Pioglitazone Tablets Special Drug Use Surveillance "Combined Use of Biguanides / Long-term Treatment": Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pioglitazone Tablets Special Drug Use Surveillance "Combined Use of Biguanides / Long-term Treatment" (NCT02003014), a registered phase not stated study. The primary outcome is Number of Participants Reporting One or More Adverse Drug Reactions.

2. Clinical Question

3. Trial Design

This completed record describes a special drug use surveillance study of pioglitazone, also identified in the titles as Actos tablets, for people with type 2 diabetes mellitus. The stated purpose was to investigate safety and efficacy during long-term treatment with pioglitazone in combination with biguanides. The summary describes the period of interest as 12 months after pioglitazone treatment began. The primary endpoint timeframe is given as baseline up to 12 months.

The registry lists one intervention, pioglitazone, and one group, Pioglitazone 15 mg to 30 mg, administered orally once daily. It does not report a randomized allocation or a like-for-like control group. The allocation method is not stated. Consequently, this record should be read as surveillance of participants receiving the described treatment, not as a randomized comparison that can isolate the treatment’s effect from other explanations for observed outcomes.

The record lists enrollment of 899 and a safety analysis set defined as all enrolled participants who completed the study. The adverse-reaction results report 880 at risk. The reason the safety-set definition and the reported at-risk count do not match the enrollment figure is not reported in the registry record. The dates listed are a study start of 2009-02 and primary completion of 2012-01; status is COMPLETED. These dates alone do not describe the timing of each participant’s follow-up.

Seven outcome measures are indicated. The listed secondary outcomes concern changes from baseline in HbA1c, fasting blood glucose, body weight, immunoreactive insulin, and HOMA-IR. Numerical results for those secondary outcomes are not reported in the registry record. The record also contains no analysis entries, so the statistical procedures used for those outcomes cannot be established from the posted information.

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: Participants with a historical diagnosis of type 2 diabetes mellitus who failed to respond adequately to treatment with biguanides were enrolled to receive pioglitazone 15 milligram (mg) - 30 mg for up to 12 months.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The safety analysis set was defined as all participants who were enrolled and 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.

Population: The safety analysis set was defined as all participants who were enrolled and completed the study.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants Reporting One or More Adverse Drug Reactions, measured over Baseline up to 12 months.

The primary endpoint is the number of participants reporting one or more adverse drug reactions, with the unit expressed as participants. The registry defines an adverse drug reaction as an adverse event that, in the investigator’s opinion, is causally related to study treatment. It defines adverse events as unfavorable and unintended signs, symptoms, or diseases temporally associated with use of a medicinal product, reported from the first dose through the last dose of study drug.

Counting participants with at least one event is a binary outcome: each participant either meets the endpoint definition or does not. This kind of endpoint can summarize how many people had an investigator-attributed reaction during the specified observation period. It does not, by itself, describe the number of reactions per participant, their duration, or their severity. Those details require separate information.

The endpoint is relevant to a safety surveillance question because it focuses on events considered related to treatment, rather than all events occurring during follow-up. That attribution is based on investigator opinion, so it is not equivalent to proof that treatment caused an event. The timeframe is baseline up to 12 months, while the summary describes surveillance for 12 months after treatment began.

Although the registry specifies the endpoint definition, a numerical result for the primary endpoint as defined is not reported in the registry record. The serious and other adverse-reaction counts posted separately should not automatically be combined or treated as the primary endpoint: the record does not establish whether those categories are mutually exclusive or provide a direct mapping to the endpoint definition.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical interpretation depends first on the outcome definition and the study design. In this record, the primary outcome is a participant-level binary measure: at least one adverse drug reaction, as judged by the investigator, during the stated period. A binary outcome differs from an event count. One person with multiple reactions may still count once for a participant-level endpoint, whereas a tally of events could count several reactions for that person. The registry’s posted serious and other reaction counts should therefore not be assumed to be participant counts for the primary endpoint unless their definitions make that clear.

