To All - Below are excerpts from a draft FDA document (not formal guidelines yet) very relevant to discussions of the SEs in R-It. (I received this coincidentally this morning for something unrelated). I have bolded some key areas. After reading this, I realized that whether the DMC analyzes SEs using a strict Type I error corrected p value (which would be more stringent than the p<.022 for the PE) depends on whether they intend to go for labeling that states that V specifically reduces [death, stroke, whatever]. Are they simply trying to get it labelled as "reduces incidence of CV events"? If the latter, the text below clearly indicates they don't have to have any formal Type I error adjusted statistical criterion for interpreting the SEs, and that the label would be very cautious about any conclusions other than overall MACE events. The MACE component SEs in this case would simply be reported as % to underscore that the effects of V were not due to one single dominant outcome, but rather across all key MACE components (This seemed to be the case in JELIS). This seems to me to be the gist of the rationale they have implied for looking at the SEs ("robust and consistent"), and the statements in the past that the SEs do NOT have to be statistically significant would also support this view. On the other hand, the fact that they have bothered to pre-specify SEs in their most recent FDA SPA might hint that they DO plan to look formally at key SEs as potential labelled indications.
While it would clearly be nice to say on the label that V reduces deaths (or any specific component), I have seen nothing from the company that would suggest they are aiming that high. Anyone else know about this? It also seems from what is stated below that their SE MACE component analyses could in theory use events after the first trigger MACE event, which would mean total event numbers for MACE components would be greater than the number of MACE events (and increase statistical power).
Excerpts Below:
The focus of this guidance is control of the Type I error rate for the planned primary and secondary endpoints of a clinical trial so that the major findings are well supported and the effects of the drug have been demonstrated. Once a trial is successful (demonstrates effectiveness or “wins” on the primary endpoint(s)), there are many other attributes of a drug’s effects that may be described. Analyses that describe these other attributes of a drug can be informative and are often included in physician labeling. Examples include: the time course of treatment effects; the full distribution of responses amongst participants; treatment effects on the components of a composite endpoint;and treatment effects amongst subgroups.
Nevertheless, it is important to understand that these descriptions with respect to additional attributes are not demonstrated additional effects of a drug unless the analyses were prespecified, and appropriate multiplicity adjustments were applied. Therefore, presenting p-values from descriptive analyses (that is, from analyses that were not prespecified and for which appropriate multiplicity adjustments were not applied) is inappropriate because doing so would imply a statistically rigorous conclusion and convey a level of certainty about the effects that is not supported by that trial.
Composite Endpoints
There are some disorders for which more than one clinical outcome in a clinical trial is important, and all outcomes are expected to be affected by the treatment. Rather than using each as a separate primary endpoint (creating multiplicity) or selecting just one to be the primary endpoint and designating the others as secondary endpoints, it may be appropriate to combine those clinical outcomes into a single variable. This is called a “composite endpoint,” where an endpoint is defined as the occurrence or realization in a patient of any one of the specified components. When the components correspond to distinct events, composite endpoints are often assessed as the time to first occurrence of any one of the components, but in diseases where a patient might have more than one event, it also may be possible to analyze total endpoint events. A single statistical test is performed on the composite endpoint; consequently, no multiplicity problem occurs and no statistical adjustment is needed.
An important reason for using a composite endpoint is that the incidence rate of each of the events may be too low to allow a study of reasonable size to have adequate power; the composite endpoint can provide a substantially higher overall event rate that allows a study with a reasonable sample size and study duration to have adequate power. Composite endpoints are often used when the goal of treatment is to prevent or delay morbid, clinically important but uncommon events (e.g., use of an anti-platelet drug in patients with coronary artery disease to prevent myocardial infarction, stroke, and death).
