Coronary heart disease (CHD) is the leading cause of death among men and women, although the risk in women temporally lags behind that of men (1). Nevertheless, the median survival for a woman age 55 to 64 with a first myocardial infarction (MI) is only 13.3 years compared to 17 years for a man [1]. Furthermore, for women aged 65 to 74 this median survival decreases to only 8.8 years [1]. Also, 64% of women who die suddenly of CHD have no previous symptoms compared to 50% for men. While most CHD risk factors are similar between men and women, they can often vary in strength and association and change throughout the lifetime [2, 3]. This is especially evident for lipids, with low high-density lipoprotein cholesterol (HDL-C) cholesterol and hypertriglyceridemia being stronger predictors of CHD in women than in men [4–6]. Also, with age and the onset of menopause, women’s lipoprotein profile becomes more atherogenic [3], which is associated with a rapid increase in cardiovascular risk. However, how to best quantify cardiovascular risk in women incorporating lipid parameters is unclear.

While low-density lipoprotein cholesterol (LDL-C), an atherogenic lipoprotein, has been recognized as the target of lipid lowering therapy, it is not always indicative of atherogenicity. Other lipoprotein parameters (i.e., remnant lipoproteins) have been shown to be more atherogenic, particularly in women, including apolipoprotein (apo) B, intermediate-density lipoprotein cholesterol (IDL-C), and subfractions of very low-density lipoprotein cholesterol (VLDL-C) [7–11]. In addition, the triglyceride to HDL-C ratio (TG/HDL-C) has been shown to be an atherogenic marker [12] and a powerful independent predictor of all cause mortality and cardiovascular events among postmenopausal women [13] and men [14, 15].

Gofman et al. first reported over 60 years ago that remnant lipoproteins were associated with the development of CHD [16]. Remnant lipoproteins have been shown to be elevated in patients with the metabolic syndrome [17], impaired glucose tolerance, and type 2 diabetes mellitus [18]. Remnant lipoproteins have also been reported to predict the extent of angiographically defined coronary artery disease (CAD) in post-menopausal women [19] and predict coronary events in patients with CAD [20, 21].

It is well recognized that HDL-C has a strong inverse association with adverse CHD events among women. However, clinical events continue to occur even in patients who achieve “target” levels. The Framingham Heart Study reported that 43% of clinical events occurred in women with an HDL-C ≥50 mg/dL [6]. Also, apo A1, the major component of HDL, has been shown to predict short-term and long-term risk in patients with normal HDL-C [22]. This finding indicates that just measuring HDL-C may not be sufficient for ideal risk assessment. Therefore, alternative or complementary risk marker assessment, such as measuring HDL components (apo A1) and composition, for determining CHD risk may need to be utilized. The purpose of this study was to evaluate a novel lipid ratio: apo A1 remnant ratio (apo A1/[VLDL₃-C+IDL-C]), which takes into account the importance of both apo A1 and remnant lipoproteins, with their association to death and myocardial infarction (MI) among peri and post-menopausal women. Furthermore, the utility of this new ratio was compared to traditional lipid parameters, apo A1, apo B, and the ratios total cholesterol to HDL-C (TChol/HDL-C), TG/HDL-C, apo B/A1, and HDL-C/(VLDL₃-C+IDL-C).

This study presents for the first time in women greater than 50 years of age that the novel apo A1 remnant ratio (apo A1/[VLDL₃-C+IDL-C]) is a strong predictor of death/MI at 1 and 3 years. This is seen among both tertiles, as well as from categories derived from recursive partitioning. Furthermore, this ratio had greater predictive ability than traditional lipid parameters, apo A1, apo B, and ratios such as TChol/HDL-C, TG/HDL-C, apo B/A1, and HDL-C/(VLDL₃-C+IDL-C). The power of this ratio is related to both the numerator (apo A1) and the remnant denominator (VLDL₃-C+IDL-C).

