Lipids as risk factors for stroke and CHD

The pathophysiology of ischemic stroke is thought to be very similar to that of atherosclerosis. They share many common risk factors and treatments, and are often grouped into a single entity: cardiovascular disease. However, differences exist.

Serum LDL has been established as one of the strongest risk factors for CHD. Treatment targeted at reducing high LDL has been shown to reduce the incidence of CHD, establishing LDL treatment as the mainstay of primary and secondary CHD prevention for decades. The same is not true for LDL and stroke.

LDL does not significantly contribute to risk prediction of incident ischemic stroke. But, is treatment targeted at the reduction of LDL effective for reducing the incidence of ischemic stroke?

Risk factors vs. therapeutic targets

Statin therapy has been shown to prevent stroke in patients with established CHD and in patients who have already suffered from transient ischemic attack, stroke or symptomatic carotid artery stenosis. However, whether statins are beneficial in the primary prevention of stroke in patients without established CHD is less clear. In ALLHAT, pravastatin failed to produce a significant reduction in first-time ischemic stroke in those without CHD, or mild to moderate hypertension and mild to moderate LDL levels (mean LDL 129 mg/dL – 148 mg/dL). However, in ASCOT-LLA, patients with treated hypertension, below average serum LDL levels (mean LDL-C ~112 mg/dL) and no known CHD, atorvastatin (Lipitor, Pfizer) did indeed reduce the incidence of CHD and ischemic stroke.

Andrew DeFilippis

Rinky Bhatia

Steven Jones

The MEGA study demonstrated a 33% reduction (P=.01) in CHD events with diet and pravastatin therapy among patients with only mildly-elevated LDL at baseline (LDL 156 mg/dL). A substudy of the MEGA trial aimed at evaluating the effect of reducing cholesterol with diet and pravastatin among Japanese patients with mildly elevated cholesterol and mildly elevated BP has been published.

This substudy of MEGA included all 3,277 patients with hypertension (clinical diagnosis by physician at baseline). Baseline characteristics of these patients included an average age of 60 years, mean systolic BP of 141 mm Hg and LDL of 156 mg/dL. Approximately 90% were on antihypertensive medications. These characteristics were similar in those randomized to treatment with diet plus pravastatin vs. diet alone.

After five years of treatment in this trial, systolic BP and diastolic BP decreased by ~2 mm Hg in both groups. Diet plus pravastatin vs. diet alone reduced LDL (-20% vs. -4%, P<.001). Statistically significant reductions in cerebral infarction (relative RR 46%, P=.044), CHD plus cerebral infarction (RRR 35%, P=.019) and all CVD events (RRR 33%, P=.012) was observed in the pravastatin group.

This substudy of the MEGA trial supports the use of a statin to prevent cerebral infarctions in patients at risk for this disease — those with hypertension in this trial. Given that serum cholesterol is a poor predictor of ischemic stroke, how and why do statins reduce the risk of cerebral infarction? Does such a mechanism also explain how statins reduce CHD events in those with “normal” LDL?

Treatment vs. treatment of target

The TNT study randomized patients with known CHD, no prior history of ischemic stroke, LDL ~100 mg /dL and mean systolic BP of 131 mm Hg to receive high-dose statin (atorvastatin 80 mg daily) vs. low-dose statin (atorvastatin 10 mg daily). After an average of five years, patients randomized to the higher dose of statin had a mean LDL of 77 mg/dL as compared with those treated with low-dose statin who achieved a mean LDL of 101 mg/dL. Those randomized to high-dose statin enjoyed a 20% reduction in major coronary events (P=.002) and a 23% reduction (P=.007) in stroke/transient ischemic attacks when compared with those treated with low-dose statin.

Recently the JUPITER trial identified patients with “low” LDL (median LDL 108 mg/dL) but increased CHD risk identified by an elevated highly sensitive C-reactive protein (hs-CRP), (median hs-CRP ~4.2 mg/L). In this trial, treatment with rosuvastatin (Crestor, AstraZeneca) significantly reduced CHD events, strokes and all-cause mortality. With regard to stroke alone, statin therapy resulted in a 48% reduction, almost all thromboembolic, with no difference in hemorrhagic stroke appreciated. In considering TNT, MEGA and JUPITER, is it that statins are effective in reducing CVD events (ischemic stroke or myocardial ischemia) in patients at risk for these events (known CHD, hypertension or elevated hs-CRP) regardless of cholesterol levels, or is our standard of “normal” cholesterol incorrect?

The relationship between reducing LDL and reductions in CV risk is supported by our understanding of the disease pathophysiology and the association with multiple treatment modalities, including LDL apheresis, bile acid sequestrants, niacin and of course, statins. Furthermore, posthoc analysis from the PROVE- IT-TIMI 22 trial demonstrates results on treatment CV event rates irrespective of the statin type/dose/potency (atorvastatin 80 mg daily or pravastatin 40 mg daily) used to achieve an LDL of 70 mg/dL and hs-CRP ≤2 mg/dL, suggesting that the level of LDL and hs-CRP achieved is responsible for the reduction in CV risk.

