| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Hypertension. 2007;49:792.)
© 2007 American Heart Association, Inc.
Original Articles |
From the Department of Clinical and Experimental Medicine (P.S., A.B., M.V., L.D.), Federico II University Medical School, Naples, Italy; the Division of Community Health Sciences (S.M.K.), St Georges, University of London, London, United Kingdom; and the Clinical Sciences Research Institute (F.P.C.), Warwick Medical School, Coventry, United Kingdom.
Correspondence to Pasquale Strazzullo, Department of Clinical and Experimental Medicine, "Federico II" University of Naples Medical School, Via S. Pansini, 5, 80131 Naples, Italy. E-mail strazzul{at}unina.it
| Abstract |
|---|
|
|
|---|
systolic blood pressure: 4.0; 95% CI: 5.8 to 2.2 mm Hg). There was a trend for lower diastolic blood pressure in patients receiving statin therapy compared with control: 0.9 mm Hg (95% CI: 2.0 to 0.2) overall and 1.2 mm Hg (95% CI: 2.6 to 0.1) in studies with a baseline diastolic blood pressure >80 mm Hg. In general, the higher the baseline blood pressure, the greater the effect of statins on blood pressure (P=0.066 for systolic blood pressure and P=0.023 for diastolic blood pressure). The blood pressure response to statins was unrelated to age, changes in serum cholesterol, or length of the trial. In conclusion, statin therapy has a relatively small but statistically significant and clinically meaningful effect on blood pressure.
Key Words: blood pressure hypertension, arterial hypercholesterolemia drug therapy statins meta-analysis
| Introduction |
|---|
|
|
|---|
Very few relatively small studies have investigated the antihypertensive effect of statins in patients with hypertension associated with hypercholesterolemia.810 The results of 2 recently published large statin studies, albeit not designed to answer this question, have attracted the interest on this subject.11,12 Many other studies, also not specifically aimed at the evaluation of the statins antihypertensive effect, have provided information concerning changes in BP during treatment with statins. Present knowledge is hampered by severe limitations, such as inadequate study design, small or very small sample size, too short treatment period, and modification of concomitant antihypertertensive therapy during the trial. Nevertheless, an effect of statins on BP is potentially important and not implausible bearing in mind the reported effects of statins on endothelial function,13 their interaction with the reninangiotensin system,14 and their ability to affect large artery compliance.15,16
We systematically reviewed the existing literature of randomized trials of statin therapy. We carried out a meta- analysis of the effect of statins on systolic and diastolic BP including only those trials in which concomitant antihypertensive therapy, if any, was kept constant during the trial.
| Methods |
|---|
|
|
|---|
Inclusion and Exclusion Criteria
We included only randomized, controlled trials published as original articles in peer-reviewed scientific journals in English provided that they used either an identical placebo or a different hypolipidemic drug. We excluded trials where we could not extract or calculate the difference between baseline and end-of-treatment systolic and diastolic BP in intervention and control groups and those that did not report any of the following variables: number of patients in both the statin and control group, length of study, description of the main relevant features of the study population, including gender, age, hypertensive status, and description of concomitant therapy, if any. If an antihypertensive treatment was given during the trial period, the study was included only if a statement that antihypertensive drug therapy (in terms of number, type, and dosage of drugs) was kept unchanged during the trial was clearly made.
Eighteen randomized, controlled trials met the inclusion criteria and were included in the meta-analysis9,10,1530 (Figure 1). Because 2 of them,25,26 however, were characterized by having 2 different pairs of control and treatment groups, we considered them as being split in 2 independent studies; thus, the final number of studies analyzed was actually 20.
|
Data Extraction
Two reviewers (A.B. and M.V.) independently extracted data from text and tables. Discrepancies about inclusion of studies and interpretation of data were resolved on arbitration (P.S.), and consensus was reached after discussion.
Statistical Methods
We used a random effect model (Stata Corp). For parallel groups trials, the intervention effect was the between-group difference in the extent of BP change from baseline. In 4 studies, the confidence intervals of the change in BP were estimated from the P value,9,16,26 and for 11 studies15,17,1922,24,2730 SDs of the BP change were estimated by assuming a correlation coefficient between baseline and follow-up BP of 0.5.31 For crossover trials, the outcome measure was given by the difference between the poststatin and the postplacebo BP. Also in this case, if the SD of the difference was not reported, we applied the same principle as for parallel group trials to the postplacebo and poststatin SD to estimate the SD of the difference. All of the trials, whether parallel groups or crossover, were analyzed together using the intervention effect and its SE for each study.
