Hypertension is a major public health challenge, and it is the most important, modifiable risk factor for CVD incidence and mortality(Reference Kjeldsen1). The global prevalence of hypertension reached 22 % in 2014 and is expected to increase to 29·2 % in 2025 if current trends persist(Reference Kearney, Whelton and Reynolds2). Among US adults, the crude prevalence of hypertension was 45·6 %, and according to the 2017 American College of Cardiology/American Heart Association (ACC/AHA) guidelines, antihypertensive medication was recommended for 36·2 % of the adult population(Reference Muntner, Carey and Gidding3). Lifestyle modification and dietary management can be an effective treatment for high blood pressure (BP), in addition to medication in advanced stages(4).
A high fruit and vegetable (FV) intake has been associated with reduced BP(Reference Wang, Manson and Gaziano5–Reference Oude Griep, Stamler and Chan8) as well as a reduction in the risk of CVD(Reference Scheffers, Boer and Verschuren9,Reference Miller, Mente and Dehghan10) , including CHD(Reference Dauchet, Amouyel and Hercberg11). A number of intervention studies have shown that increased daily intake of FV decreased BP significantly compared with a control diet(Reference Appel, Moore and Obarzanek12,Reference Appel, Champagne and Harsha13) . In contrast, some intervention studies have not demonstrated reductions in BP in response to increased FV intake(Reference Berry, Mulla and Chowienczyk14–Reference McCall, McGartland and McKinley16). These contrasting results may be related to variations in the type of FV consumed within each intervention, but whether the effect of FV on BP depends on the type of FV consumed is largely unknown. For example, FV juices, as a sub-type of FV, are generally thought to have less desirable effects than fresh FV because they contain less fibre, although similar levels of other nutrients, for example, vitamin C(Reference Oude Griep, Stamler and Chan8,Reference Liu, Xing and Chen17) . The comparative effect of juices or whole FV on BP and other CVD outcomes is relatively understudied. There is also debate regarding the effect of cooked vegetables v. raw vegetables on health and the effect of processing/cooking on the nutritional content of FV and its effect on health, including CVD risk factors and hard CVD outcomes(Reference Chan, Stamler and Brown18). For example, cross-sectional results from the International Study of Macro- and Micro-Nutrients (INTERMAP) study showed that both raw and cooked FV were associated with BP, with the association with raw vegetables being somewhat stronger than for cooked vegetables(Reference Chan, Stamler and Brown18).
The present study aimed to determine the relationship between daily portions of FV intake, either considered overall or as specific sub-types of FV (citrus fruit, other fruit, fruit juices, cooked vegetables and raw vegetables) and BP in 50- to 59-year-old men from France and Northern Ireland (NI). The overall hypothesis was that higher overall FV intake would be associated with reduced BP, but that the association may differ according to the type of FV consumed.
Methods
Study population
The PRIME (Prospective Epidemiological Study of Myocardial Infarction) study is a multi-centre, prospective cohort study examining CVD and its associated risk factors in men. The study was initiated from previous collaborative work carried out within the WHO ‘MONICA project’ (Multinational Monitoring of trends and determinants in CVD)(Reference Yarnell19). Sampling procedures, study design and primary endpoints have been described fully elsewhere(Reference Yarnell19). Between 1991 and 1994, 10 600 male participants were recruited, aged between 50 and 59 years, in four different centres (one centre in NI and three in France): Lille (n 2633), Strasbourg (n 2612) and Toulouse (n 2610) in France and Belfast (n 2745) in NI. The sample was recruited to broadly match the social class structure of the population. Written consent was obtained from all participants at baseline, and ethical approval was obtained from the Research Ethics Committee of the Faculty of Medicine, Queen’s University Belfast.
