Antioxidant state and mortality from coronary heart disease in lithuanian and swedish men: concomitant cross sectional study of men aged 50BMJ 1997; 314 doi: http://dx.doi.org/10.1136/bmj.314.7081.629 (Published 01 March 1997) Cite this as: BMJ 1997;314:629
- Margareta Kristenson, head of departmenta,
- Bo ZiedÉn, medical studentb,
- Zita KucinskienË, professorc,
- Algis Abaravicius, associate professorc,
- Laima RazinkovienË, senior chemistc,
- Liselotte SchÄfer Elinder, research fellowd,
- BjÖrn Bergdahl, associate professore,
- Birgitta Elwing, nutritionistf,
- Henrikas Calkauskas, associate professorg,
- Anders G Olsson, professore
- a Department of Health and Environment Faculty of Health Sciences S-58185 LinkÖping Sweden
- b Clinical Research Centre Faculty of Health Sciences S-58185 LinkÖping
- c Department of Physiology and Biochemistry Faculty of Medicine 2021 Vilnius Lithuania
- d Department of Medical Biochemistry and Biophysics Faculty of Medicine Karolinska Institute S-171 77 Stockholm Sweden
- e Department of Medicine and Care Faculty of Health Sciences S-58185 LinkÖping
- f Department of Preventive Medicine Centre of Public Health Sciences S- 58185 LinkÖping
- g Department of Gastroenterology and Dietetics Faculty of Medicine 2021 Vilnius Lithuania
- Correspondence to: Professor Olsson
- Accepted 12 December 1996
Objective: To investigate possible risk factors and mechanisms behind the four times higher and diverging mortality from coronary heart disease in Lithuanian compared with Swedish middle aged men.
Design: Concomitant cross sectional comparison of randomly selected 50 year old men without serious acute or chronic disease. Methods and equipment were identical or highly standardised between the centres.
Setting: LinkÖping (Sweden) and Vilnius (Lithuania).
Subjects: 101 and 109 men aged 50 in LinkÖping and Vilnius respectively.
Main outcome measures: Anthropometric data, blood pressure, smoking, plasma lipid and lipoprotein concentrations, susceptibility of low density lipoprotein to oxidation, and plasma concentrations of fat soluble antioxidant vitamins.
Results: Systolic blood pressure was higher (141v 133 mm Hg, P<0.01), smoking habits were similar, and plasma total cholesterol (5.10 v 5.49 mmol/l, P<0.01) and low density lipoprotein cholesterol (3.30 v 3.68 mmol/l, P<0.01) lower in men from Vilnius compared with those from LinkÖping. Triglyceride, high density lipoprotein cholesterol, and Lp(a) lipoprotein concentrations did not differ between the two groups. The resistance of low density lipoprotein to oxidation was lower in the men from Vilnius; lag phase was 67.6 v 79.5 minutes (P<0.001). Also lower in the men from Vilnius were mean plasma concentrations of lipid soluble antioxidant vitamins (ß carotene 377v 510 nmol/l, P<0.01; lycopene 327 v 615 nmol/l, P<0.001; and lipid adjusted tocopherol 0.25v 0.46 µmol/mmol, P<0.001. Tocopherol concentration did not differ). Regression analysis showed that the lag phase was still significantly shorter by 10 minutes in men from Vilnius when the influence of other known factors was taken into account.
Conclusions: The high mortality from coronary heart disease in Lithuania is not caused by traditional risk factors alone. Mechanisms related to antioxidant state may be important.
Mortality from coronary heart disease in 50-54 year old men is four times higher in Lithuania than in Sweden
Differences in traditional risk factors for coronary heart disease in 50 year old men in LinkÖping (Sweden) and Vilnius (Lithuania) were small–systolic blood pressure was higher in men from Vilnius, but total and low density lipoprotein cholesterol concentrations were lower and smoking habits similar
The resistance of low density lipoprotein to oxidation was lower in men from Vilnius and remained after adjustment for antioxidant vitamin concentrations
Plasma concentrations of the antioxidant vitamins, ß carotene, lycopene, and lipid adjusted tocopherol were lower in men from Vilnius; tocopherol did not differ
Mechanisms related to antioxidant state may be important in explaining the much higher mortality from coronary heart disease in Lithuanian compared with Swedish middle aged men
Mortality from coronary heart disease has increased dramatically during the past 10-15 years in eastern Europe, especially in middle aged men, but it has decreased in western Europe.1 2 3 Bobak and Marmot recently highlighted these diverging trends in mortality and the urgent need to investigate it.4 Figure 1) shows an example of these trends in middle aged men in Lithuania and Sweden. The generally held view is that traditional risk factors for coronary heart disease–that is, high blood pressure, smoking, and dyslipidaemia–have the same predictive strength in eastern and western Europe and could explain these differences in mortality.5 However, other factors may also be important. Studies have shown associations between the susceptibility of low density lipoprotein to oxidation and the severity of atherosclerosis.6 7 8 9 Furthermore, antioxidant vitamins may have a protective role in coronary heart disease.10 We compared men aged 50 in LinkÖping (Sweden) and Vilnius (Lithuania) to elucidate possible causes of the increased rate of coronary heart disease in Lithuania. We investigated traditional risk factors and other suggested mechanisms behind coronary heart disease such as the susceptibility of low density lipoprotein to oxidation and plasma concentrations of antioxidant vitamins.
