By: Dr. Geoffrey Modest
A recent study assessed the accuracy of the 2013 ACC/AHA guidelines in identifying patients in the Framingham Study who either developed cardiovascular disease (CVD) and/or coronary artery calcification (CAC), comparing the results to that of the prior 2004 ATPIII guidelines. (see JAMA. 2015;314(2):134-141).
Details:
–2435 statin-naive participants (mean age 51.3, 56% female, mean Framingham Risk Score — FRS — of 6.7%, mean LDL 121 mg/dL, mean CAC score of 95 with 42% having CAC score >0, overwhelmingly white population) were enrolled from the longitudinal offspring and third-generation cohorts of the Framingham Study. median followup 9.4 years.
Results:
–39% of patients were deemed statin-eligible by the 2013 guidelines vs 14% by ATPIII.
–There were 74 (3.0%) incident CVD events (40 nonfatal MIs, 31 nonfatal strokes, 3 fatal coronary artery disease/ CAD events).
–Those statin-eligible by 2013 guidelines had increased HR for incident CVD vs those eligible by ATP III: HR 6.8 (3.8-11.9) vs 3.1 (1.9-5.0), with p<0.001
–by the ATP III guidelines 6.9% of those statin-eligible vs 2.4% statin-ineligible had incident CVD event. by the 2013 guidelines, 6.3% of statin-eligible had CVD and 1.0% of those statin-ineligible
–for women, the numbers were smaller than for men (2.0% vs 4.4% had incident CVD). There was a nonsignificant difference between statin-eligible and statin-ineligible by the ATPIII guidelines, though there was a significant (p<0.001) difference in those groups by the 2013 guidelines
–there were similar results in those at intermediate Framingham Risk Score (FRS), 38% of the total population (3.9% developing CVD), but the CAC score was significantly higher in people with intermediate FRS who developed incident CVD. 80% of this intermediate group were statin-eligible by the 2013 guidelines vs 27% by ATP III [note: the definition of intermediate FRS is somewhat complex and depended on the LDL level and number of CAD risk factors]
–those with CAC scores>0 were more likely to be statin-eligible by 2013 guidelines than ATP III (all with p<0.001):
–CAC score >0 (n=1015) 63% vs 23%
–CAC score >100 (n=376) 80% vs 32%
–CAC score >300 (n=186) 85% vs 34%
–CAC score of 0 was associated with very low CVD risk in stain-eligible people by the 2013 new guidelines, comprising 33% of the cohort, and with a CVD rate of only 1.6%
–if all of the patients with CAC score of 0 were considered to be statin-ineligible (ie, this was a strong negative risk factor for starting statins), there was no difference in the % of incident CVD in the statin-ineligible groups by either the ATP III or 2013 guidelines, and a much higher % incident CVD in the statin-eligible group [which suggests that those with CAC=0 really would not benefit from statins, leaving even greater benefit for those with CAC>0 who were statin-eligible by either guideline]
Several points:
–FRS did not fare well in differentiating likely statin benefit as compared to the 2013 guidelines. One suggestion from the data is that perhaps a large part of the reason is that the 2013 guidelines captured more people with CAC scores>0.
–the 2013 guidelines cast a much wider net in terms of those who would be statin-eligible, which I think include a lot of people at very low risk of cardiovascular disease (see prior blogs: here , here, and here) For example, using the 2013 guideline risk calculator, a 60 year old white male nonsmoker with systolic BP on 130 on meds and an LDL of 70 and an HDL of 70 should get a moderate intensity statin. Same for a 45 year old African-American with the same clinical picture. So, it is not surprising that adding an additional 12.8 million adults who would get statins by the 2013 guidelines over the ATP III guidelines would lead to less heart disease, but would also treat lots of patients at very low risk
–CAC is the one CAD marker which is felt to be the most useful of the non-traditional ones (ie, better than C-Reactive Protein — CRP, etc) by the American Heart Assn. Many studies have shown that low CAC scores nearly exclude clinically important CAD (as in the current JAMA study, with other studies finding similarly about a 1% risk in 10 years). And those with extensive CAC have 10-fold increased risk after multivariate adjustment. This really is not so surprising: the existence of any atherosclerotic disease (eg peripheral artery disease) is associated with a significantly increased risk of CAD, so showing calcification of the coronaries themselves is at least as likely to predict risk of CAD…
— the Multi-Ethnic Study of Atherosclerosis looked at the relative risks of CRP vs CAC (see Lancet 2011; 378: 684–92), following 950 people for 5.8 years, finding no association of clinical events with CRP levels, but
–47% had initial CAC scores of 0 and had CAD event rate of 0.8/1000 person-years
–74% of all of the CAD events were in the 25% with CAC scores >100, at a rate of 20.2 events/1000 person-years
–for CAD, the 5-year NNT for CAC=0 score was 549, but was 94 for scores 1-100, and 24 for scores >100
–for CVD, the 5-year NNT for CAC=0 score was 124, but was 54 for scores 1-100, and 19 for scores >100
So, this new JAMA study by itself does not add much, other than that by really spraying statins relatively indiscriminately all over the place, there are fewer atherosclerotic events (this would likely happen by just lowering the threshold for treatment with the old ATP III guidelines). As a person who does prescribe lots of statins (always reinforcing lifestyle changes as well) to many more patients than those qualifying by ATPIII criteria (including those with PAD or any evidence of atherosclerotic disease, and often those with metabolic syndrome or glucose intolerance — as noted in prior blogs), and as someone who is pretty strongly anti-radiation exposure (eg, see here), I am leaning toward including CAC as part of my CAD risk assessment (i haven’t done so yet, but am thinking pretty strongly about doing so, if covered by insurance), since:
–CAC pretty directly correlates with coronary atherosclerosis and ultimately with cardiac events (and a score of 0, present in a good proportion of patients even at intermediate CAD risk, seems to render CVD incidence miniscule)
–the actual radiation exposure is only 0.89 mSv in the MESA study, well below both the overall background radiation (3 mSv/year) and the average dose in the now-accepted low dose lung CT scans for smokers (1.5 mSv and this is annually, though really the average annual dose is 8 mSv given the number of false positives leading to more intensive radiologic investigation: see here). And the cost in 2011 was about $100 for the CAC scan.
–and there really are harms to prescribing statins to those who do not need them: occasional really bad adverse events (I’ve seen a couple who have had bad rhabdo, renal failure, and prolonged dialysis), more frequent less bad events (diabetes, maybe cognitive decline), troublesome events (myalgias, bad dreams…), cost of meds, cost of followup exams and labs, and increasing the medicalization of the patients.