CO2 Versus Air Insufflation for Single-balloon Enteroscopy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for CO2 Versus Air Insufflation for Single-balloon Enteroscopy (NCT01758900), a randomized phase 4 study. The primary outcome is Intubation Depth.

2. Clinical Question

The trial addressed uncertainty about CO2 use in single-balloon enteroscopy. Its registry summary notes earlier use of CO2 in other endoscopic procedures while stating that its role in single-balloon enteroscopy remained uncertain. The clinical question is therefore specific to the procedure and the measured endpoint; results from other procedures are not a substitute for this randomized comparison.

3. Trial Design

This was a randomized, controlled, phase 4 trial with a parallel-group model and triple masking. Participants were allocated to CO2 insufflation or room-air insufflation. The registry record does not identify who was masked, explain how masking was maintained, or describe the randomization sequence or allocation concealment. It lists Zhizheng Ge as the sponsor.

The registry reports enrollment of 214 participants: 106 in the CO2 insufflation regulator arm and 108 in the air insufflation regulator arm. The record gives a start date of 2011-07 and primary completion in 2013-05, and lists the trial status as completed. It does not report participant flow, withdrawals, exclusions, or the numbers included in each outcome analysis.

The primary outcome was mean intubation depth in centimetres, assessed within 5 minutes after the examination. The registry also names six secondary outcomes: total enteroscopy rate, diagnostic rate, patient acceptability, procedure time, abdominal circumference, and complication rate. Numerical results are posted for total enteroscopy rate, diagnostic rate, and patient acceptability. Further numerical results for the other named secondary outcomes are not reported in the registry record.

The registry record does not report the original sample-size justification, the planned statistical analysis, interim analyses, or analysis populations. It also does not explain the masking roles or describe procedural standardization. These absences limit what can be concluded about how the trial was planned and how closely the reported analyses followed a prespecified plan.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

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 Intubation Depth, measured over Within 5 minutes after the examination.

Intubation depth is the distance, in centimetres, reached during the enteroscopy. The registry defines the primary measure as a mean and specifies measurement within 5 minutes after the examination. In general, a continuous endpoint such as this retains the scale of the measurement and allows the average measured depth to be compared between randomized groups.

For an enteroscopy, depth is a procedural measure of how far the instrument advances into the intestine. It offers a direct way to quantify the extent of insertion, rather than reducing the result to a yes-or-no classification. The registry record does not define a threshold for a successful or clinically meaningful depth, nor does it report how the depth measurement was obtained or whether it was independently assessed.

The posted values are group means with standard deviations. The comparison is expressed as a mean difference between the air and CO2 groups, with a confidence interval. The negative direction reflects the lower reported mean in the air group relative to the CO2 group. This endpoint concerns measured procedural depth; by itself, it does not establish diagnostic benefit, patient experience, safety, or any broader clinical consequence.

7. Results Posted in the Registry

• Air Insufflation Regulator: mean 238.30 versus 323.80 on control; difference -85.50 (95% CI -103.30 to -67.70), standardized effect d = -1.29; z-test p = <0.001.

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.

Total Enteroscopy Rate

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Air Insufflation Regulator: 17.6% (19/108) versus 34.9% (37/106) on control; risk difference -17.3 percentage points (95% CI -28.9 to -5.7), risk ratio 0.50 (0.31–0.82); unadjusted two-proportion z-test p = 0.004.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Air Insufflation Regulator than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Diagnostic Rate

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Air Insufflation Regulator: 47.2% (51/108) versus 58.5% (62/106) on control; risk difference -11.3 percentage points (95% CI -24.6 to 2.0), risk ratio 0.81 (0.63–1.04); unadjusted two-proportion z-test p = 0.099.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Air Insufflation Regulator and the control. The interval gives the range of effects that remain compatible with the data.

Patient's Acceptability

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Air Insufflation Regulator: 99.1% (107/108) versus 99.1% (105/106) on control; risk difference 0.0 percentage points (95% CI -2.6 to 2.6), risk ratio 1.00 (0.97–1.03); unadjusted two-proportion z-test p = 0.989.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Air Insufflation Regulator and the control. The interval gives the range of effects that remain compatible with the data.

