FDA's New Drug Safety Reporting Plan

[Published as a guest column in the October 14, 2010 edition of Law360.]

On September 28, 2010, FDA issued a Final Rule governing the safety analysis and reporting requirements for drugs under clinical investigation, even after they are on the market.  Citing the European model and seeking greater harmonization with international standards, FDA defined new terms and revised old terms.  Among its goals:  to improve patient safety by limiting the number of adverse reaction reports it receives and increasing the usefulness of each report.  Greater safety by fewer reports?  This is a paradox, of course, in a world often clamoring for more FDA oversight and greater transparency.  Yet FDA sees trouble finding real signals in the ocean of adverse event reporting it now receives. 

In this rule, FDA clarifies that the necessary reports must be filed quickly; FDA also expressly requires systematic analysis and expedited reporting of new kinds of safety information, not previously covered by the rule.  And for the first time, persons conducting bioavailability or bioequivalence studies must expedite reports of all serious adverse events occurring during the study, regardless of their expectedness or causality. 

The new rule, contained primarily within 21 C.F.R. § 312.32 and § 320.31, will take effect on March 28, 2011.  Concurrent with the publication of the Final Rule, FDA issued a draft Guidance for Industry and Investigators:  Safety Reporting Requirements for INDs and BA/BE Studies.  The comment period will remain open for 90 days.

Noise in the System

In many respects, the new rule is designed to emphasize those occurrences that actually may be caused by a drug.  FDA understands that adverse events will happen whenever relatively large groups of people are observed for any real length of time, and often there is no real possibility that the study drug had anything to do with causing the event. 

FDA always intended, it now says, for sponsors to submit adverse reactions in special or expedited form only when there was "a reasonable possibility" that the experience may have been caused by the drug.  These words in fact define those events that are subject to reporting under the historic and still-current rule, which requires reporting of serious and unexpected adverse experiences that are "associated with the use of the drug," which is defined as meaning "there is a reasonable possibility that the experience may have been caused by the drug."  Yet sponsors, FDA says, began submitting expedited adverse reaction forms for all kinds of events, often when there was no real possibility that the drug caused the event.

This "noise" in the system, FDA fears, distracts its time and attention, drowning out the true potential threats to human safety.

Sponsors, Declare Your Risk

FDA seeks to change this reality.  Now, FDA requires, sponsors must submit expedited safety reports within 15 days of determining that an adverse event occurring during a clinical trial is both "serious" and "unexpected" and also is a "suspected adverse reaction." 

These reports must be made to FDA and to all investigators to whom the sponsor is providing its drug for active clinical trials.  Numerous comments requested revised reporting formats to investigators.  FDA held its proposed ground, indicating that investigators too have become overwhelmed with difficult-to-interpret individual case reports, while affirming its belief that they too should have access to emerging safety information.  

The definition of a serious event has not changed much in the new rule, except that the determination may be made by either the investigator or the sponsor.  The definition of unexpected has not changed much either, except that FDA has emphasized that the Investigator Brochure, given to investigators by the sponsor at the outset of a clinical trial, itself must be the basis for this determination.  In other words, if the precise event type – of equal severity and specificity – has not been declared as having been observed in people taking the particular drug under study, then the event is "unexpected" under the rule. 

In keeping with this insistence that the Investigator Brochure spell out the true expected risk, sponsors now must report serious suspected adverse reactions when there is a "clinically important increase" in their rate, as compared to that listed in the brochure.

The real change, though, comes with the term "suspected adverse reaction."  This is any adverse event for which there is "a reasonable possibility" that the drug caused the adverse event.  Look familiar?  So far, of course, it is.  These same words described reportable reactions before.  The new rule goes on to define this term, stating that "reasonable possibility" means "there is evidence to suggest a causal relationship" between the drug and the event. 

What does this actually mean?  FDA clearly intends it to mean an intermediate level of suspicion, where a causal relationship has not been proved, but there is some real reason to believe the drug may be causing the events.  This may be difficult to determine, and the words themselves are subject to many interpretations.  In hindsight for example, after evidence mounts, almost any report could be viewed as having constituted "evidence" that suggested a causal relationship at an earlier time.  This could prompt sponsors, still, to err on the side of expediting reports.

What FDA intends, however, is that a sponsor will not submit an expedited safety analysis unless the sponsor itself has made a determination that the event type warrants some special consideration.  FDA hopes to change the game, so that the very filing of an expedited report is a declaration by the sponsor:  there is reason to pay attention to this event type.  Something real could be happening with our drug. 

As all good product liability lawyers know, these analyses could become Opening Statement evidence in later lawsuits.  And the expectation that the reported event types are more genuinely viewed as having serious risk potential, by the drug sponsors themselves, may increase the likelihood that the analyses can be used effectively against the sponsors.

