A Study of Metronidazole Cream in the Prevention and Treatment of Tarceva (Erlotinib)-Associated Rash: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study of Metronidazole Cream in the Prevention and Treatment of Tarceva (Erlotinib)-Associated Rash (NCT00642473), a non-randomized phase 2 study. The primary outcome is Percentage of Participants With Erlotinib Associated Rash Stratified by Severity Grade at Week 2.

2. Clinical Question

The trial addresses both prevention and treatment, but these are different clinical situations and are represented by separate cohorts. Starting cream with erlotinib is the prevention strategy; starting cream only after rash appears is the treatment strategy. A comparison between the cohorts would therefore mix differences in treatment timing and rash status with any difference associated with the cream.

The prevention cohort also has a side-specific comparison: cream is applied to the right side of the face and chest, while corresponding left-side areas receive local standard care. That structure differs from a comparison between independently randomized participant groups. The registry record does not report how side-specific assessments were analyzed or whether they were the basis of the primary endpoint analysis.

Accordingly, the clinical question is best understood as an assessment of rash severity under two metronidazole-use approaches, with a local side comparison in the prevention cohort. The available record does not establish an effect estimate for either approach. It also does not report a treatment-versus-control estimate that would support a direct causal interpretation.

3. Trial Design

The registry describes a phase II, non-randomized, unmasked, parallel trial. It enrolled 34 participants with non-squamous non-small cell lung cancer. The study was completed. The record gives no randomization procedure because allocation is classified as non-randomized, and it reports no masking.

Two cohorts are described. In the prevention cohort, metronidazole Actavis 1% cream was initiated on the same day as erlotinib. Cream was applied twice daily to the right side of the face and chest for 4 weeks. The corresponding left-side areas were treated according to local standard procedures; the study summary identifies this as non-active moisturizing cream. In the treatment cohort, participants received erlotinib and began metronidazole cream if and when rash developed. Cream was likewise applied twice daily to the right side of the face and chest for 4 weeks, with local standard procedures used on the left side.

The study summary says efficacy was evaluated at Week 2 and Week 4. The primary outcome listed in the record is rash percentage by severity grade at Week 2, after 2 weeks of metronidazole treatment. The record also describes an anticipated metronidazole treatment duration of less than 3 months and a target sample size of less than 100 individuals; these are study descriptions, not outcome results.

The registry’s parallel-trial classification does not make the cohorts interchangeable. One cohort started treatment before rash developed, while the other started after rash developed. In addition, the prevention cohort has a treated-side and local-care-side feature. These design features matter when interpreting any apparent differences, because the record does not describe randomized assignment, masking, or a statistical method for accounting for paired sides. Further details of the protocol and analysis plan 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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All enrolled participants. Here, number of participants analyzed = participants who were evaluable for this outcome in respective arms.

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 enrolled participants

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants With Erlotinib Associated Rash Stratified by Severity Grade at Week 2, measured over After 2 weeks of metronidazole treatment.

The primary endpoint is the percentage of participants with erlotinib-associated rash, categorized by severity grade at Week 2. Rash severity was assessed semi-quantitatively with CTCAE v3.0. The listed categories run from Grade 0, no rash, through Grade 4, life-threatening or disabling, and Grade 5, death related to rash; Grades 1, 2, and 3 are described as mild, moderate, and severe.

A severity classification adds information that a simple yes-or-no rash measure would not capture: it distinguishes the recorded level of severity. Reporting percentages by category can therefore describe the distribution of observed rash grades, rather than only whether any rash was recorded. However, the endpoint description warns that the same participant may be counted in more than one reported category. The categories should not be assumed to be mutually exclusive counts or added together to form a total.

The record defines the endpoint population as all enrolled participants, while specifying that the number analyzed in each arm is the number evaluable for that outcome. The actual number evaluated in each arm and the grade-specific results are not reported in the registry record. Nor does the record explain how the semi-quantitative assessments were summarized or how participants with unavailable assessments were handled.

