Avandia safety, the new Senate report, and the need for outcome based research

February 20, 2010

The U.S. Senate just released a 350-page report on the GlaxoSmithKline (GSK) drug Avandia (rosiglitazone). The report investigates whether the FDA and GSK acted properly to evaluate the safety risk of the drug. The conclusions are scathing to GSK. It asserts that GSK knew years before 2007 of the safety concerns and intentionally stifled doctors trying to make the concerns public.

Per the previous Healthcare Channel note on 2/19/2010, it does seem that the recent BARI 2D study was not part of the report, and the full RECORD data from 2009 are not incorporated into the analysis. Both studies supported the safety of Avandia, as did the ADOPT and DREAM trials.

However, the Senate report is worth reading and provides interesting insight into how Big Pharma handles a crisis threatening a blockbuster drug. On the PDF file of the report, search for the keyword “RECORD” and the internal dialogue at GSK on how to handle the underpowered nature of RECORD is revealed. The rest of the report on pages 1 through 17 is interesting as well.

The HCC continues to believe that the final verdict on Avandia safety is yet to be determined. Plenty of evidence, all insufficient alone, suggests that Avandia raises the risk for heart attack (MI), stroke and death. Also, more recent randomized controlled trials (BARI 2D and RECORD) seem to support that the drug is safe.

The Steve Nissen metanalysis that caused most of the public focus on Avandia in 2007 and the FDA advisory committee was poorly conducted and unable to make the conclusion that Avandia was unsafe. In a 2007 Annals of Internal Medicine paper, Drs. Kaul, Diamond, and Bax wrote, “A recent, widely publicized meta-analysis of 42 clinical trials concluded that rosiglitazone was associated with an approximately 43% increased risk for myocardial infarction and an approximately 64% increased risk for cardiovascular death. The sensitivity of these conclusions to several methodological choices was not assessed. The meta-analysis was not based on a comprehensive search for all studies that might yield evidence about rosiglitazone’s cardiovascular effects. Studies were combined on the basis of a lack of statistical heterogeneity, despite substantial variability in study design and outcome assessment. The meta-analytic approach that was used required the exclusion of studies with zero events in the treatment and control groups. Alternative meta-analytic approaches that use continuity corrections show lower odds ratios that are not statistically significant. We conclude that the risk for myocardial infarction and death from cardiovascular disease for diabetic patients taking rosiglitazone is uncertain: Neither increased nor decreased risk is established.”

All of the individual trials have significant flaws given that detecting low event rates such as death from a drug requires massively large populations, or “N”, usually only found in observational registries. The FDA will likely reconvene an advisory panel to determine what to do. Recall, a previous advisory committee that analyzed the studies in the recent Senate report previously voted to keep Avandia on the market.

Multiple agendas were/are at play in the name of Avandia safety. On one hand, a powerful drug company was defending a cash cow. On the other, cardiologists trying to be relevant and famous were working the room, so to speak. During a Congressional oversight hearing, some members of Congress scolded Dr. Nissen for his tactics and leaking his New England Journal of Medicine report to Congress before it was published.

Rep. Darrell Issa (R-CA), said, “Mr. Issa. OK. I am going to yield back to the gentleman. I just want to make sure something gets in the record, though. The American Enterprise Institute published something that I think says a lot about the author that we are going to hear from in a few minutes. The study’s primary author, Cleveland Clinic cardiologist, Steven Nissen, admitted to the Wall Street Journal that he was in touch with Congress while preparing his analysis. Three days after the study was submitted to the New England Journal of Medicine and before it was published, the FDA Commissioner received a letter about Avandia from members of the House Energy and Commerce Committee that seemed to reference the New England Journal of Medicine study. I just want to make sure that is in the record, and I will yield back to the gentleman.”

The revenue for Avandia has already dropped dramatically. The impact of this Senate report will likely lead to another FDA advisory committee and the drug being pulled from the market, with little impact to GSK stock.

The important point raised by the Avandia scandal, according to Dr. Sanjay Kaul of Cedars Sinai, is the need for outcome based clinical trials for new drugs that target large populations. Currently, drugs are almost always approved based on surrogate markers, such as hemoglobin A1c or cholesterol levels, and not actual outcomes such as lower death rates or fewer heart attacks. The GAO issued a report last year on this topic. The HCC interviewed the author of the report, Marcia Crosse. The video interview may be viewed here.

