ORIGINAL ARTICLE

Heart Vessels (2008) 23:96–103 © Springer 2008DOI 10.1007/s00380-007-1018-5

C. Pitsavos · C. Chrysohoou · Y. Lentzas · C. StefanadisFirst Cardiology Clinic, School of Medicine, University of Athens, Athens, Greece

D.B. Panagiotakos1 (*)Offi ce of Biostatistics and Epidemiology, Department of Nutrition-Dietetics, Harokopio University, Athens, Greece

Correspondence address:1 46 Paleon Polemiston Street, 166 74 Attica, GreeceTel. +30-210-960-3116; Fax +30-210-960-0719e-mail: d.b.panagiotakos@usa.net

Christos Pitsavos · Christina Chrysohoou Demosthenes B. Panagiotakos · Yannis Lentzas Christodoulos Stefanadis

Abdominal obesity and infl ammation predicts hypertension among prehypertensive men and women: the ATTICA Study

Abstract The aim of this work was to assess the 5-year incidence of hypertension and its predictors among prehy-pertensive adults. Under the context of the ATTICA Study, data from 1188 individuals, free of cardiovascular disease, but with defi ned high blood pressure levels (prehyperten-sion) at baseline examination (during 2001–2002) were retrieved. In 2006, the 5-year follow-up of the study was performed, and 798 of the prehypertensive participants were allocated. In this work, incidence and determinants of developing hypertension were evaluated. The 5-year age-adjusted incidence of hypertension was 18.7% in men and 24.6% in women (P = 0.05); while almost one half of pre-hypertensive individuals at the age of 55–65 years devel-oped hypertension, and approximately 6 out of 10 people over 65 years of age developed the disease. Multiple logistic regression analysis revealed that increased age (odds ratio [OR] per 1 year = 1.09, 95% confi dence interval [CI] 1.07–1.12), male sex (OR = 0.40, 95% CI 0.21–0.68), high educa-tion status (OR per 1 year of school = 0.94, 95% CI 0.88–0.98), waist circumference (OR per 1 cm = 1.04, 95% CI 1.02–1.06) and C-reactive protein (OR per 1 mg/l = 1.12, 95% CI 1.05–1.20), were positively associated with the development of hypertension. Moreover, greater adherence to Mediterranean diet seems to protect only prehyperten-sive, with abdominal obesity patients prone to develop hypertension (OR = 0.94, 95% CI 0.90–0.98). Annual inci-dence of hypertension was roughly 4% in men and women.

Older people, with low education, abdominal obesity, lower adherence to the Mediterranean diet, and increased infl am-mation, constitute a model of prehypertensive individuals that are prone to develop hypertension.

Key words Infl ammation · Abdominal obesity · Hyperten-sion · Incidence · Diet

Introduction

The level of arterial blood pressure has long been recognized as a major risk factor for several common car-diovascular diseases, including coronary heart disease, cere-brovascular disease and heart failure.1–3 The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure4 suggested a new classifi cation for borderline blood pressure levels, namely “prehypertension.” The new classifi cation describes people with blood pressures between 120 and 139 mmHg systolic or 80–89 mmHg diastolic. This “new” category between normal blood pressure and established hypertension includes a population at high risk for deve-loping hypertension and in which lifestyle modifi cations are needed. In our populations prehypertension seems to prevail more in men than in women (43% versus 35%).5

Recently, the Trial of Preventing Hypertension (TROPHY) evaluated the effi cacy of antihypertensive therapy in persons classifi ed as “prehypertensive” and revealed a cumulative incidence of hypertension of 63% among untreated prehypertensives during the four years of follow-up.6 Data regarding the incidence of hypertension among prehypertensive people, as well as factors that asso-ciated with the development of the disease in people of the Mediterranean region, are lacking in the literature. Under the context of the ATTICA Study,7 the main hypothesis tested in this work was the 5-year incidence of hypertension and its determinants among prehypertensive, cardiovascu-lar disease-free adults.

Received: March 31, 2007 / Accepted: September 12, 2007

97

Materials and methods

Study design

The ATTICA epidemiological study is a health and nutri-tion survey that has been carried out in a province of the Attica region (including 78% urban and 22% rural areas), during 2001–2002. People with cardiovascular disease or people living in institutions were excluded from the sam-pling. According to the study’s design, 4056 inhabitants from the above area were randomly selected to enroll in the study. Of them, 3042 agreed to participate; 1514 of the par-ticipants were men and 1528 were women. All participants were interviewed by trained personnel (cardiologists, dieti-cians and nurses) who used a standard questionnaire.

