Vitamin C deficiency and risk of myocardial infarction: prospective population study of men from eastern FinlandBMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7081.634 (Published 01 March 1997) Cite this as: BMJ 1997;314:634
- Kristiina NyyssÖnen, clinical biochemista,
- Markku T Parviainen, principal clinical biochemistb,
- Riitta Salonen, research scientista,
- Jaakko Tuomilehto, research scientistc,
- Jukka T Salonen, academy professora
- a Research Institute of Public Health, University of Kuopio, PO Box 1627, 70211 Kuopio, Finland
- b Department of Clinical Chemistry, Kuopio University Hospital, PO Box 1777, 70211 Kuopio
- c Department of Epidemiology and Health Promotion, National Public Health Institute of Finland, Mannerheimintie 166, 00300 Helsinki, Finland
- Correspondence to: J T Salonen
- Accepted 31 December 1996
Objective: To examine the association between plasma vitamin C concentrations and the risk of acute myocardial infarction.
Design: Prospective population study.
Setting: Eastern Finland.
Subjects: 1605 randomly selected men aged 42, 48, 54, or 60 who did not have either symptomatic coronary heart disease or ischaemia on exercise testing at entry to the Kuopio ischaemic heart disease risk factor study in between 1984 and 1989.
Main outcome measures: Number of acute myocardial infarctions; fasting plasma vitamin C concentrations at baseline.
Results: 70 of the men had a fatal or non-fatal myocardial infarction between March 1984 and December 1992. 91 men had vitamin C deficiency (plasma ascorbate <11.4 µmol/l, or 2.0 mg/l), of whom 12 (13.2%) had a myocardial infarction; 1514 men were not deficient in vitamin C, of whom 58 (3.8%) had a myocardial infarction. In a Cox proportional hazards model adjusted for age, year of examination, and season of the year examined (August to Octoberv rest of the year) men who had vitamin C deficiency had a relative risk of acute myocardial infarction of 3.5 (95% confidence interval 1.8 to 6.7, P = 0.0002) compared with those who were not deficient. In another model adjusted additionally for the strongest risk factors for myocardial infarction and for dietary intakes of tea, fibre, carotene, and saturated fats men with a plasma ascorbate concentration <11.4 µmol/l had a relative risk of 2.5 (1.3 to 5.2, P = 0.0095) compared with men with higher plasma vitamin C concentrations.
Conclusion: Vitamin C deficiency, as assessed by low plasma ascorbate concentration, is a risk factor for coronary heart disease.
This study shows that vitamin C deficiency may be associated with an increased risk of myocardial infarction
The findings also provide additional support for the role of oxidative stress and lipid peroxidation in coronary heart disease
Although deficiency in vitamin C may increase coronary risk, this study does not provide evidence in favour of the benefit of vitamin C supplements
The oxidation of low density lipoprotein has been implicated in atherogenesis and in the progression of early atherosclerotic plaques.1 2 3 Experiments in rabbits have shown that atherogenesis can be inhibited by supplementation with antioxidants.1 4 5 6 7 Epidemiological follow up studies suggest that a high intake of vitamin E might be associated with a reduced risk of coronary events.8 9 Increased body stores of the transition metals iron and mercury, which catalyse lipid peroxidation, have been associated with excess risk of myocardial infarction.10 11 12
We found that a high titre of autoantibodies against oxidised low density lipoprotein was associated with accelerated progression of carotid atherosclerosis in a prospective nested case-control study.13 Consistently, associations have been reported between raised titres of antibodies against oxidised low density lipoprotein14 and reduced resistance to oxidation of low density lipoprotein15 and increased severity of atherosclerosis. Vitamin C is an important water soluble dietary antioxidant in humans. We previously reported an association between low plasma vitamin C concentrations and enhanced progression of atherosclerosis.16 Only a few studies have investigated the relation between the use of vitamin C supplements, dietary intake of vitamin C, or plasma vitamin C concentration and the risk of coronary disease, and their results are inconsistent.3 None of them has considered the role of vitamin C deficiency.
We investigated the association of vitamin C deficiency with the risk of myocardial infarction in middle aged men free of coronary disease from eastern Finland, a population with low average vitamin C concentrations and high mortality from coronary heart disease.
Subjects and methods
The Kuopio ischaemic heart disease risk factor study is a population study designed to investigate risk factors for cardiovascular diseases, atherosclerosis, and related outcomes in men from eastern Finland.18 The baseline examinations were carried out between March 1984 and December 1989. The study sample included 3235 eligible men aged 42, 48, 54, or 60 at the baseline examination; 2682 (82.9%) participated. Men with symptomatic coronary heart disease (n = 92) or ischaemia on maximal exercise testing (n = 359),19 or both (n = 585), were excluded from the current analyses. Of the remaining 1646 men, 1605 had data on plasma ascorbate concentration at baseline.
