The Family Health Guide

A better way to predict cardiovascular risk

Since the 1980s, enormous progress has been made in understanding cardiovascular disease (CVD) in women. Recent findings help explain why CVD (heart disease and stroke) has been more difficult to diagnose and treat in women than in men: heart attack symptoms are often different; standard diagnostic tests don't predict heart problems as reliably; and women may not get as much benefit from such treatments as angioplasty and bypass surgery.

Also a lot has been learned about the biology of cardiovascular disease. It's not just a matter of cholesterol clogging the arterial plumbing. Low-level inflammation also contributes to the atherosclerotic plaques that can block blood flow to the heart and brain. Researchers at Boston's Brigham and Women's Hospital found an association between high-sensitivity C-reactive protein (hsCRP) — a marker for inflammation — and the risk of having a heart attack or stroke in healthy men and women.

Despite these advances, the usual models for predicting CVD in women don't take into account markers of inflammation. A study published in February 2007 proposes a model that incorporates hsCRP and promises greater precision in identifying women at risk.

What's wrong with the current model?

To determine the risk for cardiovascular disease, clinicians look at factors such as age, cholesterol levels, high blood pressure, smoking status, and diabetes. These factors are incorporated into the Framingham Risk Score, a risk-assessment tool. This tool is used to evaluate the 10-year risk of having a heart attack in both women and men. All high-risk women and some in the moderate-risk groups are advised to modify their diet and other lifestyle factors and possibly take medications that lower LDL "bad" cholesterol.

The trouble is that up to 20% of heart attacks occur in women without any of the major risk factors covered by the Framingham model, partly because that model doesn't include markers of inflammation or genetic predisposition, both of which are important in CVD.

The moderate-risk groups are the most puzzling. Many doctors suspect that some of these women are actually at high risk and need more intensive treatment, so they have turned to hsCRP testing (which can be performed at the time of cholesterol evaluation) to identify those at higher risk than the Framingham model would suggest. A woman who is at moderate risk by Framingham criteria but has an hsCRP level greater than 3.0 mg/L could actually be at high risk even if her cholesterol levels are normal.

Finding a new predictive model

The Brigham and Women's researchers set out to develop a risk model for women that combined newer risk markers, including hsCRP, with traditional risk factors and family history. They assessed 35 risk factors among 24,558 initially healthy women ages 45 and over.

The researchers used data from two-thirds of the women selected at random to develop a risk model that takes better account of both inflammatory biomarkers and heredity. The new model contains eight risk factors, five of which are familiar from the Framingham Risk Score — age, smoking status, systolic blood pressure, HDL ("good") cholesterol, and total cholesterol — plus hsCRP, parental history of a heart attack before age 60, and for women who have diabetes, hemoglobin A1c (a measure of blood sugar control).

To test the model, the researchers applied it to the remaining one-third of study participants and found that it was more accurate than the Framingham model. Of the women whose Framingham risk scores placed them at moderate risk, 40% to 50% were reclassified into higher- or lower-risk groups that better matched their actual rate of cardiovascular events.

This model, called the Reynolds Risk Score, could help clinicians target women who could benefit from more aggressive preventive treatment, including diet and exercise, a statin or other cholesterol-lowering medication, and possibly aspirin (which has been shown to reduce the risk of heart attack in women ages 65 and over).

Reynolds Risk Score factors

  • Age
  • Blood pressure
  • hsCRP level
  • Total cholesterol
  • HDL ("good") cholesterol
  • Smoking status
  • Family history of a heart attack before age 60
  • Hemoglobin A1c (in women who have diabetes)

What now?

You can calculate your Reynolds Risk Score at The model still needs more testing, as the authors point out. For example, most of the study participants were white professional women; the results might not apply fully to other groups.

The evergreen recommendations for reducing cardiovascular risk haven't changed: Don't smoke, actively or passively; control your weight and cholesterol with diet and exercise; and if you have high blood pressure or diabetes, do what's necessary to get it under control — including appropriate medications. If your cholesterol is normal but you have another major risk factor for cardiovascular disease, it's worth asking your doctor whether hsCRP testing could help clarify your overall risk.

June 2007 update