Dyslipidemia and Cardiovascular Risk: Lipid Ratios as Risk Factors for Cardiovascular Disease

There are extensive epidemiological data demonstrating that high blood cholesterol levels increase cardiovascular risk, and that this risk is dependent on the levels of the different blood cholesterol fractions. Moreover, the reduction of total blood cholesterol has been clearly related to a reduction in the risk of stroke, coronary disease and overall cardiovascular death. However, the traditional cholesterol measurements tend to be most accurate at predicting risk for those at the lower and higher ends of the risk spectrum. Recent data has shown LDL-Cholesterol/HDL-Cholesterol ratio and even TotalCholesterol/HDL-Cholesterol ratio, to be accurate predictors of cardiovascular risk. In fact, changes in ratios have been shown to be better indicators of successful CHD risk reduction than changes in absolute levels of lipids or lipoproteins. In the Helsinki Study, the LDLC/HDL-C ratio had more prognostic value than LDL-C or HDL-C alone (Manninen, Tenkanen, Koskinen et al, 1992). The ratio was especially accurate at predicting risk among those who also had elevated triglyceride levels. The PROSPER trial, a retrospective analysis of 6,000 patients, found that the ratio of LDL-C/HDL-C was the most powerful measure of cardiovascular disease risk in elderly people (Packard, Ford, Robertson et al, 2005). The PROCAM Study, which included almost 11,000 men aged 36 to 65 who were studied for 4 to 14 years, found a continuous and graded relationship between the LDL-C/HDL-C ratio and CVD mortality (Cullen, Schulte, Assmann et al, 1997). In addition, comparison of individual LDL-C/HDL-C ratios from subjects in the Framingham Study clearly indicates that these ratios are significantly more robust predictors of CVD than the individual levels of LDL-C or HDL-C (Kannel, 2005).


280
coexistence of important mortality indexes related to diseases whose prevalence is a demonstration to their stage of development (perinatal, nutritional and infectious health problems), with a more consistent presence of coronary disease, with increased percentage reaching 137% for males and 120% for females by the year 2010 (Yusuf, Reddy, Ounpuu et al, 2001). In epidemiological terms, coronary heart disease and cerebrovascular disease represent the most significant expressions of cardiovascular disease, and were the main causes of mortality and morbidity worldwide, accounting for one third of total mortality in the year 2001 (American Heart Association, 2003). According to the World Health Organization, every year 16 million deaths occur from cardiovascular disease, and this number is expected to rise to 20 million in the first decade of the XXI century (World Health Organization, 2002). The singular importance of coronary heart disease is extraordinarily important and it is estimated that mortality by this disease had risen to 7.2 million individuals by the year 2001 (World Health Organization, 2002). However, in recent years, there has been a trend towards a decline in this disease in Western countries, with a concomitant increase in other lands, notably in Russia and Eastern European countries. In fact, in Western countries, the number of deaths from coronary per 100,000 inhabitants was 151 in 1972, dropping to 44 in 2004, in men aged 64 or less. Similar reductions were also observed in females (36 to 11 women per 100,000). Paradoxically, in Russia, there was a marked increase of this rate, from 169/100.000, in the year 1980, to 242/100.000 deaths, in the year 2005 (Allender, Scarborough, Black et al, 2008). The onset of cardiovascular disease is consistently related to the presence of a group of cardiovascular risk factors, whose manipulation can be crucial to its prevention (see Table 1).