Potential publication bias was assessed by using a funnel plot and Eggers test.32,33 The degree of heterogeneity between trials was tested by
2. Meta-regression was used to assess whether baseline BP, length of treatment, hypertensive or diabetic status, and serum cholesterol response to statin treatment were associated with the effect of statin therapy on BP. In testing the relation of the BP effect of statin to baseline BP, we included the studies in which patients were on antihypertensive treatment at the time of recruitment, and their antihypertensive therapy was kept unchanged throughout the trial; we did not include in this analysis 2 trials28,29 in which patients were prescribed antihypertensive therapy at the same time that treatment with statins was started.
| Results |
|---|
|
|
|---|
Some studies included only hypertensive patients,10,16,17,23,28,29 others recruited either normotensive or treated hypertensive patients in reasonably good control,22,30 and a few studies included only normotensive individuals.22,23,28,30 Four studies included only type 1 or only type 2 diabetic patients,19,24,27,30 whereas 9 studies excluded patients affected by diabetes9,10,15,17,18,20,21,23,29 (Table S3).
In 7 studies, none of the participants received drugs different from statins or control medications.15,21,23,25,26 In all of the others, the patients had a variety of concomitant medication regimes (Table) and, in particular, antihypertensive treatment: in 2 such studies,28,29 antihypertensive therapy was started at the same time as statin treatment. In all of the cases, however, antihypertensive therapy was kept unchanged throughout the trial. Ten studies gave instructions for patients diet.10,17,19,20,2326 In 1 study, the patients previous diets were reviewed.13 Eight studies did not define a particular diet profile.9,16,21,22,2730
|
Most studies included only hypercholesterolemic patients (serum total cholesterol
5.2 mmol/L),912,1827,29 and 2 studies included subjects with low-density lipoprotein cholesterol >2.6 mmol/L.2830 One study enrolled normocholesterolemic subjects,16 whereas 2 studies did not use specified inclusion criteria for serum cholesterol.15,17 Details on blood lipid levels, inclusion criteria, and nonpharmacological treatments adopted (diet and lifestyle suggestions) in the single trials are reported in the Table.
Estimate of the BP Effect of Statins in the Trials Included in the Meta-Analysis
Systolic BP
The overall intervention effect for statin therapy on systolic BP (SBP) was a 1.9-mm Hg reduction (95% CI: 3.8 to 0.1), with significant heterogeneity between studies (P<0.01; Figure 2). The funnel plot did not show asymmetry consistent with publication bias, and Eggers test was not significant (P=0.58).
|
Diastolic BP
For diastolic BP (DBP), the effect of statin therapy in all of the studies was in the same direction as for SBP but less marked: 0.9 mm Hg (95% CI: 2.0 to 0.2 mm Hg; Figure 2), with significant heterogeneity between studies (P<0.01). There was no evidence of publication bias (Eggers test P=0.48).
Meta-Regression and Subgroup Analysis
The effect of statin therapy on BP was greater in studies that recruited patients with higher baseline BPs (P=0.023 for diastolic and P=0.066 for systolic BP; Figure 3). When the analysis was restricted to studies where baseline SBP was >130 mm Hg, the effect of statin therapy was 4.0 mm Hg (95% CI: 5.8 to 2.2 mm Hg), and heterogeneity was not statistically significant (P=0.26). In studies with average baseline DBP >80 mm Hg, the effect of statin therapy on DBP was 1.2 mm Hg (95% CI: 2.6 to 0.1 mm Hg; heterogeneity P=0.069; Figure 2). In those trials where SBP at baseline was
130 mm Hg and/or DBP
80 mm Hg, the average net effect of statin on systolic and diastolic BP, respectively, was negligible. In 4 trials that enrolled diabetic hypercholesterolemic patients, the weighted combined effect of statins on BP was 3.7 mm Hg (95% CI: 8 to 1.5 mm Hg)/0.8 mm Hg (95% CI: 1.9 to 0.4 mm Hg).
|
Using meta-regression analysis, no evidence was obtained that any of the following factors were significantly related to the response to statin therapy: age (P=0.86 for SBP and 0.79 for DBP), length of the trial (P=0.79 and 0.66, respectively), baseline serum cholesterol (P=0.58 and 0.91), change in serum total cholesterol (P=0.51 and 0.86), or low-density lipoprotein cholesterol (P=0.40 and 0.69), whether patients were diabetic or not (P=0.50 and 0.51) and whether antihypertensive drugs were being used by at least some patients (P=0.53 and 0.98).
