Acute Exercise Effects in Obese Pregnancy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Acute Exercise Effects in Obese Pregnancy (NCT03750695), a randomized na study. The primary outcome is Oral Glucose Insulin Sensitivity (OGIS) Index.

2. Clinical Question

3. Trial Design

This was a randomized, unmasked, crossover study of acute exercise and rest during obese pregnancy. The registry lists the study as completed, with an enrollment of 16. The study began on 2018-11-20, and primary completion was recorded as 2020-03-31. No phase applies to this study.

In a crossover design, participants receive more than one study condition, so comparisons can potentially use information from the same person under different conditions. The registry states that participants completed all interventions but in different order. It describes three sessions: a 40-minute aerobic exercise session, a 40-minute resistance exercise session, and a 40-minute quiet-rest session. The posted primary outcome results give n=15 in each condition, while safety figures list 16 participants at risk in each group. The registry record does not report why those counts differ.

Randomization and crossover describe different aspects of the design: randomization concerns assignment or ordering, while crossover means that participants experience multiple conditions. Masking was listed as none. The registry record does not report how intervention order was generated, whether periods between sessions were specified, or whether period or carryover effects were considered in the analysis. The record also does not report a sample-size justification or the original design assumptions.

The registry summary frames the study as collecting preliminary information about acute exercise effects to inform later research. Its reported primary endpoint is oral glucose insulin sensitivity. One primary endpoint and two comparisons are posted; no secondary outcome results are posted. The record does not report an analysis plan in enough detail to establish which model was intended for the crossover structure.

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: The 3 interventions were: 1) 1-hour of rest, 2) 1-hour of resistance exercise, and 3) 1-hour of aerobic exercise. Each of these interventions were separated by 1-week. N=8 participants were randomized into performing these interventions in the order of rest-aerobic-resistance and n=8 participants in the order of rest-resistance-aerobic.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Participants completed all interventions but in different order

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.

Population: Participants who completed the study. All participants were randomized to receive all interventions.

6. Primary Endpoint

The registry lists the primary outcome as Oral Glucose Insulin Sensitivity (OGIS) Index, measured over 2 hours post-intervention, up to approximately 3 hours.

The primary endpoint was the Oral Glucose Insulin Sensitivity (OGIS) Index, reported as a continuous measure in uU/mL. It was assessed during a 2-hour oral glucose tolerance test at 2 hours post-intervention, up to approximately 3 hours. The registry describes OGIS as an index of insulin sensitivity derived from glucose and insulin measurements during the test.

As a general statistical concept, an index condenses measurements into a summary intended to represent a construct that is not measured directly by a single observation. In this case, the registry describes OGIS as analogous to an insulin-sensitivity index obtained from a glucose clamp. The description says that the calculation models glucose values at 0, 90, and 120 minutes and insulin values at 0 and 90 minutes. The posted endpoint is therefore a calculated index, not a direct measurement of every aspect of glucose metabolism.

The endpoint is relevant to the study’s stated focus on maternal glucose metabolism. The posted results give a mean and standard deviation for each intervention condition. Those summaries describe the outcome values in each condition, while the reported comparisons estimate mean differences against aerobic exercise. The registry record does not report a clinically meaningful threshold for interpreting a difference in OGIS, so the numerical estimates should be read as statistical comparisons rather than as established clinical cutoffs.

7. Results Posted in the Registry

• Acute Resistance Exercise: mean 426.00 versus 405.00 on control; difference 21.00 (95% CI -18.66 to 60.66), standardized effect d = 0.38; z-test p = 0.299.
• Acute Rest: mean 418.90 versus 405.00 on control; difference 13.90 (95% CI -30.58 to 58.38), standardized effect d = 0.22; z-test p = 0.540.

8. Statistical Interpretation

9. Statistical Methods Explained

Several statistical ideas help interpret the posted estimates. They are general explanations, not claims about an unreported analysis plan. The record gives continuous outcome summaries, recalculated unadjusted mean differences, intervals, and unadjusted p-values. It does not report the investigators’ full statistical methods, so conclusions about the intended model must remain limited.

What does a mean summarize?

A mean is the arithmetic average of observed values in a group or condition. Here, each condition’s mean summarizes the OGIS values reported after that session. A mean is useful for describing the central level of a continuous outcome, but it does not show the full distribution. Two conditions can have similar means while individual values vary considerably. The standard deviation describes the spread of values around a mean; it is not the uncertainty interval for the mean difference.

