Effect of Canagliflozin in T1DM (Type 1 Diabetes Melitus) After Interruption of Continuous Subcutaneous Insulin Infusion: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effect of Canagliflozin in T1DM (Type 1 Diabetes Melitus) After Interruption of Continuous Subcutaneous Insulin Infusion (NCT02673138), a non-randomized na study. The primary outcome is Differences in Plasma Glucose Levels Following the Interruption of Basal Subcutaneous Infusion of Insulin.

2. Clinical Question

3. Trial Design

NCT02673138 is listed as a completed study sponsored by Yale University. The registry gives a study start of 2016-01 and a primary completion date of 2017-01. Its phase is listed as NA. Enrollment is 10, and the condition is Type 1 Diabetes.

The registry classifies the allocation as nonrandomized, the masking as none, and the model as single-group. The brief summary describes an open-label cross-over study in which subjects would be studied under both conditions: interruption of subcutaneous insulin infusion while receiving insulin alone, and interruption while receiving insulin with canagliflozin. The arm descriptions likewise describe overnight basal interruption on a research unit, with or without canagliflozin.

These descriptions leave important design details unclear. The registry record does not report how the order of conditions was determined, whether there was a washout period, or how the within-participant comparison was analyzed. Although both conditions are described for the same subjects, the record's single-group and nonrandomized classifications should be kept in view when interpreting the comparison. The record does not describe a randomized allocation procedure or masking.

The listed secondary outcomes concern differences in free fatty acid levels and glucagon levels following interruption of basal subcutaneous insulin infusion. The registry lists four outcome measures, but it does not report their results in the information available for this study. Details such as eligibility criteria, participant characteristics, and the actual timing of measurements are not reported in the registry 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.

Overall Study

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Differences in Plasma Glucose Levels Following the Interruption of Basal Subcutaneous Infusion of Insulin, measured over 20 hours.

The primary endpoint concerns plasma glucose levels after basal subcutaneous insulin infusion is interrupted. The registry specifies a mean continuous measure, in mg/dL, over a 20-hour timeframe. Its stated purpose is to assess differences between interruption during insulin-alone treatment and interruption during treatment with insulin plus canagliflozin or another SGLT2 inhibitor.

In general, a mean describes the average of measured values across the observations included in an analysis. It can provide a concise account of glucose levels over a measurement period, while not showing the full pattern or variability of individual measurements. The registry specifies standard error as the measure of dispersion. In general, standard error describes the precision of an estimated mean; it is not the same as the spread of individual measurements.

The outcome is relevant to the stated study question because it tracks plasma glucose after the interruption being studied. However, the registry record does not report the mean values, standard errors, or a between-condition estimate. It is therefore not possible to describe the direction, magnitude, or precision of the observed difference from this record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the specified outcome represent?

The primary outcome is a continuous measure: plasma glucose, expressed in mg/dL, after interruption of basal subcutaneous insulin infusion. The registry specifies a mean and standard error over a 20-hour timeframe. In general, a continuous outcome retains information about measured values rather than reducing them to event categories. A mean can describe the average level, while the standard error conveys uncertainty around that estimated average. The standard error should not be read as the range containing individual measurements; it describes precision of an estimate, not participant-to-participant variability.

The registry does not report the actual glucose summaries or explain exactly which measurements contributed to the mean. It also does not state whether the comparison was based on a participant-level summary, repeated measurements, or another analysis unit. These details affect what an estimate represents and how its uncertainty should be calculated.

How should the condition comparison be understood?

The study summary describes subjects undergoing both conditions, which suggests a within-participant comparison: each subject would contribute information under insulin alone and under insulin with canagliflozin. In general, paired designs can account for stable differences between participants by comparing measurements within the same participant. That can be useful when outcomes from one person are expected to be more alike than outcomes from different people.

However, the registry also identifies a single-group model and nonrandomized allocation. It does not report the sequence of conditions, whether sequence was assigned or fixed, or whether a washout period was used. It also does not explain whether the primary analysis accounted for pairing, period effects, or repeated glucose measurements. These are not details that can be inferred just from the cross-over wording. The analysis approach is not reported in the registry record, so the statistical validity of any condition contrast cannot be evaluated from the entry.

What does standard error communicate?

In general, standard error reflects the precision of an estimated mean under the assumptions of the analysis. A smaller standard error corresponds to greater precision, all else being equal. It is distinct from the standard deviation, which describes variability among observations. The registry names standard error as the dispersion measure for the primary outcome, but does not report its value. As a result, the precision of the glucose mean cannot be assessed.

A confidence interval, in general, expresses a range of estimates compatible with the data under a specified statistical procedure. Its width communicates uncertainty, and its interpretation depends on the model and assumptions used. No confidence interval is reported for the primary outcome in the registry record. No p-value or other inferential result is reported either. Without these results, it is not possible to judge the precision of a condition difference or how compatible the data are with no difference.