The registry names six secondary outcomes, but numerical comparisons are posted for total enteroscopy rate, diagnostic rate, and patient acceptability. Across these outcomes, the pattern is mixed rather than uniform: the two procedure-related rates have higher reported values in the CO2 arm, while acceptability is close between arms. The primary endpoint and these secondary measures address different aspects of the procedure. Intubation depth describes measured insertion distance; the secondary outcomes concern completion, diagnostic yield, and participant acceptability.

The secondary findings can add context to the primary result, but they do not explain the mechanism behind a difference in depth or establish that the depth difference caused any other outcome. The registry record does not report analyses linking intubation depth to the secondary outcomes, nor does it provide results for procedure time, abdominal circumference, or complication rate. Those named outcomes therefore cannot be used to confirm or qualify the observed pattern from the posted comparisons.

Unadjusted secondary results carry less weight as confirmatory evidence when several outcomes are examined. In general, considering multiple outcomes increases the chance that at least one comparison may appear unusual through random variation, even if no multiplicity adjustment is applied. The record does not report a hierarchy of secondary outcomes, a multiplicity correction, or a prespecified strategy for interpreting them. Their estimates and intervals are informative descriptions of the posted data, but they should be interpreted cautiously and in relation to the primary endpoint rather than treated as separate proof of a broad effect.

The registry's secondary outcomes also have different measurement scales and assessment windows. Total enteroscopy rate and diagnostic rate are assessed within 1 week after all examinations are finished, whereas acceptability is assessed 6 hours after the examination. These differences mean the outcomes should not be combined into one summary claim. Their separate timing and definitions matter when considering how they relate to the primary measurement taken within 5 minutes after the examination.

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

• Total Enteroscopy Rate
• Diagnostic Rate
• Patient's Acceptability
• Procedure Time
• Abdominal Circumference
• Complication Rate

9. Statistical Interpretation

10. Statistical Methods Explained

The registry reports a mean and standard deviation for the continuous primary endpoint in each arm. A mean is the arithmetic average of the observations in a group. A standard deviation describes how dispersed the observations are around that average. In general, means provide a compact summary of a continuous outcome, but they do not show the shape of the data distribution or whether a small number of unusually high or low values influenced the average. The registry record does not report distributional checks or alternative summaries.

What does the mean difference express?

The comparison is a difference between the two arm means, expressed as air minus CO2. A negative estimate therefore indicates that the observed average was lower in the air arm. The estimate is expressed in the original outcome unit, centimetres, which makes its direction and scale interpretable without converting it to another measure. The registry does not state which direction was chosen as the prespecified contrast; the direction can be read from the arm means and the posted estimate.

What does a confidence interval add?

In general, a confidence interval describes the precision of an estimate under a specified statistical procedure. A narrower interval indicates less uncertainty about the estimate than a wider interval, all else equal. An interval that includes zero is compatible with no difference as well as differences in either direction at the interval's confidence level. An interval that excludes zero is evidence against a zero difference under the model and assumptions used. It is not a range that contains every individual participant's outcome, nor does it assign a probability to a particular value of the fixed true difference after the interval has been calculated.

The registry posts a confidence interval for the primary mean difference. It does not identify the precise interval method, state whether any covariate adjustment was used, or provide a full analysis specification. The interval should therefore be interpreted as the posted comparison, not as a complete account of the analysis choices behind it.

How should the p-value be read?

In general, a p-value measures how incompatible the observed data, or more extreme data, are with a specified null hypothesis under the statistical model. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The primary comparison has an unadjusted p < 0.001. “Unadjusted” means the reported value is not corrected for considering multiple outcomes or comparisons. The registry record does not report a multiplicity plan.

How do the binary outcomes differ?

