Low‐density lipoprotein‐cholesterol lowering in individuals at intermediate cardiovascular risk: Percent reduction or target level?

Abstract

Background

Recommendations for blood cholesterol management differ across different guidelines.

Hypothesis

Lipid‐lowering strategies based on low‐density lipoprotein‐cholesterol (LDL‐c) percent reduction or target concentration may have different effects on the expected cardiovascular benefit in intermediate‐risk individuals.

Methods

We selected individuals between 40 and 75 years of age with 10‐year risk for atherosclerotic cardiovascular disease (ASCVD) between 5.0% and <7.5% who underwent a routine health screening. For every subject, we simulated a strategy based on a 40% LDL‐c reduction (S_40%) and another strategy based on achieving LDL‐c target ≤100 mg/dL (S_target‐100). The cardiovascular benefit was estimated assuming a 22% relative risk reduction in major cardiovascular events for each 39 mg/dL of LDL‐c lowered.

Results

The study comprised 1756 individuals (94% men, 52 ± 5 years old). LDL‐c and predicted 10‐year ASCVD risk would be slightly lower in S_40% compared to S_target‐100. The number needed to treat to prevent 1 major cardiovascular event in 10 years (NNT₁₀) would be 56 with S_40% and 66 with S_target‐100. S_40% would prevent more events in individuals with lower baseline LDL‐c, whereas S_target‐100 would be more protective in those with higher LDL‐c. A dual‐target strategy (40% minimum LDL‐c reduction and achievement of LDL‐c ≤100 mg/dL) would be associated with outcomes similar to those expected with the S_40% (NNT₁₀ = 55).

Conclusions

In an intermediate‐risk population, cardiovascular benefit from LDL‐c lowering may be optimized by tailoring the treatment according to the baseline LDL‐c or by setting a dual‐target strategy (fixed dose statin plus achievement of target LDL‐c concentration).

1. INTRODUCTION

It is well established that lowering levels of low‐density lipoprotein‐cholesterol (LDL‐c) is associated with a reduction in the risk of incident cardiovascular events, particularly when the cardiovascular risk is high.1, 2, 3 In subjects with lower cardiovascular risk, despite evidence of benefit,2 the absolute risk reduction with statin therapy is expected to be lower, and clinicians are often faced with the decision of whether or not to prescribe a lipid‐lowering medication.

Guidelines differ on their recommendations on LDL‐c reduction, advocating either achievement of target LDL‐c concentration, percent reduction, and/or fixed‐dose statin.4, 5, 6, 7, 8 It is not known whether a certain strategy may prevent more cardiovascular events in the population than another, and the feasibility to conduct randomized trials to address this issue is very limited.

Based on this debate, the goal of this study was to simulate different lipid‐lowering strategies, based either on LDL‐c target concentration or percent reduction, and compare their effects on the predicted cardiovascular benefit in subjects considered to be at intermediate cardiovascular risk.

2. METHODS

2.1. Population and estimation of risk for atherosclerotic cardiovascular disease

We included participants from a large database of individuals undergoing a routine health screening at the Preventive Medicine Center of the Hospital Israelita Albert Einstein, São Paulo, Brazil, from January 2006 to December 2015. Data were prospectively collected from consecutive individuals, predominantly healthy, who underwent an interview with a clinician, physical examination, and blood collection, among several procedures, as previously described.9 Patients were questioned about previous cardiovascular events and medication in use. Fasting blood was analyzed for lipids and glycemia, among other exams. LDL‐c was estimated using the Friedewald equation,10 except for the cases where triglycerides level was higher than 400 mg/dL, when LDL‐c was directly measured.

The 10‐year risk for atherosclerotic cardiovascular disease (ASCVD) was calculated by the pooled cohort equations according to the 2013 American College of Cardiology/American Heart Association (ACC/AHA) guideline recommendation.11 We included individuals between 40 and 75 years of age at intermediate ASCVD risk, defined as a calculated 10‐year risk between 5.0% and <7.5%.

