Stroke and calcium

one important issue to keep in mind is the difference between community data (eg, risk of cardiovasc dz) through predictive models such as the framingham risk score (FRS) and the actual risk of an individual (eg, by looking at actual disease present).  although the community risk scores are important for us to improve modifiable risks, we all have seen patients at remarkably high risk who have no evidence of heart dz (eg clean coronaries) for example.  there have been several models looking at the individual's risk, including ankle-brachial indices or carotid intima-media thickness (see below). in addition, CAC (coronary artery calcification) has been shown in several studies to correlate strongly with actual heart disease. in this light, recent study (see stroke cac predictor stroke 2013 in dropbox) of 4200 subjects in Germany aged 45-75, 47% men and without prior stroke, CHD, or MI were evaluated for development of stroke using CAC as well as more traditional risk factors of age, sex, systolic BP, LDL, HDL, DM, smoking, and atrial fib. results:

--CAC was an independent stroke predictor with HR 1.52 , both in men and women, esp in those <65 years old (HR 2.21 in young, nonsignif in >65 yo)

--other risk factors: age with HR 1.35  per 5 years, syst BP with HR 1.25 per 10mmHg, and smoking with HR 1.75. 

--CAC was especially useful in those in the low (<10%) and intermediate (10-20%) framingham risk score (FRS) categories. for example, those with FRS <10% but CAC in highest category had a 13 fold increased stroke risk (they do not comment on how frequent that combination was, but i suspect pretty small). those with FRS 10-20% had 8-fold increased risk in those with highest CAC score. as a whole, those with the highest CAC scores (>400) had a dramatically increased stroke risk independent of their FRS.

so, thought this might be interesting. brings up that issue of community vs individual data.  in terms of the other individual evals noted above, 

-- there was a recent review/meta-analysis of carotid intima thickness in the journal atherosclerosis (see cad cimt review athero 2013 in dropbox) which did find a slight increase in predicting cardiac events (RR 1.25), but this did not add any statistical value above assessing the traditional risk factors (ie not so useful). 

--another recent review of ankle-brachial index in the journal circulation (see cad risk ankle brachial index review circ 2012 in the dropbox) found that an ABI <=0.90 was associated with a twice the risk of total mortality, cardiovasc mortality and major coronary events compared with the overall rate in each FRS category. in men with FRS>20%, a normal ABI was found in 43% of cases and effectively reclassified them into intermediate-risk category.  in women, those at low risk (FRS<10%) and intermediate risk (between 10-20%) but with abnormal ABI were reclassified as high risk. so, the ABI adds important predictive value over the FRS. and pts with known CAD with low ABI are at higher risk of events compared to those with CAD and normal ABI.

what to do???  at this point, seems that both ABI and CAC add information, but i am not ready to use them.  CAC is particularly problematic, given the cost, radiation exposure, etc.  ABI is more practical (we can do it in the health center -- nursing visit with Peri and it takes only 10-15 minutes), but we could not handle the volume of patients who might benefit. so, just FYI, but it seems we are moving overall into a more individual-based risk assessment, here and elsewhere.