Time course for blood pressure lowering of angiotensin receptor blockers
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To analyze the temporal (hourly and 24‐hour) blood pressure (BP) lowering efficacy of angiotensin receptor blockers (ARBs).
Background
Description of the condition
Elevated blood pressure (BP), or hypertension, is a prevalent, often asymptomatic condition which affects all demographics and populations, but becomes increasingly common as people age. Hypertension can be classified as either essential (primary), that is, without any apparent etiology; or secondary, in which there is an identifiable cause. Both types are associated with an increased risk of adverse cardiovascular events such as stroke, heart failure and myocardial infarction (Kannel 1972). Pharmacological treatments have proven useful in lowering the incidence of these events in both types of hypertension, although numerous studies seem to imply that the effects differ with different drug classes in this regard (Lindholm 2000; IQWiG 2009; Wright 2009; Chen 2010; Wiysonge 2012; Xue 2015). Typically, treatment of primary hypertension is solely directed towards lowering BP, while treatment of secondary hypertension additionally involves elucidating and correcting the underlying cause if possible.
Reviews have generally focused on the BP lowering effect of different classes of drugs at a single time point (e.g. trough effects 22 to 26 hours after the last dose of the drug) (Morgan 1998). While progress has been made in recent years (Ghamami 2014), there are still very few systematic reviews assessing the temporal effects of classes of drugs on BP, and none assessing the temporal effects of angiotensin receptor blockers (ARBs) in particular. Work is currently being done in order to fill this need; for example, several other protocols for time course reviews have been published in the Cochrane Database of Systematic Reviews (Sekhon 2008; Choi 2012; Mann 2012; Wu 2012; Zhang 2012a; Zhang 2012b).
Description of the intervention
ARBs are a relatively new class of anti‐hypertensive drug. There is less evidence regarding mortality and morbidity when they are used as a first‐line drug (compared to other drug classes including thiazide diuretics, beta adrenoceptor blockers or angiotensin‐converting enzyme (ACE) inhibitors). It is therefore important to not only investigate their BP lowering efficacy, but also to better understand the time course of their anti‐hypertensive effect.
How the intervention might work
ARBs competitively block angiotensin II receptors, specifically type 1 receptors (AT1), on vascular smooth muscle and kidney tubules. This results in vasoconstriction, as well as aldosterone release and consequent water retention. The mechanism of action of ARBs is closely related to that of ACE inhibitors, which impede the synthesis of angiotensin II. There is evidence demonstrating that ARBs are less efficacious but better tolerated than ACE inhibitors (Li 2014), mainly due to a lower incidence of dry cough (because only ACE inhibitors inhibit the breakdown of bradykinin). Therefore, ARBs are typically prescribed for patients who cannot tolerate ACE inhibitor therapy. However, both drug classes may cause renal impairment in those whose kidney function depends heavily upon the renin‐angiotensin system (e.g. patients with severe congestive heart failure).
Why it is important to do this review
ARBs are one of the most commonly‐prescribed classes of anti‐hypertensive drugs and the benefits of this pharmacological therapy may be partially dependent on the time course of its BP lowering. Therefore, it is important for physicians and patients to be aware of the time course for the BP lowering effect of ARBs.
Objectives
To analyze the temporal (hourly and 24‐hour) blood pressure (BP) lowering efficacy of angiotensin receptor blockers (ARBs).
Methods
Criteria for considering studies for this review
Types of studies
Studies must be either:
-
trials that involve random allocation of participants to a standard dose (defined as one recommended for current clinical practice) of one or more ARBs, and to a parallel control group; or
-
baseline controlled trials. The reason that baseline controlled trials are eligible is because there is negligible or no placebo effect using 24‐hour ambulatory BP monitoring (Dupont 1987).
However, all trials must meet the following criteria.
-
Duration of follow‐up must be at least three weeks.
-
BP must be measured, using 24‐hour ambulatory monitoring, at baseline (following washout) and at one or more time points after week 3.
-
Parallel studies must be double‐blinded.
-
Cross‐over trials can be included if data are available before crossing over.
Types of participants
We will include adult patients (over 18 years of age) diagnosed with primary hypertension. Participants must have a baseline systolic blood pressure (SBP) of at least 140 mmHg, a baseline diastolic blood pressure (DBP) of at least 90 mmHg, or both. We are making an assumption that age, baseline risk and other comorbidities do not affect the temporal BP lowering effect of this drug class.
