Abstract
Aim To evaluate if pulse pressure (PP) is a risk predictor for atrial fibrillation (AF) in a longitudinal study of 60-year-old men and women from Stockholm (n = 4,232), free from AF at baseline, with primary end-point incident AF. Methods AF diagnoses were obtained from the national hospital discharge register. The estimated risk of AF associated with increasing PP values was calculated according to PP values above median (>52.5 mmHg) and according to 1-SD increase (14 mmHg) in PP, using a crude and an adjusted Cox proportional hazard regression model. Results During a mean follow-up of 13.6 years, 286 incident AF cases were recorded. The number of AF cases increased significantly with increasing PP quartile in men but not in women. PP values above median were associated with increased AF risk (crude HR 1.63, 95% CI 1.28–2.06; p < 0.001), but risk estimates were attenuated after adjustment for common AF risk factors. When PP was entered in the Cox regression model as a continuous variable, the risk of AF did not change by 1-SD PP increase (adjusted HR 1.04, 95% CI 0.91–1.20; p = 0.560). Conclusions PP seems not to be associated with incident AF in a Swedish population of 60-year-old men and women.
Atrial fibrillation is a highly prevalent disease with a partly unknown etiology.
A better knowledge of the risk factors associated with the risk of atrial fibrillation may improve primary prevention strategies.
Our results indicate that pulse pressure is not an independent risk factor for atrial fibrillation.
Key messages
Introduction
Atrial fibrillation (AF), the most common form of sustained cardiac arrhythmia (Citation1), increases the risk of ischemic stroke (IS) 2-6 times and is also associated with increased risk of congestive heart failure (CHF) and death, independently of other cardiovascular risk factors (Citation2,Citation3).
There are several well-known risk factors for AF besides age, including male sex, hypertension, diabetes mellitus, CHF, myocardial infarction (MI), cardiac surgery, and hyperthyroidism, which, however, explain only a small portion of incident AF cases (Citation4).
An emerging risk factor for AF is an increase in pulse pressure (PP), i.e. in the difference between systolic and diastolic blood pressure (SBP and DBP). PP increases with age and is a marker for an age-related increase in proximal aortic wall stiffness (Citation5). While a PP >50 mmHg represents an established risk marker for coronary heart disease (CHD), MI, CHF, and stroke (Citation6–8), its role as predictor of AF is controversial (Citation9). Elevated PP, in fact, has recently been found to be an independent predictive marker for new-onset AF in hypertensive patients and in diabetics, as well as in individuals with no history of cardiovascular disease and in the Framingham Heart study (Citation8,Citation10–13). On the contrary, a large clinical trial recently found that PP was not an independent predictor of new-onset AF in patients at high risk of cardiovascular events (Citation9). These contradictory results may depend upon differences in the age of the study participants, sex differences, as well as differences in the prevalence of comorbidities at baseline. Therefore, new studies are warranted in order to clarify the role of PP as an independent predictor of incident AF.
The aim of the present study was to evaluate the role of PP as risk predictor for AF in a longitudinal cardiovascular study, where all study participants were enrolled at the age of 60 years.
Material and methods
Study population
The 60-year-old men and women study from Stockholm is a prospective cohort study, designed to identify novel risk factors and biomarkers for cardiovascular events. The cohort consists of every third man and woman turning 60 between 1 July 1997 and 30 June 1998. The study subjects were randomly identified from the Swedish population register and were invited by mail to participate in a cardiovascular and metabolic health screening. Of 5,460 individuals invited to participate in the study, 4,232 agreed (response rate 78%) of which 2,039 were men and 2,193 were women. Every individual who agreed to participate was eligible for the study. All study participants completed a large, self-administered questionnaire on socio-demographics, education level, previous diseases and hospitalizations, daily habits, physical activity, diet and medications. A physical examination including anthropometric measures and a standard 12-lead resting electrocardiogram (ECG) was performed at baseline. Whole blood, serum and plasma samples were collected at baseline and stored in a biological bank. Serum levels of glucose, lipids and several other biochemical tests were analysed using enzymatic methods, described in detail elsewhere (Citation14). Exclusion criteria for the present study were: history of AF, self-reported in an open question in the questionnaire (n = 4), registered hospital diagnosis prior to enrolment (n = 12, from January 1997) or AF present on baseline ECG (n = 31), unfilled questionnaire (n = 119), missing BP measurements (n = 4), thus leaving 2,125 women and 1,937 men for the present analyses. A flow chart summarizing the inclusion and exclusion criteria is displayed in .