For outcomes recorded as counts or percentages of participants, the posted comparisons use a risk difference: the proportion in the air arm minus the proportion in the CO2 arm. In general, a risk difference is expressed on the proportion scale and indicates the absolute separation between the two group proportions. A negative value indicates a lower proportion in the air arm; a positive value indicates a higher proportion in the air arm. The interval indicates the range of differences compatible with the data and the calculation method. The registry also posts relative measures for some comparisons, but the risk difference provides an absolute-scale summary.

These binary comparisons are based on the posted event counts and group sizes. Their uncertainty depends on both the observed proportions and the numbers of participants. As with the primary endpoint, the registry does not report the full analysis specification or whether additional adjustment was applied. The p-values are unadjusted, and results across several outcomes should not be treated as independent confirmations.

What can be inferred from randomization and masking?

In general, random allocation is intended to make comparison groups similar on average with respect to factors measured and unmeasured before treatment assignment. Masking can reduce the influence of knowledge of assignment on care, measurement, or assessment, depending on who is masked and how the procedure is carried out. The record identifies the trial as randomized and triple-masked, but does not describe sequence generation, allocation concealment, or the roles of those masked. Those labels alone do not establish how successfully these safeguards were implemented.

The registry's planning illustrations use the observed effect as a starting point and apply stated assumptions: two-sided significance level 0.05, equal allocation, and a normal approximation. They show hypothetical participant requirements for specified fractions of the observed effect. They are not the original trial's sample-size calculation, because the trial's own design assumptions are not reported. The values are planning illustrations, not observed power. In general, power is a design property calculated for an assumed effect and analysis, not a retrospective measure of whether the observed result was convincing.

In general, over-dispersion matters when binary or count data vary more than the statistical model assumes: uncertainty may then be understated if the extra variation is ignored. The registry does not report an over-dispersion assessment or a model for count outcomes. It is therefore not possible to say whether over-dispersion was relevant to the analyses in this trial.

11. Confidence Intervals and Precision

The primary mean-difference interval runs from -103.3 to -67.7 cm around the estimate of -85.5 cm. Its endpoints are both negative, indicating that the posted interval is wholly in the direction of a lower mean for air than for CO2. The interval nevertheless covers a range of possible magnitudes, rather than fixing the difference at the point estimate. It is useful to consider both its direction and its width: the direction is consistent across the interval, while the width conveys remaining uncertainty about how large the mean difference may be.

In general, a confidence interval reflects sampling uncertainty under the calculation method and its assumptions. It does not include every possible source of uncertainty, such as problems in measurement, selection of participants for analysis, or departures from the planned design. The registry record does not report enough detail to assess all such sources. The posted standard deviations show variation in the measured endpoint within each arm, but the confidence interval for the difference is not an interval for individual participants.

The secondary confidence intervals vary in their relation to zero. The interval for total enteroscopy rate is on one side of zero, while the intervals for diagnostic rate and acceptability extend across zero. This distinction describes statistical precision under the reported comparisons; it does not establish that one outcome is more important than another. The intervals are unadjusted, and the record does not report whether the family of outcomes was accounted for in the analysis.

Precision and clinical relevance are separate questions. A relatively precise estimate can still lack a known practical threshold, and a wide interval can leave a broad range of effects plausible. Since no clinically important threshold for intubation depth is reported, the primary interval cannot by itself determine the practical significance of the difference.

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 illustrations give the number of participants per arm for 80% power and for 90% power. They are participant counts per group, not total event counts. The registry's planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set to the stated fraction of the observed effect. The trial's original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column represents a planning target under which a study with the assumed effect and analysis would have an 80% chance of detecting an effect at the specified significance level. The 90% column uses a higher planning target: under the same assumed conditions, it would have a 90% chance of detection. These are prospective design concepts, not probabilities that the trial result is correct or that a treatment effect exists.

The rows consider planning for 100%, 75%, and 50% of the observed effect. Smaller effects require more participants for the same target power, because they are harder to distinguish from random variation. The figures are conditional on the stated assumptions and the effect used in each row; they are not a reconstruction of the trial's actual design decision. They are not observed power.

