Abstract
Introduction. Preliminary evidence suggests that hypercholesterolemia is associated to an increased risk to develop hypertension. We aimed at evaluating the association between low-density lipoprotein cholesterol (LDL-C) level and incidence of new onset of hypertension in a large population sample.
Materials and methods. A population-based cohort of 20,074 subjects with at least one LDL-C measurement between July 2006 and June 2007 and free of antihypertensive treatment (AHT) at baseline was followed from the LDL-C date until death or 31 December 2009.
Results. During the follow-up, 10.7% of patients with LDL-C < 130 mg/dL, 13.2% of patients with LDL-C between 130 and 159 mg/dL, 12.2% of patients with LDL-C between 160 and 189 mg/dL, and 13.9% of patients with LDL-C ≥ 190 mg/dL had new-onset hypertension requiring the initiation of AHT. Compared with the LDL-C < 130 mg/dL group, the hazard ratio (HR) of initiation of AHT increased among those with LDL-C level between 130 and 159 mg/dL (HR 1.23; 95% CI: 1.08–1.40), those with LDL-C level between 160 and 189 mg/dL (HR 1.24; 95% CI: 1.01–1.51), and those with LDL-C ≥ 190 mg/dL (HR 1.45; 95% CI: 1.11–1.89).
Conclusion. A better control of cholesterolemia seems to be associated to a lower incidence of new-onset of hypertension requiring AHT in a large cohort of general population.
Key messages
Preliminary evidence suggests that hypercholesterolemia could be a risk factor for the development of hypertension.
In a large cohort of general population we observed that better control of LDL-cholesterolemia is associated to a significantly lower incidence of new antihypertensive treatment.
Introduction
Hypertension and hypercholesterolemia are well-known independent and modifiable risk factors for cardiovascular disease (Citation1). Inhibitors of 3-hydroxy-3-methylglutaryl-coenzyme A reductase (statins) are the most effective and widely used cholesterol-lowering agents in industrialized countries (Citation2). Although the long-term benefit of statin therapy is largely attributed to their cholesterol-lowering action, additional effects of these drugs are thought to be involved in the cardiovascular protection observed shortly after the initiation of treatment (Citation3). Several lines of evidence support that statins may exert a small, but clinically relevant, antihypertensive effect. In a meta-analysis of 20 randomized placebo-controlled trials involving 828 patients, statin treatment was associated with a significant reduction in systolic (mean difference: –1.9 mmHg; 95% CI: –3.8 to –0.1 mmHg) and diastolic blood pressure (mean difference: –0.9 mmHg; 95% CI: –2.0 to –0.2 mmHg). The favorable effect of statins on blood pressure control has been recently confirmed by a larger meta-analysis of 51 prospective controlled studies involving more than 40,000 patients (Citation4) and showed a small but significant reduction in systolic blood pressure (–2.62 mmHg, 95% CI: –4.41 to –1.84; P > 0.001) in patients undergoing active treatment. The effect of statins proved to be more prominent among patients showing higher baseline levels of either systolic blood pressure (i.e. >130 mmHg) or diastolic blood pressure (i.e. > 80 mmHg) (Citation4–6). In the individual studies the blood pressure-lowering effect of statins emerges only when the baseline antihypertensive treatment is adequately matched, while in other less controlled settings the effect is more controversial, thereby supporting the lack of evidence in general population (Citation7). In particular, some studies (Citation8) have indirectly demonstrated the contribution of statins to blood pressure control by reporting a comparable antihypertensive effect in statin-treated patients despite less use of very effective antihypertensive drugs (i.e. diuretics, calcium channel blockers). Moreover the blood pressure-lowering effect of statins is attenuated when blood pressure is adequately controlled by antihypertensive drugs (Citation6,Citation9). Despite this intriguing evidence of an interaction between lipid-lowering treatment and blood pressure control, few studies have been carried out to date to investigate if the cholesterol reduction per se is able to interfere favorably with the new onset of hypertension in normotensive hypercholesterolemic patients in a clinical practice setting (Citation10,Citation11).
The aim of our study was to investigate the association (if any) between hypercholesterolemia and the short-term incidence of new onset of hypertension in a large cohort of hypercholesterolemic patients with normal blood pressure control at baseline.