Baseline measurements

The baseline evaluation included: demographic characteris-tics (age, sex, mean annual income, and years of school as proxy of social status), detailed information regarding per-sonal and family history of hypertension, hypercholesterol-emia, and diabetes, dietary and other lifestyle habits, such as smoking and habitual/leisure time physical activity. In particular, the evaluation of nutritional habits was based on a validated semiquantitative food-frequency questionnaire.8 In particular, all participants were asked to report the average intake (per week or day) of several food items that they consumed (during the last 12 months). Alcohol con-sumption was measured in wineglasses (100 ml) and quanti-fi ed by ethanol intake (grams per drink). One wineglass was considered to have an equivalent of 12% ethanol concentra-tion. To describe overall diet composite scores were used, which are necessary for the evaluation of epidemiological associations. Thus, a special Mediterranean diet score was used (range 0–55) that is based on the rationale of the Mediterranean dietary pyramid.9,10 Smokers were defi ned as those who were smoking at least one cigarette per day during the past year or had recently stopped smoking (within a year); the rest of the participants were defi ned as nonsmokers. For the ascertainment of physical activity status the International Physical Activity Questionnaire was used (IPAQ11), as an index of weekly energy expenditure using frequency (times per week), duration (in minutes per time) and intensity of sports or other habits related to physi-cal activity (in expended calories per time). Participants who did not report any physical activities were defi ned as physically inactive (sedentary lifestyle). Body mass index (BMI) was measured as weight (in kilograms) divided by standing height (in meters squared). Obesity was defi ned as BMI > 29.9 kg/m2. Waist (in cm) and hip (in cm) circumfer-ences were also measured using standard procedures. Arte-rial blood pressure was measured at the end of the physical examination with subject in sitting position. All participants were at least 30 min at rest. Participants who had mean sys-tolic/diastolic blood pressures within the range of 120–139/80–89 mmHg and have never been told that they have high blood pressure levels were defi ned as prehypertensives

as has recently been suggested by the American Heart Association.4 Hypercholesterolemia was defi ned as total cholesterol levels greater than 220 mg/dl, or the use of lipid-lowering agents and diabetes mellitus as a fasting blood sugar >125 mg/dl, or the use of antidiabetic medication. High-sensitivity C-reactive protein (as marker of low-grade systemic infl ammation) was assayed by particle-enhanced immunonephelometry (N Latex, Dade-Behring Marburg GmbH, Marburg, Germany) with a range from 0.175 to 1100 mg/dl. The intra-assay and interassay coeffi cient of variation was <5%.

Follow-up

From June to August 2006, the ATTICA Study investigators performed the 5-year follow-up on 2101 participants (941 of the 3042 [31%] participants were lost to follow-up). Of the participants that did not participate in the re-examination, 21% were not found because of missing or wrong addresses and telephone numbers, and 12% refused re-examination. No signifi cant differences were observed between those who were lost to follow-up and the rest of the participants regarding sex (P = 0.89), age (P = 0.92), education level (P = 0.69), as well as presence of hyperten-sion (P = 0.45), diabetes (P = 0.34), and hypercholesterol-emia (P = 0.09). No differences were observed between those contacted by telephone and those participated in the face-to-face interviews, regarding age (45 ± 11 vs 45 ± 16 years, P = 0.77), sex (males: 55% vs 54%, P = 0.89) and place of residence (P = 0.98). Follow-up examination included detailed information from participants’ medical records, as retrieved by the study’s investigators. In some cases (15%) where individuals had no accurate records, a face-to-face interview was performed.

For testing the research hypothesis of this work, people who have been defi ned as having hypertension or were normotensive at baseline examination were excluded from the analysis. Therefore, 1188 individuals with defi ned prehy-pertension at baseline examination (in 2001) were the initial working sample of this work. Of them, 426 men and 356 women (i.e., n = 782) were found alive at the time of the follow-up, while 8 men and 8 women died due to cardiovas-cular disease (n = 9), cancer (n = 5), or because of other causes (67% participation rate in the follow-up). Thus, the remaining 782 prehypertensive individuals comprise the analytic sample of the present work.

The re-examination included information about: (a) vital status (death from any cause or due to cardiovascular disease), or development of coronary heart disease (includ-ing myocardial infarction, angina pectoris, other identifi ed forms of ischemia (WHO-ICD coding 410–414.9, 427.2, 427.6), heart failure of different types, and chronic arrhyth-mias (WHO-ICD coding 400.0–404.9, 427.0–427.5, 427.9), or development of stroke (WHO-ICD coding 430–438), (b) development of hypertension was defi ned as systolic/dia-stolic blood pressure >140 and/or 90 mmHg (average of three consecutive measurements), or use of special medica-tion, hypercholesterolemia, and diabetes, among people who did not have these disorders at baseline, as well as

98

management of these conditions, (c) assessment of body weight and height, and (d) lifestyle habits, including physical activity and smoking status, as well as various food groups intake.

