Randomised controlled trial of effect of fruit and vegetable consumption on plasma concentrations of lipids and antioxidantsBMJ 1997; 314 doi: https://doi.org/10.1136/bmj.314.7097.1787 (Published 21 June 1997) Cite this as: BMJ 1997;314:1787
- Sarah Zino, nutritionista,
- Murray Skeaff, senior lecturerb,
- Sheila Williams, research fellowc,
- Jim Mann, professor in human nutrition and medicineb
- a Department of Human Nutrition, University of Otago, Box 56, Dunedin, New Zealand
- b Department of Human Nutrition, University of Otago, Dunedin, New Zealand
- c Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand
- Correspondence to: Professor Mann
- Accepted 8 April 1997
Objectives: To determine the extent to which plasma antioxidant concentrations in people with habitual low intake of fruit and vegetables respond to increased intakes of these foods. To examine whether advice to increase fruit and vegetables will result in reduction of concentrations of total and low density lipoprotein cholesterol.
Design: Randomised controlled trial in which intervention and control groups were followed up for eight weeks. The intervention group was asked to consume eight servings of fruit and vegetables a day.
Setting: Dunedin, New Zealand.
Subjects: Eighty seven subjects with normal lipid concentrations who ate three or fewer servings of fruit and vegetables daily.
Main outcome measures: Plasma concentrations of vitamin C, retinol, α and ß carotene, α tocopherol, lipids, and lipoproteins. Dietary intake assessed with diet records over four days.
Results: The mean plasma vitamin C, α carotene, and ß carotene concentrations increased in parallel with increased dietary intake of fruit and vegetables in the intervention group. Concentrations of retinol, α tocopherol, lipids, and lipoproteins remained unchanged despite some increase in dietary vitamin E and a small reduction in saturated fat intake.
Conclusions: Following a recommendation to increase fruit and vegetable consumption produces change in plasma concentrations of vitamin C, α carotene, and ß carotene likely to reduce incidence of cancer. More specific dietary advice to modify fat intake may be necessary to reduce the risk of cardiovascular disease mediated by lipoprotein and vitamin E.
Increasing intake of fruit and vegetables raises plasma concentrations of vitamin C and α and ß carotene
These changes in plasma concentrations of antioxidants are probably associated with reduced risk of cancer
A simple recommendation to increase fruit and vegetable intake has little effect on plasma concentrations of α tocopherol, lipids, and lipoproteins
More specific dietary advice to modify fat intake may be necessary to reduce risk of cardiovascular disease associated with lipoproteins and vitamin E
Prospective studies suggest that people with high intakes of fruit and vegetables or blood antioxidant concentrations1 2 in the highest quantile of the distribution have low risks of epithelial cancers,3 coronary heart disease,4 and stroke.5 The anticarcinogenic properties of antioxidants in animal and cell culture systems and their ability to reduce oxidation of low density lipoproteins provide plausible biological explanations for the epidemiological associations.6 7
These observations have led to the recommendation that populations with high rates of cardiovascular disease and epithelial cancers should substantially increase dietary antioxidants through increased consumption of fruit and vegetables. A common response to these recommendations has been a dramatic increase in the consumption of antioxidant supplements, despite little evidence of benefit in placebo controlled trials of supplemental ß carotene, retinol, α tocopherol, and vitamin C.8 9 10 11 Vitamin E supplementation in patients with atherosclerosis may reduce risk of subsequent myocardial infarction but does not prolong life.12 The results of these intervention trials suggest that any reduction in the risk of disease associated with high antioxidant nutrient intake may result from consuming a mix of foods rich in antioxidants rather than consuming antioxidants as single nutrients.
There are few data concerning the extent to which increased intake of fruit and vegetables influences antioxidant concentration in the blood, information which is essential for rational dietary recommendations. We report the results of what we believe is the first population based study to examine the effects on blood antioxidant concentrations of increased consumption of fruit and vegetables. The dietary approach used has also enabled us to examine whether increased consumption of fruit and vegetables reduces the intake of saturated fatty acids to an extent which will facilitate lowering of concentrations of total and low density lipoprotein cholesterol.