Reversible
Smoking Concerning reversible risk factors in which intervention could be decisive, we should highlight the relative importance of smoking, arterial hypertension and hypercholesterolaemia. Although the global fight against all reversible risk factors constitute a therapeutic imperative, the elimination of hypercholesterolemia would result in the single most important benefit against the incidence of coronary heart disease as well as other atherosclerotic vascular problems (Wilson, D'Augustine, Levy et al, 1998 Regarding arterial hypertension, it always presents itself as a major risk factor, given its very high incidence and prevalence. Despite all the research carried out and considering all the remarkable therapeutic advances, the control of blood pressure levels only provides a reduction of about 40% in mortality from cerebrovascular disease and a more modest 20% reduction in mortality from coronary heart disease (Kaplan, 2002). Diabetes is another important risk factor, with the particularity of becaming the major epidemy of this century given its substantial and consistent epidemiological growth. The high cardiovascular risk that diabetes provides is well illustrated by the prognostic similarity between a diabetic patient without clinical manifestations of coronary heart disease and a patient with a history of acute coronary events (Hafnner, Lethe, Ronnema et al, 1988). Obesity is undoubtedly a major risk factor for cardiovascular disease (Higgins, Kannel, Garrison et al, 1988), now becoming a public health problem, given the alarming increase of its prevalence in industrialized countries. The pathogenic mechanisms involved in this situation are complex and not just related to the metabolic overload involved, but also determined by its close associations with arterial hypertension, type 2 diabetes, dyslipidemia, and inflammation (Higgins, Kannel, Garrison et al, 1988).
Smoking is consensually assumed as a relevant contributer to cardiovascular disease, both in the active as well as the passive form. Some studies indicate that smokers have a reduced life expectancy of about ten years (although this number is dose-dependent) and that this habit cancells the natural cardioprotection in women (Silva, 2000). In fact, the differences in cardiovascular risk amongst men and women are well known, largely documented by the classic time lag between genders, with a higher risk in males until the fifth decade of life, with a progressive increase in women of cardiovascular risk until the eighth decade of life, when the risk is similar in men and women. The explanation for this is closely linked to the production and subsequent estrogen deficiency (as a consequence of menopause) seen in different phases of a woman's life. But other factors must not be overlooked. For instance, If we consider only the lipid profile changes induced by menopause, on average a 10% increase in LDL-cholesterol, an 8% reduction in HDLcholesterol and an elevation in triglycerides are expected. Nonetheless, these changes can be normalized by hormonal replacement therapy (Stampfer, Colditz, Willet et al, 1991). Oral contraception, by contrast, tends to cause an adverse impact on lipid profile. At present, the worrying rate of young women with acute coronary events, a situation rarely seen before, has directed special attention to factors that could be blamed for this surprising finding. The association of hormonal contraception with smoking has emerged as very common in this population, likely to concur not only for atherogenic metabolic features but also for potentially thrombotic coagulation disorders (Mosca, Grundy, Judelson et al, 1999). The cardiovascular impact of alcohollic intake must also be considered. The cardiovascular impact of this behaviour is closely related to the amount of alcohol consumed. A moderate intake may confer some cardiovascular protection, particularly by raising HDL-cholesterol and reducing platelet aggregation, yet it may lead to a higher incidence of arterial hypertension and cerebrovascular disease. Patterns of high alcoholic consumption are an unusually hazardous behaviour, particularly for the heart, greatly increasing the risk of sudden death (Silva, 2000).
heterogeneity of triglyceride-rich lipoproteins (Quilomicra and VLDL) and its inseparable association with other risk factors. However, several studies have demonstrated a clear correlation between their levels and the occurrence of coronary heart disease, indicating that the presence of high levels of Triglycerides leads to a 13% increase in the risk of cardiovascular disease in men and 37% in women (Castelli, 1986;Criqui, Heiss, Cohn et al, 1993;Hokanson & Austin, 1996;Assman, Schulte & von Eckardstein, 1996). With regard to HDL-cholesterol, its inverse relationship with the risk of coronary heart disease is well accepted. In fact, this risk is 2 to 3% lower for each 1mg/dl elevation of HDL-Cholesterol (Gordon, Probstfiel, Garrison et al, 1989). The protective properties of this fraction derive not only from its involvement in reverse cholesterol transport, but are also a consequence of its anti-inflammatory capacity and protection against LDL-cholesterol oxidation (Ansell, Navab, Watson et al, 2004). On the other hand, it is recognized that individuals with very low levels of HDL-cholesterol have a higher