Earlier this year, the Mirapex drug that is used to treat Parkinson's disease and Restless Legs Syndrome became the latest example in the ongoing controversy over the full disclosure of unfavorable clinical trial data. Six years ago, cardiac failure was listed in the labeling after being noticed during clinical trials, but causation was never determined. Later, cardiac failure was cited in connection with the drug, but the full story appeared untold.
The background: A study funded by Pfizer, which compared its Dostinex drug against other meds, found a statistically significant increased risk of heart failure with Mirapex. Earlier this year, a study supported, in part, by Boehringer Ingelheim, which sells Mirapex, reached the same conclusion about both drugs. But that latest study yielded some revealing passages suggesting Boeheringer had tracked cardiac failure for some time without any public disclosure.
Briefly, the study cited "internal data" and correspondence from the head of global epidemiology at Boehringer, who provided results of a pooled analysis from randomized Mirapex trials. The study mentioned a “signal of a potential risk of heart failure arose in pooled data from 26 placebo-controlled, randomized trials” in Parkinson’s and RLS patients. But there were differences between the authors and the drugmaker, which never disclosed when it learned of the risk of cardiac failure in the pooled trials.
Last week, though, the FDA issued a warning informing the public about a possible increased risk of heart failure with Mirapex after evaluating a pooled analysis of randomized clinical trials and found that heart failure was more frequent with Mirapex than with placebo. The agency also evaluated two epidemiologic studies that suggested an increased risk of new onset of heart failure with Mirapex use, according to a statement (read here).
In discussing its explanation, the FDA noted that Boehringer provided the pooled analysis in 2008 to the agency and then updated the data two years later. Meanwhile, though, there was no opportunity for independent review, since the drugmaker never publicly disclosed the material that was used as the basis for the study that was published earlier this year in Pharmacoepidemiology & Drug Safety (back story).
Instead, the drugmaker had pointed us toward the labeling, which was revised last year and provides the same language that appeared in 2006, with the addition of one line: “In a pharmacoepidemiological study, pramipexole (Mirapex) use was associated with an increased risk of cardiac failure compared with non-use.” This appears under the post-marketing experience section. As we noted earlier, there was no mention of two studies or undisclosed data.
The FDA, however, offered a couple of caveats about its latest conclusions. For one, the agency cautioned there were study limitations that made it difficult to determine whether excess heart failure was related to Mirapex use or other influencing factors. The FDA also noted that the evaluation of the pooled analysis of randomized clinical trials showing heart failure was more frequent with Mirapex than with placebo was not statistically significant.
Some, however, may argue such a qualifier is unconvincing. Last year, for instance, US Supreme Court Justice Sonia Sotomayor noted that "both medical experts and the Food and Drug Administration rely on evidence other than statistically significant data to establish an inference of causation." She wrote that in a unanimous opinion that a drugmaker may be required to alert investors of adverse events even if statistical significance is lacking (read here).
heart attack pic thx to bart on flickr
10 Comments
2. The evidentiary threshold for FDA to require companies to update prescribing information (drug labels) remains more closely aligned with the interests of the industry it regulates than those of the public it is supposed to serve.
In this instance there is evidence from randomized clinical trials showing Mirapex increases heart failure risk versus placebo and two confirmatory studies showing an excess risk compared with other available treatments (and God knows what else). Yet Boehringer Ingelheim has decided--and the FDA has endorsed--that these facts should not be disclosed in a clear, complete, and timely manner in the product label. One can only wonder what other types of evidence the FDA could reasonably expect to obtain?
The FDA Safety Alert is a step in the right direction, but I fear the vast majority of physicians will not see this notice with the result that patients taking this drug will remain oblivious to its potentially life-threatening consequences.
After reading David Healy's excellent new book "Pharmageddon", I am convinced that until original clinical trial information is demanded as the ticket to entry on formularies, the carnage will continue. Ditto the Administration's posture towards pharma - large fines - is totally ineffective. Jail terms would work but again these seem to be not politically tenable as too much campaign money is involved.
Provided with a list of 300 categories of life-threatening adverse events, the most basic statistical analysis says that 150 of them will be numerically more common in the treatment group compared to the placebo group. Even applying the usual p < 0.05 standard (one chance in 20 that the difference between groups is due to chance), you would still on average end up with 7.5 event types on the label that differed between treatment and control groups only because of random chance (multiple hypothesis testing).
I've lived in the world that the critics here advocate, and its not helpful in anyway. Prior to 1980 or so, laboratory chemicals were simply labeled with the known and well established risks. With increasing concerns about liability, the warnings expanded until bottles of sodium chloride (table salt) came with 3 inch columns of 5 point font warning of cardiovascular risk, coma and death. Pretty much the exact same label as things that were well-established to be quite harmful.
Once the effort to include every conceivable risk on the label gains full steam, the label becomes meaningless. It creates a loss of information that is just as profound as under-labeling.
I think this is the FDAs concern. One needs to consider the full ramifications of what one is asking for.
"In assessing the safety risk posed by a product, the FDA considers factors such as “strength of the association,” “temporal relationship of product use and the event,” “consistency of findings across available data sources,” “evidence of a dose-response for the effect,” “biologic plausibility,” “seriousness of the event relative to the disease being treated,”“potential to mitigate the risk in the population,” “feasibility of further study using observational or controlled clinical study designs,” and “degree of benefit the product provides, including availability of other therapies.”" FDA, Guidance for Industry: Good Pharmacovigilance Practices and Pharmacoepidemiologic Assessment 18 (2005) http://www.fda.gov/downloads/ RegulatingInformation/Guidances/UCM126834.pdf
Here is an excerpt from the abstract of the case control study:
" The incidence rate of heart failure was increased with the current use of any dopamine agonist (RR = 1.58, 95% CI = 1.26–1.96), and particularly so for pramipexole (RR = 1.86, 95%CI = 1.21–2.85) and cabergoline (RR = 2.07, 95%CI = 1.39–3.07), compared with no use. The increase was not significant with ropinirole (RR = 1.23, 95%CI = 0.85–1.97) or pergolide (RR = 1.42, 95%CI = 0.95–2.12). Pramipexole was not associated with a significantly increased rate when compared with all other dopamine agonists collectively (RR = 1.28, 95%CI = 0.82–2.00)."
How does one say there is more risk for an agent with a RR CI of 1.21-2.85 vs one with a CI of 0.95 to 2.12?
So I can't say anything about the RR of mirapex vs other DAs from this study as it looks like any differences are likely due to chance. And I can't say anything about the overall risk of taking a DA, because the risk may come not from the DA, but from the underlying illness that leads one to take a DA.
In the case of the randomized clinical trial data, the FDA alert shows 12 cases of heart failure among 4157 mirapex treated patients and 4 among 2820 placebo patients. When I plug that into a chi square calculator, I get p = 0.31. I can hardly consider this the anchor of a "consistent pattern across studies".
The FDA comments further: "Another important limitation is that there was limited or no validation of the heart failure cases by medical chart review. Both studies had an imbalance in the percentage of patients with cardiovascular risk factors (such as a history of ischemic heart disease, chronic obstructive pulmonary disease, and arrhythmia) that were more frequent in cases than controls. "
I take 2 mg of mirapex a day, and I'm not even beginning to think about making an appointment with my doctor to get a different Rx. After all, I might get killed in an auto accident on the way to his office.
Now I'm not supporting either Boehringer or the FDA in this matter, but it must be worth exploring whether patients in the Mirapax trials were on any other medications and the extent to which these may have affected the outcome.
Nor am I suggesting that the adverse event reporting system is fool-proof - it most certainly isn't. What is clear is that we cannot rely on self-disclosed clinical trial results or voluntary adverse event reporting, and an improved system for identifying and disclosing side-effects in trials and the clinic are required if these potentially life-threatening effects of a drug are to be known to prescribers before they write the prescription.
You see lots of heart failure and MI AEs in patients who are taking blood thinners and blood pressure drugs because patients taking those drugs typically have cardiovascular problems. For the most part, the underlying disease and not the drugs are causing the AEs in these patients.
This is part of why you have to look at this Mirapex data very carefully. Everyone who takes mirapex for PD, well, has PD. Its quite possible that having PD or the risk factors for developing PD are independent risk factors for HF.