Caffeine to Reduce Mechanical Ventilation in Preterm Infants: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Caffeine to Reduce Mechanical Ventilation in Preterm Infants (NCT01751724), a randomized na study. The primary outcome is Age at First Successful Extubation.

2. Clinical Question

The trial’s stated hypothesis was that early caffeine would reduce time to first elective extubation. The registered primary outcome is titled “Age at First Successful Extubation” and is measured in days from birth until 36 weeks postmenstrual age. The record also identifies total duration of mechanical ventilation and oxygen supplementation, BPD, survival, and survival without BPD as secondary outcomes.

The registry lists enrollment of 87 and identifies the study as terminated. It does not report the reason for termination or detailed eligibility criteria. The posted primary outcome gives a median and inter-quartile range for each arm, but no sponsor-reported comparative analysis is listed. The central statistical question is therefore how to interpret the arm-specific summaries and their uncertainty without treating them as a formal estimate of a treatment effect.

3. Trial Design

This was a randomized, parallel-group trial comparing blinded caffeine citrate with blinded placebo, described as an equivalent volume of normal saline. The registry lists quadruple masking. It does not specify which groups of people were masked, how masking was maintained, or whether the masking assignment changed during the trial. The phase is listed as NA.

The sponsor was the University of Miami. The study start is recorded as 2012-12 and primary completion as 2016-01. Its status is terminated. The reason for termination, the date of termination, and whether the study stopped earlier than planned are not reported in the registry record. Enrollment is listed as 87. The outcome summaries have their own arm-specific denominators; the registry does not explain the differences between those denominators and the enrollment figure or give participant-level accounting.

The primary outcome was age at first successful extubation, in days, from birth until 36 weeks postmenstrual age. Successful extubation was defined as remaining extubated for more than 24 hours. The outcome was summarized by medians and inter-quartile ranges. Secondary outcomes included survival, total duration of mechanical ventilation, total duration of oxygen supplementation, BPD, and survival without BPD. Although the record lists 11 outcome measures, only some secondary results are posted in the results material described by the record.

The registry does not report the planned sample-size justification, the intended statistical analysis plan, the exact allocation procedure, or detailed analysis populations. No sponsor-reported analyses are listed for the posted secondary comparisons. Those omissions limit how fully the posted summaries can be connected to the original protocol or to any planned inferential framework.

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: 87 infants were enrolled. Of these, 1 died before randomization. 86 infants were randomized and assigned an intervention arm.

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 Age at First Successful Extubation, measured over From birth to until 36 weeks postmenstrual age.

The primary endpoint measures the infant’s age, in days, at the first extubation that was successful by the registry’s definition: the infant remained extubated for more than 24 hours. The observation period begins at birth and extends until 36 weeks postmenstrual age. The endpoint therefore focuses on when a sustained initial extubation occurred, rather than simply whether an infant was ever extubated.

Time to extubation is relevant to the trial’s stated objective because the registry describes a hypothesis that early caffeine use might reduce time to first elective extubation. In general, a time-based endpoint can capture both the timing of an event and the period over which it is observed. Here, however, the posted primary results are arm-specific medians and inter-quartile ranges, not a hazard ratio or a formal between-arm time-to-event analysis. The registry record does not report how infants without a successful extubation by the end of the observation period were handled in the primary analysis.

The caffeine arm’s median age was 24.0 days, with an inter-quartile range from 9.5 to 41.0 days. The placebo arm’s median was 20.0 days, with an inter-quartile range from 9.3 to 43.0 days. These summaries describe the observed distributions within each arm. On their own, they do not establish the size, precision, or statistical significance of a comparative effect. No sponsor-reported primary comparison is listed in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Clinical trial results can be presented as summaries within each arm, as direct comparisons between arms, or both. Those forms answer different questions. An arm-specific median describes the middle of the observed outcome distribution in that arm. A comparative estimate instead quantifies how outcomes differ between randomized groups, usually alongside an uncertainty interval and a statistical test. The registry posts the former for the primary endpoint but does not list a sponsor-reported primary comparison.

