Treatment Thresholds and ASCVD Risk

Vera Bittner, MD, MSPH


J Am Coll Cardiol. 2021;78(15):1508-1510. 

Abstract and Introduction


Low-density-lipoprotein cholesterol (LDL-C) is causal in the pathogenesis of atherosclerotic cardiovascular disease (ASCVD).[1] Meta-analyses of randomized lipid-lowering trials suggest a log-linear relationship between lowering of LDL-C and reduction in major adverse cardiovascular events across the spectrum of ASCVD risk, but with higher absolute risk reductions for those at greater ASCVD risk.[2] In addition to universal lifestyle counseling, the 2018 American Heart Association/American College of Cardiology Multi-Society Guidelines thus advocate a risk-based treatment algorithm, recommending high-intensity/maximally tolerated statin therapy for those with very high ASCVD risk and endorse further intensification of therapy through the addition of ezetimibe or a proprotein convertase subtilisin–kexin type 9 (PCSK9) inhibitor among those whose LDL-C remains ≥70 mg/dL while using high-intensity or maximally tolerated statin therapy.[3]

In this issue of the Journal, Oyama et al[4] analyze data from the IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial) to determine whether there was benefit to simvastatin/ezetimibe combination therapy compared with simvastatin 40 mg monotherapy among the 2,480 postacute coronary syndrome (ACS) patients who were enrolled with LDL-C levels of 50-<70 mg/dL, ie, levels below the current numeric threshold, for consideration of combination lipid-lowering therapy. The degree of LDL-C lowering and the cardiovascular and safety outcomes were compared between this group with low baseline LDL-C, those with baseline LDL-C of 70-<100 mg/dL, and participants who had LDL-C of 100–125 mg/dL. The addition of ezetimibe to simvastatin resulted in a median 17–20 mg/dL additional LDL-C lowering without differences between subgroups. Participants in all LDL-C subgroups randomized to simvastatin/ezetimibe benefited from the additional LDL-C reduction compared with those using simvastatin monotherapy, with point estimates of the hazard ratios ranging from 0.92–0.94 for the primary composite endpoint of cardiovascular death, major coronary events, or stroke. There were no statistically significant interactions in treatment benefit by baseline LDL-C stratum, and there were no safety signals. The authors concluded that these data support combination lipid-lowering therapy with simvastatin/ezetimibe in post-ACS patients even if their LDL-C while on statin monotherapy was below the current threshold of 70 mg/dL. The investigators acknowledged several limitations to their analysis, including its post-hoc nature, less than currently recommended intensity of the background statin regimen, an LDL-C distribution <70 mg/dL that was truncated at a level of 50 mg/dL, and limited ability to extrapolate their results to patients at risk for ASCVD events but without recent ACS.

The review of baseline characteristics for the 3 LDL-C subgroups shown in Table 1 of Oyama et al[4] illustrates why post-hoc subgroup analyses of randomized clinical trials should always be interpreted with some caution. Participants in the lowest LDL-C group who had an ACS despite their low LDL-C levels were older; were more likely to have diabetes, hypertension, prior coronary events and revascularization, a history of stroke, a history of heart failure; and were more likely to have been using long-term lipid-lowering therapy before their entry into the trial. They were at the highest ASCVD risk, with a Thrombolysis In Myocardial Infarction risk score for secondary prevention of ≥3 in 38% of participants, compared with 29% and 17% of participants in the middle and high LDL-C subgroups, respectively. Participants in the lowest LDL-C subgroup were also less likely to present with ST-segment elevation myocardial infarction and were least likely to be revascularized for the index event. Indeed, participants in this group randomized to moderate-intensity simvastatin monotherapy had a 7-year Kaplan-Meier event rate for the combined primary endpoint of 42.2% compared with 37.7% in the intermediate LDL-C subgroup and 29.3% in the high LDL-C subgroup. Similar gradients in risk were apparent for the "hard components" of the primary endpoint, ie, cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke, and for various secondary composite endpoints shown in the online supplement. The number of total events in the simvastatin monotherapy group, when normalized to sample size, was also highest in the low LDL-C subgroup (36 events per 100 participants) and lowest in the high LDL-C subgroup (22 events per 100 participants). The current analysis is thus not only a comparison of treatment benefits by baseline LDL-C level but also a comparison of treatment benefits by baseline ASCVD risk.

This very-high-risk, low LDL-C group randomized to simvastatin/ezetimibe combination therapy had numerically the largest absolute risk reduction for the composite primary endpoint of 3.9% (number needed to treat [NNT] = 26) compared with 1.9% (NNT = 53) and 1.6% (NNT = 63), respectively, for the middle and high LDL-C strata. A formal test for interaction of treatment by LDL-C stratum was nonsignificant, but the confidence intervals for the subgroups were rather wide, especially for the low LDL-C group, given its small sample size. Relative risk reductions were consistent across the three groups, whether assessed by Cox proportional hazards modeling (which took into account prior lipid-lowering therapy and type of index event, but did not adjust for other baseline characteristics that differed between subgroups), by competing risk analysis that took into account the impact of noncardiovascular death, by landmark analysis after month 4, and in an analysis of total events. An analysis of treatment benefit of the combined regimen with baseline LDL-C as a continuous variable also did not show any significant interaction of treatment by baseline LDL-C, but again demonstrated considerable imprecision in the estimated hazard ratios at LDL-C levels below approximately 65 mg/dL and above approximately 115 mg/dL, given the small number of individuals at risk at both extremes of the LDL-C distribution for this trial.

What can we learn from this new analysis of the IMPROVE-IT trial? First, we are reminded that patients early after ACS are heterogeneous, with substantial variation of risk for recurrent events. Formal risk stratification should thus be routine in clinical practice to appropriately target secondary prevention measures. Second, baseline LDL-C viewed in isolation is not necessarily a reliable indicator of risk for recurrent events or of potential benefits of LDL-C lowering therapies and has to be interpreted in the context of the underlying ASCVD risk. Third, the current analysis suggests that individuals at very high risk for recurrent events who have LDL-C between 50 mg/dL and <70 mg/dL while using moderate-intensity statin therapy derive a significant benefit from intensification of lipid-lowering therapy. This is consistent with current guidelines that recommend high-intensity rather than moderate-intensity statin therapy in such individuals without an LDL-C threshold. Given the design of IMPROVE-IT, the current analysis cannot answer the question whether the LDL-C threshold for adding ezetimibe or PCSK9 inhibitor should be lowered to ≥50 mg/dL among individuals at very high ASCVD risk and using high-intensity statin therapy, but it further emphasizes that we should look beyond moderate-intensity statin therapy among our very-high-risk patients who cannot tolerate high-intensity statins, even if their baseline LDL-C levels are low.