Spiriva showing a troubling pattern of serious heart risks.

Breathe Deeply: Spiriva Studies Show Troubling Pattern Of Serious Heart Risks, If You Know Where To Look

By Ed Silverman // July 14th, 2011 // 11:25 am

Last month, a widely publicized study raised concerns about the safety of the blockbuster Spiriva inhaler that is used to treat chronic obstructive pulmonary disease. The meta-analysis published in the British Medical Journal found that, when administered in the mist form known as Respimat, there was a 52 percent increase in mortality risk compared with a placebo.
Not surprisingly, Boehringer Ingelheim, which makes the inhaler, disagreed. And in disputing the findings, the drugmaker made a point of saying the results were not based on new clinical evidence and that the meta-analysis relied on already publicly available data. BI also maintained that “robust evidence” has also been published about the safety of the older, powder version known as HandiHaler (read the statement).

Specifically, Boehringer pointed to a trio of published studies, including a pooled analysis, which were offered as evidence that Spiriva - a $4.3 billion global seller - has not only been proven safe, but also reduces the risk of cardiovascular events and cardiovascular deaths. This is important because patients suffering from COPD are, generally, older and more likely to have compromised cardiovascular systems. Just the same, BI says that safety and risks are already reflected in product labeling.

However, a review of numerous published studies, briefing documents filed with the FDA and company synopses that are generally posted on the Boehringer web site tell a slightly different story, one that raises questions about the extent to which the drugmaker has willingly shared all the available data about cardiovascular events and just how difficult it can be to identify and access some of this data.

A common thread running through these various papers, including those cited by Boehringer in its press release, indicates that serious cardiovascular events, notably serious cases of tachyarrhythmia and angina, often appear at higher rates than when compared with a placebo or another drug. And the thread actually extends back as far as the fall of 2002, according to sources who have examined Spiriva data, but with one exception, asked not to be named.

In September of that year, an FDA advisory panel reviewed the Spiriva HandiHaler, which was up for agency approval. At the same time, a group of researchers published a follow-up to their 1994 paper, which was called the Lung Health Study and was funded by the National Institutes of Health, concerning Atrovent, a short-acting version of Spiriva. Their analysis found deaths and hospitalizations were more common among patients on the drug (see this).

“This has been a 10-year cover-up,” says Sonal Singh, an assistant professor at the Johns Hopkins University School of Medicine and a co-author of the recent BMJ meta-analysis. “There is a signal here for certain subgroups (of patients). But who are they and why are they experiencing these events? The signals have not been pursued by Boehringer. They’ve played hide and seek with the data.”

A great deal is at stake for all involved. Last year, Spiriva HandiHaler rang up more than $2 billion in US sales and garnered 61 percent of the market for anticholinergic drugs, according to IMS Health. The showing underscores the importance of the treatment to the drugmaker and to the large numbers of people who are prescribed the medication. The implication is that millions of Spiriva patients may be vulnerable to serious cardiovascular risks.

Meanwhile, the US labeling for HandiHaler has never been updated to reflect serious cardiovascular risk, although these events appear in various studies (Respimat has not been approved in the US, but is available in other countries). Yet experts who reviewed the studies say excess risks have been masked by the use of composite events, which Boehringer uses to suggest its inhalers actually reduce cardiovascular risk.

Here is another common thread - the studies were funded by Boehringer – and, in some cases, also by Pfizer, which co-markets Spiriva. And many of the authors were either Boehringer employees or have worked for the drugmaker as a consultant or advisory board member, or received research grants. As an example, one paper, which has been used to tout Spiriva safety and was published last year in the Chest journal, reviewed 30 Respimat and HandiHaler studies and concluded the drug was associated with a reduced risk in cardiovascular death and events.  The study was funded by both drugmakers and the authors either worked for Boehringer or had other financial ties.

For its part, Boehringer maintains that certain adverse event data in some trials may not have been captured immediately due to the nature of clinical trial work, especially in some of the older studies, but more information is actively sought as the research focus narrows and additional evidence is gathered. “As we learn more about the compound,” says John Smith, a senior vice president of clinical research and medical affairs at Boehringer, “there is more granularity available in later studies.”

“When you look at trials, things are going to shift around.There is an ebb and flow in clinical trials…And that’s one of the reasons why we don’t have precise answers. And that’s why we put multiple trials together and come up with hypotheses and then do a trial to answer that,” he continues. “I think we are fulfilling our commitment to patients and prescribers to collect information and give them the best possible information to make decisions.”

As for the FDA, the agency pointed to a January 2010 health alert that reiterated an earlier alert from 2008 that data does not suggest that Spiriva – specifically, the HandiHaler – increased the risk of stroke, heart attack, or death from a cardiovascular cause (see this). However, the agency declined to make someone available to discuss the various studies cited here.