What does a count tell the reader?

A count reports how many observations fall into a category, but its interpretation depends on the denominator and on the precise unit being counted. The adverse-reaction record gives 880 at risk and separately lists 12 serious and 36 other reactions. It does not explicitly state in the primary endpoint results how many participants had one or more reactions. Nor does it explain whether the serious and other categories overlap. The counts are useful descriptive information, but they cannot resolve those endpoint-definition questions.

A proportion is often used to describe a binary outcome by relating the number of participants meeting the definition to the number assessed. A confidence interval can then express uncertainty around that estimate. In general, uncertainty depends on both the observed data and the assumptions used to construct the interval. The registry record does not report a primary-endpoint estimate or its confidence interval, so no such precision assessment can be made from the posted primary outcome information.

Why does the study design matter?

Random allocation and an appropriate comparison group can help distinguish a treatment contrast from differences that existed before treatment or arose during follow-up. Here, the record lists one pioglitazone group, does not state an allocation method, and does not report a like-for-like control arm. As a result, the adverse-reaction counts describe events recorded in the surveillance group. They do not isolate the contribution of pioglitazone from underlying health, concomitant care, or other influences. Those are general considerations for interpreting nonrandomized surveillance, not additional findings about these participants.

The treatment summary describes pioglitazone in combination with biguanides, given orally once daily at 15 mg to 30 mg. The record does not state how dose selection was determined for individual participants or provide details that would support a dose-response analysis. A group-level descriptive count cannot establish whether event reporting differed by dose, duration, or participant characteristics when those analyses are not reported.

How should the time window be understood?

The primary endpoint timeframe is baseline up to 12 months, and the summary describes the safety and efficacy investigation as covering 12 months after the start of pioglitazone treatment. A specified window helps define which events are considered, but it does not by itself show how completely participants were observed throughout that window. The record does not report individual follow-up duration or a time-to-event analysis. The posted counts should therefore not be interpreted as rates per unit of follow-up time.

Time-to-event methods are generally useful when the timing of events and differences in follow-up are central to the question. Such methods account for when an event occurs and how long each participant is observed, subject to their assumptions. No time-to-event results or analyses are reported in this record. The serious and other counts should be read as counts, not as incidence rates or event-time estimates.

What can planning calculations establish?

Sample-size planning illustrations are not a description of the actual design unless the registry identifies them as the study’s own planning assumptions. The planning description associated with this record is explicitly an illustration based on an observed effect, using a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the target effect is expressed as a fraction of the observed effect. It is not the trial’s own design. For non-time-to-event outcomes, the description refers to participants per arm; time-to-event illustrations refer to total events.

Such calculations answer a hypothetical question about the sample size that might be needed under stated assumptions. They do not represent observed power, and they do not change the design or evidence produced by the surveillance record. Because the registry does not report a randomized two-arm comparison, a planning illustration framed around equal allocation between two arms should not be mistaken for the actual treatment structure. The record does not report the protocol’s sample-size justification.

What is the role of adjustment and dispersion?

Adjustment can account for specified participant or design characteristics when a model and suitable data are available. No adjusted analyses are reported in the registry record. The posted adverse-reaction counts are not described as model-based estimates. For count outcomes, over-dispersion means that variability exceeds what a simpler model assumes; in general, ignoring it can make uncertainty appear too small. That concept does not establish that over-dispersion occurred in this study, and the registry does not report a count model or a dispersion assessment.

Overall, the appropriate statistical reading is descriptive and cautious. The record supplies a treatment context, an endpoint definition, and adverse-reaction counts, but it does not report a numerical primary-endpoint result, confidence interval, or analysis plan. General statistical methods can explain what would be needed for stronger inference; they cannot fill gaps in the registry record.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Pioglitazone")
n      <- c(880)
events <- c(79)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Pioglitazone  880  79  9.0  7.2  11.1  7.3  11.0  7.3  11.1

11. Multiplicity, Missing Data and Analysis Populations

The record lists seven outcome measures, including a primary adverse-reaction endpoint and secondary changes from baseline in HbA1c, fasting blood glucose, body weight, immunoreactive insulin, and HOMA-IR. It does not report a statistical analysis plan specifying which comparisons or tests were intended, how outcomes were prioritized beyond the primary label, or whether any adjustment for multiple testing was planned. Those details are not reported in the registry record.