Assessment of exposure measures
Dietary data were collected via a short, self-administered sixteen-item FFQ at baseline, in each participant’s home, and were later checked by an interviewer in the clinic for consistency and missing responses. Participants were asked to indicate their usual frequency of consumption of a standard portion of fruit or vegetable based on the last weeks using the following scale: more than once per d (number per d), daily, three to four times per week, once per week, twice per month, once per month and never. The FFQ measured the frequency of consumption of sixteen food items including overall intake of FV and FV sub-types. For the purposes of the present study, frequencies of intake were converted into portions per day. The FV were then grouped into five separate sub-types (citrus fruit, fruit juice, other fruit, raw vegetables and cooked vegetables), as asked within the FFQ, and also summed to provide a measure of overall/total fruit, vegetable and FV juice (FVJ) intake. Potatoes were not included as they are not considered to be a vegetable in the UK, and they have not been included in previous PRIME analyses(Reference Dauchet, Ferrières and Arveiler20).
Assessment of outcome measures
All participants underwent a clinical examination at baseline to obtain anthropometric measurements including height (to the nearest cm), weight (to the nearest 200 g) and waist and hip circumference (to the nearest 0·5 cm). All measurements were carried out using standardised instruments and procedures. BMI was computed as weight (kg) divided by height-squared (m2). Resting systolic BP (SBP) and diastolic BP (DBP) were measured once at the end of the examination with an automatic device (Spengler SP9; Spengler), by trained staff, after a 5 min rest in the sitting position.
Assessment of other variables
At baseline, participants completed self-administered questionnaires relating to demographic and socio-economic factors(Reference Yarnell19). Participants then attended a clinic, where their questionnaires were checked for completeness. Information was collected on participants’ socio-economic status, psychosocial factors, medication and tobacco use, physical activity level and personal and family medical history(Reference Dauchet, Ferrières and Arveiler20). Socio-economic status was based on a composite score of material conditions in the household based on three proxy indicators (the type of living accommodation (rented or owned/mortgage), the number of cars/vans/motorcycles in the household and the number of baths and/or showers and toilets in the home). Composite scores were categorised into low, medium and high(Reference Wagner, Simon and Evans21).
Lifetime smoking was categorised as never smoked, smoked other than cigarettes, smoked <15 cigarette pack years, smoked ≥15 but <30 cigarette pack years and smoked ≥30 cigarette pack years. Physical activity was recorded in metabolic equivalent scores per week. A CVD screening examination was also conducted at baseline, which included a detailed history of previous CVD and asked participants to report if a doctor had ever identified them as having a given risk factor for CVD and to state any past or current treatment. In addition to this, the London School of Hygiene CVD Questionnaire for Chest Pain on Effort and Possible Infarction was used for each participant(Reference Rose, Blackburn and Gillum22). Participants were also asked to give details on any history of diabetes: diabetes was defined by the current intake of oral hypoglycaemic treatment or use of insulin. Self-reported alcohol consumption was recorded in the form of a daily diary which participants completed over a 7 d period for a number of different types of alcoholic drinks. Alcohol intake was converted into ml per week and subsequently categorised as none, 1–128, 129–265, 266–461 and ≥462 ml/week.
Blood samples were drawn at baseline from each participant after a 12 h fast. Venous blood was collected into EDTA tubes and returned to the local laboratory within 4 h of collection(Reference Yarnell19). Lipids, including total cholesterol and HDL-cholesterol, were analysed immediately, while other samples were aliquoted for long-term storage at −150°C.
