A Randomized Controlled Trial Investigating if Antibiotic Use in the First 48 Hours of Life Adversely Impacts the Preterm Infant Microbiome: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Randomized Controlled Trial Investigating if Antibiotic Use in the First 48 Hours of Life Adversely Impacts the Preterm Infant Microbiome (NCT02477423), a randomized na study. The primary outcome is Richness of the Preterm Infant Microbiome.

2. Clinical Question

The trial asks whether giving ampicillin and gentamicin during the first 48 hours of life is associated with a difference in preterm infants’ microbiome richness at 2 weeks, compared with giving placebo. The registry describes the infants in both arms as low risk and says treatment was randomized and blinded. The primary measure counts 16S rRNA gene amplicon sequence variants, used as a proxy for prokaryote species-like groupings. A higher value represents more detected groupings in a sample.

The record also identifies chronic lung disease of infancy, necrotizing enterocolitis, retinopathy of prematurity, intraventricular hemorrhage, and death as secondary outcomes. Those outcomes do not replace the primary microbiome question, and their results are not included in the posted outcome values.

3. Trial Design

This is a randomized, triple-masked, parallel-group trial. The registry lists the study as completed and records enrollment of 27. It lists two randomized and blinded arms: infants assigned to receive antibiotics and infants assigned to receive placebo. The treatment descriptions specify ampicillin and gentamicin, or saline placebo, during the first 48 hours of life.

The primary outcome is reported for 11 participants in the antibiotics arm and 11 in the placebo arm. These are the group sizes attached to the posted primary-outcome means; they should not be mistaken for the registry’s total enrollment. How the primary-outcome analysis population was selected from the enrolled participants, and why its group sizes differ from the overall enrollment, are not reported in the registry record.

The registry lists the study start as 2015-07 and primary completion as 2019-06. Its phase is listed as NA. Stool collection is described as taking place throughout hospitalization and at 18 month in the arm descriptions. The precise timing, completeness, and use of all collected samples in the primary analysis are not reported in the registry record.

The posted primary comparison is between arm means. A formal statistical analysis plan, the model used for the sponsor’s analysis, any adjustment variables, and the original design assumptions are not reported in the registry record. The posted result therefore supports a description of the recorded group summaries and comparison, but not a reconstruction of every design or analysis decision.

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 Richness of the Preterm Infant Microbiome, measured over 2 weeks.

The primary endpoint is richness of the preterm infant microbiome at 2 weeks. It is defined as the number of 16S rRNA gene amplicon sequence variants detected in each sample. The registry describes these variants as a proxy for prokaryote species-like groupings; higher richness means more species of archaea and bacteria were detected in a sample. The result is reported as a mean with standard deviation for each arm.

Richness is one way to describe the variety of detected microbial groupings in a sample. It does not, by itself, describe which organisms were present, how abundant each was, or how the measured microbiome relates to clinical outcomes. Those are distinct questions. The registry’s brief summary proposes that early antibiotics might alter microbiome development and potentially affect clinical outcomes, but the posted primary endpoint directly measures richness, not those broader outcomes.

The endpoint is useful for the trial’s stated microbiome question because it gives a quantitative summary that can be compared between randomized groups. Interpretation depends on the sample and measurement process, including which samples were available and how sequence variants were identified. Details of those procedures and of any handling of repeated samples are not reported in the registry record. The endpoint should therefore be read as the specific richness measure defined by the registry, rather than as a complete characterization of the microbiome.

7. Results Posted in the Registry

• Randomized & Blinded - Receiving Antibiotics: mean 12.80 versus 11.00 on control; difference 1.80 (95% CI -3.34 to 6.94), standardized effect d = 0.29; z-test p = 0.493.

8. Statistical Interpretation

9. Statistical Methods Explained

The result uses a continuous outcome: richness is a numeric count of detected sequence variants, and the primary summary is a mean. The comparison is a mean difference between the antibiotics and placebo groups. The registry reports each group’s mean and standard deviation, together with a confidence interval and an unadjusted p-value for the contrast. The registry does not report the full sponsor analysis method, so the exact model and any additional assumptions used in that analysis are not known.

What does a group mean describe?

