1. Department of Internal Medicine, Centralhospital, Karlstad, Sweden

Project plan07 03 14

Markers of inflammation, lipid- and glucose metabolism in relation to cardiovascular risk in two large population-based studies

Payam Khalili (PhD-student) 1,2, Johan Jendle (MD, PhD, assistant tutor) 1,2, Peter M Nilsson (MD, PhD, main tutor) 2,3

1. Department of Internal Medicine, CentralHospital, Karlstad, Sweden

2. Department of Clinical Medicine, ÖrebroUniversity, Örebro, Sweden

3. Department of Clinical Sciences Medicine, LundUniversity, UniversityHospital MAS, S-205 02 Malmö, Sweden

Background

According to a recent World Health Report from the World Health Organization (WHO), cardiovascular disease (CVD) constitutes about 30% of the overall mortality on a global scale. CVD will by 2020 become the leading cause of both mortality and disability worldwide with the number of fatalities projected to more than 20 million annually. Established risk factors for CVD are dyslipidaemia, smoking, hypertension, and diabetes mellitus. Various algorithms have been designed for risk assessment before treatment of individuals from the general population (1,2).

Assessment of cardiovascular risk

Risk-scoring systems evaluate traditional risk factors and greatly improve risk prediction. However, it is widely recognized that for example more than half of all future vascular events will occur in individuals without overt hyperlipidaemia (3). Multiple studies demonstrate that 20% to 25% of all future events occur in individuals with only one traditional risk factor (4). On the other hand, data from the global case-control study INTERHEART suggest that nine biological or lifestyle factors (six for increased and three for decreased risk) may jointly explain up to 80-90% of all cases of myocardial infarction (5). Therefore, there is substantial interest in exploring new biomarkers to identify individuals who are at increased risk for the development of CVD in general, and who could be targeted for preventive measures.

Several new biomarkers have been related to incident cardiovascular events, including levels of N-terminal brain natriuretic propeptide(NT-pro-BNP, a marker of neurohormonal acticity) (6), C-reactive protein, CRP (a marker of inflammation) (3,7), fibrinogen (a marker of thrombosis and inflammation) (8), and D-dimer (a marker of thrombosis)(9).

In an effort to evaluate whether measurement of several biomarkers simultaneously could enhance risk stratification, Wang et. al. measured ten well studied biomarkers (including CRP and NT-pro-BNP) in over 3200 individuals (10). However, the use of these biomarkers added only moderately to standard risk factors for risk assessment and prediction in individual subjects (10).

A number of clinical studies have demonstrated the importance of fat tissue distribution and especially the contribution of visceral fat accumulation in the abdominal region to the development of metabolic disorders, including glucose intolerance and dyslipidaemia. It is now generally accepted that adipose tissue in addition to its classical function as an energy storage depot, also represents an important and very active endocrine organ that produces a number of hormones with significant roles in the regulation of insulin sensivity (11). The first bioactive adipocytokine, leptin, was described in 1994 (12). Leptin was shown to be involved in the control of food intake and energy expenditure and can also induce insulin resistance, as primarily shown in animal models (13).

The role of adiponectin

Another adipocytokine, adiponectin, was discovered in 1995 (14). Reduced serum levels of adiponectin in obese compared with nonobese subjects have been reported, as well as negative correlations between adiponectin and body mass index (15-17). Furthermore, serum concentrations of adiponectin have also been found to be inversely correlated with insulin sensitivity in both non-obese (18,19) and obese (20,21) subjects, as well as in patients with type-2 diabetes mellitus (15). Prospective epidemiological studies have suggested that elevated adiponectin concentrations are associated with greater insulin sensitivity and reduced risk of development oftype-2 diabetes, independently of obesity and other potential confounders (22-24). Thus the development of interventions that raise the adiponectin levels has been proposed as a target to improve insulin sensitivity and glucose tolerance and possibly to prevent coronary heart disease (CHD) risk (25). Adiponectin has also been proposed to protect against CVD by other mechanisms. It has for example been shown that adiponectin is a strong antiinflammatory angentand down-regulates adhesion molecule expression on the surface of endothelial cells (26). However, according to a published prospective study and associated meta-analysis from 2006 any association between adiponectin levels and future CHD is unlikely to be strong (27).