So which studies merit scrutiny? With help from experts, we reviewed various studies published over the past several years of both the Spiriva HandiHaler and Respimat versions and compared specific data - incidence rates, rate ratios - for cardiovascular events found on the FDA and Boehringer web sites. The average clinician, however, is unlikely to know that some of this material exists or where to find it. Moreover, the details appear nuanced unless the studies are examined as a group.

Let’s start in 2002 and peek on page 10 of the FDA briefing document that was used to review Spiriva HandiHaler before it was approved (look here). The overview refers to six pivotal clinical studies and found that, “in the one-year, placebo-controlled studies, five of the seven deaths among (Spiriva) patients, but only one of the seven deaths in placebo patients, were attributable to cardiac ischemia or arrhythmia. And “there were subtle indications that (Spiriva) may be associated with an increased frequency of adverse cardiac effects, specifically in the category of heart rate and rhythm disorders.”

In 2007, a study published in the Canadian Respiratory Journal comparing HandiHaler with placebo did not mention any cardiovascular events, although patients who experienced a myocardial infarction within the previous six months or a cardiac arrhythmia requiring medication were excluded (read here). The study was funded by Boehringer and Pfizer, and the authors worked for the drugmakers as speakers or advisory board members.

Then turn to an unpublished version of the same study from 2005 that appears on the Boehringer web site, which states that “there were 3 percent of patients in the (Spiriva) group reporting serious cardiac disorders compared with 2.6 percent in the placebo group. Of these, the majority of cardiac events were comparable between the two groups, with the exception of myocardial infarction, which occurred in 1 percent of patients in the (Spiriva) group and 0.3 percent in the placebo group.”

In 2008, a widely cited study called UPLIFT was published in The New England Journal of Medicine that compared the HandiHaler with placebo. The study concluded that there was a reduction in cardiac adverse events associated with Spiriva HandiHaler, although a table showed the relative risks – 1.44 for serious angina and 1.25 for cardiac failure (here is the abstract).

However, a more specific breakout appears in the FDA briefing materials for the 2009 advisory panel that was held to review Spiriva Respimat – a table on page 115 shows that the incidence rate for atrial tachycardia was 0.08 compared with placebo and 0.08 again for tachyarrhythmia compared with 0.02 percent for placebo. (Again, the authors were either Boehringer employees or had financial ties to the drugmaker.)

“There’s often way more data from an FDA briefing document, for instance, such as various reviews the company sends in, all kinds of analysis, but sometimes companies don’t want to tell people that,” says Sid Wolfe, who heads Public Citizen Health Research and sat on the 2009 FDA advisory panel that reviewed the safety of the HandiHaler, although he was not asked to review the studies mentioned here. “This kind of discrepancy is unfortunate and is not exactly in the public interest, but it happens.”

So why was there more info in the FDA briefing document?

Smith says this: “The FDA briefing document is impacted by questions we’ve been asked and ongoing dialogue we have with the agency during an ongoing review period. An article may not have all the answers you’re looking for to limit words or figures or pre-specified analyses. I can understand where you’re looking for two different documents addressing the same things and they don’t look identical, but there are different audiences in different forums.”

And how does he explain the angina results? “What’s important to understand is that, while it may seem appropriate to put angina and mycocardial infaction together, they are actually different. In some studies where these are all bucketed together, there may not be the sort of granularity…In earlier trials, they may have been aggregated. Over time, they tend to get split out more.”

However, it is worth noting that in the Chest paper published last year, angina and MI are grouped together. In a more recently published study called POET, which we will get to further down below, both MI and angina are elevated, but broken out.

Meanwhile, a June 2009 synopsis from the Boehringer web site discusses a subgroup analysis and indicates that for fatal events, there was a higher rate ratio for Spiriva in patients with a history of cardiac disease - 4.03 – or arrhythmia at baseline - 8.61. And while the rate ratio for a composite endpoint of major cardiovascular events was balanced in patients with known cardiac disease (0.99), the endpoint was unbalanced for patients with known arrhythmia at baseline (1.72).

We asked Smith why there was more detail in the hard-to-find synopsis. His response: “The distinction was that this was an exploratory analysis and was not pre-specified in the protocol. As a rule, for publication, journals just rely on pre-specified analysis and table displays, and things that are exploratory are just hypotheses that require confirmation.”

Also worth noting is a table on page 158 of a 292-page briefing document submitted by Boehringer for the November 2009 FDA panel meeting (see here). This data is from an unpublished epidemiologic study called THIN-II and, like the POET study, compared Spiriva with a long-acting beta agonist and for certain cardiovascular events, such as atrial fibrillation, cardiac arrest and heart failure, Spiriva showed lower incidence and hazard ratios. But for angina, myocardial infarction, tachycardia and stroke, the Boehringer inhaler displayed higher ratios. As an example, the hazard ratio for stroke was 1.49 and was 1.38 for angina.

