Glycemic Control Using Insulin Levemir Versus Insulin NPH for Diabetes in Pregnancy: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Glycemic Control Using Insulin Levemir Versus Insulin NPH for Diabetes in Pregnancy (NCT01837680), a randomized na study. The primary outcome is Glycemic Control.

2. Clinical Question

The clinical question is comparative rather than a test of either insulin against no treatment. It concerns pregnant patients with two specified forms of diabetes and asks whether the average glucose values recorded over pregnancy differed between those assigned to Levemir and those assigned to NPH. The registry’s brief summary describes the aim as comparing glycemic control and states a hypothesis of no difference between the treatment modalities.

The primary outcome condenses measurements across visits into an overall mean. That makes the comparison about average glucose over the recorded visits, not about a single reading or the proportion of readings meeting a target. The registry also lists secondary outcomes concerning glycemic control, glucose measures, weight, delivery, neonatal weight, and maternal hypoglycemia. Their presence broadens the topics of interest, but the primary question remains the comparison of overall mean glucose.

The record does not report a more detailed eligibility definition, glucose target, or a separate clinical threshold for a meaningful difference. The statistical comparison therefore describes the difference in the posted mean values and its uncertainty; it does not, by itself, establish what size of difference would matter clinically.

3. Trial Design

The registry describes a randomized, prospective, parallel-group study comparing insulin detemir (Levemir) with insulin NPH in pregnant women with gestational diabetes or preexisting type 2 diabetes. The study is marked completed. The listed enrollment is 105. The record gives a start date of 2013-03 and a primary completion date of 2015-03, and names St. Luke's-Roosevelt Hospital Center as sponsor. Phase is listed as not applicable.

Allocation was randomized, masking was none, and the model was parallel. In a parallel design, participants are assigned to a group and outcomes are compared between groups; the registry does not describe a crossover. The lack of masking means the study was not masked according to the registry. The record does not report further details about the randomization sequence, concealment, or who knew each participant’s assignment.

The arm descriptions give the same approach to initial insulin dosing in both groups: a weight-based daily total depending on trimester, with 60% assigned to the morning and 40% to the evening. The record identifies an insulin drug intervention, while the arm descriptions specify the two insulin regimens. It does not report additional details about dose adjustment, adherence, co-interventions, or subsequent treatment changes, so those aspects cannot be characterized from the registry record.

For the primary outcome, the posted summaries include 42 participants in the NPH arm and 45 in the Levemir arm. These are the outcome-specific group counts, not necessarily the number enrolled or the number included in every other analysis. The adverse-event summaries use different at-risk counts, and the registry does not explain the differences in denominators. It also does not report the trial’s original sample-size rationale, protocol deviations, analysis populations, or a detailed statistical analysis 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 Glycemic Control, measured over up to 41 weeks.

The primary endpoint is titled “Glycemic Control.” It is the overall mean glucose value during pregnancy, measured in mg/dL, with a timeframe of up to 41 weeks. The registry defines the value as the sum of the average glucose value at each visit divided by the number of visits. Thus, the measure represents an average across visit-level averages rather than a count of participants who achieved a particular threshold.

A continuous average can be useful for comparing the overall level of a measured quantity between groups. It preserves information about the magnitude of glucose readings, while a binary endpoint would reduce each participant’s experience to whether a specified condition was met. The registry also lists a secondary outcome for the number of patients obtaining glycemic control, but it does not give the target definition in the outcome description. That outcome answers a different question from the primary mean.

For interpretation, the average should be considered alongside its spread and the uncertainty in the between-group difference. An average alone does not describe how much participants’ values varied, whether values changed over time, or whether the distribution included unusually high or low readings. The registry reports standard deviations for the group means, but it does not report the visit-level data or explain how missing visits were handled. Consequently, the posted endpoint summarizes recorded glucose averages without fully describing their longitudinal pattern or the impact of incomplete measurements.

7. Results Posted in the Registry

• Levemir: mean 107.40 versus 109.50 on control; difference -2.10 (95% CI -5.77 to 1.57), standardized effect d = -0.24; z-test p = 0.262.

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.

Number of Patients Obtaining Glycemic Control

• 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:
  • Levemir: 77.8% (35/45) versus 66.7% (28/42) on control; risk difference 11.1 percentage points (95% CI -7.6 to 29.8), risk ratio 1.17 (0.90–1.52); unadjusted two-proportion z-test p = 0.247.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Levemir and the control. The interval gives the range of effects that remain compatible with the data.