Statistical analysis

Continuous variables are presented as mean values ± stan-dard deviation. Categorical variables are presented as absolute and relative frequencies. Associations between categorical variables were tested using the chi-squared test. Comparisons of mean values of normally distributed vari-ables between those who developed hypertension and the rest of the participants were performed using the Student t-test. Since the exact time to event (i.e., development hypertension) was not available, the relative risks of devel-oping hypertension during the 5-year follow-up period, according to the participants’ baseline characteristics, were estimated using the odds ratios (OR) and their correspond-ing 95% confi dence intervals (CIs) through stepwise multi-ple logistic regression analysis. The 5% as the cutoff for the probability of entering a variable in the model and 10% as the threshold for the probability of removing a variable from the model were used. The probabilities were calcu-lated using the Wald test. Interactions between sex and other covariates were tested in all steps, and when they were signifi cant remained in the fi nal model. Deviance residuals were used to evaluate model’s goodness-of-fi t, and −2log-likelihood of the initial and fi nal model were also calculated. Finally, to further explore which of the baseline measure-

ments was the best predictor, a discriminant analysis was performed using Wilks’ lambda (lower values of the statis-tic, better discriminating ability between hypertensive and normotensive subjects). The population attributable risk (PAR) was calculated as the ratio (1 − OR)/OR. All reported P-values are based on two-sided tests and compared to a signifi cance level of 5%. SPSS version 14 (Statistical Package for Social Sciences, SPSS, Chicago, IL, USA) software was used for all the statistical calculations.

Results

During the 5-year period, 75 (18.7%) men and 85 (24.6%) women with defi ned prehypertension at baseline examina-tion (in 2001) were diagnosed as having hypertension. Fur-thermore, it could be estimated that the annual age-adjusted incidence rate is 3.74 per 100 men and 4.92 per 100 women (if the annual incidence rate ratio is considered constant). The men-to-women incidence ratio was approximately 0.8-to-1, across all age groups (P for gender differences = 0.05). A strong linear trend was observed across all age groups, irrespective of sex (P for trend <0.001). In particular, it was found that the unadjusted risk of hypertension increases by 10.7% per 1-year increase in age, among men, and by 10.2% per 1-year increase in age, among women. Furthermore, almost one half of prehypertensive individuals at age of 55–65 years developed hypertension, while approximately 6 out of 10 people over 65 years of age developed the disease (Table 1). The distribution of the baseline demographic

Table 1. Distribution of baseline characteristics among prehypertensive people who developed and in to those who did not develop hypertension within the 5-year follow-up period

Results at 5-year follow-up:

Nonhypertensive (n = 622) Hypertensive (n = 160) P

Age group (years), % <35 97 3 35–45 90 10 45–55 75 25 55–65 54 46 >65 38 62Age (years) 42 ± 11 55 ± 11 <0.001Male sex, % 56 47 0.05Years of school 12.6 ± 3.5 10.4 ± 4.0 <0.001Smoking, % 42 38 0.43Mediterranean diet score (0–55) 26.4 ± 5.5 23.5 ± 3.8 <0.001Alcohol consumption, g/day 9.2 ± 13.4 6.7 ± 9.8 0.27Physical activity, % 44 39 0.34Waist circumference, cm 89 ± 14 96 ± 12 <0.001Abnormal waist (men/women: <102/88 cm), % 51 76% <0.001Waist-to-hip ratio 0.85 ± 0.1 0.89 ± 0.09 <0.001Body mass index, kg/m2 26.1 ± 3.8 28.2 ± 4.8 <0.001Obesity, % 14 28 <0.001Total serum cholesterol, mg/dl 193 ± 40 203 ± 43 0.01Hypercholesterolemia, % 39 51 0.007Fasting blood glucose, mg/dl 91 ± 20 96 ± 22 0.01Diabetes mellitus, % 3 16 0.001Family history of hypertension, % 35 38 0.51C-reactive protein, mg/l 1.8 ± 2.3 2.5 ± 2.7 0.003Baseline systolic blood pressure, mmHg 120 ± 8 124 ± 7 0.001Baseline diastolic blood pressure, mmHg 78 ± 4 79 ± 5 0.64

99

parameters and cardiovascular risk factors levels, into those who were diagnosed as hypertensive and the rest of the participants, are presented in Table 1. People who devel-oped hypertension within the 5-year period, were at older age at baseline examination, less educated, less likely to adhere to a Mediterranean type of diet (i.e., lower Mediter-ranean diet score), had higher anthropometric indices, were 2-times more likely to be obese, had higher total cholesterol and blood glucose levels, were almost 5-times more likely to have diabetes, and had approximately 45% higher C-reactive protein levels, as compared to those who did not develop hypertension.

However, although several differences were observed between those who developed hypertension and those who did not, regarding their baseline characteristics (Table 1), residual confounding may exist. Thus, multiple logistic regression analysis revealed that only age, sex, waist circum-ference, education status, baseline blood pressure levels, and C-reactive protein were the baseline factors that predicted the development of hypertension (see fi nal model, Table 2). In particular, an increase in waist circumference was associ-ated with higher risk of hypertension, while 10 cm difference in baseline hip circumference levels was associated with 54% (95% CI 28–87%) higher risk of developing the disease. Abdominal obesity (i.e., >102 cm for men and 88 cm for women) was associated with 2.09-times (95% CI 1.31–3.23) higher risk of hypertension, irrespective of sex and other potential confounders presented in Table 2. Similarly, an increase in baseline measurements of BMI was associated with 13% higher risk of hypertension (95% CI 1.09–1.20). Further analysis showed that waist circumference was the best predictor for hypertension among all anthropometric measurements since it showed better goodness-of-fi t of the estimated models that also included age, sex, physical activ-ity status, baseline blood pressure levels, and dietary habits (i.e., −2log-likelihood for BMI, waist and hip circumferences

