Docs Misclassifying Two Thirds of Patients at High Risk for CV Events

November 24, 2010

November 24, 2010 (Chicago, Illinois) — Canadian primary-care physicians are not accurately assessing cardiovascular risk in middle-aged adults, with many patients misclassified into incorrect risk cohorts, a particular concern for high-risk patients classified as low or intermediate risk, according to the authors of a new study.

"The problem is certainly with the high-risk patients, with two-thirds of high-risk patients in Canada misclassified," lead investigator Dr Milan Gupta (McMaster University, Hamilton, ON) told heartwire . "According to our guidelines, the overwhelming majority of these patients should be treated, and so the implications of the study are that a sizable proportion of high-risk patients might not be prescribed lipid-lowering therapy because they're not perceived to be at high risk."

The findings are from the preliminary results from the Primary Care Audit of Global Risk Management (PARADIGM). The report, presented last week at the American Heart Association 2010 Scientific Sessions, was designed to evaluate primary-care physician behavior toward global cardiovascular-risk prediction in generally healthy adults and to evaluate the methods and accuracy of cardiovascular-risk stratification used by primary-care physicians in Canada.

Using the Framingham Risk Score

In the study, 3015 men and women were enrolled from 105 sites across Canada, excluding Alberta. Based on patient data, these individuals were stratified centrally by Gupta and colleagues into low, intermediate, and high risk using the Framingham Risk Score. The risk assessed by the investigators was then compared with the risk status assessed by the primary-care physicians.

Overall, patients included in the study were not candidates for lipid-lowering therapy based on LDL-cholesterol levels, with an average total- and LDL-cholesterol level of 217 mg/dL and 139 mg/dL, respectively. In addition, the middle-aged men and women had no previous history of atherosclerosis. Roughly one-third of patients had hypertension, one-third of patients were past or present smokers, and the average body-mass index was 27 kg/m2.

In an assessment of how Canadian primary-care physicians assessed cardiovascular risk, the most commonly used method was the Framingham risk score. However, just one-third of physicians reported using Framingham, while another 30% reported using clinical judgment, and a surprising 15% reported using high-sensitivity C-reactive protein (hs-CRP) to assess risk. Slightly more than 10% of doctors reported counting risk factors, while others reported using other risk scores.

Comparing the correlation between the centrally adjudicated risk of patients with the risk assessed by primary-care physicians, the researchers observed that 65.7% of low-risk patients were properly classified as low risk. Among high-risk patients, 34.2% of high-risk patients were properly classified as high risk, while it was a "crapshoot" among intermediate-risk patients. Of these intermediate patients, 50.4% were properly classified as intermediate risk.

"Among low-risk patients, the overwhelming majority of whom do not need lipid-lowering therapy, the concern is that by misclassifying them into higher-risk categories, theoretically, at least one-third of Canadians might be inappropriately treated with statins because physicians think they're at higher risk," said Gupta. "That's not a tragedy if that's really occurring out there--the drugs are generally safe and well tolerated, and even in low-risk individuals there is a benefit, although it is small."

To heartwire , Gupta said the results are likely indicative of a "best-case scenario," as the physicians selected from the 105 clinical sites were chosen for their expertise and heightened awareness of the lipid guidelines. Gupta also said that the easily identifiable high-risk patients, such as those with diabetes or a history of cardiovascular disease, were excluded from the analysis.

Interestingly, the median hs-CRP among patients was 2.6 mg/L, which is above the cut point used in the JUPITER trial, leading Gupta to speculate that if the current Canadian guidelines were applied to this cohort, a "sizable proportion would be eligible for statin therapy." He is surprised that as many as 15% of primary-care physicians are using hs-CRP to assess risk, suggesting there is an overreliance on that marker alone and that cardiovascular experts have not done a good job educating physicians about how to properly use hs-CRP in clinical practice.

The research is supported by an unrestricted, investigator-initiated research grant from AstraZeneca to Gupta.


The complete contents of Heartwire , a professional news service of WebMD, can be found at, a Web site for cardiovascular healthcare professionals.