Average Fasting Glucose

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Levemir: mean 97.30 versus 100.70 on control; difference -3.40 (95% CI -7.14 to 0.34), standardized effect d = -0.39; z-test p = 0.075.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Levemir and the control. The interval gives the range of effects that remain compatible with the data.

Post-prandial Blood Glucose

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Levemir: mean 112.90 versus 115.20 on control; difference -2.30 (95% CI -6.33 to 1.73), standardized effect d = -0.24; z-test p = 0.264.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Levemir and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broaden the assessment beyond the primary average glucose measure. They include whether participants obtained glycemic control, the time to reach it, average fasting glucose, post-prandial blood glucose, weight gain, neonatal weight, gestational age at delivery, and maternal hypoglycemia. The registry lists 15 posted secondary outcomes, although only some have numerical results in the record summarized for interpretation. The listed outcomes indicate a range of maternal, glucose, and neonatal topics, but the existence of an outcome listing does not establish that every result was reported.

Across the posted numerical glucose outcomes, the direction of the mean estimates is consistent with lower values in the Levemir group, while the count obtaining glycemic control is higher in that group. The associated unadjusted intervals include the no-difference value. This general pattern is compatible with the primary result’s uncertainty: the data do not isolate a clear between-group difference, and the secondary summaries do not provide a separate, conclusive answer.

These outcomes are related. Average glucose, fasting glucose, post-prandial glucose, and achieving glycemic control all describe aspects of glycemia, so they are not independent tests of an unrelated claim. They also use the same randomized groups. The secondary findings therefore should not be counted as multiple confirmations of the primary estimate.

Unadjusted secondary results carry less weight because the registry record does not report a prespecified hierarchy or a method for controlling multiplicity. When many outcomes are tested, the opportunity for chance findings increases. The record also lists no reported analyses, and the posted secondary comparisons are summary-based recalculations rather than documented sponsor analyses. Their appropriate role is to provide context and identify questions for further evaluation, not to override the primary outcome or to establish treatment effects on their own.

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

• Number of Patients Obtaining Glycemic Control
• Time to Achieve Glycemic Control
• Average Fasting Glucose
• Post-prandial Blood Glucose
• Weight Gain
• Neonatal Weight
• Gestational Age at Delivery
• Maternal Hypoglycemia

9. Statistical Interpretation

10. Statistical Methods Explained

The comparisons can be read as estimates with uncertainty, rather than as a contest between a “significant” and a “not significant” result. The posted group means describe the observed averages. The mean difference places the two groups on a common scale: for the primary comparison, it is reported as Levemir minus NPH. A negative difference means the posted Levemir mean is lower; a positive difference would mean it is higher. The estimate is a summary of the observed groups, not necessarily the exact difference that would occur in a broader population.

What does the confidence interval express?

A confidence interval gives a range of values compatible with the observed data under the calculation’s statistical assumptions. It communicates precision: narrower intervals indicate more precision, while wider intervals leave greater uncertainty about the size and direction of the difference. It is not a range containing every individual participant’s glucose value. Nor is it a direct probability statement that a particular fixed value lies within the interval. For these registry summaries, the primary interval extends from -5.767 to 1.567 mg/dL and includes zero, so the summary comparison does not distinguish the observed difference from no difference at conventional levels by this interval criterion.

The registry reports standard deviations for the continuous outcomes. A standard deviation describes the spread of values around a group mean. It is not the standard error of the mean and should not be read as the uncertainty interval around the group mean. The standard error reflects the precision of an estimated mean and depends on the spread and the amount of information available. The difference between group means has its own uncertainty, which is represented by the interval for that difference.

What does an unadjusted p-value tell us?

A p-value measures the compatibility of the observed data with a specified null model, conditional on the statistical assumptions used in the calculation. A small p-value can indicate that data like these would be relatively unusual under that model. It does not give the probability that the null hypothesis is true, the probability that the observed effect is due to chance, or the clinical importance of a difference. A larger p-value does not establish equivalence or prove that the groups are the same. Interpretation should therefore consider the estimated difference and its interval, not the p-value alone.

The posted p-values and comparisons are simple unadjusted recalculations from summary statistics. The registry record lists no reported analyses. As a result, these calculations should not be mistaken for a sponsor-reported model or for an analysis using participant-level data. A participant-level analysis could incorporate features that cannot be reconstructed from group summaries, but the registry does not report whether such an analysis was conducted or what methods it might have used.

