Do apolipoproteins improve coronary risk prediction in subjects with metabolic syndrome? Insights from the North Italian Brianza cohort study

Highlights

• High ApoAI levels are associated with first coronary event in non-diabetic MetS men.

• ApoAI and ApoB improve prediction in MetS subjects at intermediate risk.

• 46 tests are needed to correctly reclassify 1 case in the high-risk class.

• 1 every 6 MetS subject was reclassified in the low-risk class, none had an event.

• Rationale for apo use in MetS as second level marker to avoid unnecessary treatment.

Abstract

Objective

We assessed predictive abilities and clinical utility of CVD risk algorithms including ApoB and ApoAI among non-diabetic subjects with metabolic syndrome (MetS).

Methods

Three independent population-based cohorts (3677 35–74 years old) were enrolled in Northern Italy, adopting standardized MONICA procedures. Through Cox models, we assessed the associations between lipid measures and first coronary events, as well as the changes in discrimination and reclassification (NRI) when standard lipids or apolipoproteins were added to the CVD risk algorithm including non-lipids risk factors. Finally, the best models including lipids or apolipoproteins were compared.

Results

During the 14.5 years median follow-up time, 164 coronary events were validated. All measures showed statistically significant associations with the endpoint, while in the MetS subgroup HDL-C and ApoAI (men, HR = 1.59; 95%CI: 0.96–2.65) were not associated. Models including HDL-C plus TC and ApoB plus ApoAI for lipids and apolipoproteins, respectively, showed the best predictive values. When ApoB plus ApoAI replaced TC plus HDL-C, NRI values improved in subjects with MetS (13.8; CI95%: −5.1,53.1), significantly in those previously classified at intermediate risk (44.5; CI95% 13.8,129.6). In this subgroup, 5.5% of subjects was moved in the high (40.0% of expected events) and 17.0% in the low risk class (none had an event at 10 years).

Conclusions

ApoB and ApoAI could improve coronary risk prediction when used as second level biomarkers in non-diabetic subjects with MetS classified at intermediate risk. The absence of cases moved downward suggests the gain in avoiding treatments in non-cases and favor the use of apolipoproteins for risk assessment.

Introduction

The most recent US [1] and European [2] guidelines for the assessment of cardiovascular (CV) risk in asymptomatic adults do not recommend the use of advanced lipid measures for risk stratification. Apolipoprotein B₁₀₀ (ApoB) and apolipoprotein A-I (ApoAI) have been suggested as new biomarkers for risk assessment, since both apolipoproteins and the ApoB/ApoAI ratio showed a stronger association with CHD [3] or CVD [4] than cholesterol content measures. However, despite the strong association, several studies testing predictive abilities of CV risk score algorithms including the apolipoproteins showed a weak improvement in cardiovascular risk stratification [5], [6], [7], [8]. Recently, a pooled analysis [9] of 17 prospective studies with more than 5400 events confirmed also the lack of clinical utility, since the slight gain obtained in CVD prediction was limited by the high number of patients needed to screen to prevent one event.

However, the pooled analyses showed better results in subjects initially classified at intermediate risk, suggesting the use of apolipoproteins as secondary risk stratification markers [10]. This risk class is considered a ‘grey zone’ for treatment decisions, because of the high number of events occurring [11], which could be prevented by improving risk classification and thus treatment indications or therapeutic targets. Furthermore, the authors of the pooled analysis acknowledged that ApoAI and ApoB may be clinically useful in some subgroups at higher incidence, such as subjects with metabolic syndrome (MetS) [9]. In fact, subjects with MetS are characterized by a more complex lipoprotein profile, with modifications in lipoprotein composition and number that is more evident than that in their cholesterol content [12], [13]. A recent article showed that dysfunctional apoAI, oxidized by means of myeloperoxidase (MPO, particularly high in subjects with metabolic syndrome [14]), is more abundant in plaque than in plasma, and its plasma level showed an improvement in prediction of both CHD and CVD [15]. Therefore, this subgroup could potentially benefit in risk prediction from using ApoB and ApoAI, the main structural proteins of non-HDL and HDL particles respectively [12]. In this subgroup of subjects at cardiometabolic risk, US guidelines suggested ApoB assessment for monitoring the efficacy of LDL-lowering treatment [16].

To improve the clinical utility of apolipoproteins measurements, subgroups of subjects with MetS classified at intermediate cardiovascular risk by standard algorithms should be considered. The aim of the present study is to assess whether replacing standard lipid measures with ApoB and/or ApoAI or their ratio could be clinically useful in the prediction of 10-year risk of first coronary event in non-diabetic southern European subjects with MetS, in a primary or secondary risk assessment tool. Since there is no consensus on which lipid markers should be used in this population, in a preliminary step models including different combinations of lipoprotein cholesterol or apolipoproteins were compared with the same reference model including non-lipid risk factors only. The model with apolipoproteins that showed the highest improvement in terms of reclassification was then compared with the best model including markers of cholesterol content.