The examination protocol and measurements have been described.10 11 12 13 Plasma ascorbate concentrations were determined by a chromatographic method in deep frozen samples, which were stabilised in 5% metaphosphoric acid immediately after sampling.20 The coefficient of variation between 12 batches was 7.2%.20
Plasma ascorbate concentration was redetermined in a subsample of 401 men with hypercholesterolaemia in samples drawn 4-9 years after the baseline examination. Pearson's correlation between the baseline and the remeasurement values was 0.27. Plasma vitamin C concentration below <11.4 µmol/l (2.0 mg/l) was defined as being at high risk of developing clinical symptoms of hypovitaminosis C and is considered to be a limit for vitamin C deficiency.21 22 For these reasons we used this limit as the primary cut off value for plasma vitamin C concentration.
The main lipoprotein fractions were separated from fresh serum samples by ultracentrifugation and precipitation.23 The cholesterol and triglyceride contents of all lipoprotein fractions were measured enzymatically. Serum apolipoprotein B concentration was determined by an immunoturbidimetric method.12 Serum copper concentration was determined with an atomic-absorption spectrometric technique.24 Plasma fibrinogen concentration was determined on the basis of clotting of diluted plasma with excess thrombin.12 Hair mercury content was measured with a flow injection and amalgam system (FIAS-200) in a Perkin Elmer Zeeman 5000 Spectrometer (Norwalk, CT).12
The consumption of foods was assessed when blood was sampled by recording intake over four days with a questionnaire, which was checked at interview.25 The use of vitamin C tablets and vitamin C containing nutritional supplements was assessed in a self administered questionnaire.
The number of cigarettes, cigars, and pipefuls of tobacco currently smoked daily, the duration of regular smoking in years, and the history of myocardial infarction, angina, and other ischaemic heart disease were recorded using a self administered questionnaire, which was checked by an interviewer. A family history of coronary disease was defined as positive when the biological father, mother, sister, or brother of a subject had myocardial infarction, angina, or ischaemic heart disease.
The lifelong exposure to smoking (cigarette pack years) was estimated as the product of the number of years spent smoking and the number of tobacco products smoked daily at the time of examination. The consumption of alcohol in the previous 12 months was assessed with the Nordic alcohol consumption inventory, which contains 15 items.26 Socioeconomic status was measured with a summary index that combined measures of income, education, occupation, occupational prestige, standard of living, and housing conditions.27 Leisure time physical activity was assessed over the previous 12 months.28
Resting blood pressure was measured by one nurse with a random zero mercury sphygmomanometer.10 12 The mean of six systolic pressure values was used in these analyses. Respiratory gas exchange was measured breath by breath with the MGC 2001 system (Medical Graphics, Minneapolis, MI) during a symptom limited exercise test.19 Ischaemia in exercise electrocardiograms was coded manually by one cardiologist.10 12 Diabetes was defined as either previous diagnosis of diabetes mellitus or fasting blood glucose concentration >8.0 mmol/l.
Acute myocardial infarctions in the cohort were monitored by a registry that collected detailed diagnostic information of all such attacks in the study cohort in a prospective manner.29 30 Between March 1984 and December 1992 a fatal or non-fatal myocardial infarction was registered in 70 of the 1605 men at risk. The follow up period for individual subjects was up to 8.75 years, and the mean follow up time was about five years.
Risk factors for myocardial infarction were analysed using spss Cox proportional hazards models.31 32 All tests of significance were two sided. Relative hazards adjusted for risk factors (risks) were estimated as the antilogarithm of a coefficient for independent variables. Their confidence intervals were estimated with a macro based on the assumption of asymptotic normality of estimates.
Plasma ascorbate concentration ranged from 1.7 µmol/l to 137 µmol/l, with a mean (SD) of 47.7 (23.3) µmol/l. There was a slow average increase in the mean plasma ascorbate concentration during the six years of baseline examinations (46.0, 43.2, 43.7, 49.4, 56.8, and 48.3 µmol/l; r = 0.12, P<0.0001 for linear trend). For this reason the year of examination was adjusted for in all statistical analyses.
Plasma ascorbate concentration decreased with increasing age (r = -0.17). The age specific plasma ascorbate mean concentrations were 55.7, 52.3, 44.3, and 46.6 µmol/l for those aged 42, 48, 54, and 60, respectively (P<0.0001 for linear trend). Of dietary factors, only the intake of fruits and berries (r = 0.276) and of vegetables (r = 0.255) had any notable adjusted correlations with plasma ascorbate concentration. Plasma ascorbate concentration also correlated with the dietary intake of carotenes (r = 0.159) and the sum of C14 to C16 saturated fatty acids (r = -0.096). Men with plasma vitamin C concentrations <11.4 µmol/l differed from men with higher plasma vitamin C concentrations in many respects (table 1). Men deficient in vitamin C were older, had smoked more in their life time, and were in lower socioeconomic groups; they also had lower maximal oxygen uptake, lower dietary carotene intake, a higher blood leucocyte count, lower dietary iron intake, less moderate to vigorous physical activity during leisure time, higher systolic blood pressure, higher alcohol intake, and higher coffee consumption.