Figure 1. Flow chart of the exclusion criteria adopted in the present study.
The study was approved by Regional Ethics Review Board at Karolinska Institutet, Stockholm, Sweden.
PP measurement
Brachial PP was calculated for every study participant as the difference between the average SBP and DBP, measured at the baseline physical examination. BP was measured twice in the right arm after 5 minutes of rest, with the participant in a lying-down position using an automatic device (HEM 711; Omron Health Care, Bannockburn, IL, USA). The mean of the two values was calculated. Mean arterial pressure (MAP) was calculated as DBP plus one-third of PP in accordance with prior studies (Citation10,Citation11).
Ascertainment of AF
The primary endpoint in this study was new-onset AF. All baseline ECGs were reviewed to detect prevalent AF and/or atrial flutter at baseline. Incident AF cases were recorded through an annual matching procedure with the national hospital discharge register, by means of the participants’ personal identification number (100% follow-up). To assess diagnoses of AF, the international classification of diseases (ICD-10) codes were used. AF cases were registered when I.48.0–I.48.9 was present as main diagnosis or secondary diagnosis. The annual matching procedures have been performed since 1 January 1997 until 31 December 2012.
Statistical analyses
Anthropometric and biochemical measurements are presented as mean ± standard deviation (SD) for continuous variables and as frequencies and percentages for categorical variables. Normality of distribution was assessed for all continuous variables by analysis of the skewness of the distribution with the Kolmogorov–Smirnov test. Baseline characteristics of incident AF cases and reference group were compared, and differences between the two groups were tested by Student’s t test for continuous variables and by chi-square test for categorical variables.
PP percentiles were calculated from the PP values in the entire cohort: Q1: ≤44.0 mmHg (25th percentile); Q2: >44.0, ≤52.5 mmHg (median); Q3: >52.5, ≤62.5 mmHg; and Q4: >62.5 mmHg (75th percentile). Differences in BP, AF incident cases, and AF incidence rate across the PP quartiles were tested by ANOVA (mean values) and chi-square test (frequencies) in men and women separately.
The risk of AF associated with increasing PP values was calculated by means of a Cox proportional hazard regression analysis according to the PP quartile distribution and according to SD increase in PP. Proportionality was assessed by Schoenfeld’s test.
In order to investigate sex differences regarding PP and the risk of AF, an interaction analysis between men and women was performed by inserting an interaction term in the Cox regression model.
When PP was entered as a categorical variable in the regression model, we estimated the risk of AF expressed as hazard ratio (HR) and 95% confidence interval (CI) between different PP quartiles as well as in individuals with PP values above median (>52.5 mmHg) compared to individuals with PP values below or equal to median (≤52.5 mmHg). In the analysis stratified by sex, men with PP >62.5 mmHg were defined as exposed to elevated PP as compared to men with PP ≤62.5 mmHg. The choice of this arbitrary cut-off point was based on a comparison of baseline characteristics, AF incidence rates and HRs across PP quartiles in men.
When PP was entered in the regression model as a continuous variable, we estimated the risk of AF expressed as HR with 95% CI associated with 1-SD increase in PP (13.0 mmHg in men, 14.8 mmHg in women and 14.0 mmHg in all study participants). The competing risk of death (all-cause mortality) was considered, and additional Cox regression analyses were performed with death as well as death and/or AF as primary Endpoint.