In a recent prospective study (2007) from Uppsala, Sweden, it was reported that low levels of total adiponectin was an independent risk factor for coronary heart disease in 832 elderly 70-year old men followed 10.5 years, independent of conventional risk factors and insulin sensitivity (28). It is unclear to what degree selective survival in these elderly men could have influenced the results. Another aspect is that glucose tolerance status should ideally be evaluated with the standard method for that purpose, an oral glucose tolerance test (OGTT), not always included in health surveys.

It is therefore still of importance to further investigate the role of total adiponectin as a prospective risk marker for CVD and CHD in middle-aged subjects well classified for glucose tolerance at baseline, as glucose metabolism could either be a confounder or an intermediating variable in the putative risk association. This is possible to do in a long-term follow-up of middle-aged men included in the Malmö Preventive Project (MPP), a large health screening study between 1974 and 1992 (29).

The Värmland Health Survey

In the county of Värmland, Sweden, a large population-based screening study with high attendance rate (76%) was carried out between 1962 to1965 with a focus on a selected set of risk markers for CVD and cancer (30-32). In all, a total of 95,000 men and women from a wide age range (25-80 years) participated. Prospective follow-up analyses for cardiovascular risk were made in the late 1980´ ies bytwo independent groups; by Gunnar Lindberget al, Thesis 1992 (33), and by Sven Törnberg, et al, Thesis 1987 (34). One important finding was that sialic acid, an inflammatory marker, predicted cardiovascular risk, but the findings were somewhat difficult to interpret due to lack of data on a number of possible confounders in a large proportion of the subjects. It is, however, possible to do a repeated long-term follow-up after 40 years by use of register linkage with national registers on morbidity and mortality, focusing on risk associated with markers of inflammation and glucose metabolism. One way to do this is to use data from plasma protein electrophoresis, sampled from 6996 of the subjects.

Objectives

The primary aimof this project is to investigate the predictive and independent role for cardiovascular events and diabetes of selected markers of chronic inflammation and lipid- as well as glucose metabolism in two large population-based studies with long-term follow-up, before and after adjustment for obesity, lifestyle, smoking, social factors and family history. Whenever possible this will be done in relation to gender by stratification.

A secondary aim is to investigate cross-sectional associations with plasma adiponectin levels and other biological, social and lifestyle related variables in healthy subjects, as well in disease states.

A third aim is to investigate the influence of an objective family history, based on register data, of cardiovascular disease or diabetes for modification of long-term cardiovascular risk in offspring.

Subjects and Methods

Study populations

First cohort: A preventive case-finding programme called the Malmö Preventive Project (MPP) for detection of cardiovascular risk factors started in 1974 at the Department of Preventive medicine, University Hospital Malmö (29). The aim was to screen large strata of the adult population, in the city of Malmö, located in southern part of Sweden, in order to find high-risk individuals for preventive interventions. Subjects were invited to participate in health screening, including a physical examination and laboratory tests. Between 1974 and 1992, in total 22,444 males and 10,902 females attended the screening programme with an overall attendance rate of 71% (29). Total adiponectin levels have recently been analysed (Flyvbjerg A, Frystyk A, Medical Research Laboratories, AarhusUniversityHospital, DK-8000 Aarhus C, Denmark) in a subgroup of 4,200 middle-aged males, born in pre-specified years and living in Malmö, but not selected in any other way. The reason to choose this subset of MPP is that all individuals performed an oral glucose tolerance test (OGTT; 75 gram glucose) with multiple determinations of blood glucose levels (0´, 30´, 60´, and 120´ minutes) to make an appropriate classification of glucose tolerance and diabetes possible. In addition, all subjects´ baseline variables (lipids, urate, creatinine, lung function via spirometry) as well as a number of data from a self-administered questionnaire (29) have been collected in the database.The baseline data in the Malmö Preventive Project (MPP) cohort are stored at the Department of Clinical Sciences, University Hospital, Malmö, Sweden, and are available for further statistical analyses in this defined sub-sample of males with adiponectin data after formal request approval 061003 by the MPP Biobank Steering Committee (chair: Professor Göran Berglund) after formal application by Peter Nilsson. A similar study on the association between adiponectin and early carotid atherosclerosis has been published from the research group in Malmö (35).