were 585, 580, and 589, respectively). No difference in the predictive ability of anthropometric measurements was observed between sexes (P for interaction with sex >0.7). Furthermore, high education status was associated with lower likelihood of hypertension (see fi nal model in Table 2). Finally, high-sensitivity C-reactive protein levels pre-dicted future development of hypertension, since an increase in baseline levels was associated with higher risk of hyper-tension (Table 2). All the aforementioned relationships were independent from age, baseline systolic blood pressure, sex, total cholesterol and blood glucose levels, dietary habits, physical activity, and smoking status of the participants.

It is of interest that the effect of C-reactive protein on incidence of hypertension was affected by obesity status. Particularly, C-reactive protein levels were a highly signifi -cant predictor of hypertension among participants of normal weight (odds radio = 1.20, 95% CI 1.03–1.40), while C-reactive protein lost its signifi cance among the overweight (odds radio = 1.02, 95% CI 0.89–1.15) or obese (odds radio = 1.04, 95% CI 0.88–1.20), after adjusting for the aforemen-tioned covariates. Moreover, adherence to Mediterranean diet seems to protect prehypertensive people from develop-ing hypertension (see initial model, Table 2); however, this association is confounded by waist circumference levels. Specifi cally, among people with normal waist circumference adherence to this traditional diet does not seem to confer any protection (odds ratio = 0.97, 95% CI 0.87–1.06), while among people with abdominal obesity, greater adherence to the Mediterranean diet was associated with lower likeli-hood of developing hypertension (odds ratio = 0.94, 95% CI 0.90–0.98).

Furthermore, discriminant analysis revealed that age (Wilk’s lambda = 0.80), followed by years of school (Wilk’s lambda = 0.93), waist (Wilk’s lambda = 0.95), BMI (Wilk’s lambda = 0.95), level of adherence to Mediterranean diet

Table 2. Results from multiple logistic regression analysis that evaluated the association between baseline characteristics of the ATTICA Study participants and 5-year incidence of hypertension

Odds ratio 95% confi dence interval

Initial modelAge (per 1 year) 1.11 1.09–1.14Male vs female sex 0.51 0.29–0.86Smoking (yes/no) 1.26 0.77–1.96Physical activity (yes/no) 0.93 0.65–1.30Waist circumference (per 1 cm) 1.02 0.99–1.04Mediterranean diet score (per 1 unit) 0.94 0.90–0.99Education status (per 1 year of school) 0.94 0.89–0.99Blood glucose (per 1 mg/dl) 0.99 0.98–1.01Systolic blood pressure (per 1 mmHg) 1.02 0.99–1.06Total serum cholesterol (per 1 mg/dl) 0.99 0.98–1.00C-reactive protein (per 1 mg/l) 1.05 1.01–1.11

Final modelAge (per 1 year) 1.09 1.07–1.12Male vs female sex 0.40 0.21–0.68Education status (per 1 year of school) 0.94 0.88–0.98Waist circumference (per 1 cm) 1.04 1.02–1.06C-reactive protein (per 1 mg/l) 1.12 1.05–1.20Systolic blood pressure (per 1 mmHg) 1.03 1.00–1.06

100

(Wilk’s lambda = 0.96), waist-to-hip ratio (Wilk’s lambda = 0.97) and C-reactive protein levels (Wilk’s lambda = 0.98), were the most signifi cant predictors of hypertension. However, almost all values of Wilk’s lambda were close to 1, indicating moderate predictive ability. Additionally, cutoff point sensitivity analysis showed that the sensitivity for age greater than 45 years was 81%, the sensitivity for waist greater than 102/88 cm (men/women) was 75%, the sensitiv-ity for diet score lower than 27 was 64%, and the sensitivity of C-reactive protein greater than 0.95 mg/l was 63%.

Then an index, i.e., risk factor cumulative score, was developed based on age >45 years, waist >102/88 cm (men/women), diet score <26, and C-reactive protein >0.95 mg/l. Each individual who fulfi lled one or more of the aforemen-tioned criteria took a score from 1 to 4, or otherwise took a score of 0 (reference category). Logistic regression analy-sis revealed that 1 unit increase in the score (i.e., one additional risk factor) was associated with 45% (95% CI 12%–90%) higher risk of developing hypertension during the 5-year follow-up period, irrespective of various poten-tial confounders mentioned above. The PAR of this score per 1 unit increase was calculated 31%, indicating that 31% of the new hypertension cases could be avoided if one of the components of the index was not present.