Two multivariable Cox regression models were run: one with and one without hypertension as covariate, after adjustments by sex, body mass index (BMI), smoking, diabetes mellitus, prior MI, history of heart failure, left ventricular hypertrophy (LVH), based on prior reports (Citation4,Citation10,Citation15). BMI (kg/m2) was calculated from weight and height measurements performed at baseline. Participants were classified as current smokers or non-smokers. Information on smoking was missing in 60 individuals. Hypertension was defined as self-reported hypertension (yes/no); use of antihypertensive medication; or newly diagnosed in subjects with SBP and/or DBP ≥140/90 mmHg at the physical examination. Diabetes mellitus was defined as self-reported in the questionnaire (yes/no); use of antidiabetic medication; or fasting serum glucose of ≥7.0 mmol/L. Lab data including serum glucose were missing in two individuals, resulting in missing information about presence or absence of diabetes mellitus. Information on history of CHF and CHD, i.e. prior MI and angina pectoris, was obtained via the questionnaire (yes/no). During follow-up, subjects were classified as having history of CHD if the condition was diagnosed on or prior to the day of AF diagnosis. CHD diagnoses were collected through the annual matching procedure with the national cause of death register and hospital discharge register by using the international classification of diseases (ICD-10) codes: ICD code I.21–I.25 for MI and ICD I.20 for angina requiring hospitalization. LVH was determined using standard 12-lead resting ECG and two established criteria for LVH, the Minnesota Code and the Cornell voltage-duration product. Either criterion had to be positive to define presence of ECG-LVH. Further details on the ECG-LVH diagnoses are presented elsewhere (Citation16). Information on ECG-LVH was missing in four individuals. The missing data on covariates resulted in 66 fewer individuals included in the multivariate analysis compared to the univariate analysis (n = 3,996 versus n = 4,062). A sensitivity analysis was performed, with and without the exclusion of individuals with missing data on covariates.
A subgroup analysis was performed on hypertensive and normotensive individuals separately.
An additional multivariate analysis was performed with hypertension defined as use of antihypertensive medication.
A two-sided p value <0.05 was required for statistical significance. All calculations were performed using STATA version 13.
Results
Baseline characteristics of the study population are presented in . All continuous variables were normally distributed.
Table I. Baseline characteristics of men and women and entire cohort in incident AF cases and in the reference group.
AF incidence rate
Of 4,062 individuals included in the analysis, 286 incident AF cases were recorded (7.0%): 161 were men and 125 were women. Of 286 incident AF cases, 156 were recorded as main diagnosis and 130 as secondary diagnosis. The overall AF incidence rate was 5.2 per 1,000 person-years, 6.2 in men and 4.3 in women. Mean follow-up time was 13.6 ± 2.8 years. Death was censored. During follow-up, 551 participants died, of whom 490 were free from AF.
PP values in incident AF cases and reference group
PP was normally distributed. Mean baseline PP for the entire population was 53.9 ± 14.0 mmHg; PP was significantly higher in incident AF cases than in the reference group: 56.8 ± 14.5 mmHg versus 53.7 ± 13.9 mmHg (p = 0.0003).
AF incidence rate and risk of AF across PP quartiles
The AF cumulative incidence according to PP quartiles is illustrated in .
Figure 2. Cumulative AF incidence according to PP quartiles. Cut-off points denoting PP quartiles were at 44.0, 52.5, and 62.5 mmHg. Q1, ≤44.0 mmHg; Q2, 44.1–52.5 mmHg; Q3, 52.6–62.5 mmHg; and Q4, >62.5 mmHg.
The AF incidence rate increased from 4.0 per 1,000 person-years in the lowest PP quartile (≤44.0 mmHg) to 7.2 per 1,000 person-years in the highest quartile (>62.5 mmHg). The number of main versus secondary AF diagnoses did not differ significantly between PP quartiles in the entire cohort (data not shown), nor in men and women separately (Supplementary Table 1, available online).
