Duloxetine in Patients With Diabetic in Peripheral Neuropathic Pain With or Without Co-morbid Major Depressive Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Duloxetine in Patients With Diabetic in Peripheral Neuropathic Pain With or Without Co-morbid Major Depressive Disorder (NCT00844194), a non-randomized phase 4 study. The primary outcome is Change of Brief Pain Inventory (BPI) Average Interference Score From Baseline to Week 12.

2. Clinical Question

3. Trial Design

This phase 4 study was registered as a 12-week, open-label, parallel-group study of duloxetine in people with diabetic peripheral neuropathic pain, with or without co-morbid major depressive disorder. The registry lists non-randomized allocation and no masking. Enrollment was 108, and the study status is completed. The recorded study dates are 2009-02 for start and 2010-06 for primary completion.

The study’s main objective was to evaluate the week 12 change in BPI average interference score separately in the DPNP patients with MDD and those without MDD. The registry describes a further confirmatory evaluation in the group without MDD if statistically significant results were obtained in the group with MDD. It also lists secondary questions about pain severity, pain reduction, patient and physician impressions, functionality, quality of life, and psychological symptoms.

The registry describes four groups based on MDD status and response to 60 mg duloxetine at week 6. Responders had a pain reduction >30%; non-responders had a pain reduction <30%. These response-defined groups are not equivalent to groups assigned at random: response is observed after treatment begins, and can reflect differences among participants as well as their response during treatment.

The registry record does not report how participants entered the treatment groups, the detailed dose-adjustment schedule, or a protocol for handling departures from the intended treatment. It also does not report a randomized comparison with placebo or another treatment. The absence of random allocation and masking matters when interpreting change over time and differences between groups: neither design feature protects the comparison from all potential sources of bias.

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: All patients receiving at least one dose of study medication, having any efficacy data and dosage of duloxetine was not more than 60mg before visit 5

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: All patients receiving at least one dose of study medication (TS).

6. Primary Endpoint

The registry lists the primary outcome as Change of Brief Pain Inventory (BPI) Average Interference Score From Baseline to Week 12, measured over Baseline and Week 12.

The primary endpoint was the change in BPI average interference score from baseline to week 12. The registry defines change as the week 12 value minus the baseline value. The BPI score ranges from 0, meaning pain does not interfere, to 10, meaning pain completely interferes. The endpoint therefore focuses not only on pain intensity but on how much pain interferes with daily activities.

The study used this measure because its stated objective concerned patient-relevant interference from DPNP, and the registry defined a clinically relevant change threshold of -1.35. A negative change indicates a lower score at week 12 than at baseline, according to the stated subtraction order. The planned criterion was that the confidence interval for the mean change should not lie above -1.35. This criterion describes how the study intended to judge the result; it is not itself an observed finding.

The analysis population was defined as patients who received at least one dose of study medication, had any efficacy data, and had not received more than 60 mg before visit 5. The numerical mean change, its standard deviation, and its confidence interval are not reported in the registry record. Without those results, the direction, size, and precision of the endpoint change cannot be assessed.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Several statistical ideas help clarify what the registry’s planned endpoint would mean, even though its numerical result is not reported.

What does a mean change represent?

The primary outcome is continuous: each participant has a BPI average interference score on a 0-to-10 scale, and the analysis concerns the mean change from baseline to week 12. The individual change is calculated by subtracting the baseline score from the week 12 score. A negative change corresponds to a lower interference score at follow-up. A group mean summarizes the average change, but it does not imply that every participant changed by that amount or changed in the same direction.

In general, a mean is useful when the scientific question concerns an average level or change on a continuous scale. It can, however, conceal variation in individual responses. The registry identifies standard deviation as the measure of dispersion for the primary outcome, but does not report the observed mean change or standard deviation. The distribution of individual changes is therefore not available for interpretation.

How should the confidence interval be read?

The study’s stated decision rule concerns a confidence interval for the mean change, not only the point estimate. In general, a confidence interval describes the uncertainty around an estimate under the statistical model and its assumptions. A narrow interval indicates greater precision than a wide interval, while an interval that spans a range of values leaves more uncertainty about the mean change. A confidence interval is not a guarantee that the true value lies within its limits.

The registry’s criterion was that the interval should not lie above -1.35. The threshold is a prespecified interpretation point in the study description, not an observed effect. Because neither endpoint estimate nor interval is reported, it is not possible to determine whether the interval met that criterion or to evaluate its precision.

Why does the design affect the analysis?

Allocation was non-randomized and masking was absent. In general, random allocation helps balance known and unknown prognostic factors between treatment groups, on average. Without it, group differences may reflect baseline characteristics as well as treatment exposure. Masking can reduce the influence of treatment knowledge on behavior and outcome assessment; its absence is especially relevant when outcomes depend on participant reports. These are design considerations, not proof that bias occurred in this study.