What does a median and inter-quartile range describe?

The median is the point that divides an ordered set of observations into a lower and upper half. The inter-quartile range describes the middle half of the observed values, bounded by the first and third quartiles. It is a measure of spread, not a confidence interval for the median and not a treatment-effect interval. For age at first successful extubation, the posted medians are 24.0 days in the caffeine arm and 20.0 days in the placebo arm. The corresponding inter-quartile ranges are 9.5 to 41.0 days and 9.3 to 43.0 days.

Comparing two medians informally does not provide a test of whether the underlying distributions differ. A formal analysis could use a method suited to time-to-event data, but the exact method and its assumptions are not reported for the primary outcome. The registry describes the primary endpoint as a median-time outcome, and no two-arm recalculation was considered estimable from the posted summaries. It would not be appropriate to infer a hazard ratio or a treatment-effect confidence interval from the two medians and ranges alone.

How should a confidence interval be read?

A confidence interval expresses uncertainty around an estimate under the method used to construct it. A wider interval indicates less precision than a narrower interval, all else being equal. For the binary secondary outcomes, the posted comparisons use risk differences. The no-difference value for a risk difference is zero. In each of the three posted secondary comparisons, the risk-difference interval crosses zero. This means the interval includes values consistent with a difference in either direction, as well as no difference. It does not show that the true effect is zero, and it does not by itself establish equivalence.

What do the secondary p-values say?

The posted unadjusted p-values for survival, BPD, and survival without BPD are 0.222, 0.552, and 0.923. A p-value measures how unusual data at least as incompatible with a specified null model would be if that model and the test assumptions held. It is not the probability that the null hypothesis is true, nor does it measure the clinical importance or precision of an estimate. These p-values are described as unadjusted; no adjustment for testing multiple outcomes is reported. The registry does not provide sponsor-reported analyses for these comparisons.

For the binary secondary outcomes, arm-specific event counts and denominators are posted. The calculations express the comparison as a risk difference and include an uncertainty interval. The direction of a risk difference depends on which arm is subtracted from which; the posted comparisons use placebo as the control and caffeine as the comparison arm. The counts are useful descriptive information, but a comparison based on small event totals can have considerable uncertainty. The registry does not report adjusted estimates, covariate modeling, or subgroup analyses.

What are the planning illustrations?

The sample-size illustrations are not observed power and are not the trial’s original design justification. The registry explicitly describes them as planning calculations based on the observed effect. The stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation; each planning row considers a stated fraction of the observed effect. For the binary outcomes, the reported figures are participants per arm. They illustrate how a future study’s planning requirement changes as the target effect becomes smaller. They do not show that the trial had a particular probability of detecting an effect, and they do not repair uncertainty in the observed results.

No comparable planning figures are available for the primary median-time endpoint because the record says the values needed for a two-arm comparison were missing or not estimable. The registry does not report the original sample-size calculation, its assumptions, or whether the enrollment target changed. Consequently, the posted planning illustrations should be kept separate from the trial’s design and observed findings.

Why does over-dispersion matter in general?

For count outcomes, a statistical model that assumes variability is fully determined by the mean can be too restrictive when the observed counts vary more than that assumption allows. This is called over-dispersion; in general, accounting for it can widen uncertainty estimates and prevent overly confident inference. It is not possible to determine from the posted record whether over-dispersion was relevant to any analysis in this trial. The primary outcome is a time measure, and the posted secondary outcomes are binary counts, not a reported count-regression analysis.