We estimated the risk of AF in individuals with PP values above median (>52.5 mmHg) compared to individuals with PP values below or equal to median (≤52.5 mmHg). Individuals with PP values above median had a 63% higher risk of developing AF than individuals with PP values below median (crude HR 1.63, 95% CI 1.28–2.06; p < 0.001). In the multivariate model, PP values above median were associated with 37% increased risk of AF before hypertension was added into the model (adjusted HR 1.37, 95% CI 1.08–1.76; p = 0.010). After addition of hypertension into the model, PP values above median were no longer associated with increased risk of AF (adjusted HR 1.18, 95% CI 0.88–1.60; p = 0.270).
The interaction analysis showed no significant difference in AF risk associated with PP values above median between men and women (interaction term 0.81, 95% CI 0.50–1.31; p = 0.391).
However, since AF risk differs between men and women we also performed a stratified analysis by gender. In men, the number of AF cases increased significantly with increasing PP quartile, and the AF incidence was more than doubled in PP Q4 compared to PP Q1 (Supplementary Table 1, available online).
The estimated risk of AF in men exposed to PP values >62.5 mmHg, i.e. the highest PP quartile, was 66% higher than in men with PP values PP ≤62.5 mmHg (crude HR 1.66, 95% CI 1.20–2.29; p = 0.002). The estimated AF risk in men exposed to PP values above median (PP >52.5 mmHg) was 73% higher than in men with PP values below median (crude HR 1.73, 95% CI 1.25–2.40; p = 0.001). After adjustment for other AF risk factors, the difference in risk did not attain statistical significance in either analysis.
In women, there was no significant increase in AF cases with increasing PP quartiles (Supplementary Table 1, available online). When we estimated the risk of AF in women with PP values above median as compared to women with PP values below median, there was no significant difference in AF risk, and further analyses were therefore not performed (crude HR 1.40, 95% CI 0.99–1.99; p = 0.060).
Incremental increase in PP and AF risk
When we entered PP as a continuous variable in the univariate Cox regression model, each SD increase in PP was associated with a 24% increased risk of developing AF (). Men had a 26% increased risk of new-onset AF per SD increase in PP, and women had a 19% increased risk of new-onset AF per SD increase in PP (). The estimated risk of AF per SD increase in PP did not differ significantly between men and women (interaction term 0.99, 95% CI 0.98–1.01; p = 0.447).
Table II. Hazard ratio (HR) and 95% confidence interval (CI) for incident AF per 1-SD increase in PP.
PP did not remain significant in predicting AF in the entire population when hypertension was added into the model (). Risk factors that remained statistically significant for the entire cohort in the multivariate model were sex, BMI, hypertension and ECG-LVH. The results remained the same whether or not the individuals with missing data on covariates were included in the analysis (data not shown). We redefined hypertension as use of antihypertensive treatment; still PP was not a significant predictor of AF in the multivariate model (HR 1.09, 95% CI 0.97–1.22; p = 0.166, entire cohort).
One-SD increase in PP was associated with 17% increased risk of death (all-cause mortality) (crude HR 1.17, 95% CI 1.08–1.27; p < 0.001), and with 16% increased risk of AF and/or death (crude HR 1.16, 95% CI 1.08–1.24; p < 0.001) in the entire cohort.
Subgroup analyses
We tested PP as a predictor of AF in hypertensive and normotensive individuals separately. PP was not an independent predictor of AF in either group, neither in men nor in women separately (). There was no significant difference regarding AF risk between men and women, neither in the normotensive nor hypertensive group (interaction term 0.96, 95% CI 0.91–1.01; p = 0.153 and 1.01, 95% CI 0.99–1.03; p = 0.455, respectively).
Table III. Hazard ratio (HR) and 95% confidence interval (CI) for incident AF per 1-SD increase in PP in normotensive and hypertensive individuals, respectively.
Discussion
The main finding of this study is that a linear increase in PP and exposure to high PP are not independently associated with increased risk of developing AF in a general population of 60-year-olds with a follow-up of about 13.6 years.
