Statins for preventing major vascular events in people with hypertension
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
Primary Objective
To determine if statins, as compared to placebo or no treatment, reduce major vascular endpoints (such as total and cardiovascular mortality, stroke and myocardial infarction) in patients with elevated blood pressure.
Secondary Objectives
To determine the magnitude of the effect of statins on systolic and diastolic blood pressure in patients with elevated blood pressure.
To determine the magnitude of the effect of statin on plasma lipids (total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides) in patients with elevated blood pressure.
To determine the rate of withdrawal due to adverse effects with statins as compared to placebo or no treatment in patients with elevated blood pressure.
Background
Several epidemiological studies have already shown that the risk factors, elevated blood pressure and elevated cholesterol, often coexist in the same individual (Sander 2002; McMurray 1999). Elevated blood pressure and cholesterol are synergistic risk factors for cardiovascular events and both show a degree of cross‐correlation possibly by sharing mechanisms of pathogenesis, e.g. insulin resistance and endothelial dysfunction (Ferrario 2002). Statins are the most widely used drugs for the treatment of elevated cholesterol and have been reported to interact with elevated blood pressure (Borghi 2001; Tsiara 2003), an effect that has not been demonstrated for other lipid‐lowering drugs. The mechanisms responsible for the hypotensive effect seem to be largely independent of the effect of statins on the lipid profile, and are probably related to an interaction with endothelial function, vascular tone or angiotensin II receptors (Park 2000; Wierzbicki 2002).
The capacity of statins to improve blood pressure control could be a useful consideration for an integrated approach to better prevention of cardiovascular disease in patients with elevated blood pressure (Green 2002). Different effects of antihypertensive drugs on plasma lipid levels have been reported. For example, thiazide diuretics may cause a dose‐dependent increase in cholesterol. On the contrary, angiotensin converting enzyme inhibitors appear to have no important effect on plasma cholesterol, and also may reduce the rise in cholesterol induced by diuretics. To date is not clear whether the benefit of statins can be extended to a wide range of individuals with elevated blood pressure. This systematic review is designed to search, critically appraise and analyse the available evidence regarding the effects of statins on mortality, morbidity and blood pressure in individuals with elevated blood pressure.
Objectives
Primary Objective
To determine if statins, as compared to placebo or no treatment, reduce major vascular endpoints (such as total and cardiovascular mortality, stroke and myocardial infarction) in patients with elevated blood pressure.
Secondary Objectives
To determine the magnitude of the effect of statins on systolic and diastolic blood pressure in patients with elevated blood pressure.
To determine the magnitude of the effect of statin on plasma lipids (total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides) in patients with elevated blood pressure.
To determine the rate of withdrawal due to adverse effects with statins as compared to placebo or no treatment in patients with elevated blood pressure.
Methods
Criteria for considering studies for this review
Types of studies
Study design should meet the following criteria: randomised controlled trials comparing statins with placebo or no treatment. Baseline characteristics must be recorded, morbidity and mortality endpoints must be clearly defined and outcomes must be reported outcomes using the intention‐to‐treat principle.
Types of participants
Men and (or) women, age greater than 18 years, from all races, with baseline LDL‐cholesterol levels greater than 2.6 mmol/L (100 mg/dL) and with a baseline blood pressure of at least 140 mm Hg systolic or a diastolic blood pressure of at least 90 mm Hg measured in a standard way on at least 2 occasions. Trials must be limited to patients with elevated blood pressure or separately report outcome data on patients with elevated blood pressure as defined above.
Types of interventions
The experimental treatment: all doses of HMG CoA reductase inhibitors [statins] (e.g. atorvastatin, simvastatin, pravastatin, fluvastatin, lovastatin, cerivastatin, rosuvastatin).
The comparison intervention could include either placebo or no treatment.
Types of outcome measures
We plan to extract information on the following outcome measures:
-
Total mortality
-
Cardiovascular mortality
-
Total stroke including fatal and non‐fatal strokes
-
Total myocardial infarction including fatal and non‐fatal myocardial infarction
-
Change in systolic and diastolic blood pressure
-
Change in plasma lipids (total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides)
-
Patient withdrawal due to adverse events
-
Compliance with medication.
Search methods for identification of studies
Electronic searches
A comprehensive search of the Cochrane Controlled Trials Registry (CCTR/CENTRAL), MEDLINE (1966 to the present), and EMBASE (1974 to the present) will be completed. There will be no language or publication restrictions and no publication status restrictions. The highly sensitive search strategy to identify RCTs with additional terms relevant to the condition and intervention of interest will be used to identify the relevant articles. In case of incomplete reports, further searches will be done for connected papers, or authors will be contacted to retrieve missing information.