Discussion

In this work the 5-year incidence of hypertension among prehypertensive individuals was evaluated, in a population-based sample of adult men and women from Greece. A better understanding of factors that predict the progression from prehypertension to hypertension has clinical and public health importance because it helps guide appropriate decision-making on treatment and prevention. During the 5-year period 18.7% men and 24.6% women with defi ned prehypertension at baseline examination were diagnosed as having hypertension. Furthermore, it could be estimated that the annual age-adjusted incidence rate is 3.74 per 100 men and 4.92 per 100 women. On the other hand, in the whole population of the study 14.3% men and 13.4% women developed hypertension in the aforementioned period.12 Additionally, almost one half of prehypertensive individuals at age of 55–65 years developed hypertension, while approx-imately 6 out of 10 people over 65 years old developed the disease. Moreover, a strong linear trend regarding the incidence rate was observed across all age groups. Multi-adjusted analysis revealed that increased age (i.e., >45 years), abnormal waist, low education level, low adherence to the Mediterranean diet, and high C-reactive protein (>0.95 mg/l) levels constitute a profi le of prehypertensive people that are prone to develop hypertension within a 5-year period.

Hypertension incidence is found to be higher in prehy-pertensives compared to normotensives in our cohort, in accordance with the fi ndings of other studies.16,18,19 Further-more, prehypertension is associated with increased risk for cardiovascular morbidity and mortality13 and has been the

focus of recent investigations that aim to determine the risk factors associated with the occurrence of hypertension among prehypertensive individuals.

Aging and hypertension

Systolic blood pressure levels increase with age and the extent of this age-related increase depends on the initial value of blood pressure levels. The increase in blood pres-sure with age is mostly associated with structural changes in the arteries and especially with large artery stiffness that causes reduced aortic elasticity and systolic hypertension, particularly in older people.14 This hemodynamic pattern results from mechanical factors and other pressure-independent risk factors, such as diabetes mellitus, renal failure, obesity, and severe atherosclerosis that predomi-nate with advanced age.15 Framingham Heart Study inves-tigators have reported that patients with a normal or high normal diastolic blood pressure are two to three times more likely to progress to hypertension than individuals with optimum diastolic blood pressure.16 In our study we note that older individuals with prehypertension were more likely to progress to hypertension than younger people. This is in accordance with the fi ndings of Vasan et al., who revealed age as a predictor factor for the occurrence of hypertension over 4 years of follow-up.17 In our study the unadjusted risk of hypertension increases by 10.7% per 1-year increase in age, among men, and by 10.2% per 1-year increase in age, among women. Furthermore, almost one half of prehypertensive individuals at age of 55–65 years developed hypertension, while approximately 6 out of 10 people over 65 years old developed the disease.

Anthropometric indices and hypertension, on the role of abnormal waist

Body mass index at baseline and weight gain on follow-up were important determinants of future hypertension. These fi ndings add to the substantial data linking obesity and weight gain to the future risk of hypertension.18,19 Other studies have assessed the relationship between BMI and incidence of hypertension. Among 1427 middle-aged Japa-nese men and women, BMI values greater than 27.0 kg/m2 were associated with approximately two times greater risk of hypertension incidence over a 4-year follow-up period.19,20 However, three studies – the Atherosclerosis Risk in Com-munities Study (ARIC) (after 3 years of follow-up), the Monitoring Trends and Determinants of Cardiovascular Disease Study (POL-MONICA) (after 5 years of follow-up), and a male occupational Japanese cohort (after 6 years of follow-up) – revealed that baseline BMI was unrelated to hypertension incidence after multivariable adjustment.21 It is unclear whether these discrepant fi ndings result from differential follow-up times, or differences in multivariable adjustment. In obese subjects increased renal sodium reab-sorption and blood volume expansion are central features in the development of hypertension. Overweight is also

101

associated with increased sympathetic activity, while leptin, a protein expressed in and secreted by adipocytes, is the main factor linking central obesity, increased sympathe-tic nervous system activity, and hypertension. The rennin-angiotensin-aldosterone system has also been caus-ally implicated, as angiotensinogen is expressed in and secreted by adipose tissue; while hypoadiponectinemia, high circulating levels of free fatty acids, and increased vas-cular production of endothelin-1 have been reported as potential mechanisms for increased prevalence of hyperten-sion.22 In our study waist circumference was among the most signifi cant predictors of hypertension occurrence in prehy-pertensive individuals. This fi nding serves to emphasizes the importance of controlling central obesity in the primary prevention of hypertension.

Education status: prehypertension

The educational status has been recognized to have a sig-nifi cant role in the development and progression of various diseases, especially coronary heart disease. Furthermore, numerous studies fi nd, however, that outcomes can be improved with targeted patient education and improved physician communication skills that take into account patients’ health literacy levels.23 Gu et al.,24 in a mean of 8.2 years of follow-up of 10 525 Chinese adults free from hyper-tension at baseline, revealed that those living in urban regions had a signifi cantly increased risk for developing hypertension compared to those living in rural regions. Fur-thermore, Wang et al.25 revealed education as an important risk factor for the prevalence of hypertension and elevated blood pressure in Chinese men and women. In our study high education status was associated with lower likelihood of hypertension, since 1 school-year difference at baseline education was associated with 6% lower odds of developing the disease. This fi nding is in accordance with the increasing need for primary prevention programs to target persons of lower educational status.

