and extends in both directions. The unadjusted p-value is 0.446. The result therefore leaves substantial uncertainty about the direction and size of any group difference.

This result does not establish that the groups are equivalent. A lack of a clearly distinguishable difference in these data is not the same as evidence that the interventions have the same effect. Nor does the point estimate alone establish a benefit. The confidence interval makes clear that values consistent with lower, similar, or higher favorable-outcome probabilities remain compatible with the posted data.

The outcome is a binary classification of the Glasgow Outcome Scale-Extended at a single follow-up point. It combines moderate disability and good recovery into the favorable category, while grouping severe disability, vegetative state, and death as unfavorable. The reported counts therefore do not describe the detailed distribution of functional outcomes within those categories. They also do not provide a complete account of quality of life, recovery trajectory, or other outcomes beyond those named in the registry.

The secondary outcomes and safety results add context, but the registry does not report numerical results for the listed secondary outcomes, and the safety counts are descriptive. The record does not report enough information about missing assessments, analysis populations, or statistical planning to resolve those limitations. The most defensible reading is that the posted comparison is uncertain and should not be extended beyond the outcome, follow-up, and groups recorded.

16. Why This Trial Matters Statistically

This trial illustrates a central issue in clinical statistics: a point estimate and a decision about statistical evidence are not the same thing. The observed risk difference is 0.125, but the interval around it is wide and includes zero. A reader who considers only the point estimate may overstate what the small comparison establishes; considering the interval makes the uncertainty explicit.

Binary endpoints can be easy to communicate because they reduce an outcome to counts meeting a defined criterion. That clarity comes with a trade-off: the favorable category combines moderate disability and good recovery, and the unfavorable category combines severe disability, vegetative state, or death. The count comparison cannot show how outcomes vary within each category. This is a useful reminder to interpret the endpoint according to its definition rather than treating it as a complete account of patient experience.

The sample-size illustrations also demonstrate why planning depends on the effect one aims to detect. The figures rise as the assumed effect becomes a smaller fraction of the observed effect. Because the illustration starts from an uncertain observed result, it is not a substitute for a prospectively justified design. The trial's original sample-size assumptions are not reported in the registry record.

Finally, the record distinguishes between what can be described and what remains unknown. Randomization is recorded, but detailed procedures, missing-data handling, multiplicity strategy, and analysis populations are not reported. Separating reported findings from those gaps is essential for a measured statistical interpretation.

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 primary comparison is based on 16 participants in each group, and the confidence interval is broad. This limits the precision of the estimated difference and means the posted result is compatible with effects on either side of no difference. The registry reports that the trial was terminated but does not report the reason, so the implications of the termination for recruitment, follow-up, or analysis cannot be assessed.

The registry record does not report the number of centers, detailed eligibility criteria, baseline characteristics, randomization procedures, or which role was masked. Although the official title describes a multicenter trial and the design is listed as single-masked, the record does not identify the centers or clarify who was masked. The record also does not report adherence to the temperature protocol, achieved temperatures, or treatment duration for individual participants.

The primary result is a binary outcome at 6 months. The full score distribution is not reported, and details about outcome assessment and missing follow-up data are not reported in the registry record. The record does not provide an analysis population, adjustment strategy, prespecified statistical analysis plan, or details of how multiple outcomes were handled. Numerical results for the four listed secondary outcomes are not reported.

The serious adverse event terms are presented with percentages, but their denominator and relation to participant-level counts are not reported. Event attribution, timing, and comparative testing are also not reported. These gaps limit how far the safety figures can be interpreted beyond descriptive reporting.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: To Study the Effect of Early Cooling in Acute Subdural Hematoma Patients — NCT02064959.
• Registry record last updated 2020-10-19; status last verified 2020-09.
• Results first posted 2020-10-19.
• 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

NCT02064959 is a randomized, prospective, parallel-group trial of early hypothermia versus normothermia for patients undergoing evacuation of traumatic subdural hematoma. The registry lists single masking, enrollment of 32, and a terminated status. It describes cooling to 35°C before opening the dura, followed by maintenance at 33°C for a minimum of 48h, compared with normothermia at 37°C.

The primary outcome is favorable Glasgow Outcome Scale-Extended status at 6 months post-injury. Favorable outcomes were reported for 6 of 16 participants in the hypothermia group and 4 of 16 in the normothermia group. The risk difference is 0.125, with a 95% confidence interval from -0.193 to 0.443; the unadjusted p-value is 0.446. These results are imprecise and do not establish a clear group difference. Four secondary outcomes are listed, but numerical results for them are not reported in the registry record.