Metabolic Phenotyping During Stress Hyperglycemia in Cardiac Surgery Patients: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Metabolic Phenotyping During Stress Hyperglycemia in Cardiac Surgery Patients (NCT03743025), a randomized phase 4 study. The primary outcome is Number of Participants With Blood Glucose Levels (BG) > 140 mg/dl During Post Operative Period.

2. Clinical Question

The registry describes the study as examining exposure to dulaglutide in relation to prevention of stress hyperglycemia and the metabolic inflammatory response during the perioperative period. The primary endpoint operationalizes stress hyperglycemia as any post-CABG hospital-obtained blood glucose level above 140 mg/dL. It is a participant count outcome: a participant is counted if a qualifying measurement occurs during the hospital stay, rather than being classified by an average glucose level alone.

The registry record states that glucose measurements were obtained at bedside and by hospital laboratory. It also describes measurement during continuous insulin infusion in the intensive care unit and after transition to regular wards. The stated primary timeframe is the hospital stay, up to 12 days postoperatively. This comparison therefore concerns the occurrence of at least one threshold-crossing glucose measurement in that period, not the duration or severity of hyperglycemia. The registry record does not report a separate primary analysis of the metabolic inflammatory response.

3. Trial Design

This was a randomized, parallel-group phase 4 study sponsored by Emory University. The registry lists an enrollment of 28 and a status of terminated. It describes no masking, so the recorded design was open-label. Participants were assigned to dulaglutide or saline placebo; the registry does not report the method used to generate or conceal the allocation sequence.

The dulaglutide and placebo injections were each given once, one to three days before surgery. The study population described in the arm information consisted of participants without a history of diabetes. The main comparison was between the dulaglutide group and the placebo group. The registry lists 12.0 participants for the dulaglutide primary outcome and 13.0 for placebo. The primary analysis population is described as participants who completed the study.

The primary outcome covered the hospital stay, up to 12 days postoperatively. Secondary outcomes included glucose levels during the intensive care unit stay and hospital stay, insulin treatment and dose measures, and the duration of insulin treatment. The record also lists outcomes concerning rescue insulin therapy. Numerical results are available for some secondary outcomes, but the registry record does not report analyses for them.

The study began on 2019-03-08, had a primary completion date of 2023-05-02, and is recorded as terminated. The reason for termination, the number randomized to each group before any exclusions, and the timing and rationale for stopping are not reported in the registry record. The record also does not explain how the enrollment figure relates to the numbers included in the primary outcome analysis. Those omissions limit reconstruction of participant flow and of the study as conducted.

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:

Analysis population for the primary outcome, as defined in the registry: This analysis includes participants who completed the 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 Number of Participants With Blood Glucose Levels (BG) > 140 mg/dl During Post Operative Period, measured over During the hospital stay (up to 12 days postoperatively).

The primary endpoint counts participants with at least one hospital-obtained blood glucose measurement above 140 mg/dL after CABG during the postoperative hospital stay, with follow-up stated as up to 12 days. The registry defines stress hyperglycemia using this threshold. A participant who meets the definition is counted once for this endpoint, regardless of how many qualifying measurements were recorded; the endpoint is not a count of all elevated readings.

This kind of threshold-based outcome can summarize whether the event occurred during a clinically relevant observation period. It differs from a mean glucose outcome: a mean can describe the overall level of measurements, while a participant count captures whether any measurement crossed the prespecified threshold. The registry says measurements were collected by bedside glucose meter and hospital laboratory. During continuous insulin infusion in the intensive care unit, glucose was measured every 1-2 hours; after transition to regular wards, measurements were taken before meals and at bedtime.

The primary endpoint analysis included participants who completed the study. The registry record does not report how missing measurements were handled, whether all randomized participants were included in the analysis, or how conflicting bedside and laboratory measurements were resolved. These details matter because the definition relies on observed glucose measurements during follow-up.