Meanwhile, a study published in Respiratory Medicine in 2010, which compared the Respimat inhaler to a placebo, notes that “tachycardia has not been observed in this or other Tiotropium (Spiriva) trials. But again, the table on page 115 from this FDA briefing document, which was reviewed during the 2009 FDA advisory panel meeting to discuss Spiriva safety, shows in the largest Boehringer clinical trial to date that there was a relative risk of 8 atrial tachycardia compared with placebo. The assertion in the journal is contradictory.

The same study in Respiratory Medicine also notes the incidence rate of cardiac disorders was 0.51 percent on Respimat versus 0.22 percent for placebo. The rate ratio was 2.27 percent. (Here is the abstract). The authors, by the way, were either Boehringer employees or received fees from the drugmaker for speaking or consulting, or advisory board work.

Then there was a study that was published this past March in The New England Journal of Medicine. Called POET, the study made little mention of cardiovascular events, except for a table showing a 1.12 rate ratio for patients using Spiriva HandiHaler compared with those using the Serevent inhaler sold by GlaxoSmithKline (here is the abstract). The supplemental appendix, meanwhile, shows a 1.5 rate ratio for myocardial infarction – and a 0.55 percent incidence rate for the HandiHaler compared with 0.37 percent for Serevent.

Then visit ClinicalTrials.gov, where the listing of serious adverse events offers a more nuanced but serious picture. Here are some comparisons showing risks for patients on Spiriva versus Serevent…

For angina pectoris, 0.24 percent of Spiriva patients were at risk vs. 0.14 percent for Serevent;

For unstable angina, both products showed a 0.05 percent risk;

For post-farction angina, 1 percent of Spiriva patients were at risk vs 0 percent for Serevent;

For myocardial ischemia, 0.3 percent of Spiriva patients vs. 0.16 percent for Serevent;

For acute MI, 0.08 percent of Spiriva patients vs. 0.03 percent for Serevent patients;

For myocardial infarction, 0.54 percent of Spiriva patients vs. 0.35 percent of Serevent users.

The POET study, by the way, also excluded patients with a history of myocardial infarction within the year prior to entering the study, those with cardiac arrhythmia that required medical or surgical treatment and patients with severe cardiovascular disorders. “In doing a clinicial trial, a company may exclude patients who are likely to have an adverse reaction, because these would make the results look worse,” says Public Citizen’s Wolfe.

We asked Smith how these cardiovascular events can be explained when people with these backgrounds are not included in the study. Smith’s response: “Generally, for these clinical studies it’s easier to identify actual clinical events to put in exclusion criteria.” He went on to describe the patient population, but did not directly address the question.

As to why the information found on ClinicalTrials.gov was not published in the actual journal, he says such decisions are “driven by editorial policy and discussions between authors and journal editors and where things are placed is not something we have specific policy on.” Although, the study was funded by Boehringer and Pfizer, and most of the authors have financial ties to Boehringer in the form of consulting and lecture fees or research grants. This suggests the drugmakers had an opportunity to communicate with the editors about such decisions.

Then there is an unpublished pooled analysis of six clinical trials comparing Respimat with a placebo, which was completed in February 2010 and undertaken to evaluate fatal events. This analysis was found on the Boehringer web site and the results indicated an imbalance in mortality in patients with known cardiac rhythm disorders at randomization – the rate ratio was 3.42 (read here).

Another pooled analysis was published in the International Journal of COPD in October 2009, and this was cited in the recent Boehringer press release that was issued after the BMJ meta-analysis was released. The study, which was conducted by researchers who were either Boehringer employees or did work as consultants to the drugmaker, concluded that the Spiriva HandiHaler did not show an increased risk for death or cardiovascular morbidity. A table, however, indicates that Spiriva HandiHaler yielded a higher rate than placebo for supraventricular tachycardia. Rates for angina were not presented.

Another trial that was not sponsored by BI conform to the pattern. One was published in the American Journal of Respiratory and Critical Care Medicine in 2008 and was funded by Glaxo, which compared Serevent with HandiHaler. Table 4 showed 19 cases of cardiac disorder with Spiriva compared with nine for Serevent. And the top five most commonly reported serious adverse events that began during treatment were cardiac disorders, with 34 among Spiriva users and 13 among Serevent users.

And a study published earlier this year in Respiratory Medicine -which compared the recently approved COPD drug from Novartis with Serevent, showed Spiriva patients had a higher rate of serious adverse events, such as arrhythmia and cardiac failure than those on the other meds or placebo.

The FDA, Singh says, has failed to sufficiently follow up. “The regulators have not been asking the right questions, because there’s no inconsistency in all this data. But in some patients, tachyarrhythmia and angina can lead to cardiac death and other serious cardiac events,” says Singh, who also co-authored a 2008 meta-analysis that also found an increased risk of serious cardiovascular events, including death, heart attack and stroke. “This is an important public health question, because millions of people are using Spiriva and may be susceptible to one of these outcomes.”