This is the first large prospective cohort study that indicates that PP is not an independent predictor of AF, and it challenges results from prior reports (Citation8,Citation10–13). Differences in study design as well as in baseline characteristics of the study participants may partly explain these controversial results. The majority of the prior studies analysing the risk of AF associated with high PP were performed in individuals at high risk of cardiovascular events, e.g. patients with hypertension, LVH, and diabetes mellitus (Citation8,Citation11,Citation12), while our study population comprised relatively healthy 60-year-olds.
Two prior studies have shown that PP is an independent predictor of AF in a general population (Citation10,Citation13). In one of these studies (Citation10), AF prevalence was higher than in our study (13.1% versus 7.0%), which could explain our different results. The reason for this could be that our study comprised healthier individuals and that AF diagnoses were obtained only from hospitalized patients. Furthermore, the percentage of hypertensive individuals was lower in both prior studies than in our study, which could explain the fact that PP remained a significant predictor of AF in these studies after controlling for hypertension. The reason why our study comprised more hypertensive individuals could be that the definition of hypertension used in this study included individuals with SBP ≥140 mmHg and/or DBP ≥90 mmHg at baseline, as well as individuals with self-reported hypertension unlike prior studies. However, when we redefined hypertension as use of antihypertensive treatment only, the results regarding PP and risk of AF did not change.
We observed that increased SBP mainly drove high PP. This is in contrast to prior studies where high PP values were also driven by a decrease in DBP, and individuals with high PP had lower DBP values as compared to individuals with low PP (Citation10,Citation11). SBP, in fact, increases with age, while DBP increases until the age of 50, reaches a plateau and starts to decrease after the age of 60 (Citation7,Citation17,Citation18). In the present study, no decrease in DBP values across PP quartiles was observed. By eliminating age as a factor affecting PP, we have analysed PP independently from the effects of age on DBP and SBP. We lack repetitive BP measurements, which is in contrast to prior studies where BP was measured during follow-up as well (Citation10–12). If we had performed BP measurements during follow-up, we could possibly have seen a decrease in DBP and subsequently an increase in PP, which could have altered our results on PP as a predictor of AF. In our study, individuals could develop hypertension during follow-up as well, and later on AF—an association that we would miss.
Different observations in men and women
In men, we observed a distinct increase in AF incidence with increasing PP quartiles, which was not seen in women. In women, high PP values were not associated with increased AF risk in contrast to men. This supports the notion that the impact of a certain AF risk factor differs between men and women (Citation4). PP has been reported to be higher in women than in men (Citation10,Citation11,Citation19,Citation20); however, we observed the opposite. This may be explained by the fact that more men than women had hypertension at baseline and that men included in the study had more comorbidities like hypertension, CHD and diabetes at baseline than women, and, hence, stiffer arteries and higher PP values.
We chose to analyse men and women separately even though the interaction term showed that there was no significant difference in the AF risk associated with increasing PP between men and women. We found it important from an epidemiological and clinical point of view to report sex-specific analyses since there are no such data in the literature, and AF is less common in women whereas the effect of AF on stroke mortality is considerably higher (Citation21).
Elevated PP, hypertension and risk of AF
A standardized cut-off point has not been defined to classify exposure to elevated PP. In the present study we have used cut-offs derived from the PP distribution in our population, and we have observed that, regardless of the cut-off used (median or highest quartile), after controlling for other risk factors, the difference in AF risk across PP quartiles disappeared.
This is not in line with the results from the study by Larstorp et al. (Citation11), where extremely high PP (>87.5 mmHg) was associated with an increased risk of developing AF even after adjusting for other risk factors for AF in patients with hypertension and ECG-LVH.
We analysed PP as a continuous variable in order to examine the role of PP independently of hypertension, but the results did not change.
All these findings imply that PP is strongly associated with elevated BP and that, rather than being an independent risk factor for AF, PP mirrors the effect of high BP on the AF risk. Therefore, hypertension per se seems to be the most important factor to consider when evaluating the AF risk in a general population. The stratified analysis between PP and risk of AF in normotensive and hypertensive individuals separately also strongly argues for this.