cardiovascular risk. This population is often characterized for having concomitant hypertriglyceridemia, obesity, a sedentary lifestyle, active tobacco intoxication and decreased glucose tolerance (World Health Organization, 1999). In fact, an increased occurrence of cardiovascular events is expected for levels of HDL-cholesterol below 40 mg/dl (1.0 mmol/L) in men and less than 46 mg/dl (1.2 mmol/L) in women (UK HDL-C Consensus Group, 2004). Recent evidence further stresses the importance of determining the non-HDL-cholesterol, defined by the concentration of LDL-cholesterol + VLDL-cholesterol. This parameter can better translate the risk of cardiovascular mortality than LDL-cholesterol, as it expresses more accurately the lipoprotein atherogenicity (Cui, Blumenthal, Flaws et al, 2001). In recent years, a large number of risk factors for vascular disease have emerged from the international literature (see Table 2), demonstrating the relevance of more complex lipid disorders for the pathophysiology of atherosclerosis. Other emerging risk factors are related to inflammatory markers, as well as by the presence of metabolic changes, subtle changes in coagulation, hormonal disturbances and psychological or behavioral disorders (ILIB International Lipid Information Bureau, 2003). As we have seen, each stated factor conveys a certain risk to the affected population. However, in everyday clinical practice a large majority of patients have associations of these factors and, as such, have cardiovascular risks that express the magnitude of individual risk factors present in an exponential, rather than additive, trend (Yusuf, Giles, Croft et al, 1998;American Heart Association, 2002). An alternative option, with very promising results in the context of cardiovascular risk stratification and assessment of the effectiveness of lipid-lowering interventions, is the use o f l i p i d r a t i o s , j u s t a s t h e L D L -C h o l e s terol/HDL-Cholesterol ratio and the Total-Cholesterol/HDL-Cholesterol ratio, which have the added advantage of being easy to use in clinical practice . Changes in these relations have in fact been shown to better indicate the reduction in cardiovascular risk compared with the absolute levels of conventionally used lipid measures (Natarajan, Glick, Criqui et al, 2003;Kannel, 2005). On the other hand, the estimated LDL-Cholesterol/HDL-Cholesterol ratio translates, albeit imperfectly, an approach to the relationship of plasma apolipoproteins (apo) A-1 and apo B (Walldius & Jungner, 2005), thus enriching the lipid characterization of each patient, with the possibility of a better discrimination of cardiovascular risk, particularly among groups at intermediate cardiovascular risk . Several large studies have demonstrated that the LDL-Cholesterol/HDL-Cholesterol ratio is an excellent predictor of risk of coronary disease and an excellent way to monitor the impact of lipid-lowering therapies (Manninen, Tenkanen, Koskinen et al, 1992;Kannel, 2005;Cullen, Schulte, Assmann et al, 1997;Stampfer, Sacks, Salvini et al, 1991;Gaziano, Hennekens, O'Donnell et al, 1997). In the Helsinki Study, a clinical trial with a 5-year follow-up, involving more than 4000 middle-aged men with hyperlipidemia, the LDL-Cholesterol/HDL-Cholesterol ratio had a superior prognostic value compared with isolated values of LDL-Cholesterol and HDL-Cholesterol. The predictive ability of this ratio was particularly strong in patients with concomitant elevation of triglycerides. It was further shown that the LDL-Cholesterol/HDL-Cholesterol ratio together with the fasting triglyceride concentration, allowed the identification of a particular subgroup of patients that had a remarkable 70% reduction in the risk of coronary heart disease with gemfibrozil (lipid-lowering agent) therapy. In the PROSPER trial, a retrospective analysis of 6,000 patients, the LDL-Cholesterol/HDL-Cholesterol ratio was the stronger predictor of cardiovascular events in elderly patients (Packard, Ford, Robertson et al, 2005). From this study has emerged the recommendation of pharmacological intervention whenever the LDL-Cholesterol/HDL-Cholesterol ratio values exceed 3.3 units. Another study (PROCAM study) involving about 11,000 men aged between 36 and 65, followed over 4 to 14 years, has documented an extremely important and linear relationship between the LDL-Cholesterol/HDL-Cholesterol ratio and cardiovascular mortality (Cullen, Schulte, Assmann et al, 1997). In this study, cardiovascular mortality peaked for LDL-Cholesterol/HDL-Cholesterol values between 3.7 and 4.3 units. In line with these results is the Physician's Health Study, involving 15,000 men (40 to 84 years), where there was a 53% increase in the risk of an acute coronary event for each one-unit increase in the LDL-Cholesterol/HDL-Cholesterol ratio (Stampfer, Sacks, Salvini et al, 1991). In another mixed study, involving men and women under the age of 76, the LDL-Cholesterol/HDL-Cholesterol ratio showed a strong relationship with the risk of coronary events (Gaziano, Hennekens, O'Donnell et al, 1997), aspect reinforced in an analysis of patients from the Framingham Heart Study, where a clear superiority of LDL-Cholesterol/HDL-Cholesterol ratio in predicting cardiovascular events compared to the levels of isolated LDL-cholesterol and HDL-cholesterol was depicted (Kannel, 2005