Second cohort: During the years 1962, 1964 and 1965 a general health survey was organised in the county of Värmland, western Sweden, by the Swedish National Board of Health, called the Värmland Health Survey, VHS, (30-32). The original plan was to include the entire county populationabove the age of 25 years,116,638 individuals was invited of the total population of 250.000. In all, 88,959 (76%) first attended and completed the survey. An additional number subjects from the city of Arvika were later included in the survey in 1964. Therefore, in total, data from 95,000 subjects are available for long-term follow-up analyses. Original data have been stored at a national archive (Riksarkivet) in Stockholm, Sweden. These data have been stored on magnetic tapes which are still readableby computers.

The physical examination included measurement of body height (m), weight (kg), blood pressure (mmHg), blood sampling and urine tests were performed. The non-fasting blood samples were analyzed in Stockholm via multi-analyzer technology (Dr. Jungner, CALAB) for the following serum variables:

S-Fe

S-Creatinine

S-Lipids: total cholesterol, beta-lipoprotein

Markers of hepatic function: S-ASAT, S-ALAT, S-tymol, zinc-sulphate reaction

Markers of inflammation: S-protein-bound hexos and S-sialic acid

In addition, local laboratory analyses were carried out for:

B-Estimated sedimentation rate (ESR) (mm)

B-Haemoglobin (Hb)

U-protein, U-glucose (arbitrary units)

In a subset a plasma electrophoresis was carried out in 6996 subjects for the determination of so called M components, to be followed prospectively. In another subset, smoking data were documented after register linkage with a contemporary national survey on smoking habits, carried out during the same period in the early 1960´s.

Methods

We now plan to do the following analyses in the first cohort; the Malmö Prospective Project (MPP) by use of conventional statistical methods for description and follow-up analyses (Cox proportional hazard ratio) in collaboration with senior epidemiologist Gunnar Engström, MD, PhD, Department of Clincial Sciences, Epidemiological group, University Hospital, Malmö, Sweden:

1. Cross-sectional analysis of the association between levels of total adiponectin and markers of glucose metabolism (fasting and during an OGTT). The aim is to investigate if any specific combination of variables reflecting the glucose metabolism, adjusted for age, BMI and smoking habits, is especially closely associated with the total adiponectin levels in middle-aged men. A secondary aim is to investigate similar associations with blood pressure, lipids, hepatic function (ASAT, ALAT, GT), lung function (spirometry), and renal function (creatinine levels). Associations with lifestyle factors will be explored. If possible, also high molecular weight adiponectin will be analysed in a subset of the subjects, for independent calculations of possible associations with other risk factors. Genetic analyses will also be explored in a subset of subjects were DNA has been extracted during a later follow-up examination in 2002-2006.

2. Long-term, follow-up analyses. We will use prospective data on incident cardiovascular disease (CVD) events (coronary heart disease, CHD, and stroke) based on register linkage analyses using MPP data and national register data (The national board of health and welfare). These follow-up analyses cover a time period of 25 to 30 years. The aim is to investigate if total adiponectin, subdivided into tertiles (highest tertile as reference), is associated with long-tern risk of CVD and its components, adjusted for age, BMI, smoking habits and other covariates or confounders.

In the second cohort, the Värmland Health Survey (VHS), we plan to do the following analyses, also based upon register linkages by use of magnetic tapes containing information from VHS and national register (Socialstyrelsen), in collaboration with statistical support at the ÖrebroUniversity:

1.Prediction of CVD events and its components during long-term follow-up (40-years) in relation to markers of inflammation (hexos, sialic acid, ESR, serumelectrophoresis), lipids, renal function (creatinine, proteinuria), and hepatic function (ASAT, ALAT). Subjects with incident fatal or non-fatal CVD events (as well as CHD and stroke) will be compared with subjects without these manifestations in a case-cohort follow-up design, after stratification for age and gender. Multiple regression analyses will be applied.

