Comparison of Two Basal Insulin Therapies for Patients With Type 1 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparison of Two Basal Insulin Therapies for Patients With Type 1 Diabetes (NCT00487240), a randomized phase 3 study. The primary outcome is Change in Hemoglobin A1c (HbA1c) From Baseline to Endpoint.

2. Clinical Question

The study also listed outcomes concerning HbA1c levels and thresholds, self-monitored blood glucose, glycemic variability, hypoglycemic episodes, and body weight. These additional measures broaden the clinical questions beyond the primary HbA1c change, but their presence in the outcome list does not establish what the results were.

3. Trial Design

The registry describes a phase 3, randomized, parallel-group study with no masking. It enrolled 387 patients and is marked completed. The recorded start was 2007-06, and primary completion was 2008-08. The condition listed is type 1 diabetes. Eli Lilly and Company is named as the sponsor.

The comparison was between insulin lispro protamine suspension and insulin detemir as basal insulin in a basal-bolus treatment setting. The intervention descriptions specify twice-daily use for both basal therapies; the detemir description also specifies subcutaneous administration. Mealtime insulin therapy is part of the study context described in the registry summary.

Randomization supports a planned comparison between treatment groups, but the record does not report the allocation ratio, the method used to generate or conceal the allocation sequence, or any stratification factors. Because masking is recorded as none, the trial was open-label. The registry does not report whether outcome assessors or analysts were masked separately.

The primary outcome is change in HbA1c from baseline to endpoint, with the timeframe stated as baseline and 32 weeks. The listed analysis uses least-squares means and standard errors. The analysis population is described as randomized patients with a baseline value and at least one post-baseline value, with last observation carried forward. The record does not report the number analyzed in either group or the amount of missing outcome data.

The registry summary says a gatekeeper strategy would be used for sequential testing of secondary objectives. The order of those tests, the criteria for passing each step, and the detailed testing procedure are not reported in the registry record. The primary comparison itself is not reduced to a single reported result in the record.

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: First 4 weeks (Titration Period): acclimate patients to insulin regimen and optimize insulin dose. Final 28 weeks (Maintenance Period): Minimum of 6 months' stable insulin dosage. 456 patients were screened; 69 did not meet entry criteria and 387 were randomized; 6 patients were excluded from the 387 randomized to create the Full Analysis Set.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Number of randomized patients with baseline and at least one post-baseline value. Intent to treat population. Last observation carried forward.

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 Change in Hemoglobin A1c (HbA1c) From Baseline to Endpoint, measured over baseline and 32 weeks.

The primary endpoint is change in hemoglobin A1c, or HbA1c, from baseline to the 32-week endpoint. HbA1c is a continuous measure reported as a percent of HbA1c. Change describes how the measure moved over the study period rather than only its value at the endpoint. Comparing changes between randomized groups can address whether the groups differed in their average change over that period.

The registry specifies a least-squares mean as the summary measure and a standard error as its dispersion measure. A least-squares mean is a model-based estimate of an average; its interpretation depends on the analysis model and any adjustments, which are not described in detail in the registry record. A standard error describes the estimated uncertainty of a mean, not the spread of individual patient values. It is therefore not interchangeable with a standard deviation.

As a general statistical point, continuous outcomes retain information across the measurement scale rather than reducing observations to a yes-or-no threshold. Change from baseline can also account for where patients started, although the specific model and baseline adjustment are not reported in the registry record. The outcome is clinically relevant to the stated study question because the registry identifies HbA1c as its primary measure of glycemic change. No group means, treatment contrast, confidence interval, or primary-result p-value is reported in the registry record, so the observed size and precision of the change cannot be assessed from this record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a least-squares mean represent?

The primary outcome is described as a least-squares mean, a model-based estimate of an average. In general, a statistical model can estimate group averages while accounting for specified terms, such as baseline values or design features. Which terms were included affects the meaning of the estimate. The registry record does not report the full model specification, so it is not possible to state which factors were included or how the estimates were adjusted. The least-squares mean should not be treated as a raw arithmetic mean unless the analysis details establish that equivalence.

The record names a standard error as the dispersion measure. In general, a standard error describes the precision of an estimated mean: smaller standard errors indicate less sampling uncertainty under the model's assumptions. It does not describe how varied individual measurements are. That individual-level spread is commonly described with a standard deviation. Neither a standard error nor a standard deviation is itself a treatment effect, and the primary standard errors are not reported in the registry record.

What would a treatment contrast mean?

