BOSTON, Feb. 26 -- A risk stratification tool to predict atrial fibrillation risk may open the way for primary prevention, researchers here said.
The risk prediction model -- based on factors easily measured in a primary care practice -- had good discrimination (C=0.78) with little additional predictive value gained from adding echocardiography, Emelia J. Benjamin, M.D., of Boston University, and colleagues reported in the Feb. 28 issue of The Lancet.
The ability to identify high-risk patients represents a valuable first step toward prevention, according to an accompanying commentary by David B. Brieger, M.B.B.S., Ph.D., and S. Ben Freedman, M.B.B.S., both of the University of Sydney. "Primary prevention of atrial fibrillation has not been on our radar; even secondary prevention of the arrhythmia has not been recommended," they said, noting that many clinicians accept the condition as an unavoidable evil and focus on treating complications like stroke.
But primary prevention seemed more plausible after the ATHENA trial results published this month, which showed secondary prevention effective in reducing cardiovascular mortality and hospitalization for atrial fibrillation patients, they added. (See: Dronedarone Beneficial for Older Patients with Arrhythmias)
To provide the clinical framework, Dr. Benjamin's group developed a risk score based on retrospective analysis of the Framingham Heart Study.
Among 4,764 participants ages 45 to 95 without baseline atrial fibrillation, 10% developed the arrhythmia over 10 years of follow-up.
The significant risk factors associated with atrial fibrillation in the cohort included in the final multivariable model were: age, sex, body mass index, systolic blood pressure, treatment for hypertension, PR interval, significant cardiac murmur, and heart failure (all P<0.05 except body mass index which was P=0.08).
The researchers called this model "reasonably accurate for stratification of individuals into risk categories" (C=0.78, 95% CI 0.76 to 0.80) as a measure of discrimination.
The model indicated more than 15% risk of atrial fibrillation in only 1% of patients under age 65 but 27% of older patients.
When the risk model was applied to another Framingham Heart Study sample of 5,152 participants, the researchers found a similar predictive ability (C=0.76, 95% CI 0.74 to 0.79) and good calibration for deciles of predicted risk.
However, adding echocardiographic measurements -- left atrial diameter, left ventricular wall thickness, and left ventricular fractional shortening -- to the model did little to improve discrimination of atrial fibrillation risk.
Although the C value rose significantly (0.79 versus 0.78, P=0.005), the net number of patients who were reclassified correctly based on inclusion of echocardiography in comparison to the number incorrectly reclassified was not significant (P=0.18).
"In view of costs, use of routine echocardiography to predict risk of atrial fibrillation is unlikely to be justifiable for primary prevention screening in the general population," Dr. Benjamin's group said.
But subgroup analyses showed that individuals with valvular heart disease or heart failure had more accurate risk scores with inclusion of echocardiographic measurements (P=0.03).
Although the researchers gave preference to clinical measures that were most easily taken (such as systolic blood pressure rather than pulse pressure), Dr. Benjamin noted that risk scores are generally underutilized in clinical practice.
But "with this condition, few would argue against the assertion that an ounce of prevention is worth a pound of cure," editorialists Drs. Brieger and Freedman said.
They also wrote that the researchers have developed "a relatively simple tool [that can be] calculated by an algorithm or computer (or handheld) in the primary care setting."
They noted that therapies such as statins, fish oil, ACE inhibitors, angiotensin-receptor blockers, and various antiarrhythmics have been suggested to be effective in prevention.
In addition to finding cost-effective prophylaxis, research is needed to validate the risk score and determine how well it extrapolates to diverse populations, the researchers said.