Materials and methods
Sources of data
The present study has been carried out by using information entirely provided by the Italian National Health System (NHS) with the aim of achieving a reliable estimate of the impact of hypercholesterolemia on blood pressure control in real life. The public health in Italy is under the control of the NHS, a public system of universal character, which guarantees health care to all citizens. The NHS is mainly managed by regional administration through the local health authorities (LHU). LHUs purchase health care services according to the needs of their afferent population, and all the information collected through a network is stored in a dedicated database that routinely measures the volumes of expenditure incurred by the use of health services. The method proposed for the study is based on information retrieved by the linkage of records from the LHU database.
The Health-assisted Subjects’ Database containing patients’ demographic data, the Medications Prescription Database providing information for each outpatient prescription drug such as the name, the anatomical–therapeutic–chemical (ATC) code, the dosage, the quantity, and the prescription date, the Hospital Laboratory Investigation Database with dates and results of laboratory test, the Hospital Discharge Database containing all hospitalizations data with the discharge diagnosis codes classified according to the International Classification of Diseases, Ninth Revision (ICD-9-CM), and the Mortality Database with death data were selected.
Universal health care coverage in Italy allows completeness and comprehensiveness of the information contained in these databases, which have been used in previous epidemiological studies (Citation12).
In order to guarantee patients’ privacy each subject's code was transcoded into an anonymous univocal numeric code. No identifiers related to patients were provided to the researchers. The study complies with the Declaration of Helsinki. The local ethics committee of each LHU approved this study.
Cohort definition
We selected all patients who were 18 years or older, living in the area of the LHUs considered over the study period, with ≥ 1 low-density lipoprotein cholesterol (LDL-C) measurement between 1 July 2006 and 30 June 2007. The date of the last measurement identified in the period was considered as the beginning of observation for each individual (index date). Patients were classified into four groups according to the LDL-C baseline level: < 130 mg/dL, 130–159 mg/dL, 160–189 mg/dL, and ≥ 190 mg/dL (Citation13). The study was carried out as a retrospective cohort study with an unrestricted adjudication of the primary end-point by any physician involved and according to ESH-ESC Guidelines (Citation14).
Potential subjects had to be free of antihypertension treatment (AHT) at baseline and in the 6 months before the index date as well as free of cardiovascular (CV) and renal diseases that might justify the use of cardiovascular drugs for the treatment of diseases other than hypertension. Primary diagnosis of CV and renal disease were identified from claims data following the ICD-9-CM or by drug markers. Patients were excluded if they had hypertensive heart disease (ICD-9-CM codes 402, 404), myocardial infarction or angina (codes 410–414), stroke (codes 430–438), peripheral cardiac disease (codes 440–447), congestive heart failure (code 428), arrhythmia (code 427), renal disease (codes 580–589), or if they used any cardiovascular agent (nitrates or other cardiac agents).
Hypertension treatment assessment
As administrative databases do not contain information on blood pressure measurements, we decided to explore the problem of the relationship between hypercholesterolemia and hypertension according to a strictly conservative approach based on hypertensive drug use in detection of hypertension.
Indeed, the identification of the primary end-point based on the decision to start antihypertensive treatment according to guidelines would greatly reduce the confounding incidence of patients with transitory and/or episodic blood pressure increases apart from established hypertension (e.g. white coat hypertension). This study design would probably reduce the absolute number of subjects reaching the primary end-point, while it increases the power of the results and reduces the risk of a falsely positive observation.
We considered antihypertensive treatment the medication classified with the following groups of ATC codes: ATC C02, other antihypertensive treatments; C03, diuretics (excluded loop diuretics mainly used for hearth failure); C07, beta-blockers; C08, calcium channel blockers; and C09, agents that act on the renin-angiotensin system. Hypertension treatment was defined on the basis of the presence of at least one prescription of antihypertensive treatments from the date of the LDL-C value until death or the end of the follow-up period (31 December 2009).
Other measurements
Other baseline measures considered in this study included: age, sex, diabetes defined by ICD-9-CM code 250 or the use of insulin or oral antidiabetic drugs, and the following lipid parameters: total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C) (defined low if HDL-C < 40 mg/dL for men and HDL-C < 50 mg/dL for women), and triglycerides (TG) evaluated in the 6 months previous to the index date.