There are potentially important clinical ramifications related to this new ratio. As women age, LDL-C and HDL-C levels decline, indicating a change in HDL composition [3]. It has been shown that the specific composition of HDL alters its capacity to participate in reverse cholesterol transport and modulate inflammation. Therefore, measurements of traditional lipid parameters may not be sufficient in risk prediction. This study reinforces the growing body of evidence that among post- menopausal women, LDL-C, which is the current target of lipid therapy by the National Cholesterol Education Panel Adult Treatment Panel (NCEP ATP) III report [23], is not the most predictive lipid parameter for CHD risk [5, 24]. In addition, remnant lipoproteins are known to be elevated in postmenopausal women [8] and emphasis on remnant lipoproteins has been highlighted in the NCEP ATP III guidelines, which recommends the calculation of non-HDL-C as its measure [24]. Although non-HDL-C provides a measure of remnant lipoproteins, it also is a measure of triglyceride-rich particles, which are less atherogenic. Another advantage of the apo A1 remnant ratio is that fasting is not required for sub-fraction measurements and therefore it could be precisely calculated. Therefore, the superiority of our ratio is not surprising since it utilizes the advantages of direct measurements of apo A1 and the remnant lipoproteins over traditional lipid parameters.

Apo A1, which is the major protein component of HDL, is largely responsible for reverse cholesterol transport through the macrophage ATP-binding cassette transporter ABCA1. Oxidative mechanisms can damage apo A1, which decreases its ability to effectively promote cholesterol efflux. Such modifications can convert HDL from an anti-inflammatory to a pro-inflammatory particle. Measuring HDL-C is indicative of the size of the HDL pool, but not its composition. Therefore, an advantage of utilizing apo A1 over HDL-C is that it reflects the functioning-state of the molecule. Though epidemiological studies have consistently shown HDL-C to be inversely correlated with atherosclerotic events [25, 26], particularly among women [4], many events continue to occur despite reaching target levels [6]. Thus, emphasizing the need to determine additional measurements that can identify risk among “low-risk” populations. Further, it appears that HDL-C has a U-shaped distribution with a threshold for the upper limit of normal not being consistently defined for identifying risk, whereas apo A1 has a constant inverse relationship with CHD events [26]. Also, dysfunctional HDL is not always reflected in the HDL-C pool [27]. Unlike apo A1, very high values of HDL-C and HDL particle size have been shown to be associated with increased major adverse cardiovascular event risk [28].

We found low levels of apo A1 to be associated with an 80% increase in 1 and 3 year death/MI risk, whereas levels of HDL-C were found to not be predictive. The apo A1 remnant ratio utilizes the predictive value of apo A1, which was found to be highly predictive of death/MI by tertiles and by categories. We evaluated whether HDL-C would elicit similar results as apo A1 when divided by the remnant lipoproteins since a previous investigation found the inverse of this ratio, (VLDL₃–C+IDL-C)/HDL-C, to be predictive of CHD [29]. We found the HDL-C/(VLDL₃-C+ IDL-C) ratio, though predictive, to not achieve the same degree of predictive significance as the apo A1 remnant ratio. These results reinforce the importance of measuring apo A1, potentially above that of HDL-C, for determining death/MI risk among women >50 years.

Remnant lipoproteins are generated by lipolysis of triglyceride-rich lipoproteins and have been shown to play a greater role in CHD development than their precursors [11]. The Framingham Heart Study found remnant lipoproteins to be significantly higher in women with CHD when compared to women without CHD [8]. Among postmenopausal women, Lamon-Fava et al. reported that atherosclerosis severity is linked to a dysregulation of triglyceride/HDL metabolism [19]. They hypothesized that increases in remnant lipoproteins may be associated with increased lipid deposition in the arterial wall, plaque formation, and the generation of HDL-C subpopulations that are less efficient in the reverse cholesterol removal from the arterial wall [19]. Remnant lipoproteins, independent of cardiovascular risk factors, have also been shown to be associated with peripheral vascular endothelial dysfunction as measured by Doppler flow analysis [30], with this being more pronounced in the postprandial state [31]. Furthermore, remnant lipoproteins directly promote the initial stages of atherosclerosis through regulating the expression of ICAM-1 and VCAM-1, which are required for attachment of monocytes to the endothelial wall [32]. In addition, they have been shown to promote a pro-coagulant state and induce tissue factor in cultured endothelial cells, with large and small VLDLs supporting factor Xa and factor Xa/Va-mediated factor VII activation [32, 33]. Finally, VLDL particles generate thrombin at rates near that of activated platelets [34]. These pro-inflammatory, pro-atherogenic, and pro-thrombotic effects of remnant lipoproteins, which are independent from traditional risk factors including LDL-C, further explain the increased event rate seen in patients with CHD and increased remnant lipoprotein levels [35]. Finally, it is possible that remnant lipoproteins may be the driving factor of the increased CHD risk seen in middle aged and older woman with elevated triglycerides [36–41]. Within our study, remnant elevations were found to be associated with 1 year death/MI (HR=1.90, p=0.05), with a trend toward significance for 3 year death/MI risk (HR=1.55, p=0.09). Triglycerides were not associated with death/MI risk.