However, not all LDL-reducing modalities correlate with reductions in CV risk. Estrogen therapy, thyroxine and torcetrapib (Pfizer) all reduce LDL but all also increase CV risk. The trials which demonstrate an association between reductions in cholesterol with reduced CVD events all treat to a dose of an intervention (ie, target statin dose), not to a target LDL. This type of trial design does not allow one to prospectively examine if the level of LDL reduction or the dose of the intervention (ie: statin dose) is responsible for the reduction in CVD events.

In the AFCAPS/TexCAPs, the on-treatment LDL level did not predict CVD events. Analyses of patients on statin therapy in TNT, IDEAL and LIPID showed that LDL was a poor predictor of on-treatment CVD events. Unfortunately, in the MEGA trial, the sample size was too small to evaluate the relationship between the reduction in cerebral infraction with statin use independent of LDL reduction. Multiple biologically plausible mechanisms, independent of LDL lowering, have been identified, which may explain reductions in CHD and stroke with statin therapy.

What is the right target?

In the studies noted above and in clinical practice, LDL is measured as a concentration of cholesterol contained within LDL particles. While it is easy to measure the concentration of cholesterol in LDL, the amount of cholesterol in LDL is dynamic and varies greatly within and between patients. The concentration of cholesterol within a particle is influenced by genetics, insulin sensitivity, enzyme activity, statin therapy and factors yet to be delineated. The number of LDL particles can be measured directly by nuclear magnetic resonance (NMR) or accurately estimated by measuring apolipoprotein B100, since every LDL particle has one apolipoprotein B100. Apolipoprotein B100 is found on other arthrogenic lipoproteins [LDL, IDL, VLDL, Lp(a)], but given that LDL typically accounts for 95% of all apolipoprotein B100 particles - apolipoprotein B100 correlates well with LDL particle count by NMR. The American Diabetes Association and the American College of Cardiology Foundation recognized the superiority of apolipoprotein B100 and LDL particles over LDL concentration in the prediction of CVD events in patients with diabetes or the metabolic syndrome with LDL ≤130 mg/dL when on pharmacotherapy. The American Association for Clinical Chemistry recognizes apolipoprotein B100 and LDL particle count as superior predictors of CVD events in all individuals on statin treatment. Although apolipoprotein B100 and LDL particles may predict events in those on treatment more accurately than LDL concentration, no trials have compared the effectiveness of treatment to a particular dose of statin vs. treatment titrated to any lipid parameter, including apolipoprotein B100 and LDL particle count.

The question remains: Does available evidence support treatment (with statins) rather than to lipid goals? With the fourth Report of the Expert Panel on Detection, Evaluation, and Treatment of the High Blood Cholesterol in Adults (Adult Treatment Panel IV) underway, a simpler approach of identifying those at increased CVD risk and treating with proven therapy of diet, exercise and statin, regardless of LDL levels, should be considered.

Andrew DeFilippis, MD, is a Cardiology Fellow at The Johns Hopkins Ciccarone Center for Prevention of Heart Disease.

Rinky Bhatia, MD, and Steven Jones, MD, are Assistant Professors of Medicine, Johns Hopkins Hospital.

For more information:

• Am J Cardiol. 1998;82:1489-95.
• Biochim Biophys Acta. 2000;1529:299-309.
• Circulation. 2000;101:477-84.
• Circulation. 2002;105:1162-9.
• Circulation. 2008;117:3002-9.
• Clin Chem. 2009;55:407-19.
• Curr Pharm Des. 2009;15:479-89.
• Diabetes Care. 2008;31:811-22.
• Hypertension. 2009;53:135-41.
• JAMA. 1970;214:1303-13.
• JAMA. 1975;231:360-81.
• JAMA. 1984;251:365-74.
• JAMA. 2002;288:321-33.
• JAMA. 2002;288:2998-3007.
• Lancet. 1995;346:1647-53.
• Lancet. 2003;361:1149-58.
• Lancet. 2005;366:1267-78.
• Lancet. 2006;368:1155-63.
• N Engl J Med. 2005;352:20-8.
• N Engl J Med. 2005;352:1425-35.
• N Engl J Med. 2007;357:2109-22.
• N Engl J Med. 2008;359:2195-207.
• Ridker PM GR, on behalf of the JUPITER Trial Study Group. Rosuvastatin in the Prevention of Stroke Among Men and Women with Elevated Levels of C-Reactive Protein: The JUPITER Trial. Presented at: American Stroke Association Meeting; Feb. 14-18, 2009; Orlando, Fla.