In order to select only patients with a stable LDL-C level, the most current LDL-C measurement during the observational period (the last before the start of the AHT, or death or 31 December 2009 in case of absence of AHT) was compared with baseline LDL-C level. Patients changing LDL-C level group were excluded in order to evaluate the effect of a constant LDL level on the new prescription of AHT.
Moreover, patients included had to have no history of drug treatment for lowering lipids, and be free of these drugs during the follow-up period. Thus, drugs for lowering cholesterol were not included in the present analysis.
Statistical analysis
The univariate data analysis was based on Pearson chi-square to assess statistical significance of differences between frequencies and rates and analysis of variance for the comparison of means. Incidence rates of antihypertensive treatment by LDL-C level were computed by dividing the number of new cases of antihypertensive treatment by the total number of years that occurred before the antihypertensive treatments prescription date for the new AHT cases, or the end of the follow-up period for no new antihypertensive treatment cases (Citation15).
Incidence curves were estimated by the Kaplan–Meier method and compared with the log-rank test.
In multivariable analysis, a Cox proportional hazard model was used to estimate hazard ratios (HR) and 95% confidence intervals (CI) of antihypertensive treatment as a function of LDL-C level (LDL-C level < 130 mg/dL as reference) (Citation16). The other covariates in the model were age, gender, the presence of diabetes, HDL-C (low level as reference), and TG (< 150 mg/dL as reference). Total cholesterol was not included in the model in order to avoid over-adjustment when analyzing the effect of LDL-C. The proportional hazards assumption was not violated.
In order to verify the robustness of the results, an additional multivariable model was performed including patients with LDL-C level ‘not stable’.
Differences were considered statistically significant with P < 0.05. All statistical analyses and graphs were performed using SPSS statistical software version 15.0 (IBM Inc., Chicago, IL, USA).
Results
Among a population of 160,101 subjects living in the area of the LHU considered, we included 20,074 patients with at least one LDL-C value (12.5% of the whole cohort) as shown in . The characteristics of the categories of varying levels of LDL-C are reported in . Patients with higher levels of LDL-C were significantly older, with a lower prevalence of diabetes. From lipid parameters, they had high levels of TC, the highest proportion of patients with elevated TG levels across LDL-C levels, and the lowest proportion of patients with low HDL-C levels.
Figure 1. Flow chart of the study inclusion/exclusion criteria.
Table I. Baseline characteristics of patients according to the LDL-C level.a
During the mean follow-up of 1.6 years (maximum 3.5 years), 10.7% (n = 1166) of patients with LDL-C < 130 mg/dL, 13.2% (n = 772) of patients with LDL-C between 130 and 159 mg/dL, 12.2% (n = 319) of patients with LDL-C between 160 and 189 mg/dL, and 13.9% (n = 102) of patients with LDL-C ≥ 190 mg/dL started antihypertensive treatment. displays the overall incidence rates per 100 person-years of follow-up along the categories of LDL-C level. The LDL-C ≥ 190 mg/dL group showed a higher overall incidence rate both in the total group (11.17 versus 6.03 for LDL-C < 130 mg/dL, 9.33 for LDL-C between 130 and 159 mg/dL, 9.60 for LDL-C level between 160 and 189 mg/dL per 100 person-years; P < 0.001) and in individual variables categories except for the presence of diabetes group (4.53 versus 17.49, 14.32, 25.68 per 100 person-years; P < 0.001), low level of HDL-C group (11.43 versus 7.29, 10.76, 11.72 per 100 person-years; P < 0.001), and TG ≥ 150 mg/dL (9.92 versus 9.49, 10.87, 11.72 per 100 person years; P = 0.356), but with no significant difference. The incidence rates of antihypertensive treatment increased with age and were higher in diabetic patients, in patients with TG ≥ 150 mg/dL, and in those with low level of HDL-C.
Table II. Incidence of antihypertensive treatment per 100 person-years by LDL-C level.a
Kaplan–Meier analysis showed an increased probability of incidence of AHT during follow-up in patients with LDL-C ≥ 190 mg/dL ().
Figure 2. Incidence of antihypertensive treatment per 100 person-years by LDL-cholesterol level.