A mean is the arithmetic average of the observed outcome values in a group. It provides a single summary of the group’s central level, but it does not show the full distribution of participant values. Here, the reported means are 12.8 for antibiotics and 11.0 for placebo. The reported standard deviations, 5.3 and 6.9 respectively, describe how dispersed the measurements were around those group means. A standard deviation is not the same as a confidence interval: it describes variation among observations, whereas a confidence interval expresses uncertainty about an estimated quantity.

What does the mean difference represent?

The reported mean difference compares the antibiotics arm with the placebo arm and is 1.8. The direction follows that comparison: a positive value corresponds to a higher observed mean in the antibiotics arm. The estimate is a contrast between group summaries, not a participant-level change from before to after treatment. The record does not report a baseline richness measure or a change-from-baseline analysis.

Random assignment can support a causal comparison when the design and analysis are appropriately carried out, but interpretation still depends on the analysis population, outcome availability, and statistical methods. The registry does not report enough detail to assess those elements fully. In particular, it does not report how missing outcome measurements were handled or whether the analysis followed an intention-to-treat approach.

What does the confidence interval add?

A confidence interval gives a range of values compatible with the observed estimate under the method used to calculate it. The reported 95% confidence interval extends from -3.342 to 6.942. Its inclusion of both negative and positive values means the reported data do not precisely pin down the direction of the contrast. The interval is not a range containing 95% of individual infants’ outcomes; it concerns uncertainty in the estimated mean difference.

How should the p-value be read?

The unadjusted p-value is 0.493. In general, a p-value measures how unusual data at least as incompatible with a specified null model would be if that model were true and the test assumptions held. It does not state the probability that a treatment effect exists, nor does it measure the clinical importance of the estimate. The registry record does not report a prespecified significance threshold or the full testing procedure. The p-value should therefore be read alongside the estimate and interval, not as a stand-alone conclusion.

What role does the standardized contrast play?

A standardized effect expresses a group contrast relative to a measure of variation, which can aid comparison across measures recorded on different scales. The registry includes a standardized contrast and a pooled standard deviation, but interpretation of that standardized value depends on conventions and assumptions. It is not necessary to use the standardized statistic to understand the primary result: the mean difference and its confidence interval retain the endpoint’s original units, 16S rRNA gene amplicon sequence variants.

General statistical point: when measurements vary substantially within groups, the group means alone can give an incomplete impression of how consistently the outcome differs between groups. The standard deviations and confidence interval help convey that uncertainty and variability. They do not remedy missing information about sample selection, sequencing procedures, or the analysis plan.

10. Confidence Intervals and Precision

The reported mean difference is 1.8, with a 95% confidence interval from -3.342 to 6.942. The interval is the main indication of precision for the estimated contrast. Its width shows that the observed mean difference is not estimated narrowly. The interval includes zero, the value corresponding to no mean difference, and extends to values on both sides of zero. Consequently, the point estimate does not settle the direction or magnitude of the underlying contrast.

A confidence interval should be interpreted as uncertainty about the estimated group difference under the calculation method, not as a range of individual outcomes and not as a probability statement about the true value. In repeated use of a method with its assumptions met, confidence intervals have a long-run coverage interpretation. The specific method used to produce this registry interval is not reported in the registry record.

The standard deviations of 5.3 and 6.9 indicate variation among observed measurements in the antibiotics and placebo groups, respectively. The outcome groups each include 11 participants. With those group sizes and the observed variability, the interval conveys considerable uncertainty about the mean contrast. It does not show that the arms are equivalent, and it does not establish that any possible difference is important or unimportant.

Precision is distinct from bias. A narrow interval would not by itself rule out problems such as incomplete outcome measurement or analysis choices; a wide interval does not prove that the estimate is biased. Information about missing samples, analysis populations, and the statistical model would be needed to assess those issues more fully, and that information is not reported in the registry record.

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 events. They are illustrations based on the observed effect, not the trial’s original sample-size calculation. The planning assumptions 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 original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the number of participants per arm that the planning calculation estimates would be needed to have an 80% chance of detecting the specified effect under those assumptions, if that effect were the true one. The 90% power column gives the corresponding per-arm number for a 90% chance. These are planning probabilities under a model, not guarantees for a future study. This is not observed power.