Cochrane Central Register of Controlled Trials (CENTRAL) search strategy
1. HYDROXYMETHYLGLUTARYL‐COA REDUCTASE INHIBITORS explode all trees
2. (hydroxymethylglutaryl‐coa next reductase next inhibitors)
3. (((hmg next coa) or hmg‐coa or (hmg next co‐a)) and (reductase next inhibitor*))
4. statin*
5. fluvastatin
6. simvastatin
7. pravastatin
8. lovastatin
9. cerivastatin
10. atorvastatin
12. rosuvastatin
12. (#1or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12)
13. HYPERTENSION explode all trees
14. hypertension
15. (#13 or #14)
16. (#12 and #15)
See Appendix 1 for the MEDLINE search strategy.
Searching other resources
Reference lists of all available primary studies and review articles will be reviewed to identify additional relevant citations. Experts in the field will be contacted to know about ongoing studies or trials about to be published . We will screen abstracts of major cardiology and hypertension meetings (European Society of Cardiology, American Heart Association, American College of Cardiology) or other 'grey literature' using ISTP on the Web of Science, British Library INSIDE, ZETOC and/or SIGLE. Pharmaceutical firms will also be asked whether uncompleted or unpublished studies may have been overlooked. We will also search online registers of ongoing trials (e.g. www.current‐controlled‐trials.com).
Data collection and analysis
Selection of studies
In Phase I, all trials which appear relevant on the basis of 'Title', 'Abstract', and 'MeSH Headings' will be selected for full review by two independent reviewers (RK, OB) using an in/out sheet. Articles will only be rejected on initial screen if it can be determined from the title or the abstract that the article is not a report of a randomised controlled trial.
In Phase II, from the potentially relevant articles in Phase I, two reviewers will independently select trials (based on the full text format) for inclusion in this review . Agreement will be measured using simple agreement and kappa statistics. Disagreement will be resolved by consensus or third party adjudication (GR). Independent reviewers will document the content of each included study.
Data extraction and management
Data abstraction will be done by the two independent reviewers using a pre‐specified data extraction form and entered into the Review Manager software.
Assessment of risk of bias in included studies
The risk of bias in included studies will be assessed using individual components rather than scoring systems. The quality assessment form will include the following criteria:
-
Control of selection bias: will be assigned when the paper reports that the treatment assignment process was truly randomised (as random numbers generated by computer, table of random numbers, drawing of lots or envelopes, tossing a coin, shuffling cards, throwing a dice, etc) and that the allocation list was concealed such that the investigators were unlikely to have been able to predict (or identify) the assigned treatment prior to trial entry (a priori numbered or coded drug containers of identical appearance prepared by an independent pharmacy; central randomisation, sequentially numbered, sealed, opaque envelopes, etc).
-
Control of detection bias: will be assigned when the study reports to have been double‐blinded or triple‐blinded, such that (1) patients, (2) caregivers, and (3) investigators assessing outcome were all kept unaware of the treatment assignment.
-
Control of attrition bias: will be assigned when the report states that all patients entered in the trial and assigned treatment were included in the analysis, that any withdrawals from the trial had been listed, and the results analysed by the original treatment assignment (intention‐to‐treat analysis).
-
Control of performance bias: will be assigned when the study reports that both groups were treated equally apart from the experimental treatment (statins).
Data synthesis
Data synthesis and analyses will be done using the Cochrane Review Manager software. Quantitative analyses of outcomes will be based on intention‐to‐treat results. Relative risk ratio (RR) and a random effects model will be used to combine outcomes across trials. Test for heterogeneity of treatment effect between the trials will be made using a standard chi‐square statistic. The weighting factor for each study is the inverse of the within‐study variance plus a between‐study variance component. Thus, all pooled estimates are DerSimonian‐Laird type random effects estimators. Absolute risk reduction (ARR) = risk difference x 100 and numbers needed to treat.(NNT) = 1/risk difference will be calculated for total cardiovascular events only. The estimate with the 95% confidence intervals (CI) will be the best estimate of the mean benefit and the range of that benefit in populations with different baseline risks.
Data for the change in blood pressure and lipid levels will be combined using a weighted mean difference method. This combines a weight based on the number of individuals in the trial and the within study variance. If the trial does not report the within study variance for decrease in blood pressure, the standard deviation will be imputed from the average standard deviation from the other trials. This imputation is a limitation and to overcome this limitation, the 99% confidence interval will be reported instead of the standard 95% confidence interval reported for all other data.
Subgroup analysis and investigation of heterogeneity
We plan to perform subgroup analyses to test the effect of statins in hypertensive patients:
‐ older versus younger than 70 years
‐ male versus female
‐ with versus without manifestations of coronary heart disease
‐ with versus without elevated cholesterol
‐ with versusd without additional atherosclerosis risk factors
Sensitivity analysis
We will perform sensitivity analyses in order to explore the influence of the following factors on effect size:
‐ Repeating the analysis excluding unpublished studies (if there were any)
‐ Repeating the analysis taking into account individual components of study quality such as: allocation concealment, blinding and losses to follow‐up
‐ Repeating the analysis excluding any very long or large studies to establish how much they dominate the results.
The robustness of the results will also be tested by repeating the analysis using different measures of effects size (odds ratio) and different statistic models (random effects models).
If sufficient RCTs are identified, an attempt will be made to examine for publication bias using a funnel plot.