The study used parallel groups and divided participants by MDD status. It also described responder and non-responder groups using a post-baseline pain reduction threshold at week 6. In general, defining groups by a response measured after treatment begins can create selection effects: the resulting groups may differ in ways that are related to both the outcome and participant characteristics. A comparison between these groups is therefore descriptive unless the analysis and design support a stronger interpretation.

What does over-dispersion mean in general?

Over-dispersion occurs when observed variability is greater than the statistical model assumes. It is a concern for some count outcomes, where a model that understates variation can produce confidence intervals that are too narrow and statistical tests that appear more decisive than warranted. The primary endpoint in this study is a continuous score analyzed as a mean, not a count outcome. The registry does not report an over-dispersion assessment or identify it as part of the primary analysis.

How should planning calculations be understood?

In general, sample-size planning illustrations ask how many participants might be needed to detect a specified effect under stated assumptions. They depend on assumptions such as the significance level, allocation, outcome variability, and the effect size selected for planning. The registry describes an illustration based on the observed effect, two-sided significance level 0.05, equal allocation, and a normal approximation. It also describes examining fractions of the observed effect. The calculated planning counts are not reported in the registry record, and the observed effect required for those calculations is also not reported.

Such illustrations are planning quantities, not observed power. They do not show the probability that the completed study detected an effect, and they do not replace the estimate and confidence interval for the actual endpoint. The registry does not report the study’s own sample-size justification. The planning assumptions described in the record should therefore not be interpreted as the protocol’s original design rationale.

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("Without Major Depressive Disorder (MDD-)", "With Major Depressive Disorder (MDD+)")
n   <- c(65, 26)
m   <- c(-1.4, -2.2)
sp  <- c(1.91, 2.2)
sd  <- sp      # posted standard deviations
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  Without Major Depressive Disorder (MDD-)  65  -1.400  1.910  -1.873  -0.927
CARM  With Major Depressive Disorder (MDD+)  26  -2.200  2.200  -3.089  -1.311

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 99 outcome measures and 2 analyses. It describes multiple secondary questions, including pain severity at different time points, pain reduction, impressions of pain, functionality, quality of life, and psychological symptoms. Testing many outcomes or time points can increase the chance of finding at least one apparently notable result by chance. In general, multiplicity procedures can control or otherwise address this issue, but the registry record does not report a multiplicity adjustment plan or the details of any such procedure.

The primary objective is described separately for participants with and without MDD, with the second evaluation conditioned on statistically significant results in the first group. The registry record does not report the results of this sequence or specify further details of its statistical implementation. Without those results and methods, the extent to which the two evaluations were treated as confirmatory cannot be assessed.

The primary analysis population included patients who received at least one dose, had any efficacy data, and had not received more than 60 mg before visit 5. This is a defined analysis population, but the registry record does not report the number of patients included in the primary analysis by group, the numbers with missing endpoint data, or how missing data were handled. It also does not report whether analyses were conducted under other population definitions.

In general, missing outcome data can affect an estimate when the likelihood of missingness is related to the outcome or to participant characteristics. The consequences depend on how much data are missing and the assumptions used to handle them. Because the registry record does not describe missing-data methods or endpoint completeness, the robustness of the primary estimate to missingness cannot be evaluated.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: acute myocardial infarction (3.3%), atrial flutter (3.3%), meningioma (1.3%), carotid artery stenosis (1.3%), cerebrovascular accident (1.3%).

The registry reports serious adverse events in the two MDD groups, with 78 participants at risk in the group without MDD and 30 at risk in the group with MDD. Serious adverse events affected 4 participants in the MDD- group and 2 in the MDD+ group. Other adverse events affected 28 participants in the MDD- group and 16 in the MDD+ group. These are participant counts, not event rates adjusted for follow-up or exposure.

The registry lists serious adverse event terms and percentages, including acute myocardial infarction and atrial flutter at 3.3%, and meningioma, carotid artery stenosis, cerebrovascular accident, and transient ischaemic attack at 1.3%. The record does not give group-specific counts alongside those terms in the information available for this summary, so they should not be assigned to a particular MDD group.

Deaths are not reported in the registry record. The adverse event figures describe what was recorded during this study; they do not establish comparative safety. There was no randomized control group, and the group sizes differed. The registry information also does not provide the detail needed to assess timing, duration, severity beyond the serious-event classification, or causal attribution for each event. These counts should therefore be read as descriptive safety information rather than evidence that one group had greater or lesser risk.

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

The registry record establishes that this completed phase 4 study examined duloxetine in DPNP patients with and without co-morbid MDD, using change in BPI average interference from baseline to week 12 as its primary endpoint. It also records a clinically relevant threshold of -1.35 and a planned conditional evaluation across the two MDD groups. These details identify the question the study was designed to address.

