Treatment-risk paradox

Treatment-risk paradox highlighted


04 June 2007


Arch Intern Med 2007; 167: 1009-1016; 1019-1025


MedWire News:

Two studies in the Archives of Internal Medicine highlight concerns that high-risk heart disease patients who are most likely to benefit from evidence-based therapy are the least likely to receive it.

Andrew Yan (University of Toronto, Ontario, Canada) and colleagues studied the use of in-hospital cardiac catheterization in relation to risk among non-ST-elevation acute coronary syndrome (NSTE-ACS) patients in the prospective, multicenter, Canadian ACS Registry 1 and Registry 2, which enrolled patients between 1999 and 2001 and 2002 and 2003, respectively.

The team stratified 4414 patients into low-, intermediate-, and high-risk groups based on tertiles of the Global Registry of Acute Coronary Events (GRACE) score.

The recommendation by the American College of Cardiology and American Heart Association of an early-invasive strategy in the management of NSTE-ACS patients appears to have been effective, with in-hospital use of cardiac catheterization increasing significantly from 38.8% in ACS Registry 1 to 63.5% in Registry 2 (p<0.001). color="#000099">In the second study, Finlay McAlister (University of Alberta, Edmonton, Canada) and team explored why clinicians are less likely to prescribe new therapies in low-risk than high-risk coronary artery disease (CAD) patients.

They studied 3871 patients diagnosed with CAD by angiography at three cardiac centers in Alberta between February 2004 and December 2005. The team defined patients as being at low, medium, or high risk according to the Duke Coronary Index.

The results showed that at 1 month after an angiogram, high-risk patients were less likely to be taking angiotensin-converting enzyme inhibitors, aspirin, and statins than lower-risk patients (25.8% vs 32.3%, odds ratio=0.73; risk ratio=0.80).

A similar relationship was seen across individual drug classes, although the strongest association was seen with statins.

However, adjusting for exertional capacity and depressive symptoms caused the relationship between risk and level and use of medications to almost completely disappear.

The authors report that there was no relationship between risk and symptom-relieving anti-anginal medications in unadjusted analyses and that it is therefore unlikely that clinicians preferentially avoid treating these patients. The more likely explanation, they say, is that patients who have depression, poor functional status, or both are less likely than those without to adhere to therapy.

"The treatment-risk paradox reported in administrative database analyses is attributable to clinical factors not typically captured in these databases," the authors write.

In an accompanying editorial, John Spertus and Mark Furman (University of Massachusetts, Worcester, USA) argued: "To overcome this paradox, what is needed is a transparent method for objectively calculating the risks of an adverse prognosis in our patients and, ideally, the expected benefits of treatment."

They said that barriers to the deployment of current models include: the complexity of employing multivariable regression models; the absence of data projecting the outcomes as a function of alternative treatment strategies; the limited range of meaningful outcomes predicted; and clinical inertia in substituting complex decision aids for anecdotal decision making.

They propose the creation of a national research agenda to "instill a new model of clinical practice."

Journal