Multiplicity arises when several outcomes or hypotheses are assessed. In general, testing many hypotheses can increase the chance of at least one apparently notable result, so a prespecified strategy may be important for confirmatory interpretation. That general concern does not show that any particular finding in this surveillance record is false or that adjustment was required. Numerical results for the listed secondary outcomes are not reported, and the registry provides no analyses from which multiplicity handling could be assessed.

The safety analysis set is defined as all participants enrolled and completing the study. The record lists enrollment of 899 and an at-risk count of 880 for the adverse-reaction results. It does not explain the relationship between these figures, identify the reason for any difference, or provide a participant flow account. Missing-data rules, handling of incomplete follow-up, and the number of participants with endpoint information are not reported in the registry record.

These population details matter because a participant-level event proportion depends on who is included and who has outcome information. Without an explicit primary-endpoint count and a clear denominator for that endpoint, readers cannot reconstruct the endpoint estimate or judge the effect of missing observations. No imputation method, sensitivity analysis, or other missing-data analysis is reported. The record’s analysis entries also do not report inferential procedures.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 1% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: appendicitis (0.1%), bile duct cancer (0.1%), ovarian cancer (0.1%), diabetes mellitus inadequate control (0.1%), mania (0.1%).

The safety results list 12 serious adverse reactions and 36 other adverse reactions, with 880 at risk for each listed category. These are separate posted counts. The registry does not clarify whether participants can appear in both categories, so the counts should not be added to obtain a total number of affected participants. They also should not be substituted for the primary endpoint result, which is defined as participants reporting one or more adverse drug reactions.

The serious-event terms listed are appendicitis, bile duct cancer, ovarian cancer, diabetes mellitus inadequate control, mania, and cerebral infarction, each with a reported percentage of 0.1%. The registry does not provide a count for each term or additional detail about timing, clinical course, or attribution for each listed term. The aggregate serious-reaction count and the term-level entries are different kinds of information and should not be treated as interchangeable.

In general, serious-event classifications describe event seriousness, not necessarily whether an event was caused by treatment. This record’s adverse drug reaction definition depends on the investigator’s opinion of causal relationship. The posted counts should therefore be understood as recorded reaction information under the registry’s definitions, not as proof that treatment caused every event. No comparative serious-event rate is available because a like-for-like control group is not reported.

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

The record documents a completed special drug use surveillance study of pioglitazone in combination with biguanides for people with type 2 diabetes mellitus. It specifies an oral once-daily pioglitazone dose range of 15 mg to 30 mg and describes a 12-month period after treatment initiation. Its primary safety question concerns participants reporting one or more adverse drug reactions, where attribution is based on the investigator’s opinion.

The posted results describe 12 serious and 36 other adverse reactions among 880 at risk. The registry also lists six serious terms, each reported as 0.1%. These figures provide descriptive safety information as recorded in the surveillance group. They do not provide a complete estimate for the primary endpoint as defined, because that numerical participant count is not reported. The record does not clarify whether the serious and other categories overlap.

The design information limits causal interpretation. There is no reported like-for-like control arm, and allocation is not stated. Without a suitable comparison, observed events cannot be attributed to the treatment alone, and differences from another treatment cannot be estimated. The record also does not supply a confidence interval for the primary endpoint or analyses that quantify uncertainty around a treatment contrast.

The registry’s stated purpose includes efficacy, and it names secondary measures involving glycemic outcomes, body weight, and insulin-related measures. Their numerical results are not reported in the registry record. Thus, the purpose statement and outcome list should not be read as evidence of an efficacy finding.