For a two-group study, a treatment contrast for change in HbA1c would compare the estimated changes between the groups. In general, the contrast gives the direction and size of the estimated difference on the outcome scale. A confidence interval would express a range of values compatible with the data and the analysis assumptions, while a p-value would describe how unusual the data would be under a specified null hypothesis. Those measures answer different questions: an interval emphasizes magnitude and uncertainty, while a p-value addresses compatibility with a null model.

The registry record does not provide the two estimated changes, a between-group contrast, a confidence interval, or a primary p-value. A contrast cannot be inferred from the study title, design, or outcome definition. It is also not appropriate to reconstruct a result from the adverse-event counts or from the list of secondary outcomes.

How should the analysis population be understood?

The primary population is described as randomized patients with a baseline value and at least one post-baseline value, and is labeled intent to treat. The record also says last observation carried forward was used. In general, an intent-to-treat approach seeks to analyze participants according to randomized assignment, regardless of subsequent treatment adherence. The stated requirement for at least one post-baseline value means that the description does not indicate that every randomized patient necessarily contributed to this particular outcome analysis. The number included in each group is not reported in the registry record.

Last observation carried forward replaces a missing later value with the last available measurement. As a general method, this approach can be easy to implement, but it assumes that the carried-forward measurement is an adequate stand-in for the unobserved endpoint. That assumption may not hold if outcomes continue to change after the last measurement, or if missingness is related to health or treatment response. The record does not report how much data were missing, why values were missing, or whether sensitivity analyses examined alternative assumptions.

What can the planning assumptions tell readers?

The record describes a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between the groups, and a normal approximation. It says the effect to detect is a stated fraction of the observed effect. Such calculations are prospective illustrations of the number of participants per arm that might be needed under those assumptions; they are not evidence that the trial had a particular probability of detecting an effect after it was completed. They are not observed power.

The planning description does not provide the effect fractions or the resulting participant counts in the record. The primary observed effect is also not reported as a single comparison, so the illustration cannot be evaluated from the information recorded. These assumptions should not be mistaken for the trial's original sample-size justification. That justification is not reported in the registry record.

Why do outcome distributions and uncertainty matter?

As a general statistical point, a continuous outcome may be summarized by a mean, but the mean does not show the full distribution of individual responses. A group average can coexist with considerable variation among participants. A confidence interval around a group contrast helps express uncertainty in that average difference, but it does not show every individual outcome or establish clinical importance on its own. Assessing clinical importance requires context about meaningful changes, which is not reported in the registry record.

Likewise, event counts require denominators and a clearly defined observation period to support rate comparisons. The adverse-event record includes counts but does not report the numbers at risk. Consequently, those counts alone do not support a rate estimate or a comparison of risk between groups. Statistical interpretation must match the form of the data and the information actually recorded.

10. Multiplicity, Missing Data and Analysis Populations

The registry summary states that a gatekeeper strategy would be used to test secondary objectives sequentially. In general, gatekeeping is a way to manage the chance of false-positive findings when multiple hypotheses are tested: later tests may depend on the outcome of earlier tests. The specific test order, the decision rules, and the method for controlling the overall error rate are not reported in the registry record. The record lists numerous secondary outcomes, but does not provide a single consolidated set of results or enough detail to determine how multiplicity was handled in practice.

Missing-data handling is described for the primary endpoint as last observation carried forward. This method uses an earlier observed value in place of a missing later one. As a general matter, that choice embeds assumptions about how the outcome would have evolved after the last recorded measurement. The registry record does not report the amount or pattern of missingness, reasons for missing measurements, or sensitivity analyses under alternative missing-data assumptions. It also does not describe missing-data procedures for each secondary outcome.

The primary population is labeled intent to treat but defined as randomized patients with a baseline value and at least one post-baseline value. The number analyzed in each group is not reported in the registry record. Thus, readers cannot assess how many randomized patients contributed to the endpoint analysis or compare the analyzed population with the enrolled population. The record does not report whether other analysis populations were used, or how deviations from assigned treatment were handled.

The statistical analysis plan, including any interim analyses or stopping rules, is not reported in the registry record. The available gatekeeper statement establishes a broad intention for sequential secondary testing, not the details needed to reproduce or fully audit that procedure.

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

The registry record establishes that the trial randomized patients with type 1 diabetes to compare two basal insulin therapies within basal-bolus therapy. It identifies change in HbA1c from baseline to the 32-week endpoint as the primary outcome and specifies least-squares means and standard errors as the summary approach. Those details describe the question and analysis framework, not the outcome of the comparison.

Because the record does not report a single primary comparison, it does not show which group had a larger estimated change, how large any difference was, or how precisely it was estimated. The record therefore does not support a conclusion that one treatment had a better or worse primary outcome. It also does not permit an assessment of whether an observed contrast, if any, would be clinically important.

