By: Dr. Geoffrey Modest
Two articles came out in the New England Journal of Medicine regarding PCSK9 inhibitors and impressive decreases in LDL levels but also finding significant decreases in cardiovascular events. PCSK9 (proprotein convertase subtilisin-kexin type 9) is a protease which degrades hepatic LDL receptors (I have appended one of my blogs from 2012 at the bottom, which goes into a tad more detail). Overall these PCSK9 inhibitors work in some ways similar to statins, by increasing hepatocyte LDL receptor activity.
- Alirocumab is a monoclonal antibody which inhibits PCSK9 and reduces LDL levels. A study of quite high-risk patients with LDL>70 mg/dl on a statin at the maximum tolerated dose were assigned to biweekly subcutaneous alirocumab vs placebo over a period of 78 weeks (see N Engl J Med 2015;372:1489-99).
–2341 patients, mean age 60.5, 62% male, 93% white !!), BMI=30, 18% with heterozygous familial hypercholesterolemia, 69% with CAD, 41% with CAD-risk equivalent, 35% with diabetes, 21% smoker. 47% on high-dose statin, rest on “any statin” and 28% with “other lipid-lowering therapy”, mostly ezetimibe. baseline LDL 122 mg/dl, HDL 50 mg/dl, fasting TG 132mg/dl, lipoptrotein(a) [Lp(a)] of 21.
–at week 24 there was a dramatic 62% decrease in calculated LDL levels with the drug (p<0.001) down to 48 mg/dl (the effect persisted for the full 78 weeks, if one includes only patients who continued to take the drug), and decrease in Lp(a) of 28% (which i believe only really happens with niacin, and does not with statins). HDL increased 5%. 80% of the patients on the drug had an LDL<70 mg/dl. no diff in LDL reduction in the different groups of patients (ie, whether they had familial hypercholesterolemia or other risk groups).
–in post-hoc analysis of prespecified cardiovascular outcomes: rate of major cardiovascular events (death from CAD, nonfatal MI, fatal or nonfatal ischemic stroke, unstable angina) was 1.7% vs 3.3% [HR=0.52 (0.31-0.90), p=0.02]
–adverse effects: injection-site reactions (5.9% vs 4.2%), myalgia (5.4% vs 2.9%), neurocognitive effects of confusion/amnesia/memory impairment (1.2% vs 0.5%), eye events (2.9% vs 1.9%) Study drug/placebo discontinuation rates: 28.2%/24.5%
–of the 37% of patients on the drug who achieved an LDL<25 mg/dl, there was no difference in adverse events, including neurocognitive.
- Evolucumab, also a monoclonal antibody which inhibits PCSK9 and is administered subcutaneously biweekly,was used in another study (see N Engl J Med 2015;372:1500-9). patients on standard therapy were randomized to evolucumab or placebo and followed 11 months.
–4465 patients, mean age 58, 50% male, 86% white, 47% from the US/40% Europe/13% Asia Pacific or South Africa. risk factors: hypertension 52%, diabetes 13%, metabolic syndrome 33%, current smoker 15%, and estimated CAD risk of >10% in 10 years in 45%. 20% had prior CAD, 9% with cerebrovasc or periph vasc disease, 70% on statin (27% high-intensity), 14% on ezetimibe. baseline LDL 120 mg/dl, HDL 51 mg/dl, TG 120 mg/dl (so, this was a less sick population than the first study)
–primary endpoint: adverse events. secondary endpoint was LDL decrease. prespecified exploratory outcome with adjudicated cardiovascular events = death, coronary events (MI, unstable angina requiring hosp or revascularization), cerebrovascular events (TIA, stroke), heart failure requiring hosp.
—LDL was reduced 61% to median of 48 mg/d. these changes were consistent from 4 weeks of therapy until 48 weeks (end of study). also found were non-HDL reduction of 52.0%, apo B of 47.3%, TG of 12.6%, lipoprotein (a) of 25.5%, and HDL increase of 7%.