What can the planning assumptions tell us?

The registry's sample-size assumptions describe planning illustrations based on the observed effect, not the trial's own design. They specify a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is described as a stated fraction of the observed effect. These are planning assumptions, not evidence that the trial used those design choices. The record specifically distinguishes the illustrations from the study's own design.

In general, a planning calculation asks how much information may be needed to detect a specified effect under stated assumptions. It depends on the target effect, variability, allocation, significance level, and the method used. It is not a measure of the probability that a completed study has found a true effect. The assumptions alone do not provide the missing observed glucose results, and any planning illustration should not be interpreted as observed power. The registry record does not report the trial's own sample-size justification.

What statistical details remain unavailable?

The registry lists zero analyses. It does not identify a statistical model, a hypothesis test, an adjustment strategy, or the handling of repeated measurements. It does not report whether an intention-to-treat, per-protocol, or other analysis population was used. Missing data procedures, outlier handling, and assumptions about the distribution of the outcome are also not reported in the registry record.

In general, a statistical model should match the design and measurement structure. If repeated observations are related, treating them as independent may misstate uncertainty; models for correlated data can account for that dependence. If the data show greater variability than a model assumes, uncertainty estimates and tests can be poorly calibrated. These are general considerations, not claims about this study's data. The record does not provide enough information to determine which modeling issues arose or how they were handled.

Because the actual results and analysis methods are missing, the most defensible reading is descriptive: the registry specifies the intended endpoint and comparison, but not the estimated contrast or its uncertainty. Statistical significance, effect size, and robustness cannot be evaluated from the registry record.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Basal Interruption", "Basal Interruption With Canagliflozin")
n   <- c(10, 10)
m   <- c(197, 99)
sp  <- c(24, 13)
sd  <- sp * sqrt(n)      # standard errors converted to SDs
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Basal Interruption  10  197.000  75.895  142.708  251.292
CARM  Basal Interruption With Canagliflozin  10  99.000  41.110  69.592  128.408

11. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome concerning plasma glucose and secondary outcomes concerning free fatty acid and glucagon levels. It also lists four outcome measures. The results for these outcomes and the analyses are not reported in the registry record. A multiplicity strategy, such as an adjustment for testing several outcomes, is not reported either.

In general, testing multiple outcomes or contrasts can increase the chance of obtaining at least one apparently notable result by chance. A prespecified hierarchy or adjustment can address that concern, but the appropriate approach depends on the study plan and analysis. No such plan is described in the registry record, so it is not possible to determine how multiplicity was handled or whether the reported outcomes were interpreted as confirmatory or exploratory.

Missing-data procedures are not reported in the registry record. The entry does not state whether all enrolled participants completed both conditions, whether all planned glucose measurements were available, or how any missing measurements were treated. It also does not specify the population included in the primary analysis. In general, incomplete outcome data can affect both an estimate and its precision, particularly if missingness is related to participant characteristics or observed outcomes.

The registry record therefore leaves unanswered how the analysis population was defined, whether participants were analyzed under their assigned or completed conditions, and how missing observations were handled. These uncertainties limit interpretation of any treatment-condition comparison, especially alongside the absence of posted primary and secondary results.

12. 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 5% reporting threshold.

Most frequent non-serious adverse event terms

The registry lists safety counts for the basal-interruption condition and the basal-interruption-with-canagliflozin condition. For serious adverse events, it reports zero affected and 10 at risk in each condition. It reports zero deaths in each condition. For other adverse events, it reports zero affected among 10 at risk in the basal-interruption condition and one affected among 10 at risk in the canagliflozin condition.

These are registry counts, not estimates of comparative risk. The record does not identify the other adverse event, its severity, timing, relationship to treatment, or whether the same participants contributed safety information under both conditions. It does not report a statistical comparison of the safety counts. The event descriptions and further safety context are not reported in the registry record.

In general, a zero count in a small study does not establish that an event cannot occur; it indicates only that no event was recorded in the listed data. Likewise, a difference in descriptive counts does not by itself establish a treatment-related difference. The safety information should be read as limited reporting for this study, not as a complete assessment of risk.

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

The registry record establishes the topic and intended comparison: plasma glucose after basal subcutaneous insulin infusion is interrupted during insulin-alone treatment versus treatment with insulin and canagliflozin. It specifies a mean outcome in mg/dL over a 20-hour timeframe and identifies standard error as the dispersion measure. It also describes a nonrandomized, open-label study in which subjects were to be studied under both conditions.