Subjects–Ninety volunteers (26 men aged 19-69 years, 64 women aged 18-61 years) were recruited from 120 respondents to advertisements. Subjects had to be healthy with no history of chronic disease; have a total cholesterol concentration of <7.5 mmol/l and triglyceride concentration <3 mmol/l; not be taking dietary supplements or drugs affecting lipid metabolism; and be consuming three or fewer servings of fruit and vegetables daily. A serving equated to 1 cup raw vegetables, ½ cup cooked vegetables, ¾ cup vegetable juice, 1 medium sized whole fruit (for example, an apple), ½ cup cooked or canned fruit, or ¾ cup fruit juice.13 Smokers were not excluded.
Study design–After a two week run in during which participants consumed their usual diet, they were randomly assigned, by using random numbers, to intervention and control groups and followed up for eight weeks. Four day diet records were completed during the run in and during week 4. Unannounced 24 hour recalls were obtained at week 6 as an additional measure of compliance. Fasting venous blood samples (20 ml) were collected and body weight recorded at baseline and fortnightly intervals. Eighty seven participants completed the trial, one moved (control) and two withdrew (one intervention, one control). Ethical approval was obtained from the Southern Regional Health Authority, and all subjects gave informed consent.
Diet–The control group maintained their usual eating habits whereas the intervention group were instructed to increase consumption of fruit and vegetables to eight servings daily and not to alter consumption of nuts, oil, butter, or margarine. They were given detailed dietary instructions, menu suggestions, and recipes, reinforced during fortnightly individual interviews. Targets were achieved by incorporating vegetables into mixed dishes, soups, salads, and casseroles. Juices were substituted for other drinks and fruits for baked products.
Laboratory methods–Plasma samples were stored at -80°C. Lipids and lipoproteins were analysed as described previously for our laboratory.14 The coefficients of variation were 1.3% for cholesterol, 3.2% for triglycerides, and 3.6% for high density lipoprotein cholesterol. Concentrations of retinol, α tocopherol, and α and ß carotene were determined simultaneously with high performance liquid chromatography15; precision within batches being 2.0%, 3.9%, 7.3%, and 6.6%, and between batches 2.9%, 5.5%, 8.2%, and 7.1%, respectively. Accuracy was verified by analysis of quality control serum samples from the US National Bureau of Standards. Plasma vitamin C concentration was determined by fluorometric assay.16 The coefficients of variation within and between batches were 0.7% and 1.6%, respectively.
Statistical analysis–A previous pilot study (n=47) showed an SD of 0.27 μmol/l for ß carotene. To detect a difference of 0.2 μmol/l at 90% power (α=0.05) two groups of 39 subjects were required. For blood analysis, multiple regression was used to examine the difference between the two groups adjusted for age, sex, smoking, body mass index, cholesterol concentration, and the baseline value. For dietary analysis age, sex, and the baseline values were used as covariates. Product moment correlations determined associations between intakes and biochemical indices. Changes in plasma concentrations were calculated as week 8–baseline and changes in dietary intake as week 4–baseline. The adjusted correlations were obtained by computing two sets of residuals: residuals from regressing plasma ß carotene concentration at week 0 on that at week 8 and residuals from regressing serving number at week 0 on that at week 8. The adjusted correlations were then computed as the simple correlations between these two sets of residuals. All values are reported as means (SD) unless otherwise stated.
The mean (range) age of the 87 subjects was 30.2 (18-69) years, mean (range) body mass index was 25.4 (16.5-43.6) kg/m2, and total cholesterol and triglyceride concentrations were 5.03 (2.15-7.84) mmol/l and 1.29 (0.54-2.76) mmol/l. Table 1) shows characteristics of the two groups. Body mass index remained unchanged throughout the study.