After adjustments for the potential confounding variables () there was a significant positive association between LDL-C level and incidence of antihypertensive treatment: compared with the LDL-C < 130 mg/dL group, the antihypertensive treatment risk increased among those with LDL-C level between 130 and 159 mg/dL (HR 1.23; 95% CI: 1.08–1.40), those with LDL-C level between 160 and 189 mg/dL (HR 1.24; 95% CI: 1.01–1.51), and those with LDL-C ≥ 190 mg/dL (HR 1.45; 95% CI: 1.11–1.89). Significant HRs were also observed for age—increasing age increases the risk of new cases of AHT compared to the age group below 65 years (HR 2.82; 95% CI: 2.59–3.06)—and diabetes (HR 1.62; 95% CI: 1.39–1.88). TG ≥ 150 mg/dL also had a higher risk of incidence of AHT compared with those with TG < 150 mg/dL; whereas high level of HDL-C showed a reduced risk of new AHT cases compared with low HDL-C level. The results were confirmed when patients with LDL-C ‘not stable’ were included in the multivariable model ().
Table III. Multivariable analysis of the association between patients’ level of LDL-C and antihypertensive treatment.a
Table IV. Multivariable analysis of the association between patients’ level of LDL-C and antihypertensive treatment including patients with a ‘not stable’ LDL-C level.a
Discussion
In our study carried out on a large population sample including 20,074 patients with at least one registered LDL-C value, we observed that the incidence of new-onset hypertension based on the initiation of AHT was significantly lower in those subjects with better control of LDL-C. These findings reasonably support the concept that plasma lipid disturbances may be a significant risk factor for the new onset of hypertension, while lipid-lowering treatment can contribute to prevent the increase in blood pressure.
These data are in agreement with those just observed in the Brisighella Heart Study cohort, where cholesterol-lowering treatment was, per se, associated to a significant improvement in blood pressure control (Citation17).
Several data from the literature suggest that LDL-C leads to an increase in the stability for mRNA for AT1 receptors for angiotensin II (AT-II), leading to an increase in the density of AT1 receptors, which has been demonstrated in the platelet and at the vascular levels and is proportional to the plasma levels of LDL-C (Citation18). Moreover these overexpressed receptors have demonstrated a higher affinity for binding to AT-II when compared to normal cholesterol levels and without any concomitant involvement of AT2 receptors (Citation19). At the same time hypercholesterolemia stimulates the synthesis of several peptides of the angiotensin family and in particular of AT-II whose production in the presence of LDL-C has been described as predominantly dependent on the chymase system with only a minor involvement of angiotensin-converting enzyme (ACE) (Citation20). On the other side of the same system, AT-II promotes the oxidation of the LDL-C, and the increased availability of oxidized LDL contributes to promoting the expression of AT1 receptor proteins, which results again in an increase in the AT1 receptor density at the tissue level that is significantly enhanced when compared to native LDL-C (Citation21). The overexpression of AT1 receptors is associated with an enhanced vasoconstrictive and blood pressure response to AT-II infusion that is proportional to serum cholesterol levels and can be significantly blunted by statin treatment in concomitance with a significant decline in AT1 receptor density (Citation22,Citation23). Prospective data in humans support the hypothesis that this chronic up-regulation of AT-II receptors determines a higher new incidence of hypertension in hypercholesterolemic patients compared with normocholesterolemic subjects, especially in subjects with high plasma renin activity (Citation24).
Our observation does not support the hypothesis that statin treatment, per se, could have a positive impact on blood pressure control, independently from cholesterol reduction (Citation25). This is a variance of the results of the meta-analysis of Strazzullo et al. (Citation5) showing that the blood pressure-lowering effect of statins is unrelated to the cholesterol reduction. However, since the effect of statins on blood pressure is strongly related to the baseline cholesterol level as well as to the extent of blood pressure reduction by concomitant antihypertensive treatment (Citation6), a simple evaluation of the effects of statins might not provide reliable information about the relationship between cholesterol control and blood pressure decrease. However, although our results suggest a correlation exists between cholesterol levels and new onset of hypertension, the design is such that no inference can be made on therapeutic approaches so far. Thus, further randomized clinical trials specifically investigating the impact of cholesterolemia control on hypertension incidence are needed (Citation26).