Exclusive: Interview with Dr. Sanjay Kaul about the Crestor JUPITER label FDA meeting committee

December 16, 2009

Dr. Sanjay Kaul, MD, of Cedars Sinai Medical Center, was a voting member of the recent FDA advisory committee investigating whether to expand the indication for the statin Crestor to be used in patients with normal cholesterol levels but elevated C-reactive protein. The panel voted to approve this new indication. Dr. Kaul was the biggest critic of the JUPITER data, but voted yes to the question posed by the FDA. We asked him why.

The HCC: What was the one question voted upon?

Kaul: Has the sponsor established sufficient benefit to offset the observed risks to support the use of rosuvastatin in individuals meeting the following criteria:

1. men ≥ 50 years, women ≥ 60 years; and

2. fasting LDL < 130 mg/dL; hsCRP ≥ 2.0 mg/L; triglycerides <500 mg/dL; and

3. no prior history of cardiovascular or cerebrovascular events or CHD risk equivalent as defined by NCEP ATP-III guidelines.

The HCC: Do you think the FDA asked the right questions?

Kaul: They could have asked more direct questions specifying the types of patients eligible for rosuvastatin. But my guess is that they did not want to get entangled in that debate. Perhaps, the question is best answered by NCEP ATP IV guideline committee and/or other guideline committees of various professional societies.

The Canadian lipid guidelines recently recommend the use of hsCRP in individuals at intermediate risk for CHD as Class IIa, level B recommendation (optional, not mandatory, recommendation supported by reasonably strong quality of evidence) for initiating treatment. It will be interesting to see whether the NCEP ATP IV guidelines adopt a similar approach.

The HCC: What major points did you make?

Kaul:

1. The benefit of rosuvastatin in JUPITER was rather modest in terms of absolute risk reduction which translates into a NNT of about 82 over a 1.9 year follow-up. Attempts to extrapolate the NNT to 5 years (values ranging from 25 to 32) to compare it to other treatment interventions recommended for primary prevention (such as BP lowering agents, etc.,) is not informative and unlikely to be valid.

2. The benefit-risk of rosuvastatin is likely not reliably estimated in JUPITER because of early stoppage of the trial. Such prematurely truncated trials are notorious for overestimating benefit and underestimating risk.

3. There is substantial uncertainty in the literature with regards to the utility of hsCRP in predicting risk or treatment response to statins. There was nothing in the JUPITER trial that resolved these uncertainties. In fact, the event rates in the control arm were similar regardless of whether hsCRP was low or high indicating a lack of relationship of hsCRP and CHD risk. Furthermore, the treatment was more likely to be effective in the low hsCRP compared to the high hsCRP subgroup, suggesting an inverse dose response (which is counterintuitive).

4. The signal for increased diabetes risk with rosuvastatin is likely real and represents a class effect amongst statins. However, it was likely underestimated as the events were based on investigator reports and not adjudicated using standardized objective criteria. The lack of adjudication was rather surprising given this was a co-primary or major secondary hypothesis of the study!

5. The total mortality benefit was unlikely to be true because it was not adjudicated, the follow-up was short (KM curves begin to converge at 4.4 yrs of follow up at which time the treatment differences are no longer statistically significant), the analysis was not adjusted for multiplicity (thereby increasing the chances of a false-positive error). Furthermore, it was driven by a significant difference in cancer mortality, which is likely to be a spurious finding. Given these observations, a mortality claim could not be justified.

The HCC: Why did you vote yes even though you were not supportive of JUPITER as a reliable study given early stoppage, lack of CRP support, etc?

Kaul: Amongst the statins, only lescol (fluvastatin) does not have a label indication for primary prevention. Rosuvastatin is approved for secondary prevention, and on the basis of JUPITER, the benefit sufficiently exceeded the risk, thereby warranting approval. The key question is whether JUPITER has expanded the target population for statin treatment. Most panel members, including myself, had concerns regarding allowing an expanded indication. Most felt that the key qualifying criteria should be LDL<130 (where ATP III guidelines already offer statins as an optional treatment), hsCRP >2, plus at least 2 additional CHD risk factors (including age). This was based on the observation in JUPITER that individuals with less than 2 risk factors failed to derive significant treatment benefit.

It is important to keep in mind that a 12-4 vote in favor of rosuvastatin does not accurately reflect the level of discussion amongst the panel members regarding defining the treatment-eligible population. The FDA usually pays greater attention to the accompanying discussion rather than the vote count. Most committee members viewed their “Yes” votes as not to endorse hsCRP as a new risk factor.

Finally, I am sure the language of the label will be vigorously debated amongst the sponsor and the FDA. Let the games begin!

Sanjay Kaul, MD: The dabigatran data versus rivaroxaban

August 31

Dr. Sanjay Kaul is in Barcelona now and gave us his quick impression of the dabigatran RE-LY data and how it compares to JNJ’s Xarelto data.