Results From Other Trials Not Included in the Meta-Analysis
Two studies, composed of a total number of 64 patients, could not be included in the meta-analysis, because they provided only mean BP values at the start and end of treatment.34,35 The net effect of statin therapy on BP in these 2 studies was a mean BP reduction of 4.0 mm Hg (95% CI: 19.6 to 11.6) and of 2.0 mm Hg (95% CI: 19.9 to 15.9), respectively. The estimated combined effect in the 2 studies was a 3.1-mm Hg fall in mean BP (95% CI: 14.9 to 8.6).
| Discussion |
|---|
|
|
|---|
None of the large trials that highlighted the value of statin treatment in primary and secondary prevention of cardiovascular disease could be included in our meta-analysis, because BP values were not reported and/or concomitant antihypertensive treatment in participants with high BP was not kept constant during the study (Table S4). A recent communication (published as abstract) from the Anglo-Scandinavian Cardiac Outcomes Trial showed that between the week-6 visit and 2.5-year visit, BP was significantly lower in patients on atorvastatin than in those on placebo, with a maximum difference of 1.1/0.7 mm Hg.12 Similarly, in the University of California at San Diego Statin Study,11 both during and at the end of the trial, the patients groups treated with simvastatin or pravastatin had a significantly lower SBP and DBP compared with the placebo groups, with a maximum difference of 2.8/2.7 mm Hg. In both of these studies, the participants were prescribed antihypertensive medications that could be titrated at any time during the trial. Although both trials were, thus, not eligible for our meta-analysis, their results appear to be consistent with those of our study.
The BP effect of statins detected by our analysis was definitely lower than the average effect of regular antihypertensive drugs and reached statistical significance only for systolic BP; yet, it was clinically relevant, also considering that in most of the studies included in our meta-analysis the average BP at entry was lower than that observed in most trials of antihypertensive therapy and that, in addition, in several studies, the patients were already placed on regular treatment for high BP. In the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial, a 2 mm Hg lower on-trial systolic BP in the chlorthalidone compared with the lisinopril arm was associated with a 15% lower rate of stroke and a 19% lower rate of congestive heart failure.36 Likewise, in the Hypertension Optimal Treatment Study, a 4/4 mm Hg difference in BP was associated with a 50% reduction in the rate of cardiovascular events in diabetic participants.37
Limitations
The patients population in whom the analysis was carried out was not very large, because large-scale studies of statin therapy did not provide BP values at the start and end of the trial or did not conform to our necessarily strict inclusion criteria with respect to concomitant use of antihypertensive medications or both. Also, the studies included in this quantitative review were rather heterogeneous, because they were carried out in a variety of settings, with different methods, using various criteria and different comparative groups. Although we selected only randomized, controlled trials, few of the studies had been specifically designed to test the hypothesis evaluated by our meta-analysis. The major parts of the studies were conducted on hypercholesterolemic patients. We were unable to examine a possible doseeffect relationship or possible differences in the BP effects of different statins given the relatively small number of trials and the limited sample size available. Many studies required an estimation of the SD of the effect size; in 4 studies, the P value was used, and this may have resulted in a slight overestimate of the SD, possibly reducing the significance of the effect size. For 1 study that provided both intention-to-treat and on-treatment analysis, the intention-to-treat analysis was used; also, this choice may have led to a conservative estimate of the effect.
Possible Mechanisms of the BP Effect of Statins
There are several mechanisms whereby statins may affect BP. In experimental studies, statins increase the endothelial production of NO, an effect correlated with upregulation of endothelial NO synthase expression.3841 This effect may be reinforced by simultaneous inhibition of G proteins with reduced endothelial NO synthase mRNA degradation and, thus, increased NO bioavailability.42 Moreover, it may also be reinforced by a decrease in the release and circulating levels of the vasoconstrictor endothelin-1.43 These effects are translated in substantial improvements of endothelial function, assessed by vasodilator response to acetylcholine or as flow- mediated vasodilation, in various pathological conditions characterized by marked endothelial dysfunction.4450 These various effects manifest themselves early after statin administration and, thus, appear substantially unrelated to their cholesterol-lowering action.12,51 This notion is in keeping with the result of our meta-regression analysis indicating no significant association with the cholesterol response and no relationship with length of treatment.
Statins inhibit the production of reactive oxygen species, such as superoxide anion, and hydroxyl radicals in animal vessel models, and also this action can contribute to vasodilation.52 Statins have been shown to reduce large artery stiffness and improve systemic arterial compliance,15,16 an effect ascribed to alterations in the relative content of vascular smooth muscle cells in large arteries and to restoration of endothelial function.15 This mechanism could be expected to affect, in particular, SBP, which is in line with our results.