7. Results Posted in the Registry

• Dulaglutide: 91.7% (11/12) versus 100.0% (13/13) on control; risk difference -8.3 percentage points (95% CI -24.0 to 7.3), risk ratio 0.92 (0.77–1.09); unadjusted two-proportion z-test p = 0.288.

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.

Mean Blood Glucose Levels During the Intensive Care Unit (ICU) Stay Postoperatively

• 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:
  • Dulaglutide: mean 134.11 versus 129.79 on control; difference 4.32 (95% CI -5.74 to 14.38), standardized effect d = 0.34; z-test p = 0.400.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dulaglutide and the control. The interval gives the range of effects that remain compatible with the data.

Number of Participants Needing CII Treatment in the ICU

• 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:
  • Dulaglutide: 91.7% (11/12) versus 92.3% (12/13) on control; risk difference -0.6 percentage points (95% CI -22.0 to 20.7), risk ratio 0.99 (0.79–1.25); unadjusted two-proportion z-test p = 0.953.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dulaglutide and the control. The interval gives the range of effects that remain compatible with the data.

Mean Blood Glucose Levels During the Hospital Stay

• 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:
  • Dulaglutide: mean 119.69 versus 117.69 on control; difference 2.00 (95% CI -15.38 to 19.38), standardized effect d = 0.09; z-test p = 0.822.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dulaglutide and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes address related but distinct aspects of postoperative glucose management. They include average glucose during the intensive care unit stay, the number needing continuous insulin infusion, and average glucose during the hospital stay. Other listed outcomes concern rescue insulin, insulin amount and duration, and postoperative glucose management. The posted numeric secondary comparisons do not all point in the same direction: the group mean estimates for both glucose periods are numerically higher with dulaglutide, while the participant-count comparison for continuous insulin infusion is close to no difference by its posted risk-difference estimate.

This mixed pattern does not form a consistent secondary signal that clearly reinforces the primary comparison. Nor should each secondary result be treated as an independent confirmation or contradiction of the primary endpoint. Glucose averages, threshold-based event counts, and insulin-use outcomes measure different features of care and may respond differently to measurement schedules and clinical management. Their timeframes also differ: some cover the intensive care unit stay, while others cover the longer hospital stay.

The secondary estimates are based on small groups and have broad intervals. The registry record does not report analyses for these secondary outcomes, and it does not describe a multiplicity adjustment. Results from multiple unadjusted comparisons carry less weight because examining many outcomes creates more opportunities for apparently notable findings to arise by chance. The posted secondary findings are therefore most useful as descriptive context and as indications of what was measured, not as independent proof of an effect. Several listed outcomes have no numerical results in the registry record, so no overall conclusion can be drawn about them.

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

• Mean Blood Glucose Levels During the Intensive Care Unit (ICU) Stay Postoperatively
• Number of Participants Needing CII Treatment in the ICU
• Number of Participants With Stress-hyperglycemia Requiring Rescue Therapy With Subcutaneous Insulin After Discontinuation of CII
• Mean Blood Glucose Levels During the Hospital Stay
• Mean Units Per Hour of Insulin While in the ICU
• Mean Insulin Dose Per Day While in the ICU
• Duration of Continuous Insulin Infusion (CII)
• Days of Subcutaneous (SC) Insulin After Discontinuation of CII

9. Statistical Interpretation

10. Statistical Methods Explained

What does the risk difference represent?

A risk difference compares the event proportions between two groups by subtracting the comparator proportion from the intervention proportion. It is expressed on the probability scale, so zero represents equal observed proportions. Here, the comparison is dulaglutide relative to placebo, and the posted estimate is -0.083. Its negative sign indicates that the observed proportion was lower in the dulaglutide group. The estimate does not itself establish whether the difference is distinguishable from random variation; the confidence interval and p-value address uncertainty in different ways.

How should the confidence interval be read?

A confidence interval gives a range of values that are reasonably compatible with the data under the method used to construct it. In repeated use of a valid interval procedure, the stated confidence level describes how often intervals produced by that procedure would contain the underlying parameter. It is not a probability statement that this particular interval contains the true value. The primary interval runs from -0.240 to 0.073 and includes zero. It therefore does not rule out no difference, and its span also permits effects in either direction. The registry record does not report the method used to calculate this interval.