Statistical methods
Statistical analysis was performed using SPSS v22.0 (SPSS Inc.). Data were summarised as mean values and standard deviations. To compare categorical and continuous data between countries, χ 2 and independent-samples t tests were used, respectively. Differences in general characteristics across quartile categories of FV intakes were also assessed using descriptive statistical tests. One-way ANOVA with Tukey’s post hoc comparison was used for continuous variables, and χ 2 test was used for categorical variables. Univariate linear and logistic regression models were used to examine associations between BP and overall FV and the five FV sub-types. For linear regression, SBP and DBP were analysed as continuous variables, and FV sub-types were analysed as categorical variables (i.e. per fourth). The following cut-offs (Q1–Q4) were used: citrus fruit ≤0·07, 0·08–0·29, 0·30–0·50 and ≥0·50 portions/d; fruit juice ≤0·00, 0·01–0·07, 0·08–0·50 and ≥0·50; other fruit ≤0·14, 0·15–0·50, 0·51–1·00 and ≥1·01; raw vegetables ≤0·29, 0·30–0·50, 0·51–1·00 and ≥1·01; cooked vegetables ≤0·29, 0·30–0·50, 0·51–1·00 and ≥1·01; FVJ ≤ 1·60, 1·61–2·30, 2·31–3·57 and ≥3·58. For logistic regression, all dependent and independent variables were analysed as categorical variables. For BP, the following cut points were used to define hypertension: SBP ≥ 140 and DBP ≥ 90 mmHg. All regression analyses were adjusted for potential confounding factors which included factors that were associated with SBP, DBP and FV intake in the current analysis and also other commonly known confounders that have been previously highlighted in the literature. Model 1 was unadjusted; model 2 was adjusted for age and country; model 3 was adjusted as for model 2 plus BMI, height, smoking (five categories), physical activity, total cholesterol, HDL-cholesterol, education level (primary, secondary, technical and high), material conditions (low, medium and high) as a measure of socio-economic position, alcohol intake (five categories), diabetes and CHD history. Further sensitivity analyses were also conducted to examine potential intermediary effects of BMI and total cholesterol and HDL. For all analyses, a P value of ≤0·05 was considered statistically significant.
Results
Baseline characteristics of all PRIME participants and for participants in NI and France separately are shown in Table 1. There were significant differences in all baseline characteristics between the two countries, with the exception of SBP. Age, BMI, history of diabetes, smoking (all levels), education level (all levels), material conditions (all levels) and DBP were significantly higher in France, while alcohol intake, physical activity and SBP were significantly higher in NI compared with France. Intakes of total FVJ, citrus fruit, other fruit and raw vegetables were significantly higher in France compared with NI (all P < 0·001), while intakes of fruit juices and cooked vegetables were significantly higher in NI compared with France. Significant positive correlations were observed between intakes of FV across the various sub-groups (all P < 0·001, data not shown).
Table 1. Baseline characteristics of participants in the Prospective Epidemiological Study of Myocardial Infarction (PRIME)
(Mean values and standard deviations; numbers and percentages)*†
FV, fruit and vegetables; FVJ, fruit and vegetable juice.
* Continuous variables are presented as means and standard deviations, while categorical variables are presented as numbers and percentages.
† Differences between countries analysed using independent-samples t test for continuous variables and χ 2 test for categorical variables.
Table 2 shows the characteristics of participants across the quartiles of total FVJ intakes. Results showed a significant difference in age, incidence of diabetes, alcohol intake, physical activity, smoking, education level, material conditions and BP across the quartiles of FVJ intake.
Table 2. Characteristics of participants across quartiles of total fruit, vegetable and juice intake
(Mean values and standard deviations; numbers and percentages)* †
* Continuous variables are presented as means and standard deviations, while categorical variables are presented as numbers and percentages.
† Difference between fourths analysed using one-way ANOVA for continuous variables and χ 2 test for categorical variables.
Table 3 shows the association between SBP and FV intake, both unadjusted and adjusted for confounders. Model 1 showed that FVJ intake was significantly associated with SBP, and this remained significant when adjusted for age and country (model 2), and in the fully adjusted model (model 3). When FV categories were examined separately, increased citrus fruit intake was significantly associated with reduced SBP in both the unadjusted analyses and when adjusted for age and country. However, significance was lost in the fully adjusted model. An association was also evident between increased other fruit and raw vegetable intake and reduced SBP in all models. In contrast, intake of fruit juice and intake of cooked vegetables showed no association with SBP in all models. SBP decreased by 0·46 mmHg as intake of FVJ increased (per fourth), by 0·63 mmHg as intake of other fruit increased (per fourth) and by 1·29 mmHg as intake of raw vegetables increased (per fourth), after adjustment for potential confounders.