A further possible mechanism of the BP-lowering effect of statins is downregulation of the angiotensin IItype 1 receptor. The angiotensin IItype 1 receptor is overexpressed in hypercholesterolemic patients, and this alteration was corrected by administration of statins,14 which also markedly reduced the vasoconstrictor response to angiotensin II infusion.14
| Conclusions |
|---|
|
|
|---|
Perspectives
This study provided the first demonstration of a statistically significant effect of statin therapy on BP. Its main feature was the inclusion in the meta-analysis of only randomized, controlled trials in which concomitant antihypertensive treatment (if any) was kept constant throughout the study. Its most important limitation, on the other hand, was the relatively small patients population enrolled. Also, the study results derived mainly, although not exclusively, from the analysis of trials carried out in hypercholesterolemic patients. It remains to be clarified whether the BP effect of statins is a class effect or is a property of 1 or few substances in the class.
Although the effect on BP was moderate and statistically significant only for systolic pressure, it may be, nevertheless, relevant from the point of view of cardiovascular prevention. It can be speculated that, in hypertensive patients in whom the prescription of a statin is indicated (eg, because of concomitant hypercholesterolemia, diabetes, and/or previous myocardial infarction), this might reduce, to some extent, the dose and number of drugs required to achieve satisfactory BP control. Further studies to test the extent and the mechanism(s) of the BP effect of statins in these and other selected categories of very high-risk patients are warranted.
| Acknowledgments |
|---|
None.
Received November 30, 2006; first decision December 21, 2006; accepted January 26, 2007.
| References |
|---|
|
|
|---|
formation. J Cardiovasc Pharmacol. 2002; 40: 868874.[CrossRef][Medline]
[Order article via Infotrieve]This article has been cited by other articles:
![]() |
B. Williams The year in hypertension. J. Am. Coll. Cardiol., May 6, 2008; 51(18): 1803 - 1817. [Full Text] [PDF] |
||||
![]() |
B. A. Golomb, J. E. Dimsdale, H. L. White, J. B. Ritchie, and M. H. Criqui Reduction in Blood Pressure With Statins: Results From the UCSD Statin Study, a Randomized Trial Arch Intern Med, April 14, 2008; 168(7): 721 - 727. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. M. Ho, D. J. Magid, S. M. Shetterly, K. L. Olson, P. N. Peterson, F. A. Masoudi, and J. S. Rumsfeld Importance of Therapy Intensification and Medication Nonadherence for Blood Pressure Control in Patients With Coronary Disease Arch Intern Med, February 11, 2008; 168(3): 271 - 276. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Korantzopoulos, M. Elisaf, and H. J. Milionis Multifactorial intervention in metabolic syndrome targeting at prevention of chronic kidney disease ready for prime time? Nephrol. Dial. Transplant., October 1, 2007; 22(10): 2768 - 2774. [Full Text] [PDF] |
||||
![]() |
K. K. Koh Combination Treatment to Prevent Atherosclerosis Hypertension, September 1, 2007; 50(3): e67 - e67. [Full Text] [PDF] |
||||
![]() |
P. Strazzullo and L. D'Elia Response to Combination Treatment to Prevent Atherosclerosis Hypertension, September 1, 2007; 50(3): e68 - e68. [Full Text] [PDF] |
||||
![]() |
From the Library Br. J. Ophthalmol., August 1, 2007; 91(8): 1098 - 1098. [Full Text] [PDF] |
||||
![]() |
E. Cereda and A. Pietrobelli Statins and Blood Pressure Lowering: A Claim for Ad-Hoc Trials Hypertension, August 1, 2007; 50(2): e12 - e12. [Full Text] [PDF] |
||||
![]() |
P. Strazzullo and A. Barbato Response to Statins and Blood Pressure Lowering: A Claim for Ad-Hoc Trials Hypertension, August 1, 2007; 50(2): e13 - e13. [Full Text] [PDF] |
||||
![]() |
M.-S. Zhou Upregulation of Nitric Oxide, Inhibition of Oxidative Stress, and Antihypertensive Effects of Statins Hypertension, June 1, 2007; 49(6): e43 - e43. [Full Text] [PDF] |
||||
![]() |
P. Strazzullo, L. D'Elia, and M. Versiero Response to Upregulation of Nitric Oxide, Inhibition of Oxidative Stress, and Antihypertensive Effects of Statins Hypertension, June 1, 2007; 49(6): e44 - e44. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||