Tysabri and the Risk of Progressive
Multifocal Leukoencephalopathy (PML)2
SUMMARY
In February 2005, four months after its initial approval to treat patients with multiple sclerosis, Tysabri was withdrawn from the market because of concern about the risk of a life-threatening, frequently fatal, brain infection, PML. At the time, there was considerable uncertainty about the magnitude of the risk of PML to patients exposed to Tysabri and whether there were any identifiable risk factors that could be reliably used to identify patients at greater risk. In making its decision on whether to allow re-marketing of the drug, FDA considered whether the risk of PML (and uncertainty about the risk) outweighed the drug’s recognized substantial benefit.
__________________
1 Case studies were developed by FDA staff to inform the discussions at the IOM workshops. Text is reproduced here as originally submitted.
2 Version Date: February 10, 2014. This summary was developed for purposes of discussion at the February 12–13, 2014, IOM/FDA Public Workshop: Characterizing and Communicating Uncertainty in the Assessment of Benefits and Risks.
INTRODUCTION3
Tysabri (natalizumab) was originally approved in 2004 for relapsing forms of multiple sclerosis (MS), a progressively debilitating neurological disease. There is no known cure for MS and the disease frequently progresses to severe disability and/or death. Approximately 400,000 individuals currently live with MS in the U.S.4
Tysabri represented a novel treatment mechanism for MS, believed to work by interfering with the movement of inflammatory white blood cells from the blood vessels into the brain and spinal cord. It is administered through intravenous infusion, typically on a monthly basis. Other effective treatments such as interferons and Copaxone (glatiramer acetate) were available at the time, but a substantial number of patients with relapsing MS remained untreated for many reasons, including lack of efficacy or tolerability of existing treatments. Tysabri appeared to be substantially more effective, with a significantly greater reduction in relapse rates over these treatments.
Previously-approved drugs had required clinical trials showing evidence of benefit through two years. The results of Tysabri were so promising in the first year that the drug was granted an accelerated approval5 on the basis of 1-year interim results of the clinical trials. At the time of approval, Tysabri’s safety profile was acceptable given the demonstrated benefits.
In February 2005, four months after approval, the sponsor notified FDA of two cases of PML, a rare, frequently fatal viral infection in the brain. Both cases occurred in patients receiving Tysabri for MS. In light of this, the sponsor voluntarily withdrew Tysabri from the market and suspended all clinical trials. By that time, about 7,000 patients had received at least one dose of Tysabri.
In September 2005, after conducting an extensive safety assessment, the sponsor submitted additional efficacy and safety evidence and requested reauthorization to market the drug. FDA’s re-marketing decision hinged in large part on its conclusions about the risk of PML associated with Tysabri.
__________________
3 In addition to the links embedded in the summary text, for relevant material from FDA on this topic see http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm107198.htm (accessed August 23, 2014).
4 Figure reported by the National Multiple Sclerosis Society, see http://www.nationalmssociety.org/about-the-society/ms-prevalence/index.aspx (accessed August 23, 2014).
5 Federal regulations (21 CFR 601 subpart E) allow accelerated approval of new biologic therapies that provide meaningful benefit over existing treatments for serious or life-threatening illnesses. As a condition of the accelerated approval, the sponsor was required to verify that the demonstrated clinical benefit and safety was sustained in the second year.
In its review, FDA considered the following questions:
- What is the magnitude of risk of PML to patients exposed to Tysabri?
- Are there identifiable factors that can be used to screen for patients at higher risk for PML or otherwise mitigate the risk of PML?