The following exclusion criteria were considered: (1) Individuals in secondary prevention, with a condition considered to be coronary risk–equivalent or with a clear indication for a statin: history of clinical ASCVD, subclinical ASCVD judged to be significant by the assistant physician, aortic aneurysm, diabetes mellitus or LDL‐c ≥ 190 mg/dL; (2) LDL‐c < 70 mg/dL; (3) current use of a lipid‐lowering medication; (4) presence of variable(s) out of the recommended range to use the pooled cohort equations: total cholesterol <130 mg/dL or >320 mg/dL, high‐density lipoprotein‐cholesterol (HDL‐c) <20 mg/dL or >100 mg/dL, systolic blood pressure < 90 mmHg or >200 mmHg; and (5) missing data not allowing calculations needed for risk estimation or predicted cardiovascular benefit.

The study was approved by the research ethics committee of the Hospital Israelita Albert Einstein, São Paulo, Brazil, and the consent form was waived as the research was based on database analysis and no additional patient contact was needed.

2.2. Simulated strategies and assumptions

For every subject, we simulated a strategy based on a 40% LDL‐c reduction (S_40%) and another strategy based on an LDL‐c target ≤100 mg/dL (S_target‐100). Figure 1 details the simulation process.

Figure 1.

Criteria considered in the simulations of the strategy based on a fixed 40% LDL‐c reduction (S_40%) and the strategy based on achieving an LDL‐c target ≤100 mg/dL (S_target‐100), according to baseline LDL‐c. Abbreviations: LDL‐c, low‐density lipoprotein cholesterol; S_40%, strategy based on a fixed 40% LDL‐c reduction; S_target‐100, strategy based on achieving an LDL‐c target ≤100 mg/dL

The 40% LDL‐c reduction was chosen based on the 2013 ACC/AHA guideline recommendation, which states that it is reasonable to offer treatment with a moderate‐intensity statin (predicting an LDL‐c lowering by approximately 30% to <50%) to individuals with 10‐year ASCVD risk between 5.0% and <7.5%.4 The 40% LDL‐c reduction is approximately what is expected by daily doses of 20 mg atorvastatin or 10 mg rosuvastatin.7

In the S_target‐100, each individual would have sufficient LDL‐c reduction to achieve the LDL‐c target ≤100 mg/dL. Although arbitrarily chosen, this target level corresponds to the treatment goal for low‐, intermediate‐, and high‐risk individuals established by the National Lipid Association,5 and is close to the 115‐mg/dL target recommended for moderate‐risk subjects by the European Society of Cardiology and the European Atherosclerosis Society.6 In this simulation, the minimum treatment intensity was considered to be a 30% LDL‐c lowering (approximately corresponding to a daily 20 mg‐simvastatin dose), and the percent reduction would have progressive increments of 10% if necessary to attain the target (Figure 1). Therefore, the treatment intensity was tailored to each individual according to his/her baseline LDL‐c, resembling what would be done in clinical practice.

2.3. Estimation of cardiovascular benefit

The cardiovascular benefit from LDL‐c reduction was estimated assuming a 22% relative risk reduction (rate ratio equal to 0.78) in major cardiovascular events for each 39 mg/dL (1 mmol/L) of LDL‐c lowered, based on the statin clinical trials compiled in the Cholesterol Treatment Trialists' Collaboration meta‐analysis.1 Accordingly, if LDL‐c lowers 78 mg/dL (2 mmol/L), the relative risk reduces 39% (0.78 × 0.78 = 0.61). Therefore, the final cardiovascular risk was given by the following formula: 10‐year final ASCVD risk = 10‐year baseline ASCVD risk × 0.78ⁿ , where n is the amount of LDL‐c reduction expressed in mmol/L. The number needed to treat to prevent 1 major cardiovascular event in 10 years (NNT₁₀) was calculated as the reciprocal of the absolute difference between the baseline and the final 10‐year ASCVD risks.

2.4. Subgroup and sensitivity analyses

Once the superiority of a strategy based on LDL‐c target over percent reduction, or vice versa, is expected to be dependent on the baseline LDL‐c, we performed calculations for subgroups defined by baseline plasma LDL‐c concentration. We also performed sensitivity analyses considering more or less intensive strategies to manage LDL‐c. In the more conservative scenario, we simulated a 30% LDL‐c reduction (S_30%) and considered LDL‐c ≤130 mg/dL as the target (S_target‐130). In the more aggressive model, we simulated a 50% LDL‐c reduction (S_50%, approximately corresponding to a daily dose of 40 mg atorvastatin or 20 mg rosuvastatin) and considered LDL‐c ≤70 mg/dL as the target (S_target‐70, a common recommended level for very high‐risk patients).5, 6 Supporting Figures S1 and S2 in the online version of this article depict the simulation process adopted in these analyses.