What the record does not establish is the result of that comparison. It gives no observed glucose means, no estimated difference, no standard error value, no confidence interval, and no p-value. Thus, the posted record does not show the direction or size of any glucose difference, nor how precisely a difference was estimated. It cannot support a conclusion about whether the conditions produced different glucose levels.

The design description is not enough to reconstruct the analysis. Although the summary describes a cross-over, the registry classifies the study as single-group and nonrandomized. The order of conditions, analysis method, and handling of within-participant measurements are not reported. Participant characteristics, missing-data handling, and the analysis population are also not reported. These omissions make it difficult to assess how well any future-reported estimate would represent the planned comparison.

The safety counts are limited descriptive information. Serious adverse events and deaths are listed as zero in both conditions; one other adverse event is listed in the canagliflozin condition and none in the basal-interruption condition. The type of that event is not reported. These counts do not resolve the unanswered primary-outcome question.

In general, a trial result is best interpreted using its estimated effect, uncertainty, design, and analysis together. Here, the result and analysis components are absent from the registry record. The study description should therefore be distinguished from evidence about an outcome: it identifies what researchers intended to compare, but does not provide the posted results needed to judge the comparison.

14. Why This Trial Matters Statistically

This study raises a statistical question about measurements collected around an interruption of basal insulin infusion, comparing insulin alone with insulin plus canagliflozin. Its registry description indicates that the same subjects would be studied under both conditions. That makes the within-participant structure potentially important: observations from the same person may be related, and analysis should reflect the design rather than assume that every measurement is independent.

The record also illustrates why design labels and study descriptions need to be read together. It describes a cross-over study but classifies the model as single-group and allocation as nonrandomized. The order of conditions and the statistical handling of repeated measurements are not reported. Without those details, readers cannot assess whether period or sequence effects were considered or how a condition contrast was estimated.

The stated endpoint is a mean with standard error. This distinction matters because the average level and the precision of that average answer different questions from the variability among individual measurements. Yet neither the observed mean nor the standard error is posted. The absence of an effect estimate and uncertainty measure means that the scientific question remains statistically unresolved in the registry record.

The entry also lists secondary outcomes and sample-size planning assumptions, but does not report outcome analyses. Planning assumptions are not a substitute for a documented design justification or for observed results. This makes the record useful for understanding the intended question while limiting its value for evaluating the evidence generated.

For biostatisticians and clinical researchers, the key lesson is that a clear endpoint and planned comparison do not alone permit interpretation. Results, analysis methods, population definitions, and missing-data procedures are needed to connect the study design to a defensible estimate.

15. 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 of the registry entry is the absence of posted outcome results. Although it specifies plasma glucose as the primary outcome and identifies the measurement timeframe and summary measure, it does not provide observed values or an estimated comparison. The results for the listed secondary outcomes are also not reported in the registry record.

The design details are incomplete. The brief summary describes a cross-over in which subjects undergo both conditions, while the registry classifies the study as single-group and nonrandomized. The condition sequence, timing between conditions, and any washout period are not reported. The analysis method is not reported, including how within-participant or repeated measurements were handled.

The registry gives enrollment as 10 but does not report participant characteristics, eligibility criteria, the number contributing data to each outcome, or the analysis population. Missing-data and exclusion procedures are not reported. These omissions prevent assessment of completeness and of how representative the analyzed data may be.

The entry lists four outcome measures and describes two secondary outcome topics, but does not report a multiplicity strategy. The sample-size assumptions are described as planning illustrations based on the observed effect rather than as the trial's own design; the trial's own sample-size justification is not reported in the registry record. Safety reporting is limited to counts, without descriptions of the other adverse event or further context.

Together, these limitations mean the registry record documents the study question and selected design features, but not enough detail to evaluate the primary statistical result, its precision, or its robustness.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Effect of Canagliflozin in T1DM (Type 1 Diabetes Melitus) After Interruption of Continuous Subcutaneous Insulin Infusion — NCT02673138.
• Registry record last updated 2021-10-13; status last verified 2021-09.
• Results first posted 2018-11-29.
• 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.

19. Record Summary

NCT02673138 is a completed Yale University study of Type 1 Diabetes, with enrollment listed as 10. The registry gives a study start of 2016-01, primary completion of 2017-01, phase NA, nonrandomized allocation, no masking, and a single-group model.

The study summary describes subjects undergoing basal insulin interruption during insulin-alone treatment and during treatment with insulin plus canagliflozin. The primary outcome is mean plasma glucose in mg/dL over a 20-hour timeframe, with standard error specified as the dispersion measure. The registry also lists free fatty acid and glucagon outcomes.

Primary and secondary outcome results, analysis methods, analysis population, and missing-data procedures are not reported in the registry record. Safety counts list zero serious adverse events and zero deaths in both conditions, with one other adverse event affected in the canagliflozin condition and none in the basal-interruption condition.