Cardiovascular risk -Atherogenesis
To understand the sequence of events that occur at the vascular level, resulting in devastating clinical manifestations that are all too familiar, we must look a little closer at the physiology of this system. One of the most important organs we have without doubt is the vascular endothelium. The endothelium is the inner portion of our vessels, which can be compared to a thin membrane that carpets the blood vessels, and its integrity is fundamental for the maintenance of several potentially unstable equilibria. In this sense, a huge amount of vascular wall or circulating factors are present in close relation to the endothelium, endlessly alternating between defense and aggression, aggression, with Nitric Oxide as the key protector. As the most egregious examples of interaction near the endothelium vicinity, we have the following associations: vasodilation/vasoconstriction; anti-trombotic/pro-trombotic; anti-inflamatory/pro-inflamatory, among others. The relative hegemony of each of these interacting factors will determine the final maintenance of endothelial integrity or, conversely, its dysfunction and destruction (Houston, 2002). Endothelial dysfunction is thus the initial phase of a cascade of events that flow until the onset of clinically overt disease. In a very simplified overview, once the endothelial barrier is compromised, an association of events takes place, mainly with a lipid flooding process of the vascular wall, with the mobilization of inflammatory cells, the expression of chemotactic factors, growth and proliferation of smooth muscle and connective tissue, among others. The histologic consequence of these processes ranges from an initial lipid streak that evolves for an atherosclerotic plaque that may progress to calcification, progressively reducing the vascular lumen (Silva, 2000). Curiously, most clinical cases are not determined directly by the extreme portion of the atherosclerotic continuum. In other words, cardiovascular events do not usually stem from progressive and insidious arterial occlusion, with consequent ischemia of downstream areas. Of course, cardiovascular events tend to be characterized by their acute nature, that is, by their sudden and unpredictable occurrence. As such, the implicit pathophysiology should express facts that support real-life events. In fact, one of the most important factors in the emergence of cardiovascular events is related to the so-called "atherosclerotic plaque stability". Thus, plaques with a small lipid core, with small inflammation infiltrate, and fitted with a thick, tough outer layer will be less susceptible to disruption by various harmful factors, such as blood pressure, sympathetic activity and other vasoconstrictor stimuli. In contrast, plaques with a rich lipid core, inflammatory activity and a significant weak fibrous cap will present a higher risk of fracture and exposure of their internal contents (Ridolfi & Hutchins, 1977). This in turn will lead to the activation of several factors that promote clotting and platelet aggregation in-sito (Falk, 1991), which may also lead to a sudden reduction of the vascular lumen, or even its complete occlusion by thrombosis.
Thus, the atherosclerotic process brings with it a wide array of metabolic, inflammatory and coagulation phenomena, decisively contributing to its clinical expression. Herein lies the justification of the diverse therapeutic targets that aimed for in these patients. The importance of hypercholesterolemia as a key-player in this cascade of events is unquestioned and widely demonstrated in the published literature. A perfect expression of the interaction between research and practice is surely the publication of recommendations and guidelines that assist clinicians in the rationalization of therapeutic means available. These emerge as regular updates of successive collections of published scientific data, outlined in an admirably succinct way so they can be strategically combined and applied to the most varied health systems worldwide. Regarding the core topic of this paper, we have to address the most relevant recommendations published by the European Society of Cardiology and the National Cholesterol Education Program (NCEP). These recomendations were prepared according to an individual-risk perspective, and the therapeutic goals are defined according to the expected individual risk at long-term. Collaborative Group, 2002) showed that a reduction of 30% compared to the more restrictive goal (LDL cholesterol <100 mg/dl) was related to an additional 30% reduction in the relative risk of coronary heart disease. The PROVE IT study (Cannon, Braunwald, McCabe et al, 2004), enrolled patients who had had acute coronary events and showed that larger reductions of LDL-cholesterol, to levels lower than 100 mg/dl, could significantly provide aditional benefit in terms of future cardiovascular mortality and morbidity. According to these results one has to consider more challenging treatment goals. The aim is to reach values of LDL-cholesterol <70mg/dl in patients with very high cardiovascular risk, such as those combining several primary risk factors (with primary relevance for diabetics), in patients with primary risk factors that are poorly controlled (with special care to the ones that maintain smoking habits), in patients with multiple risk factors of the so-called metabolic syndrome (triglycerides ≥ 200 mg/dl, non-HDL-cholesterol> 130 mg/dl, HDL-cholesterol<40 mg/dl) and in patients with history of acute coronary events. The establishment of a therapeutic basis grounded in the control of cardiovascular risk factors has demonstrated its strong validity, and is further reinforced for its effectiveness in terms of cost-benefit. Improved control of risk factors almost certainly contributed to the 50% reduction in cardiovascular mortality observed in the United States of America between 1980 and 1990, with 43% attributable to the verified pharmacological advances (Hunink, Glodman, Tosteson et al, 1997). In the Netherlands, similar results were observed, and primary prevention was responsible for a 40% decline in mortality from coronary heart disease between 1978and 1985(Grobee & Bots, 1996. The adoption of dietary measures in Finland, relying on an increase in the consumption of fruits and vegetables and a reduction of saturated fats intake, has resulted in a 65% reduction in mortality from coronary heart disease in a time horizon of 20 years (Pekka, Pirjo & Ulla, 2002). Despite the promising results indicated by these data, only 35% of Americans with a formal indication for dietary or pharmacological therapy, according to the recommendations of the NCEP (2001), are complying with it (Hoerger, Bala, Bray et al, 1988). In Canada, a study carried out between 1988 and 1993, including patients at high cardiovascular risk admitted to hospitals, showed very low percentages in relation to lipid dosing prescription (28%) and early dietary (22%) or pharmacological (8%) therapy (The Clinical Quality Improvement Network (CQIN) Investigators, 1995). In Europe, results have fallen below expectations. An important follow-up study -EUROASPIRE -between 1995 and 1996, envolving nine European countries, showed that 86% of the enrolled patients had hypercholesterolaemia. Nevertheless, only 32% were on medication, and among those treated only 21% had achieved the target lipid levels (EUROASPIRE Study Group, 1997;EUROASPIRE I and II Group, 2001). In Asia and the Pacific, the outlook is not encouraging either. In patients hospitalized for acute coronary events, quite small rates of lipid profile dosing (1 to 58%) were observed, as well as for the prescription of diet (1 to 32%) or pharmacological (6 to 60%) therapy to patients with high Cholesterol levels (Asian-Pacific CHD Risk Factor Collaborative Group, 1998). The control of risk factors in clinical practice is thus a vaguely realized desideratum. The EUROASPIRE study has clarified some trends from 1995 to 2000. If the positive results have raised expectations, with an improvement seen in the control of hypercholesterolemia and hypertension, they are still accompanied by other rather disappointing indicators, such as those of smoking habits, obesity and diabetes, whose prevalence has been steadily increasing (EUROASPIRE Study Group, 1997;EUROASPIRE I and Group II, 2001). In the United States of America the results are also somewhat disappointing. In survivors of acute myocardial infarction or stroke, the control percentages for some primary risk factors are below expectations, particularly for smoking habits (18%), control of hypercholesterolemia (46%), diabetes (48%) and hypertension (53%) (Qureshi et al, 2001). As in almost all chronic conditions, the real picture lags far behind the expectations and available resources. Regarding hypercholesterolemia, the current situation is even less understandable, given its clear and strong association with the prevailing causes of death and incapacity, and the public awareness of the problem and in consideration of the demonstrated effectiveness of the available lipid-lowering drugs, that may have a quite favorable impact upon the prognosis of patients.