The curve visualizes how the planning requirement changes as the assumed effect changes. It should be read as a sensitivity illustration: changing the assumed effect changes the number of participants required. It does not show the trial's observed outcomes or establish the power the completed trial had for effects other than those used in the illustration.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -85.500; pooled standard deviation σ = 66.457.
2. n = 2 × (1.96 + 0.842)² × 66.457² / -85.500² = 9.5, rounded up to 10 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
delta <- -85.5                      # observed difference in means
sigma <- 66.4571758835349                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1758900)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 10 (80% power), 13 (90% power)
 75% of observed effect: n per arm = 17 (80% power), 23 (90% power)
 50% of observed effect: n per arm = 38 (80% power), 51 (90% power)
power.t.test: 10.5 
Simulated power at n = 10 per arm: 0.775 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-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. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome and six secondary outcomes, but does not report a statistical plan for handling multiple comparisons. It does not say whether the primary outcome was designated as the sole confirmatory test, whether secondary outcomes were ordered in a hierarchy, or whether adjusted significance levels or confidence intervals were planned. The posted p-values are unadjusted. When several outcomes are tested, unadjusted results do not control the overall chance of a false-positive finding across the set. This does not make the results unusable, but it affects how strongly individual secondary p-values should be interpreted.

The record does not report how missing outcome data were handled, the amount of missingness, or whether missingness differed by arm. It also does not provide participant flow or explain whether all randomized participants contributed to each outcome. The outcome group sizes are posted, but those values alone do not establish an intention-to-treat analysis or explain any exclusions. The analysis populations, including whether the primary comparison used all randomized participants, are not reported in the registry record.

Other details that would help evaluate the analysis are also not reported: the planned statistical models, covariate adjustments, sensitivity analyses, assumptions about the outcome distributions, and the original sample-size calculation. The registry record does not report interim analyses or stopping rules. It also does not state whether the outcomes and analysis choices were prespecified before results were known.

These gaps do not change the numerical summaries in the registry. They limit assessment of whether the posted comparisons follow a prespecified analysis and how robust the estimates are to missing data, exclusions, or alternative methods. The record identifies the study as randomized and triple-masked, but those design labels do not fill in the missing analysis and reporting details.

15. Safety Results

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

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: acute haemorrhage (0.9%), acute pancreatitis (0.9%).

The registry reports one serious adverse event among 106 participants at risk in the CO2 arm and one among 108 participants at risk in the air arm. It reports zero other adverse events in each group. The deaths field is not reported in the registry record. These counts describe what was posted; they should not be taken as a complete account of all possible harms or as evidence that the risks are identical.

The registry lists acute haemorrhage and acute pancreatitis among the top serious event terms, each reported as 0.9%. The record does not link those terms to a specific arm in the information available for these results. It also does not provide event narratives, timing, severity beyond the serious-event classification, relatedness assessments, or details of monitoring and ascertainment.

Rare events can be difficult to compare precisely when only a small number are reported. The similar posted counts do not establish equivalence in safety, and the absence of other reported adverse events does not prove that none occurred outside the registry's reported data. Safety interpretation is limited by the small number of serious events and the missing details about event collection and follow-up.

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

The posted primary result shows a difference in mean intubation depth between randomized groups, with the CO2 group having the higher reported mean. The confidence interval lies entirely in the direction of a lower mean for air relative to CO2. This is evidence about the measured endpoint in this trial and under the assumptions of the reported comparison.

It does not establish that every participant experienced the same difference, or that intubation depth is a complete measure of procedural success. The registry record does not report a threshold that defines a clinically important difference. It also does not establish that the observed difference leads to better diagnosis, improved patient experience, or a different safety profile. Those questions require separate outcomes and sufficiently complete analyses.

The posted secondary findings provide some additional context, but they are unadjusted and are not all aligned in the same way. Results for named outcomes without numerical results in the registry cannot be inferred from the primary finding. The trial's results also do not settle questions about other enteroscopy procedures, settings, or patient groups that are not described in the registry record.

