Introduction
Hypertension is a common condition, strongly associated with cardiovascular mortality and morbidity [
1,
2] and 2010 established as the leading cause behind the world’s global burden of disease [
2,
3], and as recently reviewed and put into a context of causal determinants by Fuchs and Whelton [
3]. The mechanisms behind essential hypertension are complex and yet not fully understood although both multiple biological and behavioural factors are known to be involved in hypertension development [
4]. The prevalence of hypertension often increases with age [
5], and earlier longitudinal studies also indicate that blood pressure levels late in life can be traced back to levels earlier in life, even in childhood [
6]. Still, the development of hypertension with increasing age should not be considered inexorable if proper preventive actions are taken [
3]. Correspondingly, subjects with high blood pressure, also within the normal range, run an increased risk of developing hypertension compared to subjects with lower blood pressure [
3] and there also seems to be a sliding scale of elevated cardiovascular disease risk associated with blood pressures all the way down to a level of 115/75 mmHg [
7]. European society of hypertension [
8] defines
high normal blood pressure as 130–139 mmHg systolic and/or 85–89 mmHg diastolic. Estimates of hypertension conversion rate vary, but earlier studies [
9‐
11] suggest that 2–4 individuals out of 5 are at risk to convert to hypertension in 4–10 years depending on age and level of economic development in the country for the study. Therefore, while a presumptive cardiovascular risk factor in its own right, high normal blood pressure also constitutes a risk factor for future hypertension and has sometimes even been referred to as
prehypertension. Thus, subjects of high normal blood pressure, especially in younger population strata, constitute a highly interesting and still rather poorly studied group.
Our study aim was accordingly to study the determinants for incident hypertension among a representative population sample of individuals mainly in their early middle-ages with focus on baseline blood pressure categories.
Results
Characteristics of the baseline population are presented in Table
2. Among the 1099 participants 625 (56.9%) had optimal blood pressure, 284 (25.8%) had normal blood pressure, 128 (11.6%) had high normal blood pressure and 62 (5.6%) had unstable blood pressure at the baseline survey. Mean follow-up time in this cohort was 9.7 years and during this period we observed 167 (15.2%) incident cases of hypertension during follow-up. Only 16 (2.6%) of 625 subjects with optimal baseline blood pressure converted to hypertension during follow up, while corresponding figures were 55 (19.4%) within normal, 50 (39.1%) with high normal, and 46 subjects (74.2%) with an unstable blood pressure, respectively.
Table 2
Characteristics of the study population at baseline
Age (years) | 46.7 | 10.6 |
BMI (kg/m2) | 26.2 | 3.8 |
LDL-cholesterol (mmol/L) | 3.2 | 0.9 |
HDL-cholesterol (mmol/L) | 1.3 | 0.3 |
Triglycerides (mmol/L) | 1.05 | 0.77–1.44 |
HOMAir (mmol*mU/L2) | 1.13 | 0.76–1.71 |
Alcohol (g/week) | 25.2 | 8.4–62.1 |
Sex (male) | 537 | 48.9 |
Blood pressure category | | |
Optimal | 609 | 65.3 |
Normal | 229 | 24.6 |
High normal | 78 | 8.4 |
Unstable | 16 | 1.7 |
Diabetes Mellitus type 2 | 21 | 1.9 |
Daily smoking | 154 | 14 |
Leisure time physical activity | | |
Inactive | 68 | 6.2 |
Little active | 620 | 58.3 |
Moderately active | 338 | 31.8 |
Heavily active | 38 | 3.6 |
Family history of hypertension | 441 | 40.8 |
Education | | |
≤ 9 years | 271 | 25 |
9–12 years | 428 | 40 |
> 12 years | 380 | 35 |
Accordingly, normal, high normal and unstable baseline blood pressure were all significantly associated with the development of manifest hypertension compared to optimal blood pressure OR (CI), of 7.04 (3.89–12.7), 17.1 (8.88–33.0) and 84.2 (37.4–190), respectively (Table
3), test for trend
p < 0.001. These results remained practically the same when also duration of education, leisure time of physical activity, type 2 diabetes, LDL-cholesterol and alcohol use were included in the regression (Table
3). The progression to hypertension was also significantly associated with BMI as shown in Table
3. However, age and a family history of hypertension did not contribute statistically significant to the incidence of hypertension in this model (Table
3). When Homa-ir was substituted for BMI in this model it also came out statistically significant (
p < 0.001), however, the OR’s for the other covariates remained practically the same. Level of physical activity, reported alcohol intake, level of education and concentrations of LDL cholesterol was, however, not predictive of progression to manifest hypertension in corresponding multivariate models (data not shown).