ASSESSING THE RISK OF PML (BASED ON AVAILABLE EVIDENCE IN 2006)
The sponsor’s safety assessment included a comprehensive clinical and neurological evaluation of the clinical trial subjects who had been exposed to Tysabri. FDA determined that the design, implementation, and analysis of the safety assessment were adequate. The assessment confirmed the two previously-identified PML cases but did not identify any other cases, out of 1869 MS patients treated for a median duration of 120 weeks. An additional case, however, was identified in a patient with Crohn’s disease (the sponsor had also been assessing the drug for treatment of this gastrointestinal disease). In total, 3 cases were identified in a population of ~3,000 patients exposed in the clinical trials.6
Below are some of the additional factors pertinent to FDA’s 2006 assessment of the risk of PML:
- PML is extremely rare in the general population. The death rate associated with PML in the U.S. in 2002-2005 was 0.66/1,000,000.7 The disease typically affects individuals with suppressed immune systems. PML had not previously been associated with MS.
- Tysabri’s mechanism of action raised a theoretical concern about an increase in risk of infections. In the clinical trials, the overall risk of infections (serious and non-serious) was similar for Tysabri vs. placebo. However, the drug appeared to cause an increased rate of specific types of serious atypical and opportunistic infections, including viral meningitis, herpes infections, pulmonary infections, gastrointestinal infections, and, of course, PML.
- Both MS PML cases were patients who took Tysabri in conjunction with an immunomodulating interferon drug, Avonex. However, use of interferons alone had not been associated with PML. The
__________________
6 A report of the sponsor’s safety assessment, published in 2006, reported a 95 percent confidence interval of 0.2/1000–2.8/1000. See Yousry, T.A., et al. 2006. Evaluation of patients treated with natalizumab for progressive multifocal leukoencephalopathy. N Engl J Med 354(9):924-933.
7 Christensen, K. L., et al. 2010. Progressive multifocal leukoencephalopathy deaths in the USA, 1979-2005. Neuroepidemiology 35(3):178-184.
-
Crohn’s patient also had a significant history of immunosuppressive use.
- PML is thought to be caused by the John Cunningham virus (JCV), which is carried in a dormant state by many people. At the time, there was no reliable test to detect JCV exposure.
- Diagnosis of PML requires clinical, neuroimaging, and virologic evidence. Early diagnosis of PML in MS patients can be particularly challenging because the signs and symptoms of PML are similar to those of MS.
- The majority of patients in the MS trials were enrolled for ≥2 years. Both MS PML cases occurred after approximately 2 years of exposure. However few patients in the development program had been exposed for 3+ years, resulting in uncertainty of the effect of longer-term exposure on the magnitude of the PML risk.
- In 2005, there was no known effective treatment for PML.
IMPACT OF UNCERTAINTY ON FDA’S DECISION-MAKING
The submitted additional evidence increased FDA’s confidence that the PML cases were caused by Tysabri and that the overall risk of PML was on the order of 1 in 1,000 patients. However, the assessment did little to resolve uncertainty about whether the risk was related to any underlying factors, particularly concomitant use of immunosuppressing drugs, JCV exposure, and duration of Tysabri use. No other new serious safety concerns were identified in FDA’s review to re-market Tysabri. Further, the additional efficacy evidence submitted in response to the accelerated approval requirement (the results of the 2-year study) strengthened FDA’s assessment of the drug’s benefit. Therefore, the question FDA faced was whether the risk of PML (and residual uncertainty about that risk) outweighed the substantial benefit of the drug to MS patients.
In March 2006, FDA convened an advisory committee (AC) meeting to gather expert input into the Agency’s decision. At the meeting, patients, family, and health care providers testified to the difference that Tysabri had made in the lives of MS patients, as well as the willingness of patients to continue treatment despite the risk of PML. The AC members voted unanimously to reintroduce Tysabri to the market. However, they voted unanimously that the drug should be restricted to use: (a) as monotherapy (i.e., not to be used in combination with other MS drugs), and (b) in patients with relapsing forms of MS. They were divided, however, on whether the drug should be limited to use only as a second-line therapy (i.e., only in patients who have not responded adequately to or are intolerant of other treatments).