Mediterranean type of diet and development of hypertension in prehypertensives

An extensive body of scientifi c evidence relates diet and incidence of various types of cardiovascular diseases.26 Dietary factors exert their infl uence largely through their effects on blood pressure, lipids, and lipoproteins. In particu-lar, a diet rich in fruits, vegetables, legumes, and whole grains, fi sh, nuts, and low-fat dairy products, like the Medi-terranean type of diet, seems to protect against the develop-ment and progression of cardiovascular disease.27,28 In our study, although adherence to Mediterranean diet seems to protect prehypertensive people from developing hyperten-sion, this association is confounded by waist circumference levels. Specifi cally, among people with normal waist circum-ference adherence to this traditional diet does not seem to confer any protection, while among people with abdominal obesity greater adherence to the Mediterranean diet was

associated with a lower likelihood of developing hyperten-sion. Central obesity, which refl ects disproportionate accu-mulation of visceral fat mass, has been associated with a chronic infl ammation state.29 On the other hand, the Medi-terranean type of diet has been independently associated with a reduction in various infl ammatory and coagulation markers, related to cardiovascular disease.30 Probably this benefi cial effect of diet on infl ammation state, which has been related with prehypertension,5 is more prominent in those individuals with central obesity. Furthermore, adher-ence to the Mediterranean diet seems to confer a benefi cial effect on indices of glucose homeostasis, like insulin resis-tance, especially in normoglycemic subjects.31 This may explain why the Mediterranean diet has a more pronounced benefi cial effect on prehypertensive subjects with central obesity; as those subjects have impaired insulin resistance that can be partially controlled by this type of diet, and this may lead to lower likelihood of developing hypertension.

Physical activity shows no protection

Although exercise has shown protective effects on cardio-vascular mortality,32 randomized studies have not shown that physical activity, even as part of a self-care plan for hypertension, to be effective.33 This may due to the fact that only systematic aerobic physical activity for at least 150 min per week offers signifi cant reduction in blood pressure levels,32 and this level of exercise is diffi cult to achieve in everyday life at population level. In our study few partici-pants achieved such a high level of exercise; while and the majority of prehypertensive subjects reported low levels. Additionally, 58% of them were physical inactive and showed no protective effect for the development of hyper-tension. Stephens et al.34 came to the same conclusion, reporting that 10 weeks of 30 min of aerobic exercise, three times a week at 70% VO2 peak, improved cardiorespiratory fi tness and workload achieved, but did not offer a simulta-neous reduction in systolic, diastolic, or mean arterial blood pressure in a cohort of prehypertensive African-American women. Additional studies are needed to determine the intensity and duration of exercise protocols that would be effective in lowering blood pressure in subjects prone to hypertension.

Infl ammation and prehypertension

Recent studies have suggested that low-grade systemic infl ammation participates in the pathophysiology of coro-nary heart disease, metabolic syndrome, and abnormal coagulation process.35 Prospective epidemiological studies, even in healthy subjects, showed an association between elevated C-reactive protein levels and the risk of cardiovas-cular events and peripheral vascular disease, independent of traditional risk factors.36 All risk factors have been related with increased plasma concentration of infl ammatory markers. Among them systolic and pulse pressure levels have been indicated as predictors of plasma C-reactive

102

protein levels in patients with newly diagnosed, never treated essential hypertension, indicating hypertension as a proinfl ammatory state.37 Furthermore, the presence of increased plasma concentration of hemostatic or proinfl am-matory cytokines, specifi c growth factors, and heat shock proteins in hypertensive individuals have been confi rmed in a number of studies.38,39 In a recent study we revealed an association between prehypertension status and infl amma-tory markers among cardiovascular disease-free adults, independent of other coexisting risk factors or unhealthy lifestyle behavior, suggesting that prehypertension might be a proinfl ammatory condition.5 In this work infl ammation seems to play a critical role in the development of hyperten-sion, as those prehypertensives with increased levels of high-sensitivity C-reactive protein seem to be more vulner-able in developing hypertension during the 5 years of follow-up. Although this fi nding is limited by obesity status, C-reactive protein lost its signifi cance among overweight or obese subjects, probably due to the signifi cant interaction of obesity with infl ammation, reveals the important role of suppressing infl ammation, through diet or medication, at the early stages of hypertension.

Limitations

The baseline evaluation was performed once, and may be prone to measurement error. Although three consecutive measurements were taken into account, it is known that average blood pressure of another visit may give lower values than the fi rst visit, due to regression to the mean or the familiarization of the participants to the clinic setting (“white coat effect”). Thus, the prevalence of prehyperten-sion or the levels of arterial blood pressures at baseline may be overestimated. However, our methodology is similar to those of other cross-sectional surveys and follow-up epide-miological studies, and therefore the results are comparable. The relative risks were estimated by the odds ratios through multiple logistic regression analysis. This may overestimate the reported effect. Misreporting of alcohol consumption at the baseline evaluation, due to social class, can be a poten-tial confounder. Another limitation is that we have notcompleted the dietary analysis for nutrient components (including electrolytes), so the role of dietary sodium was not evaluated in this work.