Why might over-dispersion matter?

Over-dispersion occurs when the variation in observed data is greater than a statistical model assumes. If a model underestimates variability, standard errors may be too small and intervals too narrow, making apparent precision greater than warranted. If variability is adequately represented, the uncertainty estimates are more credible. This is a general statistical concern, not a finding established for this trial. The registry provides standard deviations for the continuous outcomes, but does not provide participant-level distributions or a model assessment that would establish whether over-dispersion was present or how it was handled.

Relatedly, this primary outcome is built from glucose averages at visits. Repeated observations from the same participant are generally correlated rather than independent. A simple mean of visit averages defines the endpoint, but additional analytic choices may be needed to account for the way observations cluster within participants, especially if visit counts differ or measurements are missing. The registry’s endpoint description does not specify those choices. No particular repeated-measures model should be inferred from the reported summary alone.

How should the secondary comparisons be read?

The secondary outcomes include a participant count for obtaining glycemic control and continuous measures of fasting and post-prandial glucose, as well as outcomes concerning time to control, weight, delivery, neonatal weight, and maternal hypoglycemia. For the posted binary outcome, a risk difference is the difference between the two groups’ proportions with the outcome, expressed on an absolute scale. For a continuous outcome, the mean difference compares group averages in the outcome’s units. Each estimate has an interval representing uncertainty under the calculation used.

The registry reports no secondary analyses. The displayed secondary comparisons are therefore descriptive, unadjusted recalculations based on posted summaries. Multiple outcomes create multiple opportunities for apparently notable results, even when there is no underlying difference. An adjustment strategy or a prespecified hierarchy can help address that issue, but neither is reported in the registry record. Results from separate secondary outcomes should not be treated as independent replications when they arise from the same trial participants and related measures.

These methods explain the posted summaries, not the complete statistical conduct of the trial. The registry record does not report the original analysis plan, handling of missing data, covariate adjustment, a formal equivalence or noninferiority margin, or the sample-size rationale. The available summary measures support a cautious description of estimates and precision; they do not resolve those unreported design and analysis details.

11. Confidence Intervals and Precision

The primary mean difference is -2.1 mg/dL, and its 95% confidence interval runs from -5.767 to 1.567 mg/dL. Because the interval includes zero, the posted summary does not establish a clear separation between the groups. Its endpoints also show that the data are compatible with a difference favoring either group, rather than pointing only in the direction of the observed estimate. The interval is the key guide to the range of differences consistent with these data under the calculation used.

Precision is not the same as whether a result is statistically significant. A confidence interval can be read in terms of which effects remain compatible with the data; whether those effects matter clinically requires a meaningful-difference threshold. No such threshold is reported in the registry. Consequently, the interval cannot be classified as ruling out a clinically important difference without an externally justified threshold, and none should be assumed.

The primary outcome group means are based on 42 NPH participants and 45 Levemir participants. The registry does not explain whether all randomized participants contributed outcome data, how missing visit measurements were treated, or whether the same participants contributed throughout pregnancy. These details can affect the precision and interpretation of an average constructed from repeated visits. The reported group standard deviations offer information about spread, but do not settle those questions.

The intervals for the secondary comparisons also express uncertainty, and the intervals for the posted glucose means and the binary glycemic-control outcome include their no-difference values. These secondary estimates are unadjusted and come from multiple outcomes, so they should be considered exploratory rather than as a set of independent, confirmatory tests. The registry’s lack of reported analyses further limits the ability to assess whether precision was estimated using a prespecified participant-level model.

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 planning figures are illustrations based on the observed effects, not the trial’s original design assumptions. The stated 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. For the outcomes in this trial, the figures are participants per arm, not total event counts. The event-count convention for time-to-event outcomes is not relevant to these posted outcomes.

The 80% power column represents the planned number of participants per arm for a hypothetical study to have an 80% chance of detecting the specified effect under the listed assumptions, if that effect and those assumptions were true. The 90% power column uses a higher target chance of detection, so it requires more participants per arm for the same assumed effect. These are planning concepts, not guarantees about a future study’s outcome.

The rows use 100%, 75%, and 50% of the observed effect. The observed effect is the estimate from this trial’s posted summary, and a fraction of it is a hypothetical smaller effect to detect. Holding the other planning assumptions fixed, smaller effects require more participants because they are harder to distinguish from random variation. The displays should therefore be read as sensitivity illustrations: the required size changes when the effect of interest changes.