Finally, we analyzed the effect of dual‐target strategies, which are the scenarios where the individual would receive a medication, regardless of baseline LDL‐c, and would also have to achieve a target LDL‐c concentration. The following strategies were simulated: at least 30% LDL‐c reduction and LDL‐c target ≤130 mg/dL (S_{30% + target‐130}), at least 40% LDL‐c reduction and LDL‐c target ≤100 mg/dL (S_{40% + target‐100}), and 50% minimum LDL‐c reduction and LDL‐c ≤70 mg/dL (S_{50% + target‐70}). In these simulations, treatment would be intensified in individuals with higher LDL‐c to attain the target value (see Supporting Figures S3, S4, and S5 in the online version of this article).

2.5. Statistical analyses

Continuous variables were assumed to have a normal distribution due to the large sample size, were expressed as mean and standard deviation, and compared by t tests. Categorical variables were expressed as proportions, and the χ² test was used in the comparisons. The different strategies for cholesterol reduction were compared by multilevel mixed effects models with Bonferroni's adjustment for multiple comparisons. Analyses were performed with the use of Stata software version 14.0 (StataCorp, College Station, TX). A P value <0.05 was considered significant.

3. RESULTS

3.1. Study population and baseline characteristics

Supporting Figure S6 in the online version of this article details the individuals included and excluded in the study. From an initial population of 37 659 individuals, we first excluded 44% of the subjects due to age criteria. Among the remaining individuals, the most common exclusion criteria were a calculated 10‐year ASCVD risk <5.0% or ≥7.5% and current use of lipid‐lowering drugs. The final study population consisted of 1756 individuals (5% of the initial population).

Baseline characteristics of the initial and the study populations are shown in Table 1. Compared to the initial population, the study population was formed by individuals with more cardiovascular risk factors and higher ASCVD risk. On average, the study sample was composed by middle‐aged persons (almost exclusively white men) with overweight and high triglycerides levels; mean LDL‐c was 134 ± 27 mg/dL and mean 10‐year ASCVD risk was 6.1 ± 0.7%.

Table 1.

Baseline characteristics of the initial and the study populations

Characteristic	Initial Population, n = 37 659	Study Population, n = 1756
Age, y	42 ± 10	52 ± 5
Male sex	26 897 (71)	1646 (94)
BMI, kg/m2	26.5 ± 4.3	28.1 ± 4.2
Total cholesterol, mg/dL	197 ± 38	210 ± 30
LDL‐c, mg/dL	122 ± 34	134 ± 27
HDL‐c, mg/dL	49 ± 14	44 ± 10
Triglycerides, mg/dL	134 ± 98	164 ± 97
Fasting glycemia, mg/dL	89 ± 17	93 ± 14
Arterial hypertension	4887 (13)	442 (25)
Smoking	3894 (10)	360 (21)
10‐year ASCVD risk, %	3.0 ± 5.1	6.1 ± 0.7

Abbreviations: ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; HDL‐c, high‐density lipoprotein cholesterol; LDL‐c, low‐density lipoprotein cholesterol.

Values are expressed as mean ± standard deviation or n (%). P value is <0.001 for all variables.

3.2. Predicted reductions in LDL‐c and absolute cardiovascular risk

When the S_target‐100 was simulated to the study population, 830 (47%) subjects would have a 30% LDL‐c reduction, whereas 488 (28%) and 237 (13%) individuals would need a 40% and a 50% LDL‐c reduction, respectively, to reach the goal. Medication would not be prescribed for 201 (11%) participants.

Estimated post‐treatment LDL‐c would be lower in S_40% compared to S_target‐100 (80 ± 16 mg/dL vs 88 ± 7 mg/dL, respectively, P < 0.001) (Figure 2). Notably, LDL‐c distribution in the population would be quite diverse between S_40% and S_target‐100, with a wider dispersion in the former (Figure 2). After application of S_40%, 224 (13%) subjects would remain with LDL‐c > 100 mg/dL.