Men who had vitamin C deficiency (plasma ascorbate <11.4 µmol/l) were at an increased risk of myocardial infarction in the subsequent years (table 2). The relative risk for this lowest category compared with the highest and after adjustment for age, season, and year of examination was 4.0 (95% confidence interval 1.7 to 9.4; P = 0.0012). There were no differences in the risk in the quarters of vitamin C concentration above this limit (table 2). Of 91 men with vitamin C deficiency, 12 (13.2%) had a myocardial infarction during the follow up, compared with 58 (3.8%) of the 1514 men with higher vitamin C concentrations (P = 0.00037, Fisher's exact test). The proportions of men with infarction in all categories from the lowest to the highest was 12 out of 91, 21 out of 382, 12 out of 383, 13 out of 376, and 12 out of 373 (P = 0.0013 for linear trend in proportions).
The strongest predictors of acute myocardial infarction after adjustment only for age (in years), year of examination (covariates for individual years), and season of the year (August to October v the rest of the year) in this study cohort were maximal oxygen uptake (inversely), the number of pack years smoked, blood leucocyte count, and low plasma ascorbate concentration (table 3). Serum apolipoprotein B, plasma fibrinogen, serum copper, and triglyceride concentrations, hair mercury content, blood haemoglobin concentration, systolic blood pressure, diabetes, moderate to vigorous physical activity during leisure time (inversely), serum high density lipoprotein-2 cholesterol concentration (inversely), family history of coronary disease, and low socioenonomic group were also significantly associated with the risk of myocardial infarction. Men who had vitamin C deficiency (plasma ascorbate concentration <11.4 µmol/l) had a 3.5-fold (1.8 to 6.7, P = 0.0002) risk of myocardial infarction after adjustment for age, season, and year of examination (table 3).
To estimate the independent impact of risk factors, all factors shown in table 3, as well as age, year of examination, season, and three dietary factors, were entered simultaneously in a Cox proportional hazards model. In this multivariate model only the number of cigarette pack years smoked, low plasma vitamin C concentration, family history of coronary heart disease, low maximal oxygen uptake, and high hair mercury content (in the order of strength) had significant associations with infarction risk (table 3). When all risk factors shown in table 3, as well as dietary intakes of tea, fibre, and saturated fats, were allowed for, men with plasma ascorbate concentrations <11.4 µmol/l had a relative risk of myocardial infarction of 2.5 (1.3 to 5.2, P = 0.0095) compared with those with higher plasma ascorbate concentrations.
Only 89 (5.5%) of our study subjects took supplements containing vitamin C. The exclusion of these men did not materially change the results. In fact, among the 1516 men who had not used vitamin C containing supplements during the previous week the relative risk for having a plasma ascorbate concentration <11.4 µmol/l was 3.7 (1.9 to 7.2, P = 0.0001) after adjustment for age, season, and year of examination. The use of vitamin C supplements had no significant association with the risk of myocardial infarction.
Vitamin C as an antioxidant
Ascorbate is considered to be the most effective antioxidant in human plasma,33 and vitamin C inhibits the oxidation of low density lipoprotein in vitro.17 34 35 36 Ascorbic acid is the first antioxidant to be used up during lipid peroxidation in plasma, and detectable lipid peroxidation starts only after all ascorbate has been completely used up; researchers have even suggested that only ascorbate can prevent the initiation of lipid peroxidation.17 33 36
Ascorbate is an important physiological antioxidant that helps to regenerate reduced antioxidative tocopherol from the tocopheroxyl radical.35 36 Although it is not lipid soluble, ascorbate could theoretically also inhibit lipid peroxidation through this mechanism. Retsky and Frei have suggested that vitamin C spares, rather than regenerates, tocopherol and other endogenous antioxidants when low density lipoprotein is exposed to copper ions and that ascorbic acid can terminate lipid peroxidation, thereby protecting partially oxidised low density lipoprotein against further oxidative modification.39 Vitamin C's inhibition of lipid peroxidation in vivo has not, however, yet been confirmed in supplementation studies in humans.
