stroke: treating LDL to target

A new article from last week's Am Heart Assn meetings found that treating LDL to a lower target led to fewer subsequent cardiovasc events in those with TIA/ischemic stroke (see stroke LDL target nejm2019 in dropbox, or DOI: 10.1056/NEJMoa1910355)

Details:

--2860 patients with evidence of atherosclerotic disease, having had ischemic stroke in prior 3 months or TIA in the previous 15 days, were randomized to a target LDL of <70 mg/dL vs 90-110mg/dL, using adjustable doses of statins with ezetimibe as needed, in the Treat Stroke to Target trial from France (61 sites) and South Korea (16 sites)

    --ischemic stroke was defined as symptoms with documented  ischemic lesion on CT or MRI; TIAs had to be a motor deficit in one arm or leg, or speech disturbance >10 minutes

    --all patients were screened with nonivasive imaging (carotid duplex, CT angiography, and MR angiography) per the AHA-ASA guidelines.  Also all got trans-esophageal echo or CT angio of the aorta to detect aortic atheroma.

    --all patients had to have documented atherosclerotic disease: stenosis of extracranial or intracranial cerebral arteries, ipsilateral or contralateral to the region of imputed brain ischemia; atherosclerotic plaques of aortic arch at least 4mm thick; or known history of CAD

    --statin doses or added ezitimibe were adjusted 3 weeks after randomization to reach assigned LDL target, and were rechecked every 6 months, with adjustment of therapy as needed

--mean age 66, 67% male, 85% ischemic stroke/14% TIA, median time since index event 6 days, htn 65%/diabetes 23%/dyslipidemia 61%/current smoker 23%/former smoker 30%, CAD 17%, BP 140/80; A1c 6.3%; 55% had no prior use of statins

--mean LDL at baseline: 135 mg/dL (HDL 50 mg/dL, triglycerides 122); all had modified Rankin score post-stroke of 0-3 (0=no symptoms, 1=no disability, 2-3 needing some help in daily activities)

--primary endpoint: composite of ischemic stroke, MI, new symptoms leading to urgent coronary or carotid revascularization, or death form cardiovasc causes

--secondary endpoints: MI or urgent coronary revascularization after new symptoms; cerebral infarction or urgent revasc of a carotid or cerebral artery after TIA; cerebral infarction or TIA; any revasc of coronary, cerebral or peripheral artery; cardiovasc death; death from any cause; cerebral infarction or intracranial hemorrhage; intracranial hemorrage; newly diagnosed diabetes.

--median followup was 3.5 years (5.3 years in France and 2.0 years in S Korea); this study was stopped early because of administrative reasons (ie, they ran out of funding), so fewer endpoints occurred (277) than anticipated (385)

--study funded by French Ministry of Health and others

Results:

--lipid-lowering therapy:

    --94.0% in higher-target and 33.8% in lower-target groups were only on a statin; 5.8% and 33.8% were on statin plus ezitimibe, respectively

--mean achieved LDLs:

    --lower-target group: 65 mg/dL

        --47% of patients were in target range; 45% above target; 8% below target

    --higher target group: 96 mg/dL

        --32.2% of patients were in target range; 49% were below target; 17% above target

--composite primary endpoint:

    --lower-target group: 121 patients (8.5%)

    --higher target group: 156 patients (10.9%)

    --so, 22% reduction in primary endpoint, adjusted HR 0.78 (0.61-0.98), p=0.04

--none of the individual endpoints reached statistical significance

    --though, comparing lower vs higher LDL targets: 81 vs 100 had cerebral infarction/stroke endpoints; next most common was nonfatal acute coronary syndrome at 15 vs 23 events

--a review of their graph shows that the separation of the curves for the primary endpoint began at 6 months, and remained pretty parallel after 1 year (ie, the reduction overall happened in the first year, the majority of which was in the first 6 months, then remained the same for the rest of the study)

--no significant difference in intracranial hemorrhage or newly developed diabetes (both found to be issues in prior statin studies), though both had statistical trend to be worse in lower-target group:

    --intracranial hemorrhage in 18 patients in lower-target group (1.3%) vs 13 in higher-target (0.9%); HR 1.38 (0.66-2.82)

    --new onset diabetes in 103 (7.2%) in lower-target group vs 82 (5.7%) in higher-target with HR 1.27 (0.95-1.70)

Commentary:

--the 2006 SPARCL study (Stroke Prevention by Aggressive Reduction in Cholesterol Level)  found a 16% lower incidence of recurrent strokes with atorvastatin 80mg vs placebo, and a 33% lower incidence in those with carotid stenosis.  this was the major study leading to using statins in those with ischemic strokes or TIAs. A subsequent analysis of SPARCL found that those who achieved an LDL <70 had a 28% lower relative risk of stroke vs those at 100 mg/dL (a stroke risk reduction of 20% for each 39 mg/dL lowereing of LDL, without any lower threshold)

    --This current study is the first RCT to really test the hypothesis that those randomized to a lower LDL  would achieve a lower risk of recurrent cardiovascular events

--a few comments/issues:

    --these patients, all with atherosclerotic disease, really should have been on a baseline statin.  The average LDL was 135 and 55% of them had no prior use of statins (pretty awful for a current study)....  makes one wonder about their overall quality of care....

    --a large % did not sustain their LDL targets. also a little surprising in the setting of a clinical study.  Though it is certainly true that many patients do not adhere well to statins, at least in my pretty extensive experience there are a large number who do and repeated lipid testing confirms quite narrow disparities in LDL from year to year.