Table 3. Linear regression analysis of the association between systolic blood pressure and fruit and vegetable variables in the Prospective Epidemiological Study of Myocardial Infarction (PRIME)
(Odds ratios and 95 % confidence intervals)*
* Values represent mean differences and 95 % CI from reference category (Q1).
† Model 1 unadjusted.
‡ Model 2 adjusted for age and country.
§ Model 3 adjusted for age, country, cholesterol, BMI, height, physical activity, alcohol intake, education level, material conditions, smoking, diabetes and CHD history.
Table 4 shows the association between DBP and FV intake. The unadjusted analysis showed that FVJ intake was significantly associated with DBP, and this remained significant after adjusting for age and country, and also in the fully adjusted model. When FV categories were examined separately, other fruit and raw vegetables were significantly associated with SBP, and this remained significant when adjusted for age and country, and also in the fully adjusted model. Both citrus fruit intake and cooked vegetable intake were associated with reduced DBP in the unadjusted analyses and after adjusting for age and country, but the association became non-significant in the fully adjusted model. In contrast, fruit juice intake was not associated with DBP in all models. DBP decreased significantly, by 0·45 mmHg, as intake of FVJ increased (per fourth), by 0·56 mmHg as intake of other fruit increased (per fourth) and by 1·01 mmHg as intake of raw vegetables increased (per fourth), after adjustment for all confounders.
Table 4. Linear regression analysis of the association between diastolic blood pressure and fruit and vegetable variables in the Prospective Epidemiological Study of Myocardial Infarction (PRIME)
(Odds ratios and 95 % confidence intervals)*
* Values represent mean differences and 95 % CI from reference category (Q1).
† Model 1 unadjusted.
‡ Model 2 adjusted for age, country.
§ Model 3 adjusted for age, country, cholesterol, BMI, height, physical activity, alcohol intake, education level, material conditions, smoking, diabetes and CHD history.
Table 5 shows the association between risk of hypertension (based on SBP > 140 mmHg) and FV intake, both as overall FVJ intake and by separate FV categories. The OR of increased SBP was significantly reduced as FVJ intake increased, and this was significant in both the unadjusted and adjusted analyses. A similar pattern was evident for intake of other fruits and intake of raw vegetables. For citrus fruit, associations were significant in model 1 (unadjusted) and in model 2 (adjusted for age and country), but statistical significance was lost in the fully adjusted model. In contrast, no association was observed between fruit juice or cooked vegetable intake and SBP. The risk of hypertension decreased by 5 % as intake of FVJ increased per fourth and decreased by 7 % as other fruit and 14 % as raw vegetable intake increased per fourth.
Table 5. Logistic regression analysis of the association between hypertension (systolic blood pressure ≥ 140 mmHg) and fruits and vegetable (FV) variables in the Prospective Epidemiological Study of Myocardial Infarction (PRIME)
(Odds ratios and 95 % confidence intervals)*
* Values represent OR and 95 % CI for FV intake in comparison with Q1 (reference category).
† Model 1 unadjusted.
‡ Model 2 adjusted for age and country.
§ Model 3 adjusted for age, country, cholesterol, BMI, height, physical activity, alcohol intake, education level, material conditions, smoking, diabetes and CHD history.
Table 6 shows the association between risk of hypertension (based on DBP ≥ 90 mmHg) and FV intake, both as overall FVJ intake and by separate FV categories. The OR for having increased DBP was significantly reduced as FVJ intake increased, and this was significant in both the unadjusted and adjusted analyses. A similar pattern was evident for intakes of other fruits and intake of raw vegetables. For citrus fruit, significant associations were observed after adjusting for age and country; however, significance was lost in the fully adjusted model. In contrast, no association was observed between fruit juice or cooked vegetable intake and DBP. The risk of increased DBP decreased by 6, 10 and 17 % as intakes of FVJ, other fruit and raw vegetables increased (per fourth), respectively.
Table 6. Logistic regression analysis of the association between hypertension (diastolic blood pressure ≥ 90 mmHg) and fruit and vegetable (FV) variables in the Prospective Epidemiological Study of Myocardial Infarction (PRIME)
(Odds ratios and 95 % confidence intervals)*
* Values represent OR and 95 % CI for FV intake in comparison with Q1 (reference).