Our study has several limitations. First of all, as administrative databases do not contain data on blood pressure measurements, we used antihypertensive drug prescriptions to define cases of hypertension. With this approach we would miss those patients for whom therapy has not been prescribed, but we would also avoid those who are not hypertensive but only affected with transiently elevated blood pressure levels. Secondly, the lack of baseline blood pressure measurements did not allow us to take this parameter into account in the multivariable model. However, Halperin et al. found that higher lipid levels were independently associated with an increased risk of incidence of hypertension and that the results were confirmed when they considered a second multivariable model that additionally controlled for baseline blood pressure levels to consider the impact of lipids on the risk of hypertension through mechanisms other than blood pressure (Citation27). Thirdly, our observation is based on a limited number of mostly nominal parameters, and, given the retrospective nature of the study, we are unable to include some major risk factors for hypertension in the causal model, such as for instance obesity and smoking. However, the study has been carried out in an unselected population of subjects where the major risk factors for hypertension (e.g. salt intake, overweight and obesity, physical inactivity, etc.) are uniformly distributed in patients with and without LDL-C abnormalities. Fourthly, we are aware that some antihypertensive medication, as defined in the administrative databases, could have been prescribed for clinical indication other than hypertension. To reduce the probability of an inappropriate identification of hypertensive patients we rely on the observation that drug markers have been used previously with good correlation with the diagnosis of hypertension (Citation28). In addition, in our study the patients with known cardiovascular diseases were excluded from the study cohort to dilute any possible classification bias. This does not exclude the possibility that in patients with other conditions errors may arise through miscoding. Then, it is possible that general practitioners reinforced either antihypertensive therapy in patients in secondary prevention and in diabetics because of their higher risk: this partly explains why the regression analysis shows diabetes as predictor of incident antihypertensive treatment. Besides, the presence of diabetes was just known at the beginning of the study, and its treatment should be theoretically established irrespectively of lipid control. Therefore, this is an original observation based on the observation of administrative health data derived from a large official database, and the results are in agreement with the previous literature.
To the best of our knowledge, this is the one of the few studies investigating the effect of cholesterolemia control on new onset of hypertension based on AHT prescription in a large population sample and in the setting of daily practice. Our evidence supports the finding that an adequate control of LDL-C is associated to a lower incidence of new-onset hypertension and to significantly less prescription of new antihypertensive drugs over an average observation interval of 19 months.
Declaration of interest: The authors declare that they have no competing interests in the publication of this paper.
References
- Ferdinand KC, Kleinpeter MA. Management of hypertension and dyslipidemia. Curr Hypert Rep. 2006;8:489–96.
- Mills EJ, Wu P, Chong G, Ghement I, Singh S, Akl EA, et al. Efficacy and safety of statin treatment for cardiovascular disease: a network meta-analysis of 170,255 patients from 76 randomized trials. QJM. 2011; 104:109–24.
- Martínez-González J, Badimon L. Influence of statin use on endothelial function: from bench to clinics. Curr Pharm Des. 2007;13:1771–86.
- Briasoulis A, Agarwal V, Valachis A, Messerli FH. The BP lowering antihypertensive effects of statins: a meta-analysis of prospective controlled studies. J Clin Hypertens. 2013;15:310–20.
- Strazzullo P, Kerry SM, Barbato A, Versiero M, D’Elia L, Cappuccio FP. Do statins reduce blood pressure? A meta-analysis of randomized, controlled trials. Hypertension. 2007;49:792–8.
- Prandin MG, Cicero AF, Dormi A, Cicero AF, Kuria F, Dormi A, et al. Prospective evaluation of the effect of statins on blood pressure control in hypertensive patients in clinical practice. Nutr Metab Cardiovasc Dis. 2010;20:512–18.
- Correa V Jr, Gus M, Fuchs FD. Does the blood pressure-lowering effect of statins contribute to their beneficial cardiovascular effects?. Expert Rev Cardiovasc Ther. 2010;8:775–9.
- Mancia G, Crepaldi G, Gallus G, Parati G, Veglia F, Bilo G, et al. Long-term statin administration and ambulatory blood pressure in the mild hypertensive and hypercholesterolemic patients: the PHYLLIS study. J Hypertens. 2002;20:S4
- Mancia G, Parati G, Revera M, Bilo G, Giuliano A, Veglia F, et al. Statins, antihypertensive treatment, and blood pressure control in clinic and over 24 hours: evidence from PHYLLIS randomised double blind trial. BMJ. 2010;340:c1197.
- Freitas MPD, de Loyola Filho AI, Lima-Costa MF. Dyslipidemia and the risk of incident hypertension in a population of community-dwelling Brazilian elderly: the Bambuí cohort study of aging. Cad Saude Publica. 2011;27:351–9.