The registry reports standard deviations of 56.9 for aerobic exercise, 53.9 for resistance exercise, and 67.0 for quiet rest. These describe participant-level variability in the respective condition summaries. They do not, by themselves, describe how much each participant’s value changed between sessions. In a crossover study, the relationship between a participant’s measurements across conditions can affect the precision of a within-person comparison.

What does a mean difference represent?

A mean difference expresses the estimated contrast between two condition means in the units of the endpoint. The resistance-exercise comparison is against aerobic exercise, as is the quiet-rest comparison. A positive estimate means the first named comparison condition has a higher posted mean than the aerobic reference in these summaries. The estimates are 21.0 for resistance exercise and 13.9 for quiet rest. These are not standardized effects and do not convey how important a difference would be in practice.

The posted differences are identified as recalculated and are unadjusted. The record includes pooled standard deviations and participant counts used for the summary-statistic recalculations. Such calculations treat the available summaries as the basis for the comparison; they do not use the individual participant data needed to model each person’s repeated outcomes directly. Because this is a crossover study, a model using participant-level data could account for pairing and potentially other design features. Whether the investigators used such a model is not reported in the registry record.

What does a p-value say?

As a general explanation, a p-value measures how unusual an estimate at least as far from the null as the observed estimate would be, assuming the null hypothesis and the calculation’s assumptions. It is not the probability that the null hypothesis is true. It also does not measure the size or practical importance of an effect. The unadjusted p-values are 0.299 for resistance exercise compared with aerobic exercise and 0.54 for rest compared with aerobic exercise. They should be interpreted alongside the estimates and intervals, not as stand-alone verdicts.

“Unadjusted” matters because the record does not report an adjustment for multiple comparisons. The p-values also do not appear to account for a crossover model in the recalculated summaries. A p-value can only be interpreted relative to the assumptions of the method used to obtain it; the registry record does not report enough detail to evaluate all such assumptions.

Why does over-dispersion matter?

As a general concept, over-dispersion occurs when observed variation is greater than a model expects. It can make standard errors too small if the analysis assumes less variability than the data contain, which can make intervals too narrow and p-values too small. The endpoint is continuous, and the registry reports standard deviations rather than counts or event rates. The record does not identify an over-dispersion assessment or a model where over-dispersion was a specified issue, so it would be inappropriate to claim that it occurred. The broader point is that uncertainty calculations depend on how well the analysis represents the observed variability.

What is specific to a crossover analysis?

As a general design principle, crossover analyses can use the within-participant contrast: each participant can serve as their own reference across conditions. This can reduce the influence of stable differences between people, but inference depends on the design and analysis being handled appropriately. Period effects, intervention order, and carryover are possible considerations in crossover designs; whether they were relevant or addressed in this study is not reported in the registry record. The posted arm summaries alone do not reveal paired differences or their variability. Therefore, the recalculated unadjusted estimates are best read as descriptive comparison calculations, not as a complete account of the design-based analysis.

Finally, a confidence interval and a p-value answer related but different questions. The interval describes a range of effect estimates compatible with the data and method, while the p-value summarizes evidence against a specified null under that method. Neither supplies a clinical threshold, and neither makes the observed sample representative of every pregnancy with obesity. The available results are informative about the posted outcome summaries, but their interpretation is constrained by the small analysis counts and the absence of a reported crossover model.

10. Confidence Intervals and Precision

The comparison intervals communicate substantial uncertainty around both recalculated mean differences. For resistance exercise compared with aerobic exercise, the estimate is 21.0 and the interval runs from -18.66 to 60.66. For quiet rest compared with aerobic exercise, the estimate is 13.9 and the interval runs from -30.58 to 58.38. In each case, the interval includes zero, so the posted summaries are compatible with a difference in either direction as well as with little or no difference.

As a general explanation, a confidence interval is a range generated by a statistical method to represent uncertainty in an estimate. Its width reflects both the variability in the data and the amount of information available, along with the assumptions used in the calculation. A wide interval allows a broad range of values; a narrow interval would indicate greater precision under the same method. The registry record does not state the confidence level for these recalculated intervals, so no confidence level should be inferred from the record.

The mean differences are based on 15 participants in each condition, and the crossover design means that paired information could be relevant to an analysis. The posted summaries do not include the within-participant correlation or paired-difference standard deviation. The intervals therefore should not be interpreted as evidence that the crossover structure was fully represented. They describe the precision of the simple recalculations, not necessarily the precision of an investigator-specified crossover model.

The intervals do not define a range of likely individual responses and do not show whether a difference would be important in practice. The registry record does not report a threshold for meaningful OGIS change. Taken together with the unadjusted p-values, the intervals show that the available summary data estimate these comparisons imprecisely.