Subjects and methods
The LinkÖping-Vilnius coronary disease risk assessment study was a cross sectional study conducted concomitantly in Vilnius (600 000 inhabitants) and LinkÖping (130 000 inhabitants) from October 1993 to June 1994. It was approved by the ethics committee of LinkÖping University. A list of randomly selected men born between 1 July 1943 and 30 June 1944 was obtained from the census register in each city. The exclusion criterion was having serious acute or chronic diseases because such diseases could influence the results of investigations or make participation impossible.
The experimental protocol was thoroughly standardised between the two centres. Biochemical analyses were performed in one laboratory, at LinkÖping, except for vitamin concentrations, which were measured in Stockholm. Blood was drawn into prechilled tubes coated with ethylenediaminetetra-acetic acid (EDTA), kept cool on ice, and centrifuged after 120 minutes at 4 C. Samples were stored in a dark refrigerator at 4 C. Samples from Vilnius were sent every week to LinkÖping as express cool packages (4 C). Both sets of samples were always analysed together in random and blinded order. Thus temperatures and times were the same for samples from the two centres.
The volunteers came to the hospitals between 0730 and 0900 after having fasted and abstained from smoking for 12 hours. The morning dose of prescribed drugs was taken. We measured their body weight, height, sagittal diameter of the abdomen, and girth of waist, thigh, and hip.11 12 Blood pressure was measured twice using a mercury manometer, pulse rate was measured once after resting supine for 5 minutes, and a blood sample was taken. Smoking, alcohol, and physical activity were recorded by questionnaires. Work and leisure time physical activity were coded according to a four point scale.
Cholesterol and triglyceride concentrations were analysed by enzymatic calorimetric methods (monotest cholesterol CHOD-PAP and triglycerides GPO-PAP; Boehringer Mannheim, Germany).
Lipoproteins containing apolipoprotein B were precipitated with phosphotungstic acid and magnesium ions and the cholesterol concentration in the solution was regarded as high density lipoprotein cholesterol. Low density lipoprotein cholesterol was calculated.13 Apolipoprotein A I and B were measured by a rocket electroimmunoassay.14 Lp(a) lipoprotein was estimated by an enzyme linked immunosorbent assay (ELISA) (TintEliza, Biopool, Sweden).
The susceptibility of low density lipoprotein to oxidation was measured as described by Kleinveld et al.15 Each time two or three samples from each city that had been taken on the same day were analysed together, the time between blood sampling and analysis being eight days. Low density lipoprotein (0.75 ml) was dialysed against 3 litres of phosphate buffer 0.01 mmol/l, which had a pH of 7.4, contained sodium chloride 0.16 mmol/l, chloramphenicol 0.1 g/l, and EDTA 10 µmol/l, and was continuously bubbled through with nitrogen gas. After 20 hours of dialysis the lipoprotein was filtered (pore size 0.45 µm) and diluted in phosphate buffer that did not contain EDTA to a concentration of protein of 25 µg/ml. Total protein concentration was determined by the Lowry method with bovine albumin as the protein standard. Oxidation was initiated by copper sulphate 5 µmol/l, and its kinetics were monitored every two minutes as the change in absorbency at a wavelength of 234 nm at 30°C on a spectrophotometer (Beckman DU 640) equipped with a six position automatic sample changer. Lag phase in minutes was defined as the time between the addition of copper ions to the low density lipoprotein sample and the time point when the slope during the propagation phase reached baseline absorbency. The interassay coefficient of variation was 5%.
Resistance to oxidation in whole serum (serum lag phase) was determined according to the method of RegnstrÖm et al.7 Serum was diluted to 0.67% (volume for volume) with phosphate buffer that did not contain EDTA. The change in absorbency at a wavelength of 234 nm was monitored every two minutes for four to five hours after the addition of copper sulphate (final concentration 50 µmol/l). The serum lag phase was calculated in the same way as described above. The interassay coefficient of variation was less than 6%.