FDA concurred with the 2006 AC recommendations. As FDA con-
cluded in its 2006 decision memo,8 “in the face of these potential risks, the benefit of treatment with Tysabri clearly justifies its re-introduction into the market [with certain requirements] . . . and that physicians and patients should be given the opportunity to decide if this treatment is appropriate in any given case.” The Agency remained concerned, however, about the inability to (a) identify individual patients who are at greater risk of contracting PML, and (b) to mitigate death or other serious effects of PML. FDA cautioned that “if marketing is permitted, we fully expect that additional cases of PML, many likely to be fatal, will occur . . . and it is a fact that patients, their families, and prescribers will need to consider very seriously.”9
Thus, in June 2006, Tysabri was allowed back on the market, with a number of requirements:
- The indication was restricted to use as monotherapy and only in patients with relapsing forms of the disease. Labeling further recommended use as a second-line treatment.
- A boxed warning and detailed safety information on PML, including what was known and not known about the risk, was included in the revised (2006) product labeling.10
- The risk management plan included a restricted distribution program (called TOUCH) requiring prescribing and administration only by certified health care providers, use only in patients enrolled in the program and who met necessary conditions, documented patient counselling, and patient evaluation at 3 and 6 months after initial treatment and every 6 months thereafter.
- A prospective, observational post-marketing safety study (called TYGRIS) was required, enrolling 5,000 patients followed for 5 years to monitor for safety concerns, including PML and other serious infections.
SINCE THE 2006 RE-MARKET AUTHORIZATION
Tysabri was approved for the Crohn’s disease indication in January 2008. At that time, no new PML cases had been reported in either the post-
__________________
8 See http://www.accessdata.fda.gov/drugsatfda_docs/label/2006/125104Orig1s015SumR.pdf (accessed August 23, 2014).
9 Opening remarks by Dr. Russell Katz, Director of the Division of Neurology Products, at the 2006 FDA Advisory Committee Meeting on the subject of Tysabri remarketing. http://www.fda.gov/ohrms/dockets/ac/06/transcripts/2006-4208T1.pdf (accessed August 23, 2014).
10 For more information see http://www.accessdata.fda.gov/drugsatfda_docs/label/2006/125104s015LBL.pdf (accessed August 23, 2014).
market setting or in the sponsor’s drug development program. However, there remained uncertainty about factors that contributed to the risk of PML. Similar to the MS indication, the Crohn’s indication was limited to use as a second-line therapy and only for patients with moderate to severe symptoms. It also required restricted distribution through the TOUCH program.
The first two post-market cases of PML since the drug’s 2006 remarketing were reported to FDA in the summer of 2008. At that time, ~39,000 patients had received treatment with Tysabri worldwide, with ~12,000 patients having been treated for at least one year. Both cases occurred in MS patients in Europe, and both patients were receiving Tysabri as monotherapy. However, FDA concluded in an August 2008 advisory11 that it “still believes that Tysabri monotherapy may confer a lower risk of PML than when Tysabri is used together with other immunomodulatory medications.”
Today, the risk of PML associated with Tysabri is much better understood. As of September 3, 2013, 401 cases of PML had been reported worldwide. Table 1 presents the stratified risk estimates for the U.S. population, showing how the risk depends on length of exposure, presence of anti-JVC antibodies, and immunosuppressant use. The current product labeling12 recommends considering these factors in the context of expected benefit when initiating and continuing treatment. Additionally, an anti-JCV antibody detection test was cleared13 by FDA in January 2012.
Table 1. Estimated United States Incidence of PML Stratified by Risk Factor
| Anti-JVC Antibody Negative | TYSABRI Exposure | Anti-JVC Antibody Positive | |
| No Prior Use of Immunosuppressant | Prior Use of Immunosuppressant | ||
| <1/1,000 | 1–24 months | <1/1,000 | 1/1,000 |
| 25–48 months | 3/1,000 | 13/1,000 | |
| 49–72 months | 7/1,000 | 9/1,000 | |
The experience with Tysabri has also heightened FDA’s awareness and consideration of potential drug-induced PML risk. For example, a
__________________
11 For more information see http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm126592.htm (accessed August 23, 2014).
12 For more information see http://www.accessdata.fda.gov/drugsatfda_docs/label/2013/125104s840s847s889lbl.pdf (accessed August 23, 2014).