What does the p-value add?

The unadjusted p-value for the primary comparison is 0.288. In general, a p-value is calculated under a null hypothesis and depends on a statistical model and test. A larger p-value does not prove equality, and a smaller p-value would not by itself establish clinical importance. The registry record does not identify the analysis test, the prespecified significance threshold for this comparison, or whether the reported p-value corresponds to a planned primary analysis. The posted p-value should therefore be interpreted with the estimate and interval, not as a stand-alone conclusion.

Why can these data be imprecise?

Binary outcomes carry information through the number of participants with and without the event in each group. The primary outcome occurred in 11 of 12 dulaglutide participants and 13 of 13 placebo participants. When event counts are this high and group sizes are small, possible underlying event proportions remain uncertain. A few participants can materially affect the observed group proportions. This is one reason the interval is wide relative to the point estimate and extends across zero.

The posted secondary analyses include two continuous mean outcomes and a count of participants needing continuous insulin infusion. For a continuous outcome, a mean difference describes the difference in group averages, while its standard deviation describes the spread of individual measurements. The posted ICU and hospital glucose comparisons include standard deviations and mean-difference intervals. A standardized effect size can express a mean difference relative to a pooled standard deviation, but it does not remove uncertainty or make a small study more informative. For the binary insulin outcome, the risk difference is again an absolute comparison of event proportions.

How should the sample-size illustrations be used?

The registry record includes planning illustrations for selected secondary comparisons. These are based on the observed effects, use a two-sided significance level of 0.05, assume equal allocation between groups, and use a normal approximation. The illustrations consider detecting 100%, 75%, or 50% of the observed effect. For ICU mean glucose, the illustration for 100% of the observed effect gives 137 participants per arm for 80% power and 183 participants per arm for 90% power. For the number needing continuous insulin infusion, the corresponding figures are 28,157 and 37,694 participants per arm. For hospital mean glucose, they are 1,905 and 2,550 participants per arm.

These figures are planning illustrations based on observed estimates, not the original trial's design justification and not observed power. The registry record does not report the trial's own sample-size calculation or its assumptions. The calculations also depend on the estimated effect, so they may be unstable when the underlying study is small. The record does not provide a planning illustration for the primary endpoint. No sample-size calculation should be read as evidence that a treatment effect was or was not detected.

Over-dispersion is a general concern for count outcomes when their variability exceeds what a chosen statistical model expects. In that situation, a model that assumes less variability can understate uncertainty. The posted primary endpoint is a participant-level binary outcome, not a rate of recurrent events, and the registry does not report an over-dispersion model. It would therefore be inappropriate to assume that such a model was used or needed for the posted primary comparison.

11. Confidence Intervals and Precision

The primary risk-difference estimate is -0.083, with a 95% confidence interval from -0.240 to 0.073. The interval includes zero and allows differences in either direction. Its width shows that the posted data do not pinpoint a narrow range of plausible risk differences. The estimate should be understood as uncertain, especially given the small numbers analyzed and the high frequency of the endpoint in both groups.

The posted relative risk for the primary outcome is 0.917, with a confidence interval from 0.773 to 1.087. This interval includes 1, the value corresponding to equal relative risk. The absolute and relative comparisons express the data on different scales, but neither interval identifies a precise effect. The registry record does not state which measure was designated as the principal effect measure in the analysis plan.

For the secondary mean glucose outcomes, the posted confidence intervals for the mean differences also include zero: -5.744 to 14.38 for ICU glucose and -15.38 to 19.38 for hospital glucose. The interval for the secondary risk difference in participants needing continuous insulin infusion runs from -0.220 to 0.207 and includes zero. These intervals indicate substantial uncertainty, and each concerns a distinct outcome with its own scale and measurement period.