Figure 2.

Comparison of LDL‐c (A) and the 10‐year risk for atherosclerotic cardiovascular disease (B) at baseline and after application of the simulated strategies. *P < 0.001 vs baseline. **P < 0.001 vs S_40%. Abbreviations: ASCVD, atherosclerotic cardiovascular disease; S_40%, strategy based on a fixed 40% LDL‐c reduction; S_target‐100, strategy based on achieving an LDL‐c target ≤100 mg/dL

Predicted 10‐year ASCVD risk would also be lower in S_40% compared to S_target‐100, although the magnitude of the difference would be small (4.3 ± 0.6% vs 4.6 ± 0.9%, respectively) (P < 0.001) (Figure 2). The NNT₁₀ to prevent 1 major cardiovascular event would be 56 with S_40% and 66 with S_target‐100.

A dual‐target strategy (S_{40% + target‐100}) would be associated with outcomes similar to those expected with the S_40% (ASCVD risk reduction to 4.3 ± 0.6% and NNT₁₀ equal to 55).

3.3. Influence of baseline LDL‐c on the predicted cardiovascular benefit

In the subgroup of individuals with baseline LDL‐c between 70 and <100 mg/dL (mean LDL‐c 88 ± 8 mg/dL, ASCVD risk 6.1 ± 0.7%), S_40% would reduce LDL‐c to 53 ± 5 mg/dL and the ASCVD risk to 4.8 ± 0.6%, whereas no medication would be prescribed in S_target‐100 (Figure 3).

Figure 3.

Comparison of LDL‐c (A) and the 10‐year risk for ASCVD (B) at baseline and after application of the simulated strategies, according to baseline LDL‐c. *P < 0.001 vs baseline. **P < 0.001 vs S_40%. Abbreviations: ASCVD, atherosclerotic cardiovascular disease; S_40%, strategy based on a fixed 40% LDL‐c reduction; S_target‐100, strategy based on achieving an LDL‐c target ≤100 mg/dL

ASCVD risk would also be modestly lower with S_40%, compared to S_target‐100, in subjects with baseline LDL‐c at intermediate levels (100–<160 mg/dL) (Figure 3).

In subjects with baseline LDL‐c between 160 and <190 mg/dL (mean LDL‐c 172 ± 8 mg/dL, ASCVD risk 6.1 ± 0.7%), S_target‐100 would promote a higher reduction in LDL‐c and a slightly greater decrease in the ASCVD risk, compared to S_40% (LDL‐c: 91 ± 5 mg/dL vs 103 ± 5 mg/dL, respectively; ASCVD risk: 3.7 ± 0.5% vs 4.0 ± 0.5%, respectively) (Figure 3).

3.4. Influence of treatment intensity on the predicted cardiovascular benefit

Simulating a less intensive treatment, estimated LDL‐c would be lower in S_30% compared to S_target‐130 (94 ± 19 mg/dL vs 108 ± 12 mg/dL, respectively) (P < 0.001). As a consequence, predicted ASCVD risk would also be lower in S_30% compared to S_target‐130 (4.7 ± 0.6% vs 5.2 ± 1.0%, respectively) (P < 0.001). The NNT₁₀ to prevent 1 major cardiovascular event would be 72 with S_30% and 114 with S_target‐130. A dual‐target strategy (S_{30% + target‐130}) would have the same effect on ASCVD risk reduction as in the S_30%.

In the more aggressive LDL‐c lowering simulations, LDL‐c would reach 67 ± 13 mg/dL in the S_50% and 63 ± 4 mg/dL in the S_target‐70 (P < 0.001); predicted ASCVD risk would be 4.0 ± 0.6% and 3.9 ± 0.8% in the S_50% and S_target‐70, respectively (P < 0.001). The NNT₁₀ to prevent 1 major cardiovascular event would be 47 and 46 with S_50% and S_target‐70, respectively. A dual‐target strategy (S_{50% + target‐70}) would be associated with an ASCVD risk reduction to 3.8 ± 0.7% and NNT₁₀ equal to 44.