Table 3
Determinants of incident hypertension over 10 years in the Vara/Skövde Cohort
Covariates in model age, BMI, Family history of hypertension, BP category at baseline | | | |
Age | 1.01 | 0.99–1.03 | 0.153 |
BMI | 1.12 | 1.06–1.18 | < 0.001 |
Family Hx hypertension | 1.47 | 0.99–2.20 | 0.059 |
Blood pressure category at baseline | | | |
Optimal | 1 | | |
Normal | 7.04 | 3.89–12.7 | < 0.001 |
High normal | 17.1 | 8.88–33.0 | < 0.001 |
Unstable | 84.2 | 37.4–190 | < 0.001 |
Covariates in model age, BMI, Family history of hypertension, BP category at baseline, education, Type 2 diabetes, LDL Cholesterol, alcohol use | | | |
Age | 1.01 | 0.99–1.04 | 0.393 |
BMI | 1.11 | 1.05–1.17 | < 0.001 |
Family history of hypertension | 1.43 | 0.94–2.18 | 0.095 |
Blood pressure category at baseline | | | |
Optimal | 1 | | |
Normal | 6.58 | 3.62–12.0 | < 0.001 |
High normal | 16.0 | 8.21–31.2 | < 0.001 |
Unstable | 77.5 | 32.9–183 | < 0.001 |
Education | | | |
≤ 9 years | 1 | | |
9–12 years | 0.95 | 0.55–1.64 | 0.853 |
> 12 years | 0.69 | 0.38–1.25 | 0.222 |
Type 2 diabetes at baseline yes/no | 2.51 | 0.86–7.33 | 0.092 |
Leisure time physical activity low/high | 1.01 | 0.64–1.60 | 0.965 |
LDL cholesterol mmol L-1 | 1.05 | 0.81–1.35 | 0.722 |
Alcohol users vs non-users | 1.13 | 0.64–1.99 | 0.677 |
Discussion
Our results confirm the fact that high normal blood pressure by far constitutes the strongest risk factor for development of manifest hypertension, independent of age, BMI and family history of hypertension [
5,
17,
18]. Furthermore, the risks associated with normal, high normal and unstable baseline blood pressure increase in a dose dependent manner compared to optimal blood pressure. While overweight constituted the second strongest finding regarding the transition from high normal blood pressure to manifest hypertension nor age or a family history of hypertension contributed significantly to the risk of incident hypertension.
In other cohort-studies with similar socio-economic development as ours, OR’s for conversion from high normal blood pressure to manifest hypertension were similar to our findings using optimal blood pressure as reference [
3,
9,
11], while OR’s were correspondingly higher in unacculturated countries [
10]. Unstable blood pressure at the baseline survey was associated with the highest risk of development of hypertension as the vast majority in this category were hypertensive at follow-up. Possibilities to compare this result with previous studies are limited due to the fact that the definition of unstable blood pressure is not generally acknowledged. However, our results emphasize the importance to follow-up subjects with unstable blood pressure due to their high risk to convert into hypertension.
In epidemiological surveys a diagnosis of hypertension is generally based on a single blood pressure measurement at the study visit. However, the blood pressure is known to vary in between readings and especially between different visits to a clinic, and repeated measurements are likely to reduce the misclassification associated with this phenomenon. In this study three consecutive measurements were required for the diagnosis of hypertension, thus simulating the procedures in a more clinical setting and accordingly defining a more valid target group. This procedure is likely to reduce the proportion identified with hypertension, however, in a previous publication from this cohort we showed that the prevalence of hypertension was only six per cent higher using only the first blood pressure measurement compared to strictly applying three consecutively high blood pressures [
19]. Still, the reduced misclassification should increase the generalizability and the implications of our findings.