Many lipid and sub-lipid ratios have been reported to predict CHD risk. In the Women’s Ischemia Syndrome Evaluation (WISE) study, the TG/HDL-C ratio was found to predict cardiovascular events and mortality among women without prior MI or coronary revascularization, but non-HDL-C, the TChol/HDL-C ratio, and total cholesterol were not [13]. However, the TG/HDL-C was no longer predictive of cardiovascular events when adjusted by coronary artery disease severity. Within our study the apo A1 remnant ratio maintained its significance after adjustment by CHD severity. The TG/HDL-C ratio has also been shown to be predictive of cardiovascular events among other populations [14, 15]. However, we found the TG/HDL-C ratio to be not predictive of 1 or 3 year death/MI among women >50 years.

The apo B/A1 ratio was shown in INTERHEART, a large, international case–control study, to be one of the most predictive parameters for identifying CHD risk [9]. It was also found to show a graded relationship with MI risk, with no evidence of a threshold. Of all of the risk factors evaluated, the apo B/A1 ratio contributed the greatest to the population attributable risk, even being superior to smoking. When stratified by sex and age, similar associations persisted across all sub-groups. AMORIS showed that apo B, apo A1, and the apo B/A1 ratio are important risk factors for fatal MI, even after adjustment by age, total cholesterol, and triglycerides, with the apo B/A1 ratio being the greatest predictor of risk [7]. These results were consistent among both men and women, young (<70 years) and old (≥70 years). The apo B/A1 ratio was predictive of 1 year and 3 year death/MI when the highest tertile was compared to the lowest tertile within our study, though its predictive power was not as strong as the apo A1 remnant ratio.

In a recent meta-analysis that evaluated 233,455 subjects with 22,950 events, apo B was found to be superior to non-HDL-C, which in turn was superior to LDL-C in predicting cardiovascular risk (RR=1.43, 1.34, 1.25, respectively) [42]. Furthermore, this study found that using an apo B strategy to treat those individuals at >70th percentile in the US adult population, 500,000 more events would have been prevented than if a non-HDL-C stategy had not been used. It has also been reported that apo B is a more accurate marker of cardiovascular risk than LDL-C, and in its absence, non-HDL-C should be used since it indirectly estimates apo B [43]. Our study showed that the apo A1 remnant ratio was more predictive than apo B or non-HDL-C. One explanation for this is that remnant lipoproteins are more atherogenic than other larger apo B containing triglyceride rich lipoproteins. Remnants are smaller particles and can diffuse into the arterial wall [44] and are more likely to be bound by heparin sulfate proteoglycans within the arterial media [45]; therefore, able to promote more events.

Results of this study seem to suggest that lipids, or some of their components, might not be the best indicator of increased risk and those with a greater concentration of denser, more atherogenic particles are at higher risk than in those with abnormal lipid concentrations. Therefore, the apo A1 remnant ratio may be a parameter that may better stratify risk of women >50 years so that those found in the higher risk categories can be more aggressively treated. Interventions may include earlier introduction or more aggressive lipid-lowering therapy. In this study, there was a significant difference between the use of lipid lowering medications among the categories at baseline, though lipid lowering use was adjusted for in the multivariable model, use of these medications could have affected which stratification, and thus risk of the patient. Future studies should evaluate longitudinal use of lipid lowering therapy on apo A1 remnant ratio levels. The apo A1 remnant ratio is indicative of the number of anti-atherogenic particles over the number of atherogenic particles. Thus, particle numbers and their ratios might be more important factors than the amount of lipids carried per particle.

The new parameter, the apo A1 remnant ratio, was found to be a significant predictor of short-term and intermediate-term death/MI risk among women >50 years of age and was more strongly associated with death/MI risk than traditional lipid parameters, apo A1, apo B, and well-recognized lipid ratios. Further, this new ratio brings attention to the role of remnant lipoproteins in women greater than 50 years of age and represents a potential new measurement of risk. The apo A1 remnant ratio should be evaluated as a potential new treatment target and should be tested in a larger cohort and among other populations.