The expected cardiovascular benefit is predicted to vary substantially according to the baseline LDL‐c and the treatment intensity (Table 2). NNT₁₀ is expected to be over 100 for a 30% LDL‐c reduction in individuals with LDL‐c < 100 mg/dL, and under 40 when more aggressive strategies (S_50% and S_target‐70) are applied to those with LDL‐c > 160 mg/dL (Table 2).

Table 2.

Predicted number needed to treat to prevent 1 major cardiovascular event in 10 years, according to the lipid‐lowering strategy and baseline LDL‐c

	Strategy
	Percent Reduction			Target Based
	S30%	S40%	S50%	Starget‐130	Starget‐100	Starget‐70
LDL‐c, mg/dL
70–<100 (n = 186)	105	81	67	—	—	105
100–<130 (n = 579)	82	64	53	—	84	58
130–<160 (n = 651)	67	53	44	70	58	40
160–<190 (n = 340)	58	46	38	57	40	32
Total (n = 1756)	72	56	47	114	66	46

Abbreviations: LDL‐C, low‐density lipoprotein‐cholesterol; S_30%, strategy based on a fixed 30% LDL‐c reduction; S_40%, strategy based on a fixed 40% LDL‐c reduction; S_50%, strategy based on a fixed 50% LDL‐c reduction; S_target‐70, strategy based on achieving an LDL‐c target ≤70 mg/dL; S_target‐100, strategy based on achieving an LDL‐c target ≤100 mg/dL; S_target‐130, strategy based on achieving an LDL‐c target ≤130 mg/dL.

Nevertheless, the NNT₁₀ to prevent 1 major cardiovascular event when one of the simulated strategies is applied instead of another would usually be high (see Supporting Table S1 in the online version of this article). Even when the more aggressive approaches S_50% and S_target‐70 replace the more conservative S_target‐130, the predicted NNT₁₀ would surpass 75. High NNTs would also be observed when simulating strategy replacements in subgroups defined by baseline LDL‐c (see Supporting Table S1 in the online version of this article).

4. DISCUSSION

The main findings of this study can be summarized as follows: (1) LDL‐c and predicted 10‐year ASCVD risk would be slightly lower if all individuals with baseline 10‐year ASCVD risk between 5.0% and <7.5% have their LDL‐c reduced by 40%, compared to a strategy based on achieving LDL‐c ≤100 mg/dL; (2) baseline LDL‐c determines which modality (percent reduction or target‐based approach) prevents more events; (3) a dual‐target strategy (40% minimum LDL‐c reduction and LDL‐c target ≤100 mg/dL) may offset the influence of baseline LDL‐c on the more advantageous strategy; and (4) compared to 40% LDL‐c reduction and LDL‐c target ≤100 mg/dL, more or less aggressive approaches are expected to proportionate mild changes in the absolute ASCVD risk reduction.

Potential differences between lipid‐lowering strategies based on percent reduction or LDL‐c numerical targets are relevant for the practicing physician, but very few studies have assessed this issue.12, 13, 14 Theoretical calculations on the outcomes from different lipid‐lowering strategies were reported by Soran et al and are consistent with our findings.13 Bangalore et al and our group have explored this controversy in higher‐risk patients.12, 14 To the best of our knowledge, no comparison of different lipid‐lowering modalities has been reported in real world individuals at intermediate cardiovascular risk.

In the present study, a 40% LDL‐c reduction was slightly more advantageous than pursuing the 100‐mg/dL LDL‐c target. This occurred as a consequence of the fact that, in the target‐based strategy, most patients would not need a 40% LDL‐c reduction to reach the 100‐mg/dL target, including a minority that would not even be prescribed statins.

Of note, our 40% LDL‐c reduction simulation resembles the rosuvastatin arm in the HOPE‐3 (Heart Outcomes Prevention Evaluation‐3) randomized trial, although HOPE‐3 subjects were older and had lower mean LDL‐c compared to our sample.15 In that study, over a 5.6‐year follow‐up, the coprimary outcome occurred in 3.7% subjects receiving rosuvastatin 10 mg per day and in 4.8% individuals randomized to placebo. Extrapolating these rates to 10 years, the absolute difference in event rate would be 1.96%, meaning a 10‐year NNT of 51, which is comparable to the NNT of 56 in our 40% LDL‐c reduction simulation.