Before randomisation participants consumed 2.2 (0.9) servings of fruit and vegetables, weighing 318 (184) g. Differences between control and intervention groups according to unannounced 24 hour recalls (not reported) were similar to those calculated from diet records (table 2). The percentage of energy from total and saturated fat was lower in the intervention group; the percentage of energy from carbohydrate and dietary fibre and concentrations of ß carotene and vitamins C and E were higher (table 3). As antioxidant nutrients have a low correlation with total energy intake (ß carotene r=0.11, vitamin C r=0.12) no adjustment was made for energy.17
At baseline there were no differences between the groups in plasma antioxidant concentrations; striking differences in concentrations of vitamin C and α and ß carotene were, however, apparent throughout the trial (table 4). The control group showed no changes. The concentration of α tocopherol expressed in relation to total cholesterol concentration (μmol α tocopherol/mmol total cholesterol) showed no changes in either group (not reported). Concentrations of lipids and lipoproteins remained unchanged throughout the study (table 5).
In the intervention group, correlations between change in number of servings of fruit and vegetables and change in plasma concentrations of micronutrients were; r=0.09 for ß carotene (P=0.55), r=0.29 for α carotene (P=0.06), and r=0.25 for vitamin C (P=0.11). Change in plasma concentration of vitamin C was correlated with change in juice intake in grams (r=0.37; P<0.05). Change in plasma concentration of α carotene was correlated with change in plasma concentration of ß carotene (r=0.38; P<0.05) and with change in dietary ß carotene during the intervention (r=0.58; P<0.001). Change in dietary ß carotene and dietary vitamin C correlated poorly with corresponding changes in plasma (r=0.19; P=0.23) and (r=0.27; P=0.08), respectively. The correlation between the residuals obtained when the second variable was regressed on the first was r=0.32 (P<0.05) for serving number and plasma ß carotene.
The participants, all of whom ate few fruits and vegetables, seemed to have little difficulty in substantially increasing consumption of these foods by, on average, an extra five servings or 700 g of fruit and vegetables daily. Reported intake of ß carotene and vitamin C increased dramatically. Concomitantly, there were modest increases in consumption of total carbohydrate, dietary fibre, and vitamin E. Small decreases in consumption of total and saturated fat occurred despite the absence of specific advice concerning dietary fat.
The relation between change in plasma α carotene concentration and change in dietary ß carotene is reasonably strong. Unfortunately, data on α carotene in New Zealand foods are not available, but α and ß carotene do coexist in fruit and vegetables.18 Thus for α carotene dietary intake seems to be a major determinant of blood concentrations, whereas for plasma vitamin C concentration, juice intake is important. Although the intervention diet resulted in marked increases in plasma ß carotene and vitamin C concentrations for the intervention group, the correlation between changes in dietary and plasma concentrations of these nutrients is less striking. This finding for vitamin C is similar to that reported in patients with cancer who increased fruit and vegetable consumption.19 A wide variation in response to ingested ß carotene has been reported previously, and the presence of responders and non-responders has been suggested.20 Many dietary factors influence digestion and absorption of carotenes.21 Our data contribute to this discussion, providing some confirmation of individual variation in response and suggesting that measurement of plasma ß carotene concentration may not be a particularly good marker of an individual's daily fruit and vegetable consumption. Nevertheless, it is important to emphasise that even without specific advice to increase fruit and vegetables rich in ß carotene or vitamin C, a general recommendation to increase fruit and vegetable consumption produces marked increases in plasma concentrations of carotenes and vitamin C.
The recommended dietary change resulted in a modest change in dietary vitamin E, but this was insufficient to influence blood concentrations. Similarly, the small reduction in total and saturated fat did not influence plasma concentrations of lipids and lipoproteins.