        --so, it would have been useful to have a breakdown by the actual achieved LDL: what was the risk reduction in those who maintained the LDL levels in the tartgeted range?  this information might help us understand the real potential benefit of different LDL levels and help in encouraging patients on the importance of medication adherence

    --there are a few reasons to think that a lower LDL target might be even more beneficial than they found in this study:

        --49% of those randomized to the higher LDL target actually had lower LDLs than they were supposed to, and 45% randomized to the lower LDL target actually had higher LDLs than they were supposed to.  ie, the results are diluted a bit, since about ½ of the patients in each targeted group actually had LDLs closer to what the other targeted group was supposed to have

        --and, the fact that the study ended early without adequate sample size does dilute the possibility of finding statistically significant benefit (or adverse events) for the array of secondary outcomes evaluated

--it is reassuring that there were no significant increases in diabetes or intracranial hemorrhages with more intensive therapy, though prior studies showing increases were based on statins vs placebo, not differing intensities of statins

    --http://gmodestmedblogs.blogspot.com/2017/11/statin-use-and-diabetes-is-that-real.html is a blog which reviews the studies on statins and diabetes, including one finding that low LDL itself is associated with diabetes (eg, some people with variants of the HMG-CoA reductase gene conferring low LDL levels do have increased risk of diabetes without taking statins). also, those with A1c's nearer the diabetes range are more likely to develop diabetes on statins. not so surprising. but the protective effect of statins in diabetes (and probably in prediabetes as well, since these individuals have a significantly increased cardiovasc risk) undoubtedly outweighs the small (about 10%) increased risk of crossing our diabetes-defining A1c= 6.5% threshold to qualify them as diabetics (by the way, the 6.5% is based on that being the inflection point for developing diabetic retinopathy, and has nothing to do with developing the remarkably common and potentially fatal macrovascular cardiovascular outcomes, which are increased at a signifantly lower A1c level than 6.5%: eg see http://gmodestmedblogs.blogspot.com/2016/12/prediabetes-and-cardiovascular-risk.html)

--another issue not addressed in this study is the role of HDL: the TNT trial with differing doses of atorvastatin found that the 5-yr risk of major cardiovasc events  was similar in those with LDL >100 but HDL >55 as in those with LDL <70 but HDL 38 (see brief description and graphs in https://gmodestmedblogs.blogspot.com/2015/06/improve-it-trial-ezetimibe.html ).  would have been great to see if the achieved HDL affected the benefit of the achieved LDL in this study

--i would also add my concerns about the "high intensity" vs "low intensity" statin approach:

    --i have seen many patients have a very dramatic and in-target range response to a "low intensity" statin (one recently went from an LDL of close to 200 mg/dL down to 80 mg/dL on atorvastatin 10mg

    --and, there is great variability in the effect of statins in these intensity categories: again, i have seen many patients have much more LDL lowering on switching from atorvastatin 40-80 mg to rosuvastatin 40mg, thouugh both are considered "high intensity"

    --so, in light of the association of statin dose with adverse effects, and in the hopes in general of using the least medication as possible, i do typically start with a lower dose statin in all patients and titrate to what i think is an appropriate LDL goal (of course, this is in conjuction with helping patients achieve a more healthy lifestyle, especially in terms of diet and exercise)

--limitations of the study: the big one is that it ended prior to the anticipated ending point, leading to fewer patients recruited, and less statistically significant outcomes (by their calculations, the full study would have taken 3 more years of recruiting patients, allowing for more cardiovasc events and clearer statistical significance, as noted above). And also not enough time to assess differences in adverse events. Also, would be best to have a more diverse cohort of patients to assume generalizability of results

--other relevant blogs:

--http://gmodestmedblogs.blogspot.com/2017/03/treating-ldl-to-target-more-evidence.html is a Korean observational study finding that those with CAD who achieved an LDL <70 had a decreased progression of atherosclerotic plaques

--http://gmodestmedblogs.blogspot.com/2013/12/further-thoughts-on-am-heart-assn-2013.html has my argument that we should be treating LDL to goal, in contrast to the 2013 AHA guidelines

--http://gmodestmedblogs.blogspot.com/2017/09/new-acc-guidelines-endorse-ldl-goal.html has new ACC guidelines endorsing a targeted LDL goal, based onthe PCSK9 inhibitor studies (a reversal of their 2013 recommendations)

--http://gmodestmedblogs.blogspot.com/2015/06/improve-it-trial-ezetimibe.html critiques the ezetimibe study but also has data showing that the achieved LDL correlates with cardiovasc events

--http://gmodestmedblogs.blogspot.com/2018/08/very-low-ldl-levels-benefit-without-harm.html is a meta-analysis finding that the lower the LDL the better (includes the PCSK9 studies), finding a linear decrease in cardiovasc events down to an LDL of 21 mg/dL

so, this article reports a well-designed study finding that treating patients with cerebrovascular disease to a lower target LDL confers improved cardiovascular outcomes. it add some strong data, from an RCT, that reinforces the conclusion that in general we should use a lower LDL target for those at higher cardiovascular risk (and not just choosing a "high intensity" vs "low intensity" statin, as above), which might include adding ezetimibe to maximal statin therapy. based on the TNT trial mentioned above, i do think it is reasonable to incorporate HDL into the equation, at least to the point that those with low HDLs would probably do better with even lower LDLs.  And, per the meta-analysis of observational studies noted above, there does seem to be some benefit even with very low LDLs in those at very high cardiovasc risk.

Bottom line: I think this RCT adds important support for treating all atherosclerotic disease to a target LDL mg/dL, that being <70 for those high-risk (though lower may even be better in those at really high risk).

geoff​

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@uphams.org

to get access to all of the blogs:

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

2. click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

3. or you can just click on the magnifying glass on top right, then  type in a name in the search box and get all the blogs with that name in them

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list