Our analyses also highlight that the 100‐mg/dL target‐based strategy is expected to prevent more cardiovascular events, compared to the 40% reduction strategy, in subjects with higher baseline LDL‐c. It should not be underappreciated that more than 10% of our population would have an LDL‐c > 100 mg/dL after a 40% LDL‐c reduction. It may be argued that these individuals, under a 40% LDL‐c reduction, will have a residual risk from not reaching adequate, protective LDL‐c. In these cases, a more intense statin therapy may provide incremental long‐term benefit, which has to be balanced against a higher risk of developing diabetes mellitus.4 It is widely accepted, however, that the cardiovascular benefits from statins usually outperform the excess risk of new‐onset diabetes.3

Recent guidelines have acknowledged that treatment goals and intensity of treatment may be adjusted according to the baseline or achieved LDL‐c in higher‐risk individuals.6, 16 Our results support to expand this refinement to the intermediate‐risk stratum. Alternatively, guidelines may recommend a dual‐target strategy, where all individuals would receive a medication, but also would pursue a target LDL‐c level. It should be acknowledged, however, that the clinical impact of this approach (tailoring the lipid‐lowering strategy and intensity according to the baseline LDL‐c) is predicted to be of less importance as compared to higher‐risk individuals.

Our results have the strength of evaluating a large number of individuals within a narrow, specific range of intermediate ASCVD risk. Moreover, our results are based on real characteristics of patients commonly seen by physicians. On the other hand, the present investigation has important limitations, as it is based on simulations that do not exactly mirror what happens in real‐world patients. We considered neither the lack of adherence to lipid‐lowering therapy nor the known variability in LDL‐c response to statins.17, 18 The external validity of the study is also limited, because our sample has disproportionately more males than females.

5. CONCLUSION

A strategy based on a 40% LDL‐c reduction is predicted to prevent a slightly higher number of events, compared to a strategy based on a 100‐mg/dL LDL‐c target, in individuals at intermediate cardiovascular risk. Baseline LDL‐c should be used to guide the more appropriate strategy, with percent reduction being more protective in those with lower LDL‐c, and target achievement being more advantageous in those with higher LDL‐c. Alternatively, a dual‐target strategy (fixed‐dose statin plus achievement of target LDL‐c concentration) may be proposed. Changing the intensity of LDL‐c lowering therapy is predicted to promote a small impact on outcomes and is less important than the decision about whether or not to prescribe a medication. The results may be considered by future guidelines and by clinicians to refine therapeutic decisions in management of blood cholesterol.

Supporting information

Table S1. Predicted number needed to treat to prevent 1 major cardiovascular event in 10 years when the strategy in the first column replaces the strategy in the first line, according to baseline LDL‐c.

Figure S1. Criteria considered in the simulations of the strategy based on a 30% LDL‐c reduction (S_30%) and the strategy based on achieving an LDL‐c target ≤130 mg/dL (S_target‐130), according to baseline LDL‐c.

Figure S2. Criteria considered in the simulations of the strategy based on a 50% LDL‐c reduction (S_50%) and the strategy based on achieving an LDL‐c target ≤70 mg/dL (S_target‐70), according to baseline LDL‐c.

Figure S3. Criteria considered in the simulation of the strategy based on a 30% minimum LDL‐c reduction and achievement of an LDL‐c target ≤130 mg/dL (S_{30% + target‐130}), according to baseline LDL‐c.

Figure S4. Criteria considered in the simulation of the strategy based on a 40% minimum LDL‐c reduction and achievement of an LDL‐c target ≤100 mg/dL (S_{40% + target‐100}), according to baseline LDL‐c.

Figure S5. Criteria considered in the simulation of the strategy based on a 50% minimum LDL‐c reduction and achievement of an LDL‐c target ≤70 mg/dL (S_{50% + target‐70}), according to baseline LDL‐c.

Figure S6. Schematic flowchart depicting included and excluded subjects. ASCVD, atherosclerotic cardiovascular disease; TC, total cholesterol.

Cesena FHY, Laurinavicius AG, Valente VA, Conceição RD, Santos RD, Bittencourt MS. Low‐density lipoprotein‐cholesterol lowering in individuals at intermediate cardiovascular risk: Percent reduction or target level? Clin Cardiol. 2018;41:333–338. 10.1002/clc.22868