–rate of cardiovascular events was decreased at 1 year from 2.18% to 0.95% with drug [HR 0.47 (0.28-0.78, p=0.003]. And there was a splaying of the curves at one year, suggesting increasing efficacy of evolucumab over time
–adverse events in 69.2% of evolocumab group and 64.8% of placebo. serious adverse events in 7.5% in each group. No difference in transaminases, CKs. low level of neurocognitive events (<1%), more in the evolucumab group, but, as with the first study, these were not related to the LDL achieved.
So, these articles reveal pretty striking effects of two PCSK9 inhibitors (with remarkably similar changes of 60+% in each study in lipids even after the statin-induced decreases and clinical cardiovascular outcomes decreased by 50+%), though it is important to remember that the clinical efficacy was not a primary outcome, the studies were short-term, and we need longer-term studies both to convince us of efficacy and reveal adverse effects. But, I bring up these articles for a few reasons.
–It is clear that statins either are inadequate or not tolerated in many high-risk patients. reducing PCSK9 activity is the most promising approach I have seen for years, though it would be particularly useful to have an oral and inexpensive medication (how’s that for pie-in-the-sky?). Note that these studies did not look at PCSK9 inhibition as a single agent, just in conjunction with a statin (and it is impressive these PCSK9 inhibitors add so much to statins, given their apparently similar mechanisms of action).
–It is important that there were not many neurocognitive adverse events even with very low LDL levels, since some older studies had found increases in suicides/accidents with lipid lowering, and this was putatively attributed to changes in neuronal cell membrane lipid composition affecting neural signal transmission in the brain. The PCSK9 drugs offer the potential to look at lots of patients with very low LDL levels, and it is reassuring that there were not obvious issues in these 2 studies.
–The benefit attributed to adding PCSK9 monoclonal antibodies to statin therapy further reinforces the concept of “treating to an LDL target”, which (as those of you who have seen these blogs over time) I really support, despite the 2013 Am Heart Assn guidelines. (I might also add that the 2015 Am Diabetes Assn also continues with the “treat to target” approach, in contradistinction to the Am Heart Assn!!)
–I also want to reinforce that one needs to be careful in accepting surrogate markers as indicators of clinical efficacy (and the reason I am publishing this blog is that there were clinical endpoints in these studies, even though not the primary goals). Both LDL and HDL represent many different lipoproteins, which happen to migrate during electrophoresis to the arbitrarily-designated high vs low density regions. So, this brings up a few issues:
–there are different LDLs, some small and dense, others fluffier. the small dense ones, in a few studies, are much more atherogenic (approx 3-fold), perhaps related to the fact that they are more oxidizable, and it is the oxidized LDL which is taken up in tissue macrophages to make foam cells (though there is some controversy about this of late, many studies suggest that LDL may really be a surrogate marker for oxidized LDL, raising the possibility that simply making changes in regular LDL may not necessarily translate into changes in clinical events). In fact, every study I’ve seen which has looked at it has found that apolipoprotein B levels correlate much better with cardiac events than LDL, and apoB reflects the small, dense LDL more accurately than the total LDL.) the data on ezetimibe, which does lower LDL reasonably well, are still unclear (it is quite striking and notable that the “blockbuster trial” in the Am Heart Assn meeting of November 2014, finding clinical efficacy of ezetimibe, is still unpublished and undissectable 6 months later…)
–and, there are different HDLs. overall HDL is an important negative risk factor, incorporated into the various risk calculators, with a slew of studies showing cardioprotection (and perhaps even more so in women). but about 10-15% of people have a pro-inflammatory/pro-atherogenic variant (with apolipoprotein C-III). This is important because studies with meds which increase HDL (eg, the cholesteryl ester transfer protein inhibitors, such as torcetrapib) do fantastically well in raising HDL levels but actually may increase cardiovasc events. And, apolipoprotein A levels do correlate better with cardioprotection than total HDL levels.
–So, bottom line: we need good alternatives to statins or drugs that can be used with statins, these studies are encouraging for PCSK9 as a target, but we still need to be careful to make sure that these drugs are clinically effective, are well-tolerated, and affordable.