Conclusion

Our fi ndings indicate that a substantial proportion of par-ticipants with prehypertension develop hypertension over a 5-year period. Multi-adjusted analysis revealed that increased age (i.e., >45 years), abnormal waist, low educa-tion level, low adherence to the Mediterranean diet, and high C-reactive protein (>0.95 mg/l) levels, constitute a profi le of prehypertensive people that are prone to develop hypertension within a 5-year period. These fi ndings suggest that screening is desirable in subjects with the aforemen-tioned clinical characteristics, especially in older subjects

where the screening should be probably at least yearly; while management of those subjects should focus on life-style interventions, including diet modifi cations and control-ling excess body weight. Further studies should examine the role of medical therapy in subjects with high probability of developing hypertension that will increase substantially the overall cardiovascular risk.

Acknowledgments The ATTICA Study is supported by research grants from the Hellenic Cardiological Society (HCS2002). The authors would like to thank the fi eld investigators of the study: J. Skoumas, M. Toutouza (biochemical evaluation), I. Papaioannou (physical examina-tion), E. Tsetsekou (physical examination), A. Zeimbekis (physical examination), K. Masoura (physical examination), A. Katinioti (physi-cal examination), S. Vellas (physical examination), E. Kambaxis (nutri-tional evaluation), K. Paliou (nutritional evaluation), C. Tselika (technical support), S. Poulopoulou (technical support), M. Koukoura (technical support), K. Vassiliadou (genetic evaluation), and M. Tout-ouza (data management).

References

1. Kannel WB (2004) Hypertensive risk assessment: cardiovascular risk factors and hypertension. J Clin Hypertens (Greenwich) 6:393–399

2. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R (2002) Age-specifi c relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 pro-spective studies. Lancet 360:1903–1913

3. Wang W, Lee ET, Fabsitz RR, Devereux R, Best L, Welty TK, Howard BV (2006) A longitudinal study of hypertension risk factors and their relation to cardiovascular disease: the Strong Heart Study. Hypertension 47:403–409

4. Joint National Committee on the Prevention, Detection, Evalua-tion, and Treatment of High Blood Pressure (2005) Seventh Report of the Joint National Committee on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Hypertension 43:1–3

5. Chrysohoou C, Pitsavos C, Panagiotakos DB, Skoumas J, Stefanadis C (2004) Association between prehypertension status and infl ammatory markers related to atherosclerotic disease: the ATTICA Study. Am J Hypertens 17:568–573

6. Julius S, Nesbitt SD, Egan BM (2006) Feasibility of treating prehy-pertension with an angiotensin-receptor blocker. N Engl J Med 354:1685–1697

7. Panagiotakos DB, Pitsavos CH, Chrysohoou C, Skoumas J, Papadimitriou L, Stefanadis C, Toutouzas PK (2003) Status and management of hypertension in Greece: role of the adoption of a Mediterranean diet, the Attica study. J Hypertens 21:1483–1489

8. Katsouyanni K, Rimm EB, Gnardellis C, Trichopoulos D, Polychro-nopoulos E, Trichopoulou A (1997) Reproducibility and relative validity of an extensive semi-quantitative food frequency question-naire using dietary records and biochemical markers among Greek schoolteachers. Int J Epidemiol 26:S118–S127

9. Panagiotakos DB, Pitsavos C, Stefanadis C (2006) Dietary patterns: a Mediterranean diet score and its relation to clinical and biologi-cal markers of cardiovascular disease risk. Nutr Metab Cardiovasc Dis. 16:559–568

10. Panagiotakos DB, Pitsavos C, Arvaniti C, Stefanadis C (2007) Adherence to the Mediterranean food pattern predicts the preva-lence of hypertension, hypercholesterolemia, diabetes and obesity, among healthy adults; the accuracy of the MedDietScore. Prev Med 44:335–340

11. International Physical Activity Questionnaire (accessed at http://www.ipaq.ki.se/ on February 28, 2006)

12. Panagiotakos DB, Pitsavos C, Lentzas Y Chrysohoou C, Stefanadis C (2007) Five-year incidence of Hypertension and its determinants: the ATTICA study. J Hum Hypertens 21;686–688

13. Hsia J, Margolis KL, Eaton CB, Wenger NK, Allison M, Wu L, LaCroix AZ, Black HR, for the Women’s Health Initiative

103

Investigators (2007) Prehypertension and Cardiovascular Disease Risk in the Women’s Health Initiative. Circulation 115:855–860

14. Blacher J, Safar ME (2005) Large-artery stiffness, hypertension and cardiovascular risk in older patients. Nat Clin Pract Cardiovasc Med 2:450–455

15. Rodriguez BL, Labarthe DR, Huang B, Lopez-Gomez J (1994) Rise of blood pressure with age: new evidence of population dif-ferences. Hypertension 24:779–785

16. Chobanian A (1991) High-normal blood pressure progression to hypertension in the Framingham Heart Study. Hypertension 17:22–27