For planning at 100% of the observed effect, the figures are 184 participants per arm for 80% power and 246 for 90% power. For planning at 75% of the observed effect, they are 327 and 437 participants per arm. For planning at 50% of the observed effect, they are 734 and 982 participants per arm. The smaller the effect targeted for detection, the more participants the illustration requires. The posted outcome groups contain 11 participants per arm; the planning figures should not be confused with the observed group sizes or the study’s enrollment.

The illustration is conditional on its stated assumptions and on the observed effect being a useful basis for planning. It is not evidence that the trial had a particular chance of finding an effect, and it does not establish what effect size a future study should target.

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

Worked calculation for this trial

1. Observed difference Δ = 1.800; pooled standard deviation σ = 6.152.
2. n = 2 × (1.96 + 0.842)² × 6.152² / 1.800² = 183.4, rounded up to 184 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 <- 1.8                      # observed difference in means
sigma <- 6.15223536610881                     # 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(2477423)
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 = 184 (80% power), 246 (90% power)
 75% of observed effect: n per arm = 327 (80% power), 437 (90% power)
 50% of observed effect: n per arm = 734 (80% power), 982 (90% power)
power.t.test: 184.3 
Simulated power at n = 184 per arm: 0.799 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry identifies the primary outcome as microbiome richness at 2 weeks and lists five secondary outcomes: chronic lung disease of infancy, necrotizing enterocolitis, retinopathy of prematurity, intraventricular hemorrhage, and death. It records seven outcome measures overall. The posted information does not describe how the outcomes were grouped into those measures, the order in which they were tested, or whether any multiplicity adjustment was planned or applied. A multiplicity strategy is not reported in the registry record.

Multiplicity matters because testing several outcomes can increase the chance of at least one apparently unusual result under a collection of null hypotheses. That general concern does not mean that an unadjusted result is automatically invalid; it means that interpretation depends on which comparisons were planned and how the investigators handled them. The posted primary comparison is labelled unadjusted. No adjusted primary result or secondary-outcome results are reported in the registry record.

Information about missing data is also limited. The arm descriptions state that stool samples were to be collected throughout hospitalization and at 18 month, but the registry does not report the number of samples collected at each time, reasons for absent samples, or procedures for handling missing measurements. It does not report whether repeated samples were combined, whether only a particular sample was selected, or how the 2-week endpoint was derived when data were unavailable.

The outcome comparison includes 11 participants in each arm, while total enrollment is listed as 27. The registry does not explain the relationship between enrollment and the outcome analysis groups. It does not report the analysis population definition, such as whether all randomized participants were included according to assigned group, or whether exclusions occurred. These omissions limit assessment of selection and missing-data bias. They also prevent a complete appraisal of whether the posted contrast follows the randomized assignment for all enrolled participants.

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.

The registry reports serious adverse events for groups at risk of 12 in the antibiotics arm and 15 in the placebo arm. The number affected by a serious adverse event is 0 in each group. The reported number of other adverse events is also 0 in each group, and deaths are reported as 0 in both groups.

These are event counts for the listed groups at risk, not a measure of microbiome richness and not the primary endpoint. A count of zero means that no events of the recorded type are listed for the group in the registry results. It does not, on its own, establish that an event could not occur, or explain how adverse events were defined, monitored, or reported. Definitions, collection procedures, and the relationship between the safety groups and the primary-outcome groups are not reported in the registry record.

The safety counts should be read as the registry’s posted record, rather than as a comprehensive account of every possible harm. They do not answer the primary microbiome question and should not be used to infer a broader safety profile.

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

The posted result describes a comparison of mean microbiome richness at 2 weeks in infants randomized to receive antibiotics or placebo during the first 48 hours of life. The observed mean was 12.8 in the antibiotics group and 11.0 in the placebo group. The reported mean difference was 1.8, with a 95% confidence interval from -3.342 to 6.942 and an unadjusted p-value of 0.493. The interval indicates that the estimate is imprecise and does not determine the direction of the underlying difference.

The result does not show that the groups have identical richness. An interval that includes no difference is not an equivalence result; equivalence requires a prespecified margin and an analysis designed to assess it, neither of which is reported in the registry record. Nor does the result establish that antibiotics change richness. The observed positive mean difference is a descriptive contrast, and uncertainty around it is substantial.