Plasma concentrations of the lipophilic antioxidants and tocopherol and and ß carotene and lycopene were determined by reverse phase high performance liquid chromatography.16 Concentrations of and tocopherol were expressed relative to total triglyceride plus total cholesterol concentrations. Food intake was examined by 24 hour dietary recall.17 Total and percentage energy from food constituents were calculated according to national food tables.18 19 Twenty per cent of the volunteers were interviewed on a Monday and the others from Tuesday to Friday.
We used the statistical package for social sciences (spss) for Macintosh 6 for statistical analyses. Student'st test was used to test differences between groups. Dichotomous data were tested by a 2 test. The Mann-Whitney U test was applied when the data had a skewed distribution, but it did not alter the test results in any substantial way. Thus, only the results from Student'st test are given in the tables. P values of 0.01 or less were regarded as significant. Multiple regression analysis was performed to study the dependence of lag phase on different variables.
The participation rate was 83% in both cities. In Vilnius 109 men participated; 131 were invited, 18 did not answer, and four refused. In LinkÖping 101 men participated; 122 were invited, 16 did not answer, three refused, and two were excluded. The number of participants with a cardiovascular diagnosis was similar. In each city 10 men were receiving treatment for hypertension. Five men in Vilnius had had a myocardial infarction and one had had a stroke; the corresponding figures in LinkÖping were four and two respectively.
As shown in table 1), mean body weight did not differ between the men in the two cities, but men in Vilnius had a higher mean body mass index. However, abdominal sagittal diameter and the ratio of waist to hip girth did not differ between the groups. Systolic blood pressure was higher in the men from Vilnius, but diastolic blood pressures did not differ. The number of current smokers was similar. More men in Vilnius reported low rates of physical activity during leisure time, but there was no difference during work time. The total energy intake did not differ between the two groups (table 2)). Men in Vilnius had a higher relative fat intake than those in LinkÖping. Energy contribution from alcohol was similar (table 3)).
Mean total and low density lipoprotein cholesterol concentrations were lower in the men from Vilnius (table 4)), and the mean ratio between low and high density lipoprotein and apolipoprotein B concentrations were insignificantly lower. Plasma triglycerides, high density lipoprotein cholesterol, and Lp(a) lipoprotein concentrations did not differ.
Men from Vilnius had an appreciably shorter lag phase for low density lipoprotein after oxidative stress (table 4)). Figure 2) shows the cumulative distributions of the lag phase for low density lipoprotein; the whole distribution is moved towards shorter times in the Lithuanians. Lag phase in whole serum was also shorter in the men from Vilnius.
The plasma concentrations of fat soluble antioxidants are given in table 5). Plasma concentrations of ß carotene, lycopene, and tocopherol (corrected for lipid concentration) were lower in the men from Vilnius; carotene and tocopherol corrected for lipid concentration did not differ between the two groups.
We studied the dependence of the lag phase on and tocopherol, ß carotene, lycopene, low density lipoprotein cholesterol, high density lipoprotein cholesterol, body mass index, energy percentage fat intake, systolic blood pressure, and city in a forward stepwise multiple regression model. The variables with independent significant explanatory relations to lag phase were tocopherol, low density lipoprotein cholesterol, and city. R2 for the equation was 0.30–that is, 30% of the variation in lag phase could be explained by these three factors. Thus after correction for tocopherol and low density lipoprotein cholesterol concentration the difference between the cities was still significant by 10.3 (SE 1.8) minutes (P<0.001).
We studied comparatively small, randomly selected and homogeneous population samples in two countries by using identical methods concomitantly. This approach has previously shown significant differences in the distribution of risk factors which could help to explain mechanisms behind observed differences in mortality from coronary heart disease between populations.20 21 22
The differences in traditional risk factors in the men from Vilnius and LinkÖping were small. This is in agreement with the World Health Organisation's monitoring trends and determinants in cardiovascular disease study (MONICA) performed in 1983-6, which showed that Lithuanian men from Kaunas had higher body mass index and moderately higher systolic blood pressure than Swedes from Gothenburg, whereas plasma cholesterol concentrations and smoking habits were similar.23 In addition, the Kaunas-Rotterdam intervention study showed that in 1972-4 Lithuanian men had a more advantageous cardiovascular risk profile with less smoking, lower cholesterol concentration, and higher rates of physical activity than Dutch men.5 Obviously, data from these two studies could not predict the dramatic increase in mortality from coronary heart disease in Lithuania over the past 10 years.