Types of interventions
Intervention ‐ ARBs including: azilsartan, candesartan, eprosartan, irbesartan, losartan, olmesartan, telmisartan, valsartan. If there is more than one dose being studied in a single randomized controlled trial, we will use the highest dose within the recommended dose range to maximize the chance of finding a difference in effect at different times.
Control ‐ Placebo or no treatment.
Types of outcome measures
Primary outcomes
End point systolic and diastolic hourly blood pressures using a 24‐hour ambulatory BP device (or other technique).
Secondary outcomes
Hourly pulse rate.
Search methods for identification of studies
Electronic searches
The Cochrane Hypertension Information Specialist will search the following databases from date of inception for published, unpublished, and ongoing studies:
-
the Cochrane Hypertension Specialised Register via the Cochrane Register of Studies (CRS‐Web);
-
the Cochrane Central Register of Controlled Trials (CENTRAL) via the Cochrane Register of Studies (CRS‐Web);
-
MEDLINE Ovid (from 1946 onwards), MEDLINE Ovid Epub Ahead of Print, and MEDLINE Ovid In‐Process & Other Non‐Indexed Citations;
-
Embase Ovid (from 1974 onwards);
-
ClinicalTrials.gov (www.clinicaltrials.gov); and
-
World Health Organization International Clinical Trials Registry Platform (www.who.it.trialsearch).
The subject strategies for databases will be modeled on the search strategy designed for MEDLINE in Appendix 1. Where appropriate, these will be combined with subject strategy adaptations of the highly sensitive search strategy designed by Cochrane for identifying randomised controlled (as described in the Cochrane Handbook for Systematic Reviews of Interventions, Box 6.4.b. (Higgins 2011a)).
Searching other resources
-
The Cochrane Hypertension Information Specialist will search the Hypertension Specialised Register segment (which includes searches of MEDLINE and Epistemonikos for systematic reviews) to retrieve published systematic reviews related to this review title, so that we can scan their reference lists to identify additional relevant trials.
-
We will check the bibliographies of included studies and any relevant systematic reviews identified for further references to relevant trials.
-
We will contact experts/organizations in the field to obtain additional information on relevant trials.
-
We may contact original authors for clarification and further data if trial reports are unclear.
-
We will not perform a separate search for adverse effects of interventions used for the treatment of hypertension. We will consider adverse effects described in included studies only.
Data collection and analysis
Selection of studies
Studies will initially be screened for relevance based on abstracts and titles. If a study does not meet the inclusion criteria or it fulfils any exclusion criteria, it will be rejected. Full texts of selected studies from the initial screening will be reviewed for their overall applicability based on the inclusion criteria. Two authors will independently assess the selected studies for inclusion. A third author will resolve any discrepancies.
Data extraction and management
Data will be entered into a data extraction form and data extraction will be independently cross‐checked by two authors. A third author will double‐check all the interpolations and calculations performed. The investigators of the specific trials will be directly contacted in case any missing data need to be obtained.
Hourly BP data are often presented in graphical format. Data will be extracted either manually or using suitable data extraction software and cross‐checked by two authors. After resolving any significant discrepancies (defined as greater than 1 mmHg), an average of the two values obtained by the authors will be taken and used in the meta‐analysis.
In studies where the same patient cohort is treated with either different doses of the same drug or with the same drug but for different lengths of time, data from the highest dose and longest treatment period will be used in the meta‐analysis. The logic behind this is that higher dosage and longer treatment are likely to produce the greatest change in BP.
Where ambulatory BP was measured and reported at less than one hourly interval (for example, every 30 or 15 minutes), the average hourly BP will be calculated and used in the meta‐analysis.
Angiotensin receptor blockers are commonly administered once or twice daily, with the first/only dose given in the morning. In analysing ambulatory BP data, the time at which the morning dose was given will be taken as 'Hour 0'. Where the time of drug administration is not specified in the study, 9 am will be assigned as 'Hour 0'. When given as twice daily dosing, the time at which the afternoon dose was given will be taken as 'Hour x', where x is the number of hours since the morning dose was given, not 'Hour 0'. The reason behind this is that while we anticipate angiotensin receptor blockers will have a differential effect on BP depending both on the time of day and the time since drug administration, the emphasis of this review is on their circadian effect, not their duration of effect after a dose.
Assessment of risk of bias in included studies
The risk of bias will be assessed following the methodology described in the Cochrane Handbook Chapter 8 (Higgins 2011b), under the subheadings: sequence generation