From meeting presentation to peer reviewed journal---POISE hits the pages of Lancet

The POISE (Perioperative Ischemic Evaluation) beta blocker study was presented at AHA last November. I blogged about it in a year end wrap up of developments in perioperative medicine. Last Tuesday the study was published in Lancet. The spin regarding this study has been largely negative. Missing from the discussion is the fact that the beta blocker use was associated with benefit for the primary endpoint. From the Lancet:

The primary endpoint was a composite of cardiovascular death, non-fatal myocardial infarction, and non-fatal cardiac arrest……Fewer patients in the metoprolol group than in the placebo group reached the primary endpoint (244 [5·8%] patients in the metoprolol group vs 290 [6·9%] in the placebo group; hazard ratio 0·84, 95% CI 0·70–0·99; p=0·0399). Fewer patients in the metoprolol group than in the placebo group had a myocardial infarction (176 [4·2%] vs 239 [5·7%] patients; 0·73, 0·60–0·89; p=0·0017).

But wait a minute. Secondary endpoints pointed to harm associated with metoprolol:

However, there were more deaths in the metoprolol group than in the placebo group (129 [3·1%] vs 97 [2·3%] patients; 1·33, 1·03–1·74; p=0·0317). More patients in the metoprolol group than in the placebo group had a stroke (41 [1·0%] vs 19 [0·5%] patients; 2·17, 1·26–3·74; p=0·0053).

Clinical trial experts criticize conclusions based on secondary endpoints when such endpoints are positive, as seen in the PROACTIVE controversy.

So, how should we view the POISE results? In the case of POISE the secondary endpoints were those that mattered more. The authors concluded:

Our results highlight the risk in assuming a perioperative β-blocker regimen has benefit without substantial harm, and the importance and need for large randomised trials in the perioperative setting. Patients are unlikely to accept the risks associated with perioperative extended-release metoprolol.

Understated, perhaps, but true. The second sentence was interesting in that it acknowledged the role of patient participation in decision making.

DB blogged about the study Friday as an important lesson for the performance and safety movement:

Second, the performance indicator and safety movement which has a "ready, fire, aim" philosophy, must reevaluate their strategies. One must wonder if our current push towards performance measurement has caused strokes and death in some patients. Finally, 30 years after finishing my residency, I continue to reshape my medical knowledge. As new knowledge appears we must quickly adjust our practice. I am interested in the ability of the performance movement to adjust. Of course, they will shrug off the unintended consequences that they caused here.

It gets you thinking---who makes up these safety and performance measures anyway? What are their credentials other than being effective cheerleaders (“champions” is the trendy term, I believe)?

DB also linked to Medscape’s Heartwire report about the trial. Dr Philip J Devereaux, who originally presented the POISE data at AHA, said in the report:

"If even only 10% of physicians followed these guidelines—which incidentally in the United States are used in quality assessments, where you have people going around ranking hospitals in terms of whether or not they are giving perioperative beta blockers—and if the POISE data are true, then in the past decade 800,000 people would have died prematurely and 500,000 would have had a major stroke perioperatively because we gave beta blockers….”

Where’s the press coverage of all these deaths? While a few reports such as this one hyped the results by and large the media response was modest. They didn’t have Big Pharma to pick on. What if, instead of an old, cheap off-patent medication the drug in question had been a new expensive block buster which had been heavily promoted by one of the big drug companies? I can only imagine.