Increased PP is a risk marker for cardiovascular mortality and total mortality (Citation22,Citation23). In our study, increasing PP was associated with increased risk of death, and death and/or AF combined, although the estimated risk for these end-points was almost the same and lower than the estimated risk of AF alone. Our interpretation of this result is that an increased risk of death from all causes in individuals exposed to high PP might not fully explain the lack of association between increased PP and AF risk observed in our population.
There are several potential limitations of our study. PP was calculated from one single BP measurement, and a substantial portion of the study participants had antihypertensive treatment at baseline that affected the BP values. Misclassification of individuals with temporarily elevated BP due to nervousness or anxiety at the time of the physical examination should be considered. More exact BP measurements like central BP values or pulse wave velocity (PWV) were unfortunately not available for this study. It should be mentioned that BP was measured in a lying-down position, which was according to practising routines in Sweden at the time when the study was conducted. The recommendation nowadays is that BP should be taken in a sitting position. Another limitation is that we lack a reference group of normotensive individuals with a high PP.
AF diagnoses were based on hospital discharge codes as well as self-reports and a standard ECG at baseline. This means that a certain number of prevalent and incident AF cases will remain undetected. An important fact to consider in this matter is that approximately one-third of all AF cases are silent (Citation24), which means that a substantial portion of incident AF cases remain undetected regardless of method to ascertain AF diagnoses. AF diagnoses from public health care were not available for the present study. However, it has been recently reported that AF diagnoses recorded in primary care represents only 12% of all AF diagnoses in Stockholm County between the years 2006 and 2010 (Citation25). Another limitation is that we did not have information about presence of valvular disease from the physical examination, a well-known risk factor for AF.
In conclusion, our data show that PP seems not to be an independent predictor of incident AF, diagnosed during hospital stay, in a population-based cohort of 60-year-old men and women from Sweden.
Declaration of interest
The authors declare no potential conflict of interest.
This work was supported by a research grant from Stockholm Council County (ALF) to B.G., from Loo och Hans Ostermans stiftelse to B.G., and the Swedish Heart and Lung Foundation to U.d.F.
References
- Markides V, Schilling RJ. Atrial fibrillation: classification, pathophysiology, mechanisms and drug treatment. Heart. 2003;89:939–43.
- Benjamin EJ, Wolf PA, D'Agostino RB, Silbershatz H, Kannel WB, Levy D. Impact of atrial fibrillation on the risk of death: the Framingham Heart Study. Circulation. 1998;98:946–52.
- Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke. 1991;22:983–8.
- Benjamin EJ, Levy D, Vaziri SM, D'Agostino RB, Belanger AJ, Wolf PA. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA. 1994;271:840–4.
- Safar ME, Nilsson PM, Blacher J, Mimran A. Pulse pressure, arterial stiffness, and end-organ damage. Curr Hypertens Rep. 2012;14:339–44.
- Madhavan S, Ooi WL, Cohen H, Alderman MH. Relation of pulse pressure and blood pressure reduction to the incidence of myocardial infarction. Hypertension. 1994;23:395–401.
- Vaccarino V, Berger AK, Abramson J, Black HR, Setaro JF, Davey JA, et al. Pulse pressure and risk of cardiovascular events in the systolic hypertension in the elderly program. Am J Cardiol. 2001;88:980–6.
- Ciaroni S, Bloch A, Lemaire MC, Fournet D, Bettoni M. Prognostic value of 24-hour ambulatory blood pressure measurement for the onset of atrial fibrillation in treated patients with essential hypertension. Am J Cardiol. 2004;94:1566–9.
- Verdecchia P, Dagenais G, Healey J, Gao P, Dans AL, Chazova I, et al. Blood pressure and other determinants of new-onset atrial fibrillation in patients at high cardiovascular risk in the Ongoing Telmisartan Alone and in Combination With Ramipril Global Endpoint Trial/Telmisartan Randomized AssessmeNt Study in ACE iNtolerant subjects with cardiovascular Disease studies. J Hypertens. 2012;30:1004–14.