† Model 1 unadjusted.
‡ Model 2 adjusted for age and country.
§ Model 3 adjusted for age, country, cholesterol, BMI, height, physical activity, alcohol intake, education level, material conditions, smoking, diabetes and CHD history.
Further exploratory analyses were conducted to examine potential intermediary effects of BMI and also HDL and total cholesterol. The results of this analysis did not alter the findings observed (data not shown).
Discussion
Using data collected from populations in NI and France, SBP and DBP were significantly inversely associated with intake of overall FVJ, but, when considering sub-types, were only associated with other fruit and raw vegetables. Increased intakes of the two sub-types were consistently associated with reduced BP and reduced risk of hypertension. There was no association between increased intake of citrus fruit, fruit juice or cooked vegetables and either SBP, DBP or risk of hypertension. Most previous studies have considered FV together, without taking into consideration the FV sub-types(Reference Appel, Moore and Obarzanek12,Reference McCall, McGartland and McKinley16,Reference Moore, Conlin and Ard23) . Some studies also include juice within the overall FV variable, while others do not(Reference Hills and Armitage24). Unlike other studies, the present study considered intakes of FV sub-types, as well as overall FVJ intake, with adjustment for confounders.
These results are in line with a number of studies that reported an inverse association between overall FV intake and BP(Reference Woodside, Young and McKinley25). A recent meta-analysis of observational studies concluded that increasing FV intake was associated with reduced BP(Reference Li, Li and Wang26). Individual studies, not included in the meta-analyses, also reported similar findings(Reference Ascherio, Hennekens and Willett27–Reference Steffen, Kroenke and Yu29). The beneficial effect of overall FV intake on BP reduction is suggested to be due to the effect of the combination of nutrients and other components found in FV (e.g. fibre, antioxidants, other vitamins and minerals) potentially acting synergistically to improve the vascular phenotype(Reference Appel, Moore and Obarzanek12). Determining the effect of any single nutrient within FV over other nutrients is very difficult, but examining the effect of overall FV as a food group may be more reflective and relevant to our habitual diet(Reference Appel, Moore and Obarzanek12). Our results support the notion that total FVJ intake has a protective effect against hypertension, likely due to the effect of the combination of many nutrients such as phytochemicals, vitamins and minerals(Reference Srinath Reddy and Katan30).
Analysis of FV sub-types indicated that other fruit (not citrus and not fruit juice) and raw vegetables were significantly associated with BP. In contrast, citrus fruit, fruit juice and cooked vegetables were not associated with BP outcomes. There is limited research on the association between sub-types of FV such as citrus fruit and BP. Our results are consistent with other studies which reported no association between fruit juice and BP(Reference Kivimäki, Siltari and Ehlers31–Reference Sahebkar, Ferri and Giorgini33). As suggested by previous studies, this lack of association is possibly due to the low fibre and high sugar (either as added sugar or fructose) content(Reference Oude Griep, Stamler and Chan8,Reference Chan, Stamler and Brown18) which have been associated with high BP(Reference Pase, Grima and Cockerell34). Conflicting evidence, however, comes from a number of previous studies that examined the effect of a single type of fruit juice on BP, with some reporting positive effects of specific types of fruit juices, for example, cherry juice, berry juice and pomegranate juice on BP(Reference Keane, George and Constantinou32,Reference Seshadri, Beiser and Kelly-Hayes35,Reference Tjelle, Holtung and Bøhn36) . It is possible that the specific type of fruit juice may be important in terms of its effect on BP and may relate to the presence of other bioactive compounds(Reference Sahebkar, Ferri and Giorgini33), or processing conditions(Reference Wallace, McEvoy and Hunter37).