These figures do not describe the actual enrollment target or explain how the trial was designed. The original sample-size justification and design assumptions are not reported in the registry record. They also do not show the power achieved by the completed trial. This is not observed power; it is a prospective planning illustration calculated from the observed effect and stated assumptions.

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 = -2.100; pooled standard deviation σ = 8.621.
2. n = 2 × (1.96 + 0.842)² × 8.621² / -2.100² = 264.6, rounded up to 265 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 <- -2.09999999999999                      # observed difference in means
sigma <- 8.62147796801344                     # 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(1837680)
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 = 265 (80% power), 355 (90% power)
 75% of observed effect: n per arm = 471 (80% power), 630 (90% power)
 50% of observed effect: n per arm = 1,059 (80% power), 1,417 (90% power)
power.t.test: 265.5 
Simulated power at n = 265 per arm: 0.801 (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 lists 16 outcome measures and posts 15 secondary outcomes. Evaluating many outcomes raises a multiplicity concern: even when all underlying effects are absent, testing many comparisons can increase the chance that at least one appears unusual by chance. This is why secondary p-values are usually interpreted in light of a prespecified analysis plan, outcome hierarchy, or adjustment strategy. The registry record does not report such a plan or state whether any multiplicity adjustment was performed. The posted secondary comparisons are unadjusted recalculations and should be treated accordingly.

The record also does not report how missing data were handled. This matters for an outcome defined from average glucose at each visit, because participants may contribute different numbers of visits or have missing measurements. The endpoint definition describes dividing the sum of visit-level average glucose values by the number of visits, but does not clarify which visits were included for each participant or how incomplete records were treated. The posted means alone cannot reveal whether missingness affected the estimates.

Analysis-population details are also not reported. The registry does not specify whether the primary comparison followed an intention-to-treat approach, included participants according to treatment received, or used another population definition. It does not describe exclusions after randomization, adherence requirements, protocol deviations, or how treatment changes were handled. Those choices can affect the meaning of a randomized comparison and the generalizability of its estimate.

Finally, the registry record lists no reported analyses and gives no detailed statistical analysis plan. It is therefore not possible to determine whether analyses were adjusted for baseline variables, how repeated measurements were handled, whether interim analyses were planned, or whether outcome definitions changed. These are unknowns, not evidence that particular methods were used or omitted during trial conduct.

15. 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 non-serious adverse event terms

The adverse-event summary reports serious adverse events affecting 0 participants in each group. The corresponding at-risk counts are 52 for NPH and 53 for Levemir. The record also reports 0 deaths in each group. These figures describe the categories and counts recorded in the registry; they do not establish that no event of any kind occurred.

For other adverse events, the registry reports 27 participants affected in the NPH group and 45 in the Levemir group, with at-risk counts of 52 and 53, respectively. These are counts of participants recorded as affected, not event rates calculated. The registry does not provide event descriptions, severity breakdowns for these other events, timing, or an analysis comparing groups. The serious-event terms list is empty, so the record does not identify specific serious-event types.

Safety counts should be interpreted alongside the registry’s at-risk denominators and the limited reporting detail. The at-risk counts differ from the outcome-specific counts for the primary glucose comparison, and the record does not explain the denominator differences. These summaries alone do not establish whether the groups differ in safety, and they do not support conclusions about events that were not reported.

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

The posted primary result describes average glucose during pregnancy in the two randomized groups. The Levemir group mean was 107.4 mg/dL and the NPH group mean was 109.5 mg/dL. The recalculated difference was -2.1 mg/dL, with a 95% confidence interval from -5.767 to 1.567. The interval includes zero, so the result remains uncertain about the direction and size of a difference. It does not show that the groups are identical.

Nor does the result establish that one treatment works better or worse. The record does not report a clinical threshold for an important glucose difference, a formal equivalence or noninferiority margin, or a sponsor-reported analysis. A failure to distinguish the groups in this summary is not proof of equivalence. To support equivalence, for example, the study would need an appropriate prespecified margin and an analysis designed to assess it; the registry does not report those elements.

The primary endpoint is an average across visit-level averages. It does not, on its own, indicate how glucose changed throughout pregnancy, how many readings were above or below a target, or how many participants experienced a particular pattern. The separate secondary outcome for obtaining glycemic control is also not fully defined in the registry description. The posted summaries should therefore be understood according to the outcome definitions that are recorded, without attributing unreported clinical meaning to them.