- Mitchell GF, Vasan RS, Keyes MJ, Parise H, Wang TJ, Larson MG, et al. Pulse pressure and risk of new-onset atrial fibrillation. JAMA. 2007;297:709–15.
- Larstorp AC, Ariansen I, Gjesdal K, Olsen MH, Ibsen H, Devereux RB, et al. Association of pulse pressure with new-onset atrial fibrillation in patients with hypertension and left ventricular hypertrophy: the Losartan Intervention For Endpoint (LIFE) reduction in hypertension study. Hypertension. 2012;60:347–53.
- Valbusa F, Bonapace S, Bertolini L, Zenari L, Arcaro G, Targher G. Increased pulse pressure independently predicts incident atrial fibrillation in patients with type 2 diabetes. Diabetes Care. 2012;35:2337–9.
- Roetker NS, Chen LY, Heckbert SR, Nazarian S, Soliman EZ, Bluemke DA, et al. Relation of systolic, diastolic, and pulse pressures and aortic distensibility with atrial fibrillation (from the Multi-Ethnic Study of Atherosclerosis). Am J Cardiol. 2014;114:587–92.
- Halldin M, Rosell M, de Faire U, Hellenius ML. The metabolic syndrome: prevalence and association to leisure-time and work-related physical activity in 60-year-old men and women. Nutr Metab Cardiovasc Dis. 2007;17:349–57.
- Ahmad Y, Lip GY, Lane DA. Recent developments in understanding epidemiology and risk determinants of atrial fibrillation as a cause of stroke. Can J Cardiol. 2013;29(7 Suppl):S4–13.
- Halldin M, Fahlstadius P, de Faire U, Vikstrom M, Hellenius ML. The metabolic syndrome and left ventricular hypertrophy–the influence of gender and physical activity. Blood Press. 2012;21:153–60.
- Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, et al. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289:2560–72.
- Joffres M, Falaschetti E, Gillespie C, Robitaille C, Loustalot F, Poulter N, et al. Hypertension prevalence, awareness, treatment and control in national surveys from England, the USA and Canada, and correlation with stroke and ischaemic heart disease mortality: a cross-sectional study. BMJ Open. 2013;3:e003423.
- Franklin SS, Khan SA, Wong ND, Larson MG, Levy D. Is pulse pressure useful in predicting risk for coronary heart Disease? The Framingham heart study. Circulation. 1999;100:354–60.
- Winston GJ, Palmas W, Lima J, Polak JF, Bertoni AG, Burke G, et al. Pulse pressure and subclinical cardiovascular disease in the multi-ethnic study of atherosclerosis. Am J Hypertens. 2013;26:636–42.
- Friberg J, Scharling H, Gadsboll N, Truelsen T, Jensen GB, Copenhagen City Heart Study. Comparison of the impact of atrial fibrillation on the risk of stroke and cardiovascular death in women versus men (The Copenhagen City Heart Study). Am J Cardiol. 2004;94:889–94.
- Benetos A, Rudnichi A, Safar M, Guize L. Pulse pressure and cardiovascular mortality in normotensive and hypertensive subjects. Hypertension. 1998;32:560–4.
- Glynn RJ, Chae CU, Guralnik JM, Taylor JO, Hennekens CH. Pulse pressure and mortality in older people. Arch Intern Med. 2000;160:2765–72.
- Camm AJ, Corbucci G, Padeletti L. Usefulness of continuous electrocardiographic monitoring for atrial fibrillation. Am J Cardiol. 2012;110:270–6.
- Forslund T, Wettermark B, Wandell P, von Euler M, Hasselstrom J, Hjemdahl P. Risk scoring and thromboprophylactic treatment of patients with atrial fibrillation with and without access to primary healthcare data: experience from the Stockholm health care system. Int J Cardiol. 2013;170:208–14.