Interestingly, our findings showed a negative association between other fruit (not citrus and not fruit juice) and BP outcomes. Some studies have found inverse associations between single types of fruit, such as apple and tart cherry, and BP(Reference Oude Griep, Stamler and Chan8,Reference Keane, George and Constantinou32) . For example, Keane et al. concluded that Montmorency tart cherry intake acutely reduced SBP in men with early hypertension(Reference Keane, George and Constantinou32), while Oude Griep et al. reported a positive relationship between DBP and apple intake in East Asian consumers, although this was not found in other countries(Reference Oude Griep, Stamler and Chan8).
Finally, when comparing the association of cooked vegetables v. raw vegetables on BP, our results indicated a significant association between raw vegetable intake and BP, but no significant association between cooked vegetables and BP, after adjusting for confounding factors. The lack of association between cooked vegetables and BP in our study may be explained by the effect of the cooking method on the nutritional value of the vegetables. Similar findings were reported in the cross-sectional, US-based INTERMAP study which was conducted in 2195 males and females aged 40–59 years. In the present study, an inverse association was noted between both raw vegetable intake and BP, and cooked vegetable intake and BP, but the association was stronger for raw vegetables(Reference McCall, McGartland and McKinley16). The results of the present study were potentially explained by the effect of cooking, which could significantly change the chemical composition of vegetables and influence the concentration and bioavailability of bioactive compounds, such as antioxidant, water-soluble and heat-sensitive nutrients(Reference Miglio, Chiavaro and Visconti38). The effect also depends on cooking conditions (such as cooking duration and method) and morphological and nutritional characteristics of vegetable species, in addition to the interaction with other dietary factors that can affect nutrient absorption(Reference Miglio, Chiavaro and Visconti38,Reference Jiménez-Monreal, García-Diz and Martínez-Tomé39) .
Study strengths and limitations
Strengths of the present study are that it considered the sub-type of FV as well as overall FV intake. The analysis was also carried out, using the same methodology, on pooled data collected in France and NI, two countries with significant differences in lifestyle behaviours. In addition, unlike other studies, the PRIME study sample was large and included a wide range of confounding factors.
Limitations of the current analysis include the specific age group and sex of the population (males, aged 50–59 years); therefore, it is difficult to generalise the findings to women or younger age groups. In addition, assessment of lifestyle behaviours relied on self-report rather than objective measures. The use of a non-validated FFQ to assess dietary intake is also a limitation. Although widely used in epidemiological studies, FFQ is prone to recall bias, thus limiting their accuracy in assessing dietary intake. In the present study, a short sixteen-item FFQ was used to assess dietary intake and the number of specific types of FV and fruit juices; therefore, further exploration of the association between further sub-types of FV and BP was not possible. Similarly, detailed information about vegetable cooking methods and processing were not available. However, previous results by Dauchet et al. demonstrated that this FFQ was suitable for discriminating between low and high consumers of FV. In their analysis, they noted strong positive correlations between the self-reported FV intakes from the FFQ and biomarkers of FV status, specifically B-cryptoxanthin, vitamin C and α- and β-carotene(Reference Dauchet, Ferrières and Arveiler20). Furthermore, a previous study reported that dietary questionnaires with restricted number of items do not overly affect the ability to rank individuals according to their FV intake(Reference Thompson, Subar and Smith40). However, we cannot rule out the possibility that the observed associations between FV intakes may be explained by compensatory changes in other food intakes that we were unable to explore. The retrospective nature of the FFQ is also a limiting factor in that it only reflected food consumption over the previous 7 d period and therefore did not capture potential seasonal variation in food intake. Although our analyses were adjusted for country, there may have also been differences between the two countries in terms of overall dietary pattern. Indeed, a further analysis showed that when the two countries were analysed separately, the findings became stronger for France, while those for NI became attenuated, however, for NI, this may simply have been due to lack of statistical power. The assessment of BP also had limitations in that only one BP measurement was performed, and therefore, results should be viewed with