13 For more information see http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm288471.htm (accessed August 23, 2014).
central question in a 2013 AC meeting on vedolizumab, another drug indicated for Crohn’s disease with a similar mechanism of action, was whether the applicant had adequately characterized the potential risk of PML.14
FDA’S COMMUNICATIONS ABOUT PML15
FDA first issued a public health advisory16 on Feb. 28, 2005, announcing the first two PML cases and Tysabri’s market withdrawal. FDA issued a press release17 at the time re-marketing was approved, which outlined the narrower indication and the TOUCH program. In addition to the 2008 advisory18 described above, FDA has issued three Drug Safety Communications: Feb. 5, 2010,19 confirming that PML risk increases with the number of Tysabri infusions received; April 22, 2011,20 confirming that PML risk is increased in patients who have taken other drugs that suppress the immune system; and Jan. 20, 2012,21 confirming that PML risk is increased in people who test positive for JVC.
__________________
14 For more information see FDA’s background document for the December 9, 2013, Advisory Committee meeting on the subject of vedolizumab: http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/GastrointestinalDrugsAdvisoryCommittee/UCM377618.pdf (accessed August 23, 2014).
15 See http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm107198.htm (accessed August 23, 2014) for all communications.
16 See http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2005/ucm108413.htm (accessed August 23, 2014).
17 See http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2006/ucm108662.htm (accessed August 23, 2014).
18 See http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm126592.htm (accessed August 23, 2014).
19 See http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm199872.htm (accessed August 23, 2014).
20 See http://www.fda.gov/Drugs/DrugSafety/ucm252045.htm (accessed August 23, 2014).
21 See http://www.fda.gov/Drugs/DrugSafety/ucm288186.htm (accessed August 23, 2014).
Uncertainty About CV Risk Associated with LAMA Drugs for COPD22
SUMMARY
In December 2013, FDA approved a novel combination product, Anoro Ellipta, as a long-term maintenance treatment for patients with chronic obstructive pulmonary disease (COPD). One of its agents, umeclidinium, is a member of a class of long-acting antimuscarinic agents (LAMAs), which has been the subject of concern since 2007, when pooled analyses suggested an increased risk of stroke, cardiovascular death, and myocardial infarction (MI) associated with tiotropium, one drug in this class. Since that time, various meta-analyses and randomized clinical trials have drawn conflicting conclusions about the cardiovascular (CV) risk of inhaled antimuscarinic agents, and this uncertainty has influenced FDA’s decision-making regarding other drugs in the class. The low numbers of major adverse cardiac events (MACE) observed in the Anoro Ellipta premarket clinical trials made it difficult to draw definitive conclusions about CV risk for this specific product. Therefore, an important question in the Anoro Ellipta approval decision was whether to require a post-market study to further assess potential CV risk. This case highlights the impact uncertainty can have on decisions regarding a class of drugs, as well as the issues presented by conflicting evidence from multiple sources.
INTRODUCTION
In December 2013, FDA approved Anoro Ellipta (umeclidinium and vilanterol inhalation powder) as a long-term maintenance treatment for patients with chronic obstructive pulmonary disease (COPD). COPD refers to a group of progressive, debilitating respiratory conditions, including emphysema and chronic bronchitis. It is the third leading cause of death in the U.S. There are several pharmaceutical options to treat COPD, including inhaled bronchodilators and steroids. The approved dose for Anoro Ellipta is one inhalation (umeclidinium/vilanterol 62.5 mcg/25 mcg) once daily.
Anoro Ellipta is a combination bronchodilator inhalation product, with two active ingredients. Umeclidinium is a long-acting antimuscarinic agent (LAMA) and vilanterol is a long-acting beta2-adrenergic agonist
__________________
22 Version Date: February 5, 2014. This summary was developed for purposes of discussion at the February 12–13, 2014 IOM/FD, Public Workshop: Characterizing and Communicating Uncertainty in the Assessment of Benefits and Risks.
(LABA). Neither component is currently approved in the U.S. as a single-ingredient product; however, vilanterol is available as part of a related combination product, Breo Ellipta (fluticasone furoate and vilanterol inhalation powder).
The development program for Anoro Ellipta included 8,138 patients with COPD in four 6-month trials, one 12-month long-term safety trial, and 9 trials of shorter duration. Over 2,400 COPD patients were treated with either the 62.5 mcg/25 mcg umeclidinium/vilanterol combination or a higher dose of the combination (125 mcg/25 mcg umeclidinium/vilanterol). The product showed improved bronchodilation over placebo and over either individual component. A major safety concern for Anoro Ellipta, as for the class of LABAs, is an increased risk of asthma-related death, and the approved product labeling23 includes a boxed warning highlighting this risk.
This case study focuses on another potential safety concern identified by FDA in its review of Anoro Ellipta. LAMAs have been the subject of concern since 2007, when a pooled analysis of controlled clinical trial data suggested an increased risk of stroke, cardiovascular death, and myocardial infarction (MI) associated with tiotropium, one drug in this class. Since that time, various meta-analyses and randomized clinical trials have drawn conflicting conclusions about the cardiovascular (CV) risk of inhaled antimuscarinic agents. This uncertainty has influenced FDA’s decision-making regarding other drugs in this class (described below).
The low numbers of major adverse cardiac events (MACE) observed in the Anoro Ellipta pre-market clinical trials made it difficult to draw definitive conclusions about CV risk of this product. Therefore, an important question in FDA’s approval decision for Anoro Ellipta was whether to require a post-market study to further assess potential CV risk. As FDA explained in the Anoro Ellipta decision memo24 (p. 13):
[T]he question becomes at what point we feel that a hypothesized safety issue caused by a class of drugs has been answered and further exploration is not necessary? Put another way, when a concern is discovered by meta-analysis, when is there enough data from randomized trials to assuage this concern?
__________________
23 See http://www.accessdata.fda.gov/drugsatfda_docs/label/2013/203975s000lbl.pdf (accessed August 23, 2014).
24 See http://www.accessdata.fda.gov/drugsatfda_docs/nda/2013/203975Orig1s000ODMemo.pdf (accessed August 23, 2014).
A LOOK AT THE PAST: ASSESSING CV RISK FOR OTHER LAMAS
Understanding FDA’s decision-making in the Anoro Ellipta case requires a look at how uncertainty was addressed in previous LAMA cases.
Tiotropium–Spiriva HandiHaler
Spiriva (tiotropium), delivered by the HandiHaler device, was the first LAMA to be approved by FDA, in 2004. At the time of approval, the available safety data raised the possibility of increased risk of “heart rate and rhythm disorders,” but risk of major CV events was not a major concern. The first indication of a possible CV risk associated with tiotropium was reported to FDA in 2007 by the drug’s sponsor. Results of a separate meta-analysis (Singh 2008)25 showed an increased relative risk of CV events with tiotropium and another short-acting antimuscarinic agent over control (RR 1.60 (95% CI (1.22-2.10)). However, subsequent results26 of a large (~6,000 patients), 4-year, randomized, controlled trial (called UPLIFT), did not show an increase in CV events associated with tiotropium, contradicting the meta-analysis results.
FDA conducted a comprehensive review to evaluate the strength of the 2007 pooled analysis, the 2008 meta-analysis, and the UPLIFT clinical trial. It identified a number of methodological limitations of the pooled and meta-analyses including potential study selection bias and potential differences in patient characteristics in treatment arms (e.g., sicker patients on placebo may have dropped out more). FDA’s assessment and conclusion on the CV risk of Spiriva HandiHaler is explained in a New England Journal of Medicine (NEJM) Perspectives article.27 There, FDA stated (p. 1099):
Because of the strength of the UPLIFT data, the absence of a strong signal related to stroke or cardiovascular events with tiotropium, and the potential methodologic limitations of the Singh meta-analysis, the FDA concluded that current data do not support the conclusion that there is an increased risk of stroke, heart attack, or death associated with tiotropium HandiHaler.
__________________
25 Singh, S., Y. K. Loke, and C. D. Furberg 2008. JAMA 300:1439-1450. [Erratum, 2009. JAMA 301:1227-1230.]
26 Tashkin, D. P., B. Celli, S. Senn, et al. 2008. N Engl J Med 359:1543-1554.
27 Michele, T. M., et al. 2010. N Engl J Med 363:1097-1099.
Tiotropium–Spiriva Respimat
At the same time the Spiriva HandiHaler CV risk signal was being assessed, FDA was also reviewing for market approval another tiotropium product, Spiriva Respimat, which had a different formulation and delivery device. In this case, three pre-market clinical trials showed an imbalance of deaths, of varying causes, including cardiac failure and MI, between the drug and placebo. Further, because the delivery of locally-acting drugs like inhaled bronchodilators differ depending on their formulation and device, FDA was unable to draw clear conclusions about the safety of Spiriva Respimat on the basis of the available evidence for Spiriva HandiHaler (and vice versa). Spiriva Respimat is not approved for marketing in the U.S., although the drug is approved elsewhere.
Subsequently, the drug’s sponsor conducted a large (>17,000 patients), prospective safety trial (TIOSPIR) to better assess the safety of Spiriva Respimat compared to Spiriva HandiHaler. The TIOSPIR results were published in October 201328 and indicate that the risk of death and of major cardiovascular adverse events is similar for the two tiotropium products. While the TIOSPIR results appear to be reassuring, it should be noted that the data from this trial have not yet been submitted to FDA for review.
Tudorza Pressair
Tudorza Pressair (aclidinium bromide inhalation powder) was approved in 2012. The pre-market evidence included a safety database limited in size, which hindered FDA’s ability to draw definitive conclusions.29 Because of this, and given the lingering uncertainty of possible disparate safety profiles of tiotropium delivered by different devices (the TIOSPIR trial was not yet published), a post-market CV outcome trial was required as a condition of the drug’s approval. In the Anoro Ellipta decision memo30 (p. 13), Dr. Curtis Rosebraugh explained the thinking about aclidinium at that time: “At the time of the review for aclidinium, I did not believe we had enough data to exonerate tiotropium, thereby casting a shadow on the LAMA class.”
__________________
28 Wise, R. A., et al. 2013. N Engl J Med 369:1491-1501.
29 The safety database included 733 patients exposed to the proposed 400 mcg dose of aclidinium for approximately 6 months, and 329 patients for approximately 1 year. It is important to note that these numbers do not represent unique patients; that is, the 329 patients listed as having an exposure for 1 year are also included among the 733 patients listed as having an exposure for 6 months.
30 See http://www.accessdata.fda.gov/drugsatfda_docs/nda/2013/203975Orig1s000ODMemo.pdf (accessed August 23, 2014).
ASSESSMENT OF THE SUBMITTED EVIDENCE FOR ANORO ELLIPTA
In its review of Anoro Ellipta, FDA conducted a comprehensive evaluation of the pre-market data submitted by the drug’s sponsor to support approval. However, there were limitations that hindered FDA’s ability to draw conclusions with respect to Anoro Ellipta’s CV safety profile. Like most pre-market clinical trials, the Anoro Ellipta trials were not designed (e.g., in terms of size or duration) to conclusively assess CV risk. In addition, the Anoro Ellipta program was also hampered by another issue common to clinical trials, namely, the generalizability of data gathered on a specific patient population in a highly controlled setting to the broader COPD population in the “real-world.” The limitations of the pre-market data are highlighted in the Anoro Ellipta decision memo,31 which states (p.9) that “any results are fragile at best and conclusion would be tenuous. The most that can be said is that there is not a consistent trend of MACE events indicating harm with drug use compared to placebo (or dose-response increase).”
IMPACT OF UNCERTAINTY ON FDA’S DECISION-MAKING FOR ANORO ELLIPTA
FDA concluded that substantial evidence of efficacy and overall safety had been demonstrated and that the benefits outweighed the risks of Anoro Ellipta. However, the question remained whether additional postmarket studies would be required to further assess CV safety for this drug.
The drug sponsor had proposed conducting observational studies to explore any possible CV risks. However, at this time, FDA does not generally consider observational studies to be adequate to quantitatively answer questions regarding a drug’s potential to cause increased CV events. Therefore, if additional evidence was necessary, it would have to be in the form of a large, randomized clinical trial.
After considering the totality of evidence, FDA concluded that a postmarket CV clinical trial was not necessary. In the Anoro Ellipta decision memo32 (p. 13), Dr. Rosebraugh explained:
[If the published results for TIOSPIR withstand FDA scrutiny], it would seem that the realization needs to be made that the advice for outcome trials for LAMA agents was based on a meta-analysis for tiotropium/
__________________
31 See http://www.accessdata.fda.gov/drugsatfda_docs/nda/2013/203975Orig1s000ODMemo.pdf (accessed August 23, 2014).
32 See http://www.accessdata.fda.gov/drugsatfda_docs/nda/2013/203975Orig1s000ODMemo.pdf (accessed August 23, 2014).
ipratropium (and possible concerns of class effects), which has proven to be a false signal. If this is a logical progression of thought, then contemplation must occur regarding at what point CV outcome trials should no longer be required of LAMA agents. I believe we are at that point.
BROADER RELEVANCE TO FDA’S CONSIDERATION OF UNCERTAINTY
This case highlights the impact uncertainty can have on decisions regarding a class of drugs, as well as the issues surrounding the uncertainty presented by conflicting evidence from multiple sources. These issues are not isolated to this class of drugs. For example, both issues are important in cases concerning the assessment of CV risk in Avandia (rosiglitazone) and other drugs to treat Type-2 diabetes mellitus.33
This case also highlights a challenge in determining the role of meta-analysis in FDA’s assessment of benefits and risks of drugs. FDA considers meta-analysis to be an important tool for safety assessment in the regulation of pharmaceutical products.34 However, meta-analyses present challenges in their design, evaluation, and interpretation. For example, in its 2010 NEJM article (p. 1099),35 FDA commented:
We have entered an era of increasingly frequent publication of meta-analyses, some of which identify potential safety signals. Such publication commonly leads to urgent calls to take immediate regulatory action, without acknowledgment of potential pitfalls in the interpretation of data from meta-analyses and pooled analyses, such as those encountered in the tiotropium evaluation. We must use measured restraint during our evaluations to ensure that safe drugs remain on the market and that their use is not restricted in a way that unnecessarily denies beneficial interventions to patients who need them.
FDA’S COMMUNICATIONS TO THE PUBLIC
FDA’s communications on this topic have primarily focused on the potential CV risk in tiotropium.36 In March, 2008, FDA issued an Early
__________________
33 The comparison of rosiglitazone to the case of LAMAs is discussed in the Anoro Ellipta decision memo.
34 Given its importance, FDA recently conducted a public workshop on the topic of meta-analysis. See http://www.fda.gov/ForIndustry/UserFees/PrescriptionDrugUserFee/ucm360080.htm (accessed August 23, 2014).
35 Michele, T. M., et al. 2010. N Engl J Med 363:1097-1099.
36 FDA’s communications on tiotropium are available at http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm107272.htm (accessed August 23, 2014).
Communication (now archived37), which notified the public of the sponsor’s pooled analysis first suggesting an association between Spiriva HandiHaler and stroke. FDA updated this communication in October, 2008, announcing the publication of the Singh 2008 meta-analysis and the completion of the clinical trial UPLIFT. In 2010, in addition to its NEJM article, FDA posted on its website another update38 explaining its conclusions that the available data do not support an association between Spiriva HandiHaler and CV risk.
__________________
37 See http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/DrugSafetyInformationforHeathcareProfessionals/ucm070651.htm (accessed August 23, 2014).
38 See http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/DrugSafetyInformationforHeathcareProfessionals/ucm197429.htm (accessed August 23, 2014).