Aim
Given the demonstrated role-playing of blood cholesterol in the atherosclerotic continuum, we designed two studies to ascertain the usefulness of the LDL-cholesterol/HDLcholesterol, Triglycerides/HDL-cholesterol and Total-cholesterol/HDL-cholesterol ratios in predicting cardiovascular risk, through its relation to cardiovascular events and peripheral arterial disease (PAD) in two different clinical and experimental settings.

Study 1 -Usefulness of the lipidic ratios predicting peripheral artery disease in hypertensive patients: A retrospective analysis
The importance of the lipidic profile is well established in atherosclerotic processes related to coronary artery disease. Its relation with atherosclerosis in other vascular territories, particularly the inferior limbs has also received strong support from several experimental settings and in different clinical contexts. In order to address wether the lipid ratios can predict the occurrence of obstructive peripheral artery disease (PAD) we conducted a crosssectional study in a sample of hypertensive patients. The study population consisted of 920 Portuguese nationals, aged between 20 and 91 years (mean 64.23 + 12.30 years).

Methods
A total of 920 hypertensive patients (51.3% female, age 64.22 ± 12.01 years) were consecutively included in the study. None of the patients were taking drugs or were in situations known to affect lipoprotein metabolism. Total cholesterol, triglycerides and HDL cholesterol were measured. LDL cholesterol was obtained by Friedewald's formula (if triglycerides <3.39 mmol/l) or by ultracentrifugation. The LDL-Cholesterol/HDL-Cholesterol, Total Cholesterol/HDL-Cholesterol and Triglycerides/HDL-Cholesterol ratios were calculated in all patients. Blood pressure and heart rate were measured in standard conditions. Ankle-Brachial index (ABI) was estimated bilaterally as the ratio of ankle (left and right) systolic blood pressure and brachial (highest upper limb) systolic blood pressure. The normal range for ABI was 0.9-1.3 mmHg, and individuals with ABI<0.9 were classified as having peripheral arterial disease. All data was processed using STATA for Windows, version 11.1. The distribution of the variables was tested for normality using the Kolmogorov-Smirnov test, and for homogeneity of variance by Levene's test. Simple descriptive statistics were used to characterize the sample and the distribution of variables. Logistic regression analysis was used to determine the influence of the lipidic parameters on the occurrence of PAD.