The blood pressure category in focus here, high normal blood pressure, has in some contexts been labelled prehypertension based on its increased risk of hypertension within near future [
5,
17]. Actually, the latest expert guidelines in the US have included this segment of the blood pressure distribution in the diagnosis of hypertension [
20]. While lifestyle modifications are the recommended intervention, also pharmacological treatment can be advocated when the global risk is estimated to be high. The association between overweight and the development of hypertension is well established before [
21]. Hypertension is also known to cluster within families, even so family history has seldom been considered in previous studies [
9,
22‐
24]. We found that a history of hypertension among first degree relatives was not associated with development of hypertension, however, it was close to significance (OR 1.47 (CI 0.99–2.20,
p = 0.057)), and a type 2 error may be at hand. Therefore, we suggest that the clinical implications of heritability should be further studied in the future. Indeed, part of the effects of heritability might be expressed through the impact of baseline blood pressure categories as observed in this study.
The results obtained are in accordance with previous studies and confirm the fact that also a mild blood pressure elevation, already in middle age, constitutes a serious risk factor for future hypertension. CVD-risk grading tools, such as SCORE [
25], are important instruments for cardiovascular disease- and death-risk calculation and can serve as support in approaching question of treatment initiation. We believe, however, that in order to make prevention of cardiovascular disease more effective, employing early prevention of hypertension is essential as that may in a large part reduce the vascular consequences usually attributed to aging and further limit the population burden of blood pressure-related CVD as expressed by Fuch and Whelton [
3]. That age in itself is not the dominant factor is also illustrated by the increased risk of conversion from high normal blood pressure to manifest hypertension also generally seen among younger population strata [
3,
9‐
11]. In fact, most participants in our cohort were aged 30–50 years at baseline and thus illustrates the development of hypertension from adulthood into early middle-ages before age-related chronic conditions take overhand. Results obtained from this study can be instantly implemented into clinical practice in primary care through doctors emphasizing current blood pressure and BMI as major factors when identifying subjects in need of preventive measures to reduce the risk of developing hypertension. The major importance of BMI is in accordance with general recommendations of the maintenance of a healthy body weight by means of attention to caloric intake and physical activity [
26] and findings in previous interventions to reduce blood pressure using a strict diet on low sodium and the DASH-diet, respectively [
27,
28]. The implications for clinical practice and strategies of prevention are obvious.
Among the strengths of this study are the representative population-based sampling and the comparably high participation rate, that confers a strong generalizability. It should be emphasized that our study population derives from two smaller communities, Vara with about 16,000 residents and Skövde with about 50,000 residents, and thus exposed to a higher prevalence of overweight and obesity as there is a strong gradient along the population density for these conditions as recently published by Hemmingsson et al. [
29]. Other differences in lifestyle and in the environment should also be considered when comparing the outcome to that in bigger cities, however, the strong basic determinants should still be the same. Further strengths are the utilization of standard methods and the strict repeating of the same study protocol. Limitations of this study include the natural risk for inaccuracy regarding self-reported information and losses to follow-up. Furthermore, the study did not take into account diet and especially not salt consumption, an important factor in hypertension aetiology and blood pressure regulation, thus a potential residual confounder [
27]. Though length of education was not a significant factor in the incidence of hypertension in our study other information on socioeconomic aspects could also have been valuable in the analysis.
In conclusion, high normal blood pressure is common in the general population and these individuals run a considerable risk of progression to manifest hypertension already in the middle-ages, as do individuals with one or two blood pressures within the hypertension range. They should be easy to identify among the flow of patients in primary care applying the present findings and proper intervention should be in accordance with general preventive strategies with a focus on diet and starting in young adults. However, controlled clinical trials comprising also cost benefit analyses are called for in this area.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.