Confidence intervals describe statistical precision under the methods used; they do not capture every source of uncertainty. They do not address unreported missing-data handling, possible differences between completers and all randomized participants, or the implications of examining multiple outcomes. Since the registry record does not report interval methods or a detailed analysis plan, the intervals are best treated as the posted estimates of precision rather than as a complete account of uncertainty.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary endpoint and multiple secondary outcome measures. It does not report a plan for controlling multiplicity, such as a hierarchy of hypotheses or adjustment of secondary p-values. Without such a plan, each unadjusted secondary p-value is a separate exploratory comparison. In general, testing many outcomes without adjustment raises the chance that at least one comparison will appear notable by chance, even when no corresponding underlying effect exists. The registry record does not provide enough information to determine how the investigators intended to interpret the outcome family.

The primary endpoint population is described as participants who completed the study. The arm-level primary results include 12.0 participants for dulaglutide and 13.0 for placebo, while enrollment is listed as 28. The registry record does not explain participant flow, the difference between enrollment and analyzed participants, or the reasons any participants may not have contributed to this outcome. It does not report whether the primary analysis followed assignment regardless of treatment received, whether any other analysis population was used, or whether exclusions occurred after randomization.

Missing-data methods are not reported in the registry record. In particular, the record does not say whether missing glucose measurements or incomplete follow-up were treated as absence of the event, excluded, imputed, or handled in another way. It also does not describe sensitivity analyses for missing outcomes. Because the endpoint is defined by any qualifying measurement during follow-up, incomplete observation could affect whether a participant is classified as having the event.

The registry record reports no analyses in its analysis summary and does not describe interim analyses, stopping rules, or a prespecified statistical plan. The study status is terminated, but the reason for termination and whether termination was related to results are not reported in the registry record. These omissions prevent a full assessment of outcome selection, analysis decisions, and the risks of bias associated with incomplete follow-up or selective reporting.

13. 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

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hospital readmission due to nausea and vomitting, acute kidney injury, and dehydration (8.3%).

The adverse-event record lists one serious adverse event among 12 participants at risk in the dulaglutide group and none among 13 participants at risk in the placebo group. The listed serious term is Hospital readmission due to nausea and vomitting, acute kidney injury, and dehydration, recorded at 8.3%. No deaths are listed in either group.

The record also lists other, non-serious adverse events for 4 participants in the dulaglutide group and 2 in the placebo group. These counts should not be combined with the serious-event counts: the categories differ in severity, and the registry record does not explain whether participants could contribute to more than one category or how events were coded. The serious term includes several conditions in one description, so the registry entry does not establish how many distinct clinical events or diagnoses were represented beyond what it reports.

These figures are descriptive. The groups are small, and the registry record does not provide a detailed safety analysis, event timing, relatedness assessment, or a statistical comparison. The counts alone cannot establish whether an event pattern is caused by treatment or how it would generalize to a larger population.

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

The primary data show that qualifying postoperative blood glucose readings were recorded for 11 of 12 analyzed dulaglutide participants and 13 of 13 analyzed placebo participants. The posted risk difference is negative, but its confidence interval includes zero and spans values on either side. The p-value is 0.288. These results do not precisely establish the direction or size of a group difference.

They also do not demonstrate that the groups are equivalent. A nonzero difference may remain compatible with the interval, and a failure to distinguish the estimate from zero is not proof that no effect exists. Conversely, the negative point estimate by itself is not proof that dulaglutide reduced the event. Interpretation must account for uncertainty, the small analysis groups, the endpoint's high observed frequency, and the fact that only study completers are specified as the primary analysis population.

The endpoint captures any qualifying blood glucose level during the hospital stay, not glucose control over time, the number of high readings, or the inflammatory response. Secondary glucose means and insulin-use measures address related but nonidentical questions. Their unadjusted results should not be taken as a substitute for a prespecified primary analysis, particularly when the statistical plan and multiplicity approach are not reported.

The registry record does not report the reason the study was terminated, the handling of missing data, or the relationship between enrollment and the outcome analysis population. Those gaps restrict conclusions about the completeness and representativeness of the analysis. The findings describe the posted data for this study; they do not establish a treatment recommendation or a general result for other populations or settings.