Vitamin C and blood pressure
Vitamin C deficiency was associated in our study with raised blood pressure. There is some evidence in favour of a role of oxidative stress in the aetiology of hypertension.40 We observed in a population study an association between lowered plasma concentrations of ascorbic acid and raised resting blood pressure both in normotensive and in hypertensive men.41
In a small clinical trial to investigate the effect of antioxidant supplementation on blood pressure in 40 middle aged men the mean systolic blood pressure decreased by 12.5 (SE 2.5) mm Hg in the supplemented group and by 5.2 (1.9) mm Hg in the placebo group (P = 0.027 for difference).42 In addition, the reduction in blood pressure correlated strongly with the increase in plasma ascorbic acid concentration.42
Vitamin C and coronary heart disease
In a prospective population study, the national health and nutrition examination survey epidemiologic follow up study, men and women with the highest vitamin C intakes (>50 mg/day and regular vitamin C containing supplements) had lower coronary mortality than subjects with the lowest vitamin C intake (by 45% and 25%, respectively).44 In the health professionals' follow up study and in the nurses' health study the use of vitamin C supplements was not significantly associated with the risk of coronary events,8 9 although there was a non-significant trend towards a reduced risk among those taking supplements.
In the 12 year follow up of the prospective Basel study a low plasma concentration of both vitamin C (<22.7 µmol/l) and carotene (<0.23 µmol/l) was associated with a twofold increase in the risk of coronary heart disease (P = 0.022).45 46 The excess mortality from heart disease associated with low plasma ascorbate concentration alone was not, however, significant in this study of 2974 Swiss men with comparatively high average antioxidant intakes.
In a Finnish study a low dietary intake of vitamin C was associated with increased coronary mortality among women but not among men in a cohort of 5133 Finns who did not have heart disease and were followed up for 14 years.47 In a 20 year follow up study in Britain by Gale et al of 730 elderly men and women who were free of major cardiovascular disease at baseline the mortality from cerebrovascular stroke was highest in those with the lowest vitamin C intake.48 People in the highest third of plasma vitamin C concentration had a 30% reduced risk of stroke compared with those in the lowest third after adjustment for a number of cardiovascular risk factors. No association was found between vitamin C concentration and mortality from coronary heart disease.48
Riemersma et al observed an association between low plasma concentrations of vitamin C and an increased risk of angina in a population based case-control study.49 The unadjusted risk was 2.4 in the lowest fifth of plasma vitamin C concentration (<13.1 µmol/l). This odds ratio was weakened to 1.6 and was not significant when smoking and other coronary risk factors were adjusted for.
Confounding is a major problem in all non-experimental studies of associations between nutrient intakes and the risk of disease. An important potential source of bias in studies of the association between plasma vitamin concentrations or the use of supplements and the risk of coronary heart disease are other differences between people with a low and a high intake of vitamins. Although people taking vitamin supplements are generally more health conscious than others, those with a vitamin deficiency may be ill or have detrimental health habits. We tried to assess differences between subjects with low and high plasma ascorbate concentrations in every respect that was relevant to coronary risk. However, some unmeasured factors might have counfounded the relation between vitamin C concentration and risk of infarction or the statistical control for confounding may not have been perfect because of imprecision in measurement or an imperfect fit of the model used. Plasma concentration of vitamin C varied considerably with season within subjects. This variability was most probably random and thus was likely to attenuate the observed relation with risk of infarction. It also implies that a single measurement of plasma vitamin C concentration may classify subjects incorrectly and may be insufficient for clinical practice.
Common to all previous prospective studies is a narrow range and a high overall population concentration of vitamin C. We previously reported that the average plasma ascorbate concentration is comparatively low in eastern Finnish men and that during winter it falls below the limit of deficiency (2.0 mg/l).22
To our knowledge, our current results are the first empirical evidence in humans to show that vitamin C deficiency, as measured by low plasma ascorbate concentration, is a risk factor for coronary heart disease. In our cohort plasma vitamin C concentration above the limit of deficiency was not associated with the risk of acute myocardial infarction. Thus, high intakes of vitamin C or vitamin C supplements would probably not reduce the risk of acute myocardial infarction. Our findings suggest, instead, that if a minimal necessary requirement of vitamin C is not met the risk of myocardial infarction is increased.
We thank Dr Rainer Rauramaa of the Kuopio Research Institute of Exercise Medicine for collecting data; Ms Merja Ihanainen for food recordings; Dr Jaakko ErÄnen for coding the exercise electrocardiograms and Drs Esko Taskinen, Juha VenÄlÄinen, and Hannu Litmanen for supervising parts of the maximal exercise tests; Mrs Marjatta Kantola, Dr Sirpa Suntioinen, and Mr Kari SeppÄnen for performing some of the chemical analyses; and Mr Kimmo Ronkainen for carrying out the data analyses.
Funding Finnish Academy and the Finnish Ministry of Education.
Conflict of interest None.