Reduction in risk of cancer
Our main aim was to determine whether increased fruit and vegetable consumption could produce alterations in concentrations of plasma antioxidants which might be expected to produce clinical benefit. At least 10 prospective studies have shown that high intakes of fruit and vegetables confer protection against cancer, cardiovascular disease, and stroke.3 5 22 23 Several studies have examined the relation between plasma antioxidant concentrations and subsequent risk of disease, most notably for lung cancer for which six out seven prospective studies have reported reduced risk for those in the highest quantile of the distribution of plasma ß carotene concentrations.3 22 Different approaches have been used, and the results of the various studies are not directly comparable with our results.
Nevertheless, it is of interest to note that in the Basle study, which used a nested case-control approach, plasma ß carotene concentrations of patients with lung cancer were comparable with our baseline measurements, whereas those in the Basle control group were remarkably similar to concentrations reached by the participants who increased their consumption of fruit and vegetables.24 Results from the Japan-Hawaii cancer study suggest that the effect of the intervention achieved in our study might reduce the relative risk of lung cancer from 2.4 to 1.2.25 Failure to show a beneficial effect of ß carotene supplementation in several large randomised controlled trials suggests the possibility that ß carotene may not be the key or sole protective factor but rather a marker for one or more closely related factors.8 9 10 Should this be the case, the general dietary recommendation made in this study remains likely to produce considerable benefit in terms of reducing risk of cancer, especially epithelial cancers. Furthermore, on the basis of prospective trials, it would move antioxidant intake of people with low fruit and vegetable consumption into the range of intake associated with reduced risk of disease.1 4 On the basis of recent findings, Levine et al advocate an appreciable increase in the recommended daily allowance for vitamin C from 60 mg to 200 mg, an intake readily achieved by participants in the intervention group.26
Other health benefits
Vitamin E has a powerful effect in reducing oxidation of low density lipoprotein,27 and several cohort studies have shown that high intakes are associated with reduced risk of cardiovascular disease.28 29 30 Nested case-control studies in which concentrations of α tocopherol have been measured in plasma samples have not provided confirmation of the dietary associations23; the Cambridge heart antioxidant study of α tocopherol supplementation in people with atherosclerosis, however, suggested that substantial amounts used in this trial could reduce the risk of non-fatal myocardial infarction.12 While the precise role of α tocopherol in the aetiology and the possible protection against the clinical consequences of coronary artery disease remains to be established, it is clear that a general recommendation to increase fruit and vegetable intake will have little effect on plasma concentrations of α tocopherol, even though the bulk of the dietary intake in the United States is derived from this food group.31 To increase plasma concentrations of α tocopherol through modification of food intake rather than supplements it seems to be necessary to increase substantially consumption of foods rich in vitamin E such as nuts, margarine, and oils. Similarly, a simple recommendation to increase fruit and vegetable intake is unlikely to cause a large enough reduction in saturated fat intake to reduce concentrations of total and low density lipoprotein cholesterol; more specific dietary advice to modify fat intake is required.
It is of interest to note that changes in concentrations of vitamin C and α and ß carotene occurred within the first two weeks of the intervention and remained virtually unchanged for the duration of the trial. This suggests that future studies exploring the relations between dietary intake and plasma concentrations of these antioxidants can be of relatively short duration.
In conclusion, we have shown that people with low consumption of fruit and vegetables can appreciably increase plasma concentrations of α carotene, ß carotene, and vitamin C when they follow recommendations to increase their consumption of fruit and vegetables substantially. On the basis of comparison with epidemiological data, the concentrations of plasma antioxidants achieved might be expected to reduce the risk of cancer. On the other hand, plasma α tocopherol concentration is not increased, and plasma concentrations of lipids and lipoproteins are not altered by this recommendation. More specific dietary advice to modify intake of fat sources is probably necessary to reduce the risk of cardiovascular disease associated with lipoproteins and vitamin E.
We gratefully acknowledge the cooperation of participants in the study and the excellent technical and research assistance of Margaret Waldron, Ashley Duncan, and Steven Tiszavari
Funding: New Zealand Lotteries Grants Board. United Fresh Limited contributed funds towards the provision of foods for the intervention group.
Conflict of interest: None.