Statistical evidence and clinical meaning are related but distinct. A p-value addresses compatibility with a statistical null under a model; it does not indicate whether an effect is large enough to matter in practice. The confidence interval helps convey the range of estimates compatible with the data, but it cannot replace a prespecified clinical threshold or additional information about measurement and outcomes. The registry record does not report those elements.

The result should therefore be read narrowly: the posted group means and comparison indicate a difference in measured intubation depth in this randomized trial. The available registry information does not support broader claims about overall benefit, safety equivalence, or the complete balance of outcomes.

17. Why This Trial Matters Statistically

This trial illustrates how a procedural question can be expressed as a quantitative endpoint. Rather than relying only on a broad judgment about examination success, it records intubation depth as a continuous measure and compares group means. The estimate remains in centimetres, the unit clinicians and researchers can connect to the underlying measurement, while the confidence interval communicates uncertainty around the group difference.

It also demonstrates why an endpoint should be interpreted in context. A difference in insertion depth is not interchangeable with a difference in diagnostic rate, completion, acceptability, or complications. The registry lists these as distinct outcomes, and the available results do not show that one necessarily explains another. Separating endpoints helps prevent a precise result on one measure from being used to make unsupported claims about an entire procedure.

The record is also a useful example of the distinction between trial results and the statistical plan needed to evaluate them fully. Randomization and masking are important design features, but the registry does not report enough detail about allocation, analysis populations, missing data, or multiplicity to reconstruct the full inferential approach. A numerical estimate can be informative while important reporting gaps remain.

Finally, planning illustrations based on an observed effect should not be mistaken for the original sample-size rationale. They show how assumed effect size and target power affect participant requirements, not whether the completed trial was well powered for the question it was designed to answer. This distinction matters whenever retrospective planning calculations accompany posted trial results.

18. 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 eligibility criteria, baseline characteristics, participant flow, withdrawals, or reasons for exclusion. Without these details, it is difficult to assess the population represented by the analyzed outcomes or whether the groups remained comparable in the reported analyses. The enrollment and arm sizes are available, but the analysis populations are not.

The record gives limited detail about the intervention and procedure. It does not describe CO2 delivery settings, how room-air insufflation was applied, procedural standardization, operator experience, or how intubation depth was measured. It also does not identify who was masked despite labeling the trial triple-masked. These gaps make it harder to assess reproducibility and the possibility of measurement or performance differences between arms.

Only some named secondary outcomes have numerical results posted. Procedure time, abdominal circumference, and complication rate are listed, but further numerical results for them are not reported in the registry record. The registry also does not provide outcome-level missingness, event definitions, a multiplicity approach, or a detailed statistical analysis plan. The posted secondary comparisons are unadjusted.

Finally, there is no reported threshold for a clinically important change in intubation depth. The result therefore supports a statistical comparison of the group means, but the registry information is insufficient to determine the practical importance of the difference or its relationship to broader clinical outcomes.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: CO2 Versus Air Insufflation for Single-balloon Enteroscopy — NCT01758900.
• Registry record last updated 2016-11-25; status last verified 2016-10.
• Results first posted 2016-07-14.
• 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.

22. Record Summary

NCT01758900 is a completed, randomized, controlled, phase 4, parallel-group trial comparing CO2 insufflation with room-air insufflation during single-balloon enteroscopy. The registry lists triple masking and enrollment of 214 participants, with 106 in the CO2 arm and 108 in the air arm.

The primary outcome was mean intubation depth in centimetres, measured within 5 minutes after the examination. The posted mean was 323.8 cm for CO2 and 238.3 cm for air; the estimated mean difference for air relative to CO2 was -85.5 cm, with a 95% confidence interval from -103.3 to -67.7 cm and unadjusted p < 0.001.

Numerical secondary comparisons are posted for total enteroscopy rate, diagnostic rate, and patient acceptability. The registry does not report a complete statistical plan, original sample-size assumptions, analysis populations, or missing-data handling. Additional named secondary outcomes do not have numerical results reported in the registry record.