Two major findings
A striking finding in our study was the lower resistance of low density lipoprotein to oxidation in men from Vilnius. Several studies have suggested that low density lipoprotein may be oxidised in the arterial wall and thus initiate and promote atherosclerosis.6 A short lag phase for the oxidation of low density lipoprotein is associated with coronary atherosclerosis in patients with coronary heart disease.7 9 Susceptibility of low density lipoprotein to oxidation has been related to progression of atherosclerosis in carotid and femoral arteries, and a higher proportion of partially oxidised low density lipoprotein was found in patients with progression of atherosclerotic plaques.8
Another finding of interest was the lower plasma concentrations of ß carotene, lycopene, and tocopherol in men from Vilnius. No difference was found in the mean concentrations of tocopherol corrected for lipid concentrations. The relevance of these findings is supported by a recent study showing that Swedish men with coronary heart disease had reduced serum concentrations of tocopherol but not tocopherol.24 Furthermore, in a European multicentre case-control study of patients with coronary heart disease the mean ß carotene concentration in adipose tissue was significantly lower in cases, but no difference was found in tocopherol concentration.10 However, Riemersma et al found that plasma concentrations of vitamin E were independently and inversely related to the risk of angina.25 The health professionals follow up study found that high intakes of vitamin E were associated with decreased rates of coronary heart disease only when the high concentrations were obtained through dietary supplements such as vitamin pills.26 In line with this, it was recently shown in English patients with coronary heart disease that 400 to 800 IU daily of tocopherol decreased the number of cardiovascular deaths and non-fatal infarctions by 47% after a mean follow up of 510 days.27
Some studies support the view that ß carotene has a protective role in coronary heart disease. A case-control study showed an increased risk of subsequent myocardial infarction with low concentrations of ß carotene in blood samples drawn 7-14 years before the infarction.28 Furthermore, a 12 year follow up study found that mortality from cardiovascular disease was correlated to low baseline concentrations of ß carotene.29 In the health professionals follow up study ß carotene intake was associated with a lower risk of coronary heart disease among current smokers.26 On the other hand, two recent interventional studies with ß carotene treatment for up to 12 years did not show any effect on cardiovascular disease.30 31
Relation between lag phase and serum vitamins
The dependence of lag phase on antioxidants and other factors in a multiple regression model showed that 30% of the variability in the lag phase could be explained by tocopherol and low density lipoprotein cholesterol concentrations. Furthermore, the city also contributed to the explanation, indicating that other factors that we did not measure are also important. The dependence of lag phase on tocopherol concentration was unexpected as the two populations did not differ in the lipid adjusted concentrations of this antioxidant. However, tocopherol was the most abundant lipid soluble antioxidant in all of the men and should therefore have an impact on the susceptibility of low density lipoprotein to oxidation. As only part of the lag phase was explained by the measured variables, other hitherto unknown factors must account for the difference in lag phase between the cities.
The finding of lower antioxidant concentrations in Vilnius may reflect differences in dietary habits between the two populations. Unfortunately, the food tables used in Lithuania are not reliable enough to determine the intake of micronutrients and the type of fat. Men in Vilnius got more of their energy from fat than did the men in LinkÖping, but their total and low density lipoprotein cholesterol concentrations were lower, which may indicate a higher intake of polyunsaturated fatty acids.32 Kardinaal et al showed that low concentrations of ß carotene in adipose tissue are particularly related to risk of coronary heart disease when intakes of polyunsaturated fatty acids are high.33 This favours the view that the low concentrations of ß carotene found in Lithuania may be a risk factor.
In conclusion, the results from our study suggest that traditional risk factors cannot explain the recent large increase in mortality from coronary heart disease in Lithuania compared with Sweden. The suggestion recently put forward that mechanisms other than traditional risk factors are responsible for the diverging trends in coronary heart disease in eastern and western Europe4 is in line with our results. The differences in lag phase of low density lipoprotein oxidation and plasma concentrations of lipid soluble antioxidant vitamins in randomly selected Lithuanian and Swedish men are similar to differences recorded in men with and without coronary heart disease. This leads to the hypothesis that the increase in coronary heart disease in Lithuania may be related to the antioxidant status.
We thank nurses Susanne Warjerstam-Elg and Regina Milasiene for skilful technical help, and Dr Carin Kullberg for help with data handling.
Funding Supported by grants from the Swedish Medical Research Council (No 06962), the Royal Swedish Academy of Sciences, the Swedish Institute, and the LinkÖping University and Procordia Research Fund.
Conflict of interest None.