Richness is a particular summary of detected microbial groupings. It does not characterize every aspect of the microbiome, and the primary endpoint does not directly establish effects on the five listed clinical outcomes. Results for those secondary outcomes are not reported in the registry record. The registry’s purpose statement describes a broader interest in microbiome development and clinical outcomes, but that broader question cannot be answered from the primary richness comparison alone.

The study concerns low-risk preterm infants and antibiotic or placebo assignment during the first 48 hours of life. The posted result should not be generalized automatically to other populations, treatment settings, durations, or outcomes. The registry does not provide enough information about outcome availability, the analysis population, or the full statistical methods to resolve those limitations.

16. Why This Trial Matters Statistically

This trial raises a statistical question in which the endpoint is a biological measure with variation between infants and uncertainty in how samples represent the microbiome. A group mean offers a compact comparison, while the standard deviations and confidence interval show why the mean difference alone is insufficient. The posted mean difference of 1.8 is accompanied by a 95% confidence interval from -3.342 to 6.942, a range that includes both directions of contrast.

The result also illustrates the distinction between an observed estimate and the information available to estimate it precisely. The outcome comparison has 11 participants in each arm, and the registry reports group standard deviations of 5.3 and 6.9. The sample-size illustrations indicate that detecting smaller fractions of the observed effect would require more participants per arm under their stated planning assumptions. Those calculations are not a reconstruction of the trial’s original design.

Another important issue is the gap between a focused primary endpoint and a broader scientific motivation. Richness captures the number of detected sequence variants, but does not itself describe all microbiome features or establish effects on clinical outcomes. The registry lists five secondary outcomes, yet their results and the approach to multiple testing are not reported. This makes it important to distinguish the primary estimate from unreported outcomes and analyses.

Finally, a randomized design is only fully interpretable when the analysis population, outcome completeness, and analysis procedures are clear. The registry does not report those details. That does not negate the posted summaries, but it limits how thoroughly readers can assess their robustness and scope.

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 outcome comparison includes 11 participants per arm, while the registry lists total enrollment of 27. The registry does not explain the difference between those counts or describe the analysis population. As a result, the extent to which the posted outcome includes all randomized participants cannot be assessed from the registry record.

The record describes planned stool collection throughout hospitalization and at 18 month, but does not report the number or timing of samples available for the 2-week primary endpoint. Missing-data handling, sample selection, and any analysis of repeated measurements are not reported. These details matter for understanding how the endpoint was obtained and whether observed measurements represent the randomized groups comparably.

The outcome is a richness measure based on sequence variants and serves as a proxy for species-like groupings. The registry does not report sequencing and processing details needed to evaluate measurement choices. Richness alone also cannot represent every feature of the microbiome or establish clinical consequences.

The sponsor’s full analysis method, original sample-size rationale, multiplicity plan, and results for the listed secondary outcomes are not reported in the registry record. The posted confidence interval is wide, so the primary contrast is imprecise. These limitations constrain assessment of the result’s precision, robustness, and broader applicability without changing the descriptive values that were posted.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: A Randomized Controlled Trial Investigating if Antibiotic Use in the First 48 Hours of Life Adversely Impacts the Preterm Infant Microbiome — NCT02477423.
• Registry record last updated 2020-10-08; status last verified 2020-09.
• Results first posted 2020-09-03.
• 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.

21. Record Summary

The registry record identifies a completed, randomized, triple-masked, parallel-group trial with enrollment of 27. It compares ampicillin and gentamicin during the first 48 hours of life with saline placebo in low-risk preterm infants. The primary endpoint is mean microbiome richness at 2 weeks, measured as 16S rRNA gene amplicon sequence variants.

The posted primary outcome includes 11 participants in each arm. Mean richness is 12.8 with antibiotics and 11.0 with placebo; the reported mean difference is 1.8, with a 95% confidence interval from -3.342 to 6.942 and an unadjusted p-value of 0.493. The record lists five secondary outcomes, but their results are not reported. Details of the original design assumptions, analysis population, missing-data methods, multiplicity plan, and full analysis method are not reported in the registry record.