- Laaksonen DE, Niskanen L, Nyyssonen K, Lakka TA, Laukkanen JA, Salonen JT. Dyslipidaemia as a predictor of hypertension in middle-aged men. Eur Heart J. 2008;29:2561–8.
- Degli Esposti L, Saragoni S, Batacchi P, Benemei S, Geppetti P, Sturani A. Adherence to statin treatment and health outcomes among newly treated patients: results from an administrative database analysis. Clin Ther. 2012;34:190–9.
- Shalev V, Chodick G, Silber H, Kokia E, Jan J, Heymann AD. Continuation of statin treatment and all-cause mortality: a population-based cohort study. Arch Intern Med. 2009;169:260–8.
- Mancia G, De Backer G, Dominiczak A, Cifkova R, Fagard R, Germano G, et al.; Management of Arterial Hypertension of the European Society Of Hypertension; European Society of Cardiology. 2007 Guidelines for the management of arterial hypertension: The Task Force for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens. 2007;25:1105–87.
- Sun Z, Zheng L, Detrano R, Zhang X, Xu C, Li J, et al. Incidence and predictors of hypertension among rural Chinese adults: results from Liaoning province. Ann Fam Med. 2010;8:19–24.
- Marubini E, Valsecchi MG, Emmerson M. Analysing survival data from clinical trials and observational studies. New York: Wiley; 1995. p. 103–59.
- Borghi C, Dormi A, Veronesi M, Sangiorgi Z, Gaddi A; Brisighella Heart Study Working Party. Association between different lipid-lowering treatment strategies and blood pressure control in the Brisighella Heart Study. Am Heart J. 2004;148:285–92.
- Strehlow K, Wassmann S, Böhm M, Nickenig G. Angiotensin AT1 receptor over-expression in hypercholesterolaemia. Ann Med. 2000;32: 386–9.
- Nickenig G, Wassmann S, Böhm M. Regulation of the angiotensin AT1 receptor by hypercholesterolaemia. Diabetes Obes Metab. 2000; 2:223–8.
- Murakami K, Uehara Y, Abe S, Inoue Y, Ideishi M, Saku K, et al. Positive correlation between chymase-like angiotensin II-forming activity in mononuclear cells and serum cholesterol level. J Cardiol. 2007;50: 291–8.
- Niebauer J, Tsao PS, Lin PS, Pratt RE, Cooke JP. Cholesterol-induced upregulation of angiotensin II and its effects on monocyte-endothelial interaction and superoxide production. Vasc Med. 2001;6:133–8.
- van der Harst P, Wagenaar LJ, Buikema H, Voors AA, Plokker HW, Morshuis WJ, et al. Effect of intensive versus moderate lipid lowering on endothelial function and vascular responsiveness to angiotensin II in stable coronary artery disease. Am J Cardiol. 2005;96:1361–4.
- Nickenig G, Bäumer AT, Temur Y, Kebben D, Jockenhövel F, Böhm M. Statin-sensitive dysregulated AT1 receptor function and density in hypercholesterolemic men. Circulation. 1999;100:2131–4.
- Borghi C, Veronesi M, Cosentino E, Cicero AF, Kuria F, Dormi A, et al. Interaction between serum cholesterol levels and the renin-angiotensin system on the new onset of arterial hypertension in subjects with high-normal blood pressure. J Hypertens. 2007;25:2051–7.
- Mihos CG, Salas MJ, Santana O. The pleiotropic effects of the hydroxy-methyl-glutaryl-CoA reductase inhibitors in cardiovascular disease: a comprehensive review. Cardiol Rev. 2010;18:298–304.
- Borghi C. Interactions between hypercholesterolemia and hypertension: implications for therapy. Curr Opin Nephrol Hypertens. 2002;11: 489–96.
- Halperin RO, Sesso HD, Ma J, Buring JE, Stampfer MJ, Gaziano JM. Dyslipidemia and the risk of incident hypertension in men. Hypertension. 2006;47:45–50.
- Glynn RJ, Brock DB, Harris T, Havlik RJ, Chrischilles EA, Ostfeld AM, et al. Use of antihypertensive drugs and trends in blood pressure in the elderly. Arch Intern Med. 1995;155:1855–60.