2.Prediction of diabetes-related outcomes, including retinopathy, nephropathy, or other diabetes-related causes of hospitalisation indicated on the hospital discharge summary classification. The similar predictors will be used, after stratification for age and gender.

3. Family burden of CVD and diabetes and long-term risk in offspring. This approach will be carried out after register linkage between VHS and the Multi-Generation Register (MGR), containing extensive data on parents and siblings of Swedish citizens going back to 1930 (Statistics Sweden, Örebro). The importance of family risk of CVD and diabetes for development of end-points in the next generation, adjusted for individual risk factors in offspring obtained at screening, can be ascertained after register linkage between MGR data on VHS subjects and national registers (Socialstyrelsen), as previously used in the MPP to show separate effects of maternal versus paternal influences on CVD risk in offspring (36).

It is also possible, resources permitting, to do similar register linkages between probands in VHS and the Medical Birth Register (MBR) on pregnancy outcomes (started on a national scale in 1973) as well as the national census register (“Folk- och Bostadsräkning” in 1960) with information on social background (occupation, education, family size, living conditions) for analyses of the contribution of social factors for the prediction of long-term CVD risk. Are, for example, elevated levels of total cholesterol or blood pressure equally predictive of CVD risk in subjects from different occupational categories (manual/non-manual) or not?

Data from MBR can be used in many ways, but will be subject to a forthcoming revision and extension of this project plan.

Ethical aspects

The follow-up analyses of CVD events and diabetes in the MPP cohort has been fully approved by the Regional Ethical Committee, Lund, Sweden in 2004 (Dnr. 85/2004).

Previous follow-up analyses in the VHS study have been ethically approved in the 1980´s, but a new application for ethical approval of long-term, follow-up analyses will be made during Spring 2007 in collaboration with cancer epidemiologists Sven Törnberg, MD, PhD, and John Carstensen, MD, PhD, Karolinska Institute, Stockholm, Sweden, including the new set of epidemiological analyses as outlined in this project plan. This application is currently underway.

Time plan

Planning of the projectSpring 2007

Analyses of adiponectin levels in MPPMarch 2007

Application for ethical approval in VHSMarch 2007

Register linkages with national registerAutumn 2007

Analyses of outcomes in MPP and VHS2008

Publications2008-2010

Summary of project

The unique opportunities in Sweden to link data derived from large-scale population-based screening studies with outcome data from national registers is depending on the existence of a 10-digit personal identification number. This is the prerequisite for the statistical analyses outlined in this research plan. From the MPP in Malmö, several publications have reported on the impact of conventional risk factors (29,37) as well as markers of inflammation (38-40) for risk of CVD and its components. The current approach to investigate the putative role of adiponectin for CVD risk in middle-aged men is in line with previous investigations. In a recent study from Uppsala, Sweden, low levels of total adiponectin was shown to be an independent risk factor for CHD in elderly men (28), but as the MPP men were middle-aged at baseline this will avoid potential bias from selective survival, and in addition take more precise measures of glucose tolerance into account as this is an important covariate of adiponectin for association with early manifestations of atherosclerosis, e.g. withart. carotid intima media thickness (35).

In the VHS, previous follow-up studies after 25 years revealed that the serum sialic acid concentration is a strong predictor of cardiovascular mortality. A possible explanation of this finding is that the

serum sialic acid concentration may reflect the existence or the activity of an atherosclerotic process (41). In another report from VHS it was stated that a mortality trend for ischaemic heart disease after 20 years of follow-up was associated with increasing cholesterol levels for people younger than 65 years, and was more pronounced for men than women (42). For people 65 years or older, there was a weak trend for men, but not even a tendency for women. Regarding acute myocardial infarction(MI), significant trends were observed for males as well as females, for young as well as old people. It will therefore be of considerable interest to expand on these associations after 40 years of follow-up, including the addition of new exposures at baseline as well as new outcome measures. As Värmland is a part of Sweden known for a high incidence of CVD and lower mean life expectancy (43) due to both known (44,45) and unknown circumstances, this is of special importance to further explore.

References

1. Wilson PW, D'Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837-47.

2. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterolin Adults. Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of HighBlood Cholesterol In Adults (Adult Treatment Panel III). JAMA 2001;285:2486-97.