Similarly, the adverse-event counts should not be read as comparative rates. The numbers at risk are not reported in the registry record, and the record does not report deaths. The listed serious adverse-event terms and percentages provide limited descriptive information, but do not supply a complete assessment of comparative safety. The secondary outcomes listed in the registry do not establish results for glycemic variability, hypoglycemia, self-monitored glucose, or body weight.

As a general statistical principle, absence of a reported comparison is not evidence that the groups were the same. It means that the available registry information is insufficient to estimate the contrast. Conversely, the study's randomized design does not by itself reveal the result or remove uncertainty caused by incomplete reporting. A reader would need the group-level estimates, treatment contrast, uncertainty measures, and analysis details to judge the primary finding.

The record's completed status indicates that the trial is marked completed in the registry. Completion alone does not establish efficacy, safety, or precision. The most defensible interpretation is limited to the design and outcomes recorded, with the primary numerical comparison not reported in the registry record.

12. Why This Trial Matters Statistically

This trial is statistically informative as an example of the difference between a registered research question and a reportable treatment-effect estimate. The record specifies a randomized, parallel comparison, identifies an HbA1c change endpoint, and names a model-based summary. Yet it does not provide the single primary comparison needed to interpret the result. Design information and outcome definitions are essential, but they cannot stand in for effect estimates and their uncertainty.

The analysis population description also illustrates why labels need accompanying operational definitions. The primary population is called intent to treat while requiring baseline and post-baseline measurements. The number analyzed by group and the extent of missing data are not reported in the registry record. Without those details, readers cannot assess how closely the analyzed data represent all randomized participants.

The planned gatekeeper strategy is another important statistical feature. Sequential testing can help address multiplicity across secondary outcomes, but the description of an intention to use such a strategy is not enough to judge its implementation. The order of hypotheses, decision rules, and results are not reported in the registry record.

Finally, the adverse-event information underscores the difference between counts and rates. Counts alone do not give comparative risk when denominators are absent. Across these elements, the record shows why transparent reporting requires not only naming outcomes and methods, but also presenting estimates, uncertainty, analysis populations, and missing-data context. That distinction is central to interpreting clinical trial evidence responsibly.

13. Limitations

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

The primary limitation for interpretation is that the registry record does not report a single primary comparison for change in HbA1c. It therefore lacks the group estimates and treatment contrast needed to assess direction, magnitude, and precision. The primary outcome is also described as having several categories or time points and is not reduced to one comparison, leaving the result difficult to interpret as a unified primary finding.

Important analysis details are not reported in the registry record. These include the full model specification, the number analyzed in each group, the amount and pattern of missing primary-outcome data, and the detailed procedures for secondary-outcome testing. Although last observation carried forward is named, no sensitivity analyses or missing-data diagnostics are described. The sample-size justification is not reported in the registry record.

The trial was unmasked, and the registry does not report whether outcome assessors or analysts were masked. Randomization is recorded, but the allocation ratio, sequence generation, concealment procedures, and stratification factors are not reported. These omissions limit the ability to assess implementation and possible sources of bias from the registry record alone.

Safety interpretation is also limited. Counts of serious and other adverse events are recorded, but the numbers at risk are not reported. The record does not report deaths, and the serious-event terms listed at 0.5% each are not assigned to treatment groups in the information recorded. Thus, these data do not provide a complete basis for comparing event rates between groups.

These limitations concern the detail available in the registry record; they do not establish that the trial itself failed to collect or analyze the relevant information.

14. Related Tutorials

15. Related Statistical Calculators

16. Sources

• ClinicalTrials.gov: Comparison of Two Basal Insulin Therapies for Patients With Type 1 Diabetes — NCT00487240.
• Registry record last updated 2010-11-04; status last verified 2010-10.
• Results first posted 2009-09-25.
• 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.

17. Record Summary

The record describes NCT00487240, a completed phase 3 randomized parallel-group study with no masking, sponsored by Eli Lilly and Company. It enrolled 387 patients with type 1 diabetes. The recorded study start was 2007-06, and primary completion was 2008-08.

The comparison was insulin lispro protamine suspension twice daily versus insulin detemir, described as subcutaneous twice daily, in basal-bolus therapy. The primary outcome was change in HbA1c from baseline to the 32-week endpoint, summarized using least-squares means and standard errors. The primary analysis population is described as randomized patients with baseline and at least one post-baseline value, using last observation carried forward.

The record lists secondary glycemic, hypoglycemia, and body-weight outcomes and states that secondary objectives would be tested using a gatekeeper strategy. A single primary comparison, detailed testing procedures, and analysis-group sizes are not reported in the registry record. Adverse-event counts are recorded, but numbers at risk are not.