Dyslipidemia and Cardiovascular Risk: Lipid Ratios as Risk Factors for Cardiovascular Disease 289
Groups were compared using the χ2 test for categorical variables and the Student's t test (2 groups) or ANOVA with the post-hoc Tukey test (3 groups) for quantitative variables. A value of P≤0.05 was taken as the criterion of statistical significance for a 95% confidence interval.

Results
The general characteristics of the studied population are summarized in Table 4 Table 5. Lipid profile of the study population. Figure 1 further ilustrates the differences in the lipidic ratios among patients with and without PAD, with all three considered ratios presenting significant differences between the considered groups. A multivariable logistic regression analysis was also performed considering PAD as the dependent variable (dichotomized in normal/abnormal), and forcing each lipidic parameter (either individual lipis or lipid ratios) in a model adjusted for the conventional Framingham cardiovascular risk factors (age, sex, diabetes, blood pressure, smoking status and body mass index). The observed Odds Ratios (OR) with 95% confidence intervals is depicted in figure 2. Although there's an appreciable tendency of association with PAD in all lipid variables, it reaches statistical significancy only for the lipidic ratios.

Study 2 -Usefulness of the lipidic ratios in a low-to-moderate cardiovascular risk population: A sub-analysis of the EDIVA (Estudo de Distensibilidade Vascular) project
The EDIVA project was an epidemiological study assessing cardiovascular risk through sequential Pulse Wave Velocity measurement (Maldonado, Pereira, Polónia et al, 2011), but since serum lipids were available for all the included patients, we re-analyzed the EDIVA database aiming to address the delineated objective: to ascertain the usefulness of The LDL-Cholesterol/HDL-Cholesterol, Total Cholesterol/HDL-Cholesterol and Triglycerides/HDL-Cholesterol ratios in the general population. The study population consisted of 2200 Portuguese nationals (1290 men and 910 women), aged between 18 and 91 years (mean 46.33±13.76 years). Of these, 668 had low cardiovascular risk, and 1532 were patients with hypertension, diabetes and/or dyslipidemia. Individuals defined as having low cardiovascular risk were those who had had no chronic disease, had never been prescribed chronic pharmacological therapy, and had a normal physical exam, electrocardiogram, blood and urine tests, these characteristics having remained unchanged for at least two annual assessments. The patient group was under pharmacological therapy for at least one of the above pathologies.

Methods
The study's aims were explained to all participants and their informed consent was obtained. The methodology used to collect the data was approved by the Portuguese Data Protection Commission and the study was approved by the Ethics Committees of the hospitals involved. Mean follow-up was 2 years. This was a prospective, multicenter, observational study monitoring the occurrence of major adverse cardiovascular events (MACE) -death, stroke, transient ischemic attack, myocardial infarction, unstable angina, peripheral arterial disease, revascularization or renal failure. Follow-up of the patients consisted of annual assessments including, blood pressure (BP) measurement, laboratory tests, including serum lipids, and clinical observation. Total cholesterol, triglycerides and HDL cholesterol were measured. LDL cholesterol was obtained by Friedewald's formula (if triglycerides <3.39 mmol/l) or by ultracentrifugation. The LDL-Cholesterol/HDL-Cholesterol, Total Cholesterol/HDL-Cholesterol and Triglycerides/HDL-Cholesterol ratios were calculated in all patients. At each consultation, the subjects' weight and height were measured and body mass index (BMI) was calculated in kg/m2. Blood pressure and heart rate were measured in standard conditions, in a supine position and after a 10-minute resting period, by an experienced operator and using a clinically validated (class A) sphygmomanometer (Colson MAM BP 3AA1-2®; Colson, Paris) (Pereira & Maldonado, 2005).Three measurements were taken and the arithmetic mean was used in the analysis. All participants underwent routine fasting laboratory tests. At the first consultation they filled out a questionnaire concerning relevant personal and family history, smoking habits, alcohol consumption and medication. Data from the sample subjects were processed using STATA for Windows, version 11.1. The distribution of the variables was tested for normality using the Kolmogorov-Smirnov test, and for homogeneity of variance by Levene's test. Simple descriptive statistics were used to characterize the sample and the distribution of variables. Cox proportional hazards analysis was used to determine the influence of the lipidic parameters on the occurrence of the specified cardiovascular events. C-Statistics was calculated to address the reliability of the lipidic parameters as prognostic variables.
Groups were compared using the χ2 test for categorical variables and the Student's t test (2 groups) or ANOVA with the post-hoc Tukey test (3 groups) for quantitative variables. A value of P≤0.05 was taken as the criterion of statistical significance for a 95% confidence interval.