This endpoint is a measure of rash severity at a specified assessment time. By itself, it does not establish why a rash occurred, whether a recorded grade changed from an earlier assessment, or whether differences were caused by metronidazole. Such conclusions would require suitable comparisons and analysis details, which are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods determine how observations are converted into estimates and how much uncertainty surrounds those estimates. The registry record identifies a primary endpoint but does not describe the statistical analysis used to evaluate it. The explanations below are general; they do not imply that any particular method was used in this trial.

What does a percentage by severity grade describe?

A percentage expresses the share of an analysis group assigned to a particular outcome category, using a stated denominator. For a severity endpoint, each percentage is meaningful only alongside the definition of the category, the assessment time, and the participants included in its denominator. Here, the endpoint is described as percentage of participants by rash grade at Week 2, and the population description says that the number analyzed is the number evaluable in each arm.

The registry record also says a participant may be counted in more than one reported category. This matters because a set of category percentages need not represent a single partition of participants. Readers should not assume the categories sum to a total or that each person contributes to only one grade. The record does not explain the circumstances in which a participant could appear in multiple categories or how the analysis handled such reporting.

Why do the cohort definitions matter?

In a randomized comparison, assignment by chance is intended to make treatment groups comparable on average, including characteristics that are not measured. This trial is classified as non-randomized, so that design protection is not present. Cohort membership may reflect differences in timing or clinical circumstances, and those differences can affect outcomes. The prevention cohort starts cream with erlotinib, whereas the treatment cohort starts cream when rash develops. A contrast between these cohorts would not by itself distinguish treatment effects from differences related to cohort entry or timing.

The side-specific arrangement in the prevention cohort also creates a paired structure: two sites on the same participant receive different topical procedures. In general, observations from the same person may be more alike than observations from different people. Treating paired observations as independent can misstate uncertainty. A suitable analysis might account for within-person pairing, but the registry record does not report whether paired-side data were analyzed, nor does it describe the method.

What would a confidence interval add?

A confidence interval describes uncertainty around an estimated quantity under the assumptions of the analysis. It helps distinguish a precise estimate from one that could vary substantially across plausible values. The interval does not repair weaknesses in study design, remove confounding, or establish causation. For this trial, primary outcome estimates and confidence intervals are not reported in the registry record, so the precision of the rash findings cannot be evaluated.

A p-value, if reported, would address how compatible the observed data are with a specified statistical model or null hypothesis. It would not measure the size or clinical importance of an effect, and it would not establish that a treatment caused an outcome. No primary endpoint p-value or statistical test is reported in the registry record. A missing p-value should not be interpreted as proof that a comparison was nonsignificant.

How should multiple outcomes or categories be considered?

When an analysis examines multiple outcomes, time points, groups, or severity categories, the chance of at least one apparently unusual result can increase if each is tested separately without adjustment. Statistical approaches to multiplicity can control or describe that issue, but the appropriate choice depends on which questions were prespecified and how results are intended to be interpreted. The record lists two outcome measures and notes efficacy assessments at Week 2 and Week 4, but it does not report a multiplicity strategy, the hierarchy of analyses, or whether all outcomes and time points were analyzed.

Severity categories also require careful interpretation even without formal hypothesis tests. The endpoint description permits the same participant to be counted in more than one category. Thus, category-specific results would need to be understood according to their exact definitions and denominators, rather than read as mutually exclusive groups unless the analysis explicitly established that structure.

What does a planning calculation mean?

In general, a sample-size calculation uses assumptions about the effect to detect, variability or event frequency, significance level, allocation, and desired power. It estimates how many observations a future design would need under those assumptions; it does not describe how much power the completed trial actually had. The registry’s planning assumptions describe a two-sided significance level of 0.05, equal allocation, a normal approximation, and effects defined as fractions of an observed effect. Any resulting values are planning illustrations, not observed power. The calculated sample-size values and the observed primary effect are not reported in the registry record, so no particular planning estimate can be interpreted from this information.

For a binary percentage endpoint, the observed estimate is based on event counts and an analysis denominator. Small denominators can make estimates less stable, and unequal or outcome-dependent evaluability can complicate comparisons. These are general considerations, not claims about the unreported primary outcome results in this trial.

10. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a multiplicity plan. It describes efficacy evaluation at Week 2 and Week 4, lists two outcome measures, and names a Week 2 rash-severity endpoint as primary. Without an analysis plan, it is not possible to determine which comparisons were prespecified, whether the Week 4 assessment was secondary, or whether any adjustment was planned for multiple outcomes, time points, grades, or cohorts. Multiple testing is a general concern when many analyses are considered: results can appear noteworthy by chance if the number of tests is not accounted for. No claim about such an adjustment can be made from this record.

Missing-data procedures are not reported in the registry record. The endpoint population is described as all enrolled participants, but the number analyzed is defined as participants evaluable for the outcome in their respective arms. The record does not give those evaluable counts for the primary endpoint, describe why assessments might be unavailable, or say whether missing observations were excluded, imputed, or addressed in a sensitivity analysis. Consequently, the extent and possible influence of missing outcome data cannot be assessed.

The analysis population is not described in greater detail. The record does not specify an intention-to-treat analysis, a per-protocol analysis, or another named population for the primary endpoint. It also does not explain whether all enrolled participants were included in safety reporting or how treatment exposure affected inclusion. These details matter because an analysis restricted to evaluable participants can differ from one that includes all assigned or enrolled participants, particularly if evaluability is related to outcome or treatment.

There are no posted primary analyses in the record. The design description and endpoint definition therefore should not be mistaken for evidence that particular missing-data or multiplicity methods were applied.

11. 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: death (18.8%), dyspnea (6.2%), falling (6.2%), diarrhea (5.6%).

The registry lists serious adverse events for two groups. In the prevention group, 3 of 18 participants at risk were affected; in the treatment group, 5 of 16 were affected. Other adverse events are listed for 6 of 18 and 1 of 16, respectively. These are descriptive counts as recorded, not estimates from a randomized comparison. The registry does not report an uncertainty interval or adjusted comparison for these figures.

The serious-event terms include Death at 18.8%, Dyspnea at 6.2%, Falling at 6.2%, and Diarrhea at 5.6%. The registry record does not state the denominator used for these term percentages or how the terms relate to the group-level counts. It lists the number of deaths as not reported. Therefore, the Death term percentage should not be read as a confirmed count of deaths or combined with the group-level event counts to infer one.

These safety figures do not establish that an event was caused by metronidazole or erlotinib. The registry record does not provide event-level attribution, timing, duration, or a causal assessment. Differences in the descriptive counts between the cohorts also cannot be interpreted as treatment effects, because allocation was non-randomized and the cohorts followed different treatment-start strategies. The counts describe what is posted, with those limitations.

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

The record documents a phase II study of topical metronidazole used either from the start of erlotinib treatment or after rash developed. It specifies rash severity at Week 2 as the primary endpoint and describes a CTCAE v3.0 grading approach. It also provides adverse-event summaries for the two groups. These registry entries clarify the study question and the structure of the reported safety data.

The record does not report the primary endpoint percentages, grade-specific participant counts, or statistical analyses. It therefore does not show how rash severity was distributed, how outcomes compared between strategies, or how precisely any estimate was measured. The absence of posted results is not evidence that the cream had no effect; it means the effect cannot be quantified from the registry record.

Nor does the design support a simple causal comparison between the prevention and treatment cohorts. Their treatment timing differs, and allocation was not randomized. The prevention cohort’s side-specific cream and local-care arrangement may provide a within-participant contrast, but no results or analysis details for that contrast are reported. Without these results and methods, a causal effect cannot be inferred.

The adverse-event counts are descriptive and should be read separately from the primary rash endpoint. The record does not report causality assessments or a formal comparison. The listed Death term percentage does not resolve the unreported death count, because the relevant denominator and relationship to that count are unspecified.

In general, an outcome estimate is most informative when accompanied by a clear analysis population, denominator, measure of uncertainty, and prespecified comparison. Those elements are not available for the primary endpoint in the registry record. The sound interpretation is limited to the documented design, outcome definition, and posted safety counts; no claim that metronidazole prevents or treats erlotinib-associated rash is supported by the posted results as described.