15. Why This Trial Matters Statistically

This trial illustrates why an effect estimate should be read together with its uncertainty. The primary point estimate favors a lower event proportion in the dulaglutide group, but its confidence interval crosses zero and permits effects in either direction. The p-value alone cannot resolve that uncertainty, and the observed event counts are small. The main statistical lesson is not a binary declaration of success or failure, but the limited precision of the available comparison.

The endpoint also highlights how outcome definitions shape interpretation. Counting participants with any measurement above a threshold differs from comparing average glucose or recording insulin use. Those measures can tell different parts of the clinical story, so agreement across them cannot be assumed. Here, the registry provides several kinds of outcomes but no reported multiplicity strategy, which makes careful separation of primary and secondary evidence important.

The record is additionally useful for showing how incomplete reporting constrains statistical assessment. It identifies a completer analysis population and a terminated status, but does not explain participant flow, missing-data handling, stopping rationale, or the original sample-size justification. These omissions make it difficult to reconstruct the analysis fully or judge how representative the analyzed participants are of everyone enrolled or randomized.

Finally, planning illustrations based on observed effects can help convey how sample size and precision relate, but they are not evidence of observed power and are not the trial's design rationale. In a small study, such illustrations depend strongly on uncertain estimates. Keeping that distinction clear prevents a retrospective planning calculation from being mistaken for a prespecified design feature or a conclusion about treatment effect.

16. 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 analysis is restricted in the registry description to participants who completed the study. The record does not report the number randomized by group before exclusions, the reasons for noncompletion, or a participant flow account. As a result, it is not possible to determine how the analyzed population relates to all enrolled or randomized participants. Enrollment is listed as 28, but the registry record does not explain the primary outcome analysis counts.

The study was open-label, and the registry does not report whether outcome assessment or glucose measurement procedures were blinded. Blood glucose was measured by bedside meter and hospital laboratory, but the record does not explain how results from the two sources were reconciled or whether measurement methods were applied consistently across groups. These are relevant considerations for a threshold-based endpoint.

The study is recorded as terminated, but the reason and timing in relation to outcome collection are not reported. The record does not provide the original sample-size justification, a detailed statistical analysis plan, interim-analysis information, or methods for missing data. It also does not report adjustment for multiple secondary outcomes. These gaps limit assessment of whether analyses and outcomes followed prespecified decisions.

Only a limited set of secondary comparisons has numerical summaries in the analysis record, and the registry does not report analyses for them. The small group sizes and wide intervals constrain precision. The primary endpoint reflects whether a qualifying measurement occurred during hospitalization; it does not quantify the frequency of elevated readings or the broader metabolic inflammatory response described in the study aim. The findings therefore provide a narrow and uncertain view of the study question.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Metabolic Phenotyping During Stress Hyperglycemia in Cardiac Surgery Patients — NCT03743025.
• Registry record last updated 2024-05-30; status last verified 2024-05.
• Results first posted 2024-05-30.
• 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.

20. Record Summary

NCT03743025 is a randomized, open-label, parallel-group phase 4 study of a single preoperative dulaglutide injection versus saline placebo in participants without a history of diabetes undergoing cardiac surgery. The registry lists enrollment of 28 and status as terminated. The primary outcome counted participants with a hospital blood glucose level above 140 mg/dL during the postoperative stay, up to 12 days.

The primary analysis included study completers: 11 of 12 participants in the dulaglutide group and 13 of 13 in the placebo group had the event. The posted risk difference was -0.083, with a 95% confidence interval from -0.240 to 0.073 and an unadjusted p-value of 0.288. The interval includes zero and indicates substantial uncertainty about the size and direction of the difference.

The registry record does not report the reason for termination, missing-data procedures, a detailed statistical plan, or the relationship between enrollment and the analysis population. Secondary outcomes address glucose levels and insulin use, but their unadjusted findings have limited evidential weight.