Results
The general characteristics of the study population are summarized in Table 6. Mean age was 46.33±13.77, indicating a relatively young sample, with similar proportions of men and women (59% and 41%, respectively). With regard to cardiovascular risk factors, 52% of the patients were hypertensive, 33% had dyslipidemia and 11% were diabetic; 17% were smokers and 15% had a family history of cardiovascular events. About 37% were medicated for cardiovascular pathologies, with 13.6% of the patients undertaking statins. This factor was controlled in all the multivariable analysis. Mean follow-up is currently 21.42±10.76 months. A total of 50 non-fatal MACE (2.2% of the sample) were recorded, including 27 cases of stroke, 19 of coronary events, 2 of renal failure and 2 of occlusive peripheral arterial disease.   Table 7. Lipid profile of the study cohort, stratified for the presence or absence of of MACE.

No
In the multivariable model analysis, adjusting for all conventional Framingham cardiovascular risk factors (age, sex, diabetes, blood pressure, smoking status and body mass index), the lipids ratios were associated with MACE, with stronger associations than the ones observed for the individual lipidic variables. Overall, the Total-Cholesterol/HDL-Cholesterol was found to be the best single predictor of MACE. In figure 3 we plot the hazard ratios for quintiles of the lipid ratios. A linear increase of the hazard ratios across quintiles of the Total-Cholesterol/HDL-Cholesterol is clearly depicted, while for the other ratios only the upper-extreme quintiles showed an important association with cardiovascular events. Comparative data of risk association for those in the extreme quintiles of each lipidic variable is presented in figure 4. Of note, one can see that the combination of two individual lipidic components into a single variable provides stronger association with cardiovascular risk, as expressed by the depicted hazard ratios for the lipid ratios. On the other hand, the lipid ratio with the strongest association was the Total-Cholesterol/HDL-Cholesterol ratio, in line with the data depicted in figure 3. The ROC curve analysis provided the Areas-Under-the-Curve (AUC, equivalent to the Cstatistics) for the different lipid parameter considered in the analysis.