17. Vasan RS, Larson MG, Leip EP, Kannel WB, Levy D (2001) Assess-ment of frequency of progression to hypertension in nonhyperten-sive participants in the Framingham Heart Study: a cohort study Lancet 358:1682–1686

18. Stamler J (1991) Epidemiologic fi ndings on body mass and blood pressure in adults. Ann Epidemiol 1:347–362

19. Nakanishi N, Nakamura K, Ichikawa S, Suzuki K, Kawashimo H, Tatara K (1998) Risk factors for the development of hypertension: a 6-year longitudinal study of middle-aged Japanese men. J Hyper-tens 16:753–759

20. Lee JS, Kawakubo K, Kashihara H, Mori K (2004) Effect of long-term body weight change on the incidence of hypertension in Japanese men and women. Int J Obes Relat Metab Disord 28:391–395

21. Rywik SL, Williams OD, Pajak A, Broda G, Davis CE, Kawalec E, Manolio TA, Piotrowski W, Hutchinson R (2000) Incidence and correlates of hypertension in the Atherosclerosis Risk in Commu-nities (ARIC) study and the Monitoring Trends and Determinants of Cardiovascular Disease (POL-MONICA) project. J Hypertens 18:999–1006

22. Francischetti EA, Genelhu VA (2007) Obesity-hypertension: an ongoing pandemic. Int J Clin Pract 61:269–280

23. Safeer RS, Cooke CE, Keenan J (2006) The impact of health literacy on cardiovascular disease. Vasc Health Risk Manage 2:457–464

24. Gu D, Wildman RP, Wu X, Reynolds K, Huang J, Chen CS, He J (2007) Incidence and predictors of hypertension over 8 years among Chinese men and women. J Hypertens 25:517–523

25. Wang Y, Chen J, Wang K, Edwards CL (2006) Education as an important risk factor for the prevalence of hypertension and ele-vated blood pressure in Chinese men and women. J Hum Hyper-tens 20:898–900

26. Keys A, Menotti A, Karvonen MJ, Aravanis C, Blackburn H, Buzina R, Djordjevic BS, Dontas AS, Fidanza F, Keys MH (1986)

The diet and 15-year death rate in the Seven Countries Study. Am J Epidemiol 124:903–915

27. De Lorgeril M, Salen P, Martin JL, Monjaud I, Delaye J, Mamelle N (1999) Mediterranean diet, traditional risk factors and the rate of cardiovascular complications after myocardial infarction. Final report of the Lyon Diet Heart Study. Circulation 99:779–785

28. Trichopoulou A, Costacou T, Bamia C, Trichopoulos D (2003) Adherence to a Mediterranean diet and survival in a Greek popu-lation. N Engl J Med 348:2599–2608

29. Festa A, D’Agostino R, Willimas K, Karter AJ, Mayer-Davis EJ, Tracy RP, Haffner SM (2001) The relation of body fat mass and distribution to markers of chronic infl ammation. Int J Obes 25:1407–1415

30. Chrysohoou C, Panagiotakos DB Pitsavos C, Das UN, Stefanadis C (2004) Adherence to the Mediterranean diet attenuates infl am-mation and coagulation process in healthy adults: The ATTICA Study. J Am Coll Cardiol 44:152–158

31. Panagiotakos DB, Tzima N, Pitsavos C, Chrysohoou C, Zampelas A, Toussoulis D, Stefanadis C (2007) The association between adherence to the Mediterranean diet and fasting indices of glucose homoeostasis: the ATTICA Study. J Am Coll Nutr 26:32–38

32. Miller TD, Fletcher GF (1998) Exercise and coronary artery disease prevention. Cardiologia 43:43–51

33. Viera AJ, Jamieson B, Dealleaume LJ (2007) How effective are hypertension self-care interventions? Fam Pract 56:229–231

34. Stephens Q, Kirby T, Buckworth J, Devor S, Hamlin R (2007) Aerobic exercise improves cardiorespiratory fi tness but does not reduce blood pressure in prehypertensive African American women. Ethn Dis 17:55–58

35. Mizia-Stec K, Gasior Z, Zahorska-Markiewicz B, Holecki M, Kumor P (2006) Infl ammatory markers in a 2-year follow-up of coronary artery disease. Heart Vessels 21:302–308

36. Libby P (2002) Infl ammation in atherosclerosis. Nature 420:868–874

37. Schillaci G, Pirro M, Gemelli F, Pasqualini L, Vaudo G, Marchesi S, Siepi D, Bagaglia F, Mannarino E (2003) Increased C-reactive protein concentrations in never-treated hypertension: the role of systolic and pulse pressure. J Hypertens 21:1841–1846

38. Derhaschnig U, Shehata M, Herkner H, Bur A, Woiserschlaer C, Laggner AN, Hirscl MM (2002) Increased levels of transforming growth factor-beta1 in essential hypertension. Am J Hypertens 15:207–211

39. Pockley AG, De Faire U, Kiessling R, Lemne C, Thulin T, Frost-egard J (2002) Circulating heat shock protein antibody levels in established hypertension. J Hypertens 20:1815–1820