The choice of the components of a composite endpoint should be made carefully. Because the occurrence of any one of the individual components is considered to be an endpoint event, each of the components is of equal importance in the analysis of the composite. The treatment effect on the composite rate can be interpreted as characterizing the overall clinical effect when the individual events all have reasonably similar clinical importance. The effect on the composite endpoint, however, will not be a reasonable indicator of the effect on all of the components or an accurate description of the drug’s benefit, if the clinical importance of different components is substantially different and the drug effect is chiefly on the least important event. Furthermore, it is possible that a component with greater importance may appear to be adversely affected by the treatment, even if one or more event types of lesser importance are favorably affected, so that although the overall outcome still has a favorable statistical result, doubt may arise about the treatment’s clinical value. In this case, although the overall statistical analysis indicates the treatment is successful, careful examination of the data may call this conclusion into question. For this reason, as well as for a greater depth of understanding of the treatment’s effects, analyses of the components of the composite endpoint are important (see section III.D) and can influence interpretation of the overall study results.
For composite endpoints whose components correspond to events, an event is usually defined as the first occurrence of any of the designated component events. Such composites can be analyzed either with comparisons of proportions between study groups at the end of the study or using time-to-event analyses. The time-to-event method of analysis is the more common method when, within the study’s timeframe of observation, the duration of being event-free is clinically meaningful. Although there is an expectation that the drug will have a favorable effect on all the components of a composite endpoint, that is not a certainty. Results for each component event should therefore be individually examined and should always be included in study reports. These analyses will not alter a conclusion about the statistical significance of the composite primary endpoint and are considered descriptive analyses, not tests of hypotheses. If there is, however, an interest in analyzing one or more of the components of a composite endpoint as distinct hypotheses to demonstrate effects of the drug, the hypotheses should be part of the prospectively specified statistical analysis plan that accounts for the multiplicity this analysis will entail, as described above for mortality.
In analyzing the contribution of each component of a composite endpoint, there are two approaches that differ in how patients who experience more than one of the event-types are considered.
One approach considers only the initial event in each patient. This method displays the incidence of each type of component event only when it was the first event for a patient. The sum of the first events across all categories will equal the total events for the composite endpoint. The other approach considers the events of each type in each patient. With this method, each of the components can also be treated as a distinct endpoint, irrespective of the order of occurrence, giving the numbers of patients who ever experienced an event of each type. In this case, each patient can be included in the event counts for more than one component, and the sum of events on all component types will be greater than the total number of composite events using only the first events.
The different components of a composite endpoint are selected because they are all clinically important; however, because each one is not necessarily equally affected by the drug, it is relevant and important to examine the effects of the drug on the individual components as well as on the overall endpoint. Presenting only data on the composite might imply meaningful treatment effects on all of the individual components, when a composite effect may in fact be established with little or no evidence of effect on some of the individual components. On the other hand, showing the results of the analysis for each of the individual components may imply an effect on an individual component when an appropriate statistical analysis would not support that conclusion. Thus, it is important to present descriptive analyses of between-group differences for the components in a way that does not overstate the conclusions.
It is common for one component of a composite endpoint to overly influence the treatment effect, but even if that is not so, and all components contribute, the inclusion of a particular component in a composite does not usually support an independent conclusion of efficacy on that component. FDA’s guidance for industry Clinical Studies Section of Labeling for Human Prescription Drug and Biological Products — Content and Format calls for presentation in labeling of the components of a composite endpoint but without a statistical analysis of the separate components unless the components were prespecified as separate endpoints and assessed with a prospectively defined hypothesis and statistical analysis plan. In such a case, the statistical analysis will usually consider all events of each type, not just first-occurring events. Only findings on prespecified endpoints that are statistically significant, with adjustment for multiplicity, are considered demonstrated effects of a drug. All other findings are considered descriptive and would require further study to demonstrate that they are true effects of the drug.
To demonstrate an effect on a specific component or components of a composite endpoint, the component or components should be included prospectively as a secondary endpoint for the study or possibly as an additional primary endpoint (see section III.C.5), with appropriate Type I error rate control. If control of the Type I error rate is ensured with respect to the individual component or components, in addition to control for the composite, a trial will be potentially able to support conclusions regarding drug effects on the individual component or components as well as the composite.
AI