caution. Given the high variability of general BP measures, the use of one BP measurement limits identification of cases of hypertension and, in particular, limits continuous analyses with BP. In addition, lifestyle behaviours were only assessed at one time point, and data collection for the present study took place from 1991 to 1993; therefore, we cannot rule out the possibility of change in lifestyle behaviours, including dietary habits and food products consumed over time. The cross-sectional design of the study is also a key limitation in that both the exposure and outcome measures were simultaneously assessed, thus ruling out evidence of a temporal relationship. The cross-sectional design raises the issue of reverse causality. Indeed, the associations observed in the present study do by no means indicate, nor prove, that FV reduce BP or hypertension. Reverse causality may weaken any true association between FV intake and BP. Without longitudinal data, it is not possible to establish a true cause and effect relationship. It is also possible that other confounding factors not accounted for in the present study may be masking the true effect of FV intake on BP. Finally, while the results are interesting, the lack of validation of the FFQ together with the limited assessment of BP means that the overall results need to be interpreted with caution. This cohort will have included men at baseline who had been diagnosed with hypertension and were being managed by antihypertensive medication. These participants may have been classified as non-hypertensive, but that will have been due to the prescribing of antihypertensive medication and the control of their BP. In the present study, we were unable to adjust for use of BP medication due to the lack of availability of antihypertensive medication data for the French cohort. However, re-analysis of the Belfast cohort with exclusion of those who reported use of antihypertensive medication at baseline did not alter findings (data not shown).
Conclusion
In conclusion, after adjusting for potential confounding factors, the results of cross-sectional analysis suggested that overall FVJ intake may be associated with reduced BP and reduced risk of hypertension. When FV were analysed separately, the association with BP depended on the FV sub-type, with other fruit and raw vegetable intake being inversely associated with BP, but not fruit juice, citrus fruit or cooked vegetables. These results suggest that the strength of the association between FV sub-types and BP might be related to the type of FV consumed, or to processing or cooking-related factors. A more defined classification of FV consumed during dietary data collection may provide more valuable information when studying associations with health outcomes. Further intervention studies to examine the dose–response effects of specific FV on BP are recommended, with a consideration of the possible effect of factors, such as storage, processing and cooking, that will impact on overall nutrient profile.
Acknowledgements
The PRIME Study is organised under an agreement between INSERM and the Merck, Sharpe and Dohme-Chibret Laboratory, with the following participating laboratories: The Strasbourg MONICA Project, Strasbourg, France (M. M., D. Arveiler, B. Haas); The Toulouse MONICA Project, INSERM U558, Toulouse, France (J. F., J. B. Ruidavets); The Lille MONICA Project, INSERM U744, Lille, France (P. A., M. Montaye); The Department of Epidemiology and Public Health, Queen’s University Belfast, NI (A. Evans, J. Yarnell, F. K.); The Department of Atherosclerosis, INSERM U545, Lille, France (G. Luc, J. M. Bard); The Laboratory of Hematology, La Timone Hospital, Marseille, France (I. Juhan-Vague); The Laboratory of Endocrinology, INSERM U326, Toulouse, France (B. Perret); The Vitamin Research Unit, The University of Bern, Bern, Switzerland (F. Gey); The Trace Element Laboratory, Department of Medicine, Queen’s University Belfast, NI (J. V. W., I. Young); The DNA Bank, INSERM U525, Paris, France (F. Cambien); The Coordinating Center, INSERM U909, Villejuif, France (P. Ducimetière, A. Bingham).
The present analysis was financially supported by University of Petra, Jordan. info@uop.edu.jo
The authors’ responsibilities were as follows: N. A. E. prepared the initial draft of the manuscript and carried out the statistical analysis; C. C. P. directed the statistical analysis; C. E. N. assisted with statistical analysis and drafting the manuscript; F. K. and G. J. L. were co-investigators of the Belfast centre of the PRIME study; M. M. and K. B. were co-investigators of the Strasbourg centre of the PRIME study; P. A. and J. D. were co-investigators of the Lille centre of the PRIME study; V. B. and J. F. were co-investigators of the Toulouse centre of the PRIME study. All of the authors contributed to the drafts, revisions and proof reading of the manuscript. J. V. W. had the primary responsibility for the final content.
The authors declare there are no conflicts of interest.