The record does not report the numerical primary outcome, its confidence interval, or the primary analysis sample size by group. As a result, it is not possible to say whether the mean interference score decreased, whether the change met the study’s criterion, or how precisely any estimated change was determined. The study’s objective and decision rule must not be mistaken for evidence that the rule was met.

The non-randomized, open-label design further limits causal interpretation. A change from baseline may occur during treatment, but without a randomized control group it cannot be separated from changes that might have occurred over time for other reasons. Group comparisons may reflect differences in depression status, response classification, or other participant characteristics. Knowledge of treatment may also affect reporting of a subjective pain-interference outcome. These are limitations of inference, not findings that bias necessarily occurred.

The registry’s safety counts offer descriptive information about participants affected by adverse events, but do not provide a controlled estimate of comparative risk. The missing numerical efficacy results, incomplete detail about missing data, and absent information about the analysis methods prevent a full assessment of the strength and uncertainty of the findings. The record is therefore informative about the study’s design, population, and intended endpoint, but does not permit a conclusion about the observed primary treatment outcome.

14. Why This Trial Matters Statistically

This trial raises several important statistical issues in a setting where the outcome is a patient-reported measure and the comparison is not randomized. Its primary endpoint focuses on interference from pain rather than only pain intensity. That distinction matters because symptom severity and the extent to which symptoms disrupt daily life are related but different constructs. The BPI scale gives the study a defined measure and direction of change, while the threshold of -1.35 makes explicit the study’s intended criterion for clinical relevance.

The study also illustrates why a stated objective is not a result. A protocol-level threshold and a conditional analysis plan describe how investigators intended to interpret evidence. To assess whether the criterion was met, readers need the mean change and its confidence interval. Those numerical results are not reported in the registry record, so the planned criterion cannot be evaluated from the information recorded.

Its group structure highlights a distinction between randomized comparisons and comparisons based on participant characteristics or post-treatment response. MDD status was not randomized, and response status was defined using pain reduction after treatment began. Such groups can be useful for describing outcomes, but differences between them do not isolate the effect of depression status or treatment response.

Finally, the record emphasizes that a large set of listed outcomes brings questions about multiplicity, missing data, and analysis populations. These are not technical details detached from interpretation: they help determine how much confidence to place in apparent patterns. The registry does not report key numerical efficacy results or several methods needed to evaluate those issues, making transparent reporting of estimates and uncertainty especially important.

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 central limitation for interpreting efficacy is that the registry record does not report the primary endpoint estimate or confidence interval. Without these, neither the direction and size of the mean change nor its precision can be assessed, and the stated threshold of -1.35 cannot be evaluated. Numerical results for the secondary outcomes are also not reported in the record.

The study was non-randomized and unmasked. Comparisons between MDD+ and MDD- participants may be influenced by differences between those groups that are not due to duloxetine. The responder and non-responder classifications depend on pain reduction after treatment began, so they do not create randomized comparison groups. In addition, an unmasked design may influence reporting or assessment of subjective outcomes, although the record does not establish that this occurred.

The registry does not report detailed group-level primary analysis counts, the amount or pattern of missing endpoint data, or the methods used to handle missingness. It also does not report the detailed statistical model, covariate adjustment, or the study’s own sample-size justification. Although 99 outcome measures and 2 analyses are listed, the record does not describe a multiplicity adjustment plan.

Safety interpretation is limited by the descriptive event counts, the absence of a randomized control group, and the lack of detailed information about event timing and attribution. Deaths are not reported in the registry record. Together, these limitations mean the registry information can describe the study’s plan and recorded safety counts but cannot support a comprehensive appraisal of comparative efficacy or safety.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Duloxetine in Patients With Diabetic in Peripheral Neuropathic Pain With or Without Co-morbid Major Depressive Disorder — NCT00844194.
• Registry record last updated 2014-05-15; status last verified 2013-12.
• Results first posted 2011-10-31.
• 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 phase 4 study enrolled 108 participants with diabetic peripheral neuropathic pain, with or without co-morbid major depressive disorder. It was a 12-week open-label, non-randomized, parallel-group study of oral duloxetine. The primary endpoint was change in BPI average interference score from baseline to week 12, with a stated clinically relevant threshold of -1.35.

The primary numerical estimate and confidence interval are not reported in the registry record, so whether the stated criterion was met cannot be determined. The record lists 99 outcome measures and 2 analyses, but does not report key details about multiplicity or missing-data handling. Serious adverse events affected 4 participants in the MDD- group and 2 in the MDD+ group; deaths are not reported. The design and absent numerical efficacy results limit conclusions about treatment effects.