 "Evidence to Suggest…"

The rule includes examples of the kinds of events that may indicate a causal relationship.  Similar concepts have long been included in signal detection guides, but their incorporation into the rule is new.  "Evidence to suggest a causal relationship" may include the single occurrence of an uncommon event known to be strongly associated with drug exposure, such as Stevens-Johnson Syndrome; one or more occurrences of an event that is not commonly associated with drug exposure, but is otherwise uncommon in the study population, like tendon rupture; and an aggregate analysis of specific observed events that indicates they occur more frequently in the drug group than in the placebo or historical control groups. 

This sounds straightforward, and FDA estimates that completing a safety report would take only 4 to 12 hours.  If only things were so easy.  It can be hard enough to interpret the particular circumstances of an individual adverse event, or to interpret multiple events in large trial groups.

As FDA itself has acknowledged, there also are serious difficulties in interpreting pooled data.  When specific events are aggregated across multiple trials, there are many opportunities for statistical uncertainty and instability to be introduced.  In the same month this Final Rule was announced, FDA personnel published a Perspective in the New England Journal of Medicine highlighting a situation where pooled analyses led to a major concern that the drug Tiotropium increased stroke and death in people with chronic obstructive pulmonary disease.  A large, four-year randomized clinical trial found no such risk.  In concluding, the FDA authors noted that findings from pooled and meta-analyses often lead to urgent calls for regulatory action, without acknowledgement of the potential pitfalls in these analyses. 

The Guidance document offers many suggestions for interpreting pooled data.  We increasingly live in a world where many people will have access to huge aggregated event databases.  Unfortunately, there is no clean answer, yet, to the vast uncertainties introduced.  We remain in serious danger that the noise will appear, convincingly, to have true "risk" sound.

Keeping the Blind?

During the comment period for the draft rule, sponsors raised the concern that the expectation of a contextual analysis to evaluate most adverse reactions may lead to treatment assignment unblinding that could jeopardize trial results.   

FDA sought to address this concern by separately treating study endpoints.  When trials are designed specifically to study the drug's effect on a particular outcome, such as heart attack, stroke, or death, FDA says the event reports typically should be made as set forth in the study protocol.  When pre-specified study endpoint events occur, premature unblinding normally would not be required, unless there is evidence suggesting a causal relationship.  When might this occur?  The example included in the rule is death from anaphylaxis in a trial with all-cause mortality as a study endpoint.  Again, it is easy to imagine that in hindsight, judgments made in less clear situations may be challenged.

And again, FDA asks sponsors to evaluate their data and take a stand.  Do these data suggest a potentially "real" reaction?  Is this something new, which we did not expect?  If not, then no reporting is required.  If a report is filed, however, an unexpected risk has been suspected.

The issue of unblinding may be even more difficult in the context of non-endpoint events.  It has long been true that unblinding may occur for serious events, particularly when information about the drug could be crucial to treating the immediate medical needs.  Yet it is not difficult to anticipate that the focus on contextual evaluations, and the expectation that a sponsor declare a suspected risk, may lead to more frequent unblinding of potentially related events. 

A Significant Risk

Whenever a potentially "real" risk is suggested, through whatever means, the new rule requires expedited reporting to FDA and to all study investigators. 

The new rule expressly states broader duties for clinical trial sponsors to monitor their safety data, from any source, and to report the discovered risks quickly and with narrative, contextual analysis.  Sponsors must submit expedited reports within 15 days of determining that other safety information from any source, including epidemiologic studies, aggregate analyses from multiple studies and findings from animal or in vitro testing, "suggest a significant risk" in people who take the drug.   

The phrase "suggest a significant risk" is not defined, but the FDA offers this context:  ordinarily, such a finding would result in a safety-related change in the protocol, leading to revised risk disclosures to investigators and study participants, and perhaps affecting the overall conduct of the trial.  

Why Monitor Generics?

In contrast to the large-scale clinical trials often done to investigate new drugs, the bioavailability and bioequivalence studies for generic drugs typically are small, and brief, and conducted in healthy volunteers.  This means that all serious adverse events are of interest, regardless of apparent causality and even when the same type of event has been observed with the reference listed drug. 

For this reason FDA welcomes, and now requires, the "noise" of these reports. 

The new rule governing these studies is more limited, requiring only the submission of serious adverse event reports to FDA and all study investigators.  Yet this is a new development, applying to studies conducted within the United States even when they are exempt from IND requirements.  In addition, experience from foreign studies should be included in the ANDA submission, as part of the information required to establish that the proposed drug product can be expected to have the same therapeutic effect as the reference listed product. 

This is Not: The End

The Final Rule contains additional details and the accompanying Guidance document offers further nuance.  Investigational drug sponsors will need to evaluate the new requirements carefully and be prepared to implement them quickly. 

To the extent companies could improve the situation outlined in the Guidance document, it may be wise to submit comments to FDA.  Although this rule reflects the agency's rejection of many expressed concerns, there were multiple ways the rule was shaped by comments received.