The registry record does not report the full analysis plan, analysis populations, handling of missing outcomes, censoring rules, or adjustments for multiple comparisons. Those details affect how a reader should interpret formal inference. In their absence, the posted summaries and unadjusted calculations should be read as limited descriptive and exploratory evidence rather than a complete account of the planned statistical analysis.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 11 outcome measures, while the posted summaries include a primary endpoint and selected secondary outcomes. It does not report a prespecified testing hierarchy, a multiplicity-control strategy, or which secondary outcomes were intended to support formal claims. The secondary p-values are unadjusted. When multiple outcomes are examined, the chance of at least one apparently unusual result can increase; methods such as a hierarchy or adjustment can address that issue in a planned analysis. Whether any such method was planned or used for this trial is not reported in the registry record.

There are no sponsor-reported analyses listed for the posted secondary comparisons, and no sponsor-reported comparative analysis is listed for the primary endpoint. The registry does not report whether interim analyses were planned or performed. It also does not report a statistical analysis plan, the exact analysis populations, or how protocol deviations were handled. These omissions make it difficult to assess whether the posted summaries match the intended analysis or whether other planned analyses were conducted.

Missing-data handling is not reported in the registry record. The outcome denominators differ across posted measures, but the registry does not explain why. No assumptions or methods are given for missing outcomes, withdrawal, or exclusions from an analysis. For the primary time-based outcome, the registry also does not describe how infants without a successful extubation by the end of follow-up were handled, or how censoring was applied. These details matter because different analysis choices can affect both estimates and uncertainty.

The study status is terminated, but the reason and timing are not reported. Termination alone does not identify whether missingness, recruitment, safety, or another issue affected the analysis. The registry does not report whether the termination altered outcome collection or analysis plans. In the absence of these details, the posted evidence should be read with caution: the record supports descriptive reporting, but it does not permit a full audit of multiplicity, missing data, or analysis-population decisions.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: seizures (0.0%), tachycardia (0.0%).

The registry reports serious adverse event counts separately by arm. In the caffeine arm, 0 participants were recorded as affected among 41 at risk; in the placebo arm, 0 were recorded as affected among 42 at risk. The listed serious-event terms Seizures and Tachycardia each have a reported frequency of 0.0%. Other adverse events are also reported as affecting 0 participants, with 41 at risk in the caffeine arm and 41 at risk in the placebo arm.

These figures describe the events recorded in the registry, not proof that no such events occurred beyond the recorded reporting period or under other definitions. The record does not provide event-level detail, severity grading beyond the serious-event category, exposure-adjusted rates, or a formal safety comparison. The denominators for other adverse events differ from those for serious adverse events, and the reason is not reported.

Deaths are listed separately: 9 in the caffeine arm and 5 in the placebo arm. These counts should not be assumed to be adverse events, attributed to treatment, or interpreted as a comparative safety result without additional information. The registry does not report causes of death or a sponsor analysis of mortality. The posted safety figures are therefore a limited account of recorded event counts, not a complete basis for causal conclusions.

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

The registry documents a randomized, quadruple-masked comparison of caffeine citrate and placebo in mechanically ventilated preterm infants. The primary outcome is age at first successful extubation, with success defined as remaining extubated for more than 24 hours. Arm-specific medians and inter-quartile ranges are posted, but no sponsor-reported comparative analysis is listed. The summaries alone do not quantify a treatment effect or establish whether the groups differ.

The secondary outcomes offer additional descriptive information, including survival, BPD, and survival without BPD. Their posted unadjusted comparisons are imprecise: the risk-difference intervals cross zero, and the record lists no sponsor-reported analyses. This does not demonstrate that the treatments have the same effect, just as the observed arm differences do not demonstrate a causal effect. The trial’s posted secondary evidence should be treated as exploratory rather than definitive.

The record also does not provide the original sample-size justification, full analysis plan, handling of missing data, multiplicity approach, or explanation for termination. The planning illustrations based on observed effects are not observed power and are not evidence about what the trial was designed to detect. These missing details constrain both interpretation of the results and assessment of the study’s statistical strength.