Discussion and conclusions
As previously mentioned, cardiovascular disease, as an expression of atherosclerotic processes, is the leading cause of death in industrialized countries. The key role played by cholesterol in essential pathophysiologic processes that lead to the occurrence of clinically significant cardiovascular events is well recognized. In contemporary clinical practice, this notion is well entrenched, and the individual cardiovascular risk definition incorporates, among other factors, the lipid profile, including the Total Cholesterol, LDL Cholesterol, HDL Cholesterol and triglycerides. A practical evidence of the aforementioned is the fact that the major cardiovascular risk tables currently available (e.g. the Framingham score or the EuroSCORE), incorporate lipid parameters in the definition of thresholds of risk. On the other hand, therapeutic decisions and monitoring have been largely centered on the conventional lipid profile. Even the international recommendations (such as those issued by the National Cholesterol Education Program -NCEP, 2001) recommend target levels of LDL and HDL -cholesterol to determine cardiovascular risk and evaluate the effectiveness of lipid-lowering therapies. However, some studies have indicated important limitations of these parameters in the prediction of cardiovascular risk, particularly in patients with intermediate cardiovascular risk . However, more recent evidence has suggested other lipidic components to optimize the definition of cardiovascular risk in clinical practice. In fact, several studies have expressed the superiority of the levels of apolipoprotein (apo) B, apo A-1 and its ratio, both in predicting cardiovascular events and in the evaluation of treatment efficacy (Packard & Marcovina, 2006;Yusuf , Hawken, Ounpuu, et al, 2004;Meisinger, Loewel, Mraz et al, 2005;Barter, Ballantyne, Carmena et al, 2006;Kim, Chang, Choi et al, 2005). In fact, considering that each lipidic particle contains one molecule of the atherogenic apo B, then its levels are a direct measure of the number of potentially atherogenic particles in the different  (Walldius & Junger, 2006). In contrast, the concentration of apo A-1 translates the number of anti-atherogenic particles contained in the HDLcholesterol, thus enclosing the conceptual framework of apoB/apoA-1 ratio as a measure of the ratio of atherogenic particles versus anti-atherogenic particles transported in the blood.
Despite the growing enthusiasm about the potential of these emerging parameters for their best performance in the definition of cardiovascular risk, there still remain some questions that limit their dissemination in clinical practice. The central question is very practical, and focuses on the cost-benefit relation associated with a change in the traditional clinical approach. In fact, it is not yet clear whether the superiority of these new lipid parameters over the more conventional ones for risk stratification is enough to justify the additional cost inherent to their laboratory determination (Pischon, Girman, Sacks et al, 2005). Furthermore, despite the current literature supporting apolipoproteins as better predictors of cardiovascular events, its use may not be the most practical operational perspective. Moreover, it is not yet clear whether the replacement of conventional parameters for emerging ones will translate into clear clinical benefit, or if, conversely, it will confuse the various protagonists over the clinical decision frame. In contrast to this line of argument, several studies have also emerged affirming quite clearly the advantages of using lipid ratios, based on conventional parameters, such as those studied in this work. This is based on the fact that, on the one hand, they add cardiovascular risk discriminative capacity to the individual lipid parameters, and on the other, they are more favorable than the apolipoproteins considering cost and immediate operationalization . As mentioned earlier, several studies have shown fairly consistently that changes in these ratios are favorable indicators of cardiovascular disease risk, above the absolute levels of individual lipids (Natarajan, Glick, Criqui et al, 2003;Kannel, 2005). Accumulating evidence in this regard is quite broad, spreading over several clinical frameworks (Manninen, Tenkanen, Koskinen et al, 1992;Kannel, 2005;Cullen, Schulte, Assmann et al, 1997;Stampfer, Sacks, Salvini et al, 1991;Gaziano, Hennekens, O'Donnell et al, 1997;Packard, Ford, Robertson et al, 2005). The results presented here clearly fall into this line, reinforcing the belief in the superiority of the lipid ratios, particularly the Total-Cholesterol/HDL-Cholesterol and the LDL-Cholesterol/HDL-Cholesterol ratios, over the classic lipid parameters, predicting peripheral arterial disease in hypertensive patients (in a high cardiovascular risk) and predicting future major cardiovascular events (including stroke and myocardial infarction) in a low-to-intermediate cardiovascular risk population. One of the curious aspects extracted from the second presented study was the existence of a linear relationship for the Total-Cholesterol/HDL-Cholesterol ratio with the risk of MACE, something not apparent in the LDL-Cholesterol/HDL-Cholesterol ratio. This same result was reproduced in the Quebec Cardiovascular Study, in which more than 2.000 middleaged men were followed for 5 years, monitoring the occurrence of major cardiovascular events (Lemieux, Lamarche, Couillard et al, 2001). The lipid parameters with better performance in predicting risk in this study were the Total-Cholesterol/HDL-Cholesterol ratio and the LDL-Cholesterol/HDL-Cholesterol ratio, although only the first stated ratio expressed a linear relationship with risk. One possible explenation for this result is metabolic in nature. In fact, it is well documented that patients with dyslipidemia showing high triglycerides and low HDL-cholesterol (generally patients with abdominal obesity and insulin resistance), often have marginal or even normal levels of LDL-Cholesterol (Lamarche, Després, Moorjani et al, 1996). Moreover, LDL-Cholesterol concentrations are often estimated indirectly from 3 measurements (Total-Cholesterol, Triglycerides and HDL-Cholesterol), which may include a variation that can reach 25% (Schectman & Sasse, 1993), with a potential and quite significant impact in the LDL-Cholesterol/HDL-Cholesterol ratio, eventually under-estimated. By contrast, the two components included in the Total-Cholesterol/HDL-Cholesterol ratio are measured directly. Supporting the superiority of these ratios over the isolated lipid parameters, is their unique ability to reflect the bidirectional cholesterol traffic (in and outward) through the arterial intima in a way that the individual LDL and HDL-Cholesterol levels cannot reach (Kannel, 2005). Consistent with this assumption, another recent cohort prospective study, involving over 15.000 women followed over a period of 10 years, demonstrated that the Total-Cholesterol/HDL-Cholesterol ratio alongside the non-HDL Cholesterol were predictors of future cardiovascular events, as good or better than apolipoprotein fractions (Ridker, Rifai, Cook et al, 2005). Of course, there are still unresolved issues, such as the definition of a cut-off in these ratios from which lipid-lowering therapy should be considered. The current guidelines of the NCEP (2001) recommend a cut-off of 2.5 for the ratio LDL-cholesterol/HDL-cholesterol. However, recent studies suggest that the risk of cardiovascular events begins to have significant expression for values between 3.3-3.7 (Cullen, Assmann & Schulte, 1997), in line with the results we reported here. Given all the data currently available, as long as the fundamental reservations to the routine use of apolipoproteins are not exceeded, the use of lipid ratios in clinical practice is strongly advised, both in risk stratification and therapeutic decision and in monitoring its effectiveness.

Acknowledgments
The author sincerely thanks Dr Markus Carpenter for the linguistic assistance.