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 sample-size figures are planning illustrations, not counts of participants observed in the trial. For these continuous-outcome comparisons, figures are participants per arm, not total events. The planning values use two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. The effect targeted is a stated proportion of the observed effect. The trial’s own original design assumptions and sample-size justification are not reported in the registry record.

The 80% power column means the planned number of participants per arm under those assumptions for a design intended to have an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column gives the corresponding planning number for a design intended to have a 90% chance under the same assumptions. Higher planned power generally requires more participants when other assumptions are held fixed. Smaller target effects require more participants because they are harder to distinguish from ordinary sampling variation.

The displayed planning rows consider the full observed effect and smaller proportions of that effect for each comparison. These are conditional illustrations: if the actual effect differs from the observed estimate, or if the analysis model and variability differ from the planning assumptions, the calculated requirements would not describe that alternative design. The figures are not evidence that the trial had a particular power, and they are not observed power. They should be read as planning scenarios based on the stated assumptions, rather than as a reconstruction of the study’s original design.

The curve presents the same general planning relationship across effect sizes: targeting a smaller effect entails a larger participant requirement under otherwise fixed assumptions. It does not turn the small study’s estimates into reliable estimates of what a future study will observe. The observed effect can be uncertain, especially when based on a small analysis sample, so planning based on it is illustrative rather than definitive.

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 = 21.000; pooled standard deviation σ = 55.420.
2. n = 2 × (1.96 + 0.842)² × 55.420² / 21.000² = 109.3, rounded up to 110 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 <- 21                      # observed difference in means
sigma <- 55.4203031388317                     # 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(3750695)
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 = 110 (80% power), 147 (90% power)
 75% of observed effect: n per arm = 195 (80% power), 261 (90% power)
 50% of observed effect: n per arm = 438 (80% power), 586 (90% power)
power.t.test: 110.3 
Simulated power at n = 110 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 record posts one primary endpoint and two comparisons: resistance exercise versus aerobic exercise, and quiet rest versus aerobic exercise. It does not report whether the analyses were prespecified in this form, whether adjustment for multiple comparisons was used, or how multiplicity was handled in an investigator analysis. The displayed p-values are unadjusted. When several comparisons are examined, the chance of at least one apparently unusual result can increase; as a general statistical principle, multiplicity procedures can address that issue, but no particular procedure can be attributed to this study.

Missing-data handling is not reported in the registry record. The primary outcome results list 15 participants for each condition, whereas enrollment was 16 and the safety groups list 16 participants at risk. The registry does not explain the difference, identify missing outcome measurements, or state whether any imputation or other missing-data method was used. It also does not report whether all randomized participants were included in an analysis population or how exclusions, if any, were handled.

The record says participants completed all interventions but in different order. It does not report whether every participant contributed a complete outcome for every condition, or whether the analysis population was defined in another way. That distinction matters in a crossover study because incomplete sets of repeated measurements can affect which comparisons are available and how they should be analyzed.

The analysis model, handling of intervention order or period, and any adjustment for baseline or other covariates are not reported in the registry record. Interim analyses are also not reported. These omissions do not establish that such methods were absent from the full study materials; they mean only that the registry record does not describe them. Consequently, the unadjusted recalculations should not be read as a complete account of missing-data, multiplicity, or crossover-analysis decisions.

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 lists 16 participants at risk in each of the resistance-exercise, aerobic-exercise, and resting-session groups. For each group, it reports zero serious adverse events and zero deaths. It also reports zero other adverse events in each group. No serious adverse-event terms are listed.

These are counts recorded for the study conditions, not estimates of the risk in all pregnant women with obesity or in other settings. Because participants completed all interventions in different order, the group labels refer to study sessions or conditions; the registry record does not explain in detail how adverse events were attributed across a participant’s multiple sessions. The numbers at risk should not be treated as the primary-outcome analysis counts: the OGIS results list 15 participants per condition, while safety figures list 16.

As a general statistical point, zero recorded events in a small study does not establish that an event cannot occur or that a condition has no safety risks. The registry record does not provide event definitions, the adverse-event collection window, or additional detail about ascertainment. These safety counts should therefore be read narrowly as the posted event totals for the study, without extrapolation beyond the recorded participants and conditions.

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

The results describe OGIS means after three acute sessions in pregnant women with obesity. The posted mean was 405.0 uU/mL after aerobic exercise, 426.0 uU/mL after resistance exercise, and 418.9 uU/mL after quiet rest. The recalculated comparisons estimate a mean difference of 21.0 for resistance exercise versus aerobic exercise and 13.9 for quiet rest versus aerobic exercise. Their intervals are broad and include zero, and the p-values are unadjusted.

These findings do not establish that one session produces a meaningful change in insulin sensitivity relative to another. The registry does not report a threshold for a meaningful OGIS difference, and the posted calculations do not account visibly for the repeated measurements within the crossover design. Nor do the results establish equivalence: a failure to demonstrate a clear difference is not proof that conditions are the same. The precision of the estimates is limited by the small outcome counts and the information available in aggregate summaries.

The study concerns acute sessions, not a repeated exercise program. Its posted endpoint is an index measured after an oral glucose tolerance test, not a long-term maternal or offspring health outcome. The registry summary describes potential relevance to metabolic and vascular health, but the posted primary outcome results do not establish effects on gestational diabetes, preeclampsia, offspring outcomes, or long-term health.

The study also does not answer whether the estimates apply to people outside the described population. The record concerns obese pregnancy and does not report outcome results for normal-weight or overweight groups. Its main contribution is preliminary descriptive information about OGIS after the specified sessions, alongside uncertainty that should be considered when interpreting the comparisons or planning further research.

16. Why This Trial Matters Statistically

This study is statistically interesting because its crossover design and its posted summaries invite different kinds of interpretation. A crossover can support comparisons within the same participants, which may be useful when outcomes vary substantially between people. But a suitable analysis generally needs to account for the repeated observations and relevant design features. The registry record says participants completed all conditions in different order, yet the posted aggregate summaries do not provide paired contrasts or enough information to assess the intended model.

The difference between enrollment and outcome counts also matters for transparent interpretation. Enrollment was 16, primary outcome results report 15 per condition, and safety counts list 16 at risk in each condition. The record does not explain this pattern. Reporting such counts alongside an account of missing outcome data and analysis populations helps readers understand precisely which observations contribute to each result.

The intervals are important because they make the uncertainty visible. Both recalculated comparisons include zero and span values in either direction. The mean estimates alone might suggest a directional pattern, but the intervals show why the posted summaries do not settle the comparisons. This is a useful reminder that an estimate’s apparent direction is not the same as a precise or established effect.

The planning illustrations likewise demonstrate how assumptions shape future sample-size requirements. They use fractions of observed effects and stated design assumptions, not an original protocol justification. In a small preliminary study, an observed effect can be uncertain; planning should therefore be interpreted as conditional rather than definitive. Overall, the record illustrates why statistical interpretation depends not only on point estimates, but also on design, analysis details, missingness, and precision.

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 results list 15 participants per condition, while the total enrollment and safety counts are 16. The registry record does not explain this difference or state how many participants contributed complete outcome data across all sessions. That uncertainty limits assessment of the analysis population and missingness.

The study is small, and the comparison intervals are wide. The posted recalculations use unadjusted aggregate summaries. Because participants completed multiple conditions, within-person association could affect the uncertainty of a crossover comparison; the registry does not provide paired outcome summaries, correlations, or the model used to account for repeated measurements. The posted calculations therefore cannot establish how a full crossover analysis would differ.

The record does not report whether intervention order, period effects, carryover, or other design features were incorporated into the analysis. It also does not report an adjustment for the two comparisons, a missing-data strategy, or the investigators’ analysis populations. These are reporting limitations of the registry record; they do not prove that the full study materials contained no such methods.

The endpoint concerns OGIS after an acute session and a 2-hour oral glucose tolerance test. No secondary outcome results are posted. The registry record does not report a clinically meaningful OGIS threshold, so the practical importance of the numerical differences cannot be judged from these statistics alone. The study’s results are specific to the described pregnant population with obesity and do not provide comparative results for other weight groups. The registry summary presents the work as preliminary; the posted results do not address effects of sustained exercise or long-term outcomes.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Acute Exercise Effects in Obese Pregnancy — NCT03750695.
• Registry record last updated 2021-07-16; status last verified 2021-05.
• Results first posted 2021-07-16.
• 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

NCT03750695, titled “Acute Exercise Effects in Obese Pregnancy,” is a completed, randomized, unmasked crossover study sponsored by Duke University. Enrollment was 16. Participants completed acute aerobic exercise, resistance exercise, and quiet rest in different orders, with each session lasting 40 minutes.

The primary endpoint was the Oral Glucose Insulin Sensitivity Index, measured after intervention during a 2-hour oral glucose tolerance test. The registry posts means and standard deviations for 15 participants in each condition. Recalculated, unadjusted comparisons against aerobic exercise have intervals that include zero. The record lists no serious adverse events, other adverse events, or deaths in the three safety groups. It does not report the original sample-size justification, detailed crossover analysis, missing-data handling, or multiplicity approach.