Accordingly, the posted information can be used to understand what outcomes were measured and what summary values were recorded. It does not support a firm conclusion about the comparative effect of caffeine on the primary endpoint, nor does it establish the absence of an effect. Clinical decisions should not be inferred from these statistical summaries alone.

13. Why This Trial Matters Statistically

This trial raises a common statistical distinction: a clinically motivated primary endpoint can be measured and summarized without a corresponding comparative analysis being available in the registry. Age at first successful extubation is directly aligned with the stated hypothesis, but only arm-specific medians and inter-quartile ranges are posted. Without a comparative estimate and its uncertainty, the central question cannot be answered as precisely as the trial’s randomized design might otherwise permit.

The secondary findings also illustrate why uncertainty matters. Counts and unadjusted comparisons can look directionally different between arms, but broad intervals that include no difference do not resolve whether a true difference exists or how large it might be. A nonsignificant or imprecise comparison is not evidence of equivalence. Conversely, observed differences in counts are not proof of a treatment effect.

Another statistical issue is the distinction between a planning illustration and a trial design calculation. The registry’s planning values are based on observed effects and stated assumptions; they do not describe the original design or observed power. Keeping these concepts separate prevents a retrospective calculation from being mistaken for a prespecified justification.

Finally, the record’s missing details about termination, analysis populations, missing data, and multiplicity affect reproducibility and interpretation. These are not minor technicalities: they help readers understand which participants and outcomes contributed to each result, which comparisons were planned, and how uncertainty was handled. This record is therefore useful as an example of why transparent reporting of estimates, analysis methods, and participant flow is essential alongside outcome counts.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The primary limitation is the absence of a reported comparative analysis for the primary endpoint. The registry provides median age and inter-quartile range by arm, but no treatment-effect estimate, confidence interval, or p-value for that outcome. The two medians cannot substitute for a formal analysis of the time-to-extubation distributions.

Several elements of the trial’s conduct and analysis are not reported in the registry record: detailed eligibility criteria, baseline characteristics, dose and schedule, allocation procedure, masking details, protocol deviations, analysis populations, and missing-data methods. The record also does not explain the varying denominators across outcome summaries. Without participant-level flow or an explanation of exclusions, readers cannot determine why the available outcome counts differ.

The registry status is terminated, but the reason and timing are not reported. The original sample-size calculation and any changes to the target are also not reported. Planning illustrations based on observed effects cannot fill that gap. The statistical analysis plan, interim analyses, multiplicity strategy, and handling of censoring for the primary endpoint are not reported either.

For the secondary outcomes, the posted comparisons are unadjusted and no sponsor-reported analyses are listed. The intervals are broad and cross the no-difference value, limiting precision. Results for some listed secondary measures, including total duration of mechanical ventilation and oxygen supplementation, are not included in the posted summaries described by the record. These limitations restrict the conclusions that can be drawn from the available registry information.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Caffeine to Reduce Mechanical Ventilation in Preterm Infants — NCT01751724.
• Registry record last updated 2017-06-19; status last verified 2017-05.
• Results first posted 2017-06-19.
• 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.

18. Record Summary

NCT01751724 is a randomized, parallel, quadruple-masked trial sponsored by the University of Miami. It compared blinded caffeine citrate with blinded placebo, described as an equivalent volume of normal saline, in mechanically ventilated preterm infants. The registry lists enrollment of 87, a start of 2012-12, primary completion of 2016-01, and status as terminated.

The primary outcome was age at first successful extubation, measured in days from birth until 36 weeks postmenstrual age. The posted arm-specific medians were 24.0 days for caffeine and 20.0 days for placebo. A sponsor-reported primary comparison is not listed. Posted secondary comparisons cover survival, BPD, and survival without BPD; they are unadjusted, and their risk-difference intervals cross zero.

The record does not report the reason for termination, the original sample-size justification, or key details of the statistical analysis plan, missing-data handling, and analysis populations. The available results are descriptive and do not establish a precise comparative effect.