13. Why This Trial Matters Statistically

This trial raises several statistical issues that are common in studies of treatment-related symptoms. First, prevention and treatment are different questions. Starting an intervention before an event occurs is not the same strategy as starting it after the event appears. Combining these settings or comparing their outcomes without accounting for timing can obscure what a result means.

Second, the non-randomized design affects interpretation. Randomization is one way to reduce systematic differences between groups; without it, observed differences may reflect participant or clinical differences as well as treatment. The registry does not report methods that would address such differences. This is especially relevant because the two cohorts were defined by different treatment-start circumstances.

Third, the prevention cohort uses a side-specific comparison, which raises a dependence issue. Measurements from two body sites on one participant are generally not equivalent to measurements from separate, unrelated people. A statistical method needs to respect the paired structure if side-level results are compared. The registry record does not say whether this was done.

Finally, outcome reporting and precision are essential. A severity-grade endpoint can convey more detail than a yes-or-no event measure, but its interpretation depends on the category definitions, denominators, and rules for participants who may appear in more than one category. The registry states that multiple-category counting may occur but does not post the endpoint results or analysis method. Without those details, readers cannot judge the size or precision of the rash findings.

These issues matter because a study’s statistical conclusion depends not only on the outcome label, but also on the comparison, analysis population, handling of repeated or paired observations, and uncertainty around estimates. The record documents design choices that make those considerations especially important, while leaving the corresponding analysis results unreported.

14. 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 for interpreting efficacy is that the registry record does not report the primary endpoint results or any analyses. No grade-specific percentages, evaluable counts, effect estimates, confidence intervals, or p-values are listed. Consequently, the observed rash distribution and its statistical precision cannot be assessed.

Allocation was non-randomized, and masking was absent. The two cohorts also differed in treatment timing: cream began with erlotinib in the prevention cohort and only after rash developed in the treatment cohort. These features limit the interpretation of any between-cohort contrast. The record does not report adjustment methods for potential differences between participants or cohorts.

The prevention strategy included metronidazole on the right side and local standard care on corresponding left-side areas. The registry record does not report side-specific results or explain whether the analysis accounted for pairing within participants. It also does not clarify how side-level observations relate to the stated participant-level primary endpoint.

The primary endpoint description allows a participant to be counted in more than one severity category. The record does not explain how category overlap was handled in the analysis. The number evaluable for the endpoint in each arm, reasons for non-evaluability, and missing-data methods are also not reported. Multiplicity handling across the listed outcomes and assessment times is not reported.

Safety reporting is limited to group-level counts and a list of serious-event terms. Denominators for the term percentages, event-level details, attribution, and the number of deaths are not reported in the registry record. These limitations prevent a detailed safety comparison and preclude causal conclusions from the descriptive counts.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: A Study of Metronidazole Cream in the Prevention and Treatment of Tarceva (Erlotinib)-Associated Rash — NCT00642473.
• Registry record last updated 2016-05-23; status last verified 2016-04.
• Results first posted 2016-05-23.
• 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.

18. Record Summary

The registry record describes a completed phase II, non-randomized, unmasked, parallel study enrolling 34 participants with non-squamous non-small cell lung cancer. It evaluates metronidazole Actavis 1% topical cream alongside erlotinib in a prevention cohort and a treatment cohort. Prevention treatment began with erlotinib; treatment-cohort cream began if rash developed. Cream was applied twice daily to the right side of the face and chest, with local standard procedures on corresponding left-side areas.

The primary endpoint is the percentage of participants by erlotinib-associated rash severity grade at Week 2, assessed using CTCAE v3.0. The registry record does not report the endpoint results or statistical analyses. It lists serious adverse events affecting 3 of 18 participants in the prevention group and 5 of 16 in the treatment group, along with other adverse-event counts. The serious-event term percentages are listed, but their denominators and the number of deaths are not reported. The posted information supports a description of the design and reported safety figures, not a quantified conclusion about rash prevention or treatment.