In general, surveillance records can describe events observed in a treatment setting and can help identify what was recorded under specified definitions. Their interpretation depends on the population assessed, follow-up, outcome ascertainment, and analysis details. Several of those details are not reported in the registry record. The posted information supports a limited descriptive account; it does not establish that treatment caused or prevented any particular event, nor does it provide a comparative judgment about treatment effects.

14. Why This Trial Matters Statistically

This record illustrates a basic statistical distinction between describing events and estimating a treatment effect. It identifies a treatment context, defines an adverse drug reaction through investigator attribution, and reports serious and other reaction counts. Those features make the record relevant to understanding how safety outcomes were recorded during pioglitazone use with biguanides. But counts observed in one listed group do not by themselves answer what would have happened under a different treatment.

It also shows why an endpoint’s unit and its reported result need to match. The primary endpoint is the number of participants with one or more reactions, while the posted safety information separates serious and other reactions. Unless category overlap and endpoint mapping are clear, those quantities cannot be combined or assumed to represent the primary measure. Clear definitions and denominators are essential for interpreting binary outcomes.

The record further highlights the importance of the analysis population. Enrollment is listed as 899, the safety set is defined by enrollment and completion, and the adverse-reaction results report 880 at risk. Without an explanation of the relationship among these figures, the reader cannot determine why the reported populations differ. That is a reporting limitation, not a basis for assuming a particular reason.

Finally, the record lists several secondary outcomes but does not report their numerical results or statistical analyses. A stated objective and an outcome list describe what was of interest; they are not a substitute for estimates, uncertainty intervals, and a transparent analysis plan. Statistically, the record is most informative as a limited description of recorded adverse reactions and as an example of how design and reporting details constrain inference.

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 main limitation is that the record does not report a numerical result for its defined primary endpoint, the number of participants with one or more adverse drug reactions. Instead, the posted safety information lists serious and other reaction counts. The registry does not explain whether these categories overlap or how they correspond to the primary endpoint, so they cannot be used to reconstruct that result.

The study is not reported as a randomized comparison, no like-for-like control arm is listed, and allocation is not stated. These features prevent a comparative treatment-effect interpretation from the posted counts. The record also does not provide adjusted estimates, confidence intervals for the primary endpoint, or an analysis plan. Reasons for attributing events to treatment are described generally as investigator opinion, but event-specific attribution details are not reported.

Enrollment is listed as 899, while the adverse-reaction results use 880 at risk and the safety analysis set is defined as participants who were enrolled and completed the study. The reason for the difference among these population descriptions is not reported in the registry record. Participant flow, missing-data handling, and completeness of outcome ascertainment are also not reported. These gaps make it difficult to assess the denominator and potential influence of incomplete follow-up.

Although secondary outcomes are named, numerical results are not reported. The record does not describe multiplicity handling, subgroup analyses, dose-specific results, or time-to-event methods. Nor does the list of serious terms provide event-level counts or clinical details. These limitations restrict interpretation to the posted descriptive information and prevent broader conclusions about comparative safety or efficacy.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Pioglitazone Tablets Special Drug Use Surveillance "Combined Use of Biguanides / Long-term Treatment" — NCT02003014.
• Registry record last updated 2016-11-10; status last verified 2016-09.
• Results first posted 2016-11-10.
• 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

This completed special drug use surveillance record concerns pioglitazone, taken orally once daily at 15 mg to 30 mg, in combination with biguanides for people with type 2 diabetes mellitus. The stated follow-up period is 12 months after treatment began. The primary endpoint is participants reporting one or more investigator-attributed adverse drug reactions.

The record lists enrollment of 899 and defines the safety analysis set as enrolled participants who completed the study. The adverse-reaction results list 880 at risk, 12 serious reactions, and 36 other reactions. Six serious terms are each listed at 0.1%. A numerical result for the primary endpoint itself is not reported in the registry record.

No randomized control comparison, allocation method, inferential analysis, or numerical results for the listed secondary outcomes are reported. The adverse-reaction counts therefore provide descriptive information, not a comparative or causal estimate.