These findings apply to the trial population and reported measurements, not automatically to every pregnant person with diabetes or every insulin regimen. The registry describes gestational diabetes and preexisting type 2 diabetes, randomized parallel groups, and no masking. It does not provide detailed eligibility criteria, adherence data, or the handling of missing measurements. Those omissions limit the conclusions that can be drawn about the populations and circumstances to which the estimates apply.

The responsible interpretation is that the registry’s summary-based primary comparison is imprecise enough to leave differences in either direction compatible with the data. Secondary outcomes offer related context but do not resolve that uncertainty. This is a statistical description of the posted record, not medical advice or a treatment recommendation.

17. Why This Trial Matters Statistically

This trial is statistically informative as an example of why an estimated difference, its uncertainty, and the analysis context should be read together. The primary point estimate alone suggests a lower average glucose in the Levemir group, but its confidence interval crosses zero and spans values in both directions. Focusing only on the direction of the estimate would omit the uncertainty; focusing only on the p-value would omit the range and scale of effects compatible with the data.

The record also illustrates the difference between a trial’s observed outcome summaries and a complete statistical report. Group means and standard deviations allow basic comparisons, but do not reveal how repeated visits were modeled, how missing values were handled, whether baseline adjustment was used, or which analysis population was intended. Those choices matter for interpreting precision and the role of randomization.

Its collection of secondary outcomes raises a further general lesson: multiple related measures can provide useful context, but they do not automatically constitute independent confirmation. Without a reported multiplicity plan or analysis hierarchy, unadjusted secondary findings should be considered less conclusive than a prespecified primary analysis.

Finally, planning illustrations based on observed effects can show how sample-size needs change with the effect a future study aims to detect. They cannot reconstruct this trial’s original design rationale or describe power achieved after completion. Separating those planning calculations from observed results prevents an illustrative calculation from being mistaken for a property of the completed study.

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.

Several limitations are specific to what the registry record makes available. The posted primary comparison is based on group summary statistics, and the record lists no reported analyses. The recalculated unadjusted difference and p-value may not reflect a participant-level analysis, adjustment for baseline measures, or handling of repeated glucose observations. Whether such analyses were conducted is not reported in the registry record.

The outcome aggregates visit-level glucose averages, but the record does not describe the number or timing of visits for each participant, how the averaging was performed when visits were missing, or whether all randomized participants contributed. The primary outcome counts are 42 for NPH and 45 for Levemir, while the adverse-event summaries use at-risk counts of 52 and 53. The registry does not explain these different denominators, so they should not be treated as interchangeable or reconciled by assumption.

The record does not give detailed eligibility criteria, glucose targets, dose-adjustment rules beyond the initial dosing description, adherence, treatment changes, or co-interventions. It also does not report a clinically important difference, equivalence margin, original sample-size justification, analysis-population definition, or a multiplicity strategy. These missing details prevent assessment of several design and interpretation questions.

Secondary outcome reporting is incomplete in the available record: 15 secondary outcomes are posted, but numerical results are not available for every listed outcome in the summaries considered. No secondary analyses are reported. The adverse-event record gives counts but limited event detail and no comparative analysis. Accordingly, conclusions should stay close to the outcomes and summaries actually recorded, without assuming unreported methods or results.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Glycemic Control Using Insulin Levemir Versus Insulin NPH for Diabetes in Pregnancy — NCT01837680.
• Registry record last updated 2017-05-30; status last verified 2017-05.
• Results first posted 2017-05-30.
• 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

NCT01837680 is a completed randomized, prospective, parallel-group comparison of insulin detemir (Levemir) and insulin NPH in pregnant women with gestational diabetes or preexisting type 2 diabetes. The registry lists enrollment of 105 and identifies St. Luke's-Roosevelt Hospital Center as sponsor. Masking is listed as none.

The primary outcome is overall mean glucose in mg/dL up to 41 weeks. The posted group means are 109.5 for NPH and 107.4 for Levemir. The recalculated unadjusted mean difference is -2.1, with a 95% confidence interval from -5.767 to 1.567. The interval includes zero, leaving uncertainty about the direction and size of the difference.

The registry lists 16 outcome measures, posts 15 secondary outcomes, and lists no reported analyses. It does not report the original sample-size rationale, missing-data methods, analysis populations, or multiplicity plan. The posted summaries support a cautious description of estimates and their uncertainty, but not conclusions beyond the outcomes and methods recorded.