He wrote, “The ROCKET AF study with rivaroxaban 20 mg once daily dose vs.

warfarin titrated to INR of 2.5 is expected to come out in December 2009.

Rivaroxaban works upstream in the coagulation cascade inhibiting factor Xa while dabigatran is a direct thrombin inhibitor. Thrombin acts downstream in the coagulation cascade catalyzing the conversion of fibrinogen to fibrin.

The bleeding complication with rivaroxaban is dose-dependent with the highest risk observed with 20 mg daily dose, the same dose being evaluated in ROCKET AF trial. The results of RELY suggest that dabigatran is associated with lower bleeding at 110 mg dose compared with warfarin and similar bleeding at 150 mg dose.

The safety data with regards to hepatotoxicity looks good for dabigatran. Although no major liver safety concern was observed with rivaroxaban in the RECORD studies or ATLAS phase 2 study, the treatment duration was rather short. Long-term safety data will have to await the results of RECORD AF trial.

One potential advantage for JNJ’s Xarelto (rivaroxaban) appears to be once daily dose versus the twice daily dose with dabigatran which might impact compliance rates.

Finally, only head to head studies with rivaroxaban and dabigatran will provide conclusive evidence of efficacy and safety differences between the two treatments.”

Prasugrel approved with restricted label

July 10, 2009

Eli Lilly’s drug prasugrel, to be branded as Effient, was approved with a restricted label as we predicted (label attached).

• It seems the label will not allow a claim of superiority over Plavix. Recall, the TIMI trial was designed to show this.
• Black box warnings were added that were not voted on by the advisory panel (sans Dr. Kaul and the OSE). Importantly, the frail elderly woman type of patient (older than 75 and less than 60KG), is a black box warning. Patients that might need heart surgery, CABG, are also in the warning.

· Connection to cancer risk is in the label with post-approval data collection required

It is no surprise to us that this news was released on a Friday with no company conference call to our knowledge. A common tactic is to release bad news on Friday evenings. In addition, per our previous note today, the FDA response to the congressional oversight investigation into the prasugrel advisory panel and admitted mistakes in handling of the matter.

FDA NEWS RELEASE

For Immediate Release: July 10, 2009

FDA Approves Effient to Reduce the Risk of Heart Attack in Angioplasty Patients

The U.S. Food and Drug Administration has approved the blood-thinning drug Effient tablets (prasugrel) to reduce the risk of blood clots from forming in patients who undergo angioplasty, a common procedure to unblock a clogged coronary artery.

During an angioplasty, a balloon is used to open the artery that has been narrowed by atherosclerotic plaque. Often, a tiny wire mesh scaffold (stent) is inserted into the blood vessel to help keep the artery open after the procedure. Platelets in the blood can clump around the procedure site, causing clots that can lead to heart attack, stroke, and death.

Effient was studied in a 13,608-patient trial comparing it to the blood-thinning drug, Plavix (clopidogrel), in patients with a threatened heart attack or an actual heart attack who were about to undergo angioplasty.

The fraction of patients who had subsequent non-fatal heart attacks was reduced from 9.1 percent in patients who received Plavix to 7.0 percent in patients who received Effient. While the numbers of deaths and strokes were similar with both drugs, patients with a history of stroke were more likely to have another stroke while taking Effient. In addition, there was a greater risk of significant, sometimes fatal, bleeding seen in patients who took Effient.

“Effient offers physicians an alternative treatment (The HCC note: not a superior treatment compared to Plavix) for preventing dangerous blood clots from forming and causing a heart attack or stroke during or after an angioplasty procedure,” said John Jenkins, M.D., director of the Office of New Drugs, in the FDA’s Center for Drug Evaluation and Research. “Physicians must carefully weigh the potential benefits and risks of Effient as they decide which patients should receive the drug.”

The drug’s labeling will include a boxed warning alerting physicians that the drug can cause significant, sometimes fatal, bleeding. The drug should not be used in patients with active pathological bleeding, a history of mini-strokes (transient ischemic attacks) or stroke, or urgent need for surgery, including coronary artery bypass graft surgery.

Effient is manufactured by Eli Lilly and Company of Indianapolis, in partnership with Tokyo-based Daiichi Sankyo Ltd.

Cedars Sinai Panel: the SYNTAX stent v CABG trial

A trio of cardiac experts from Cedars Sinai Medical Center in Los Angeles discusses the SYNTAX trial that compared the safety end efficacy of drug eluting stenting versus CABG surgery in multivessel disease. Sanjay Kaul moderates cardiac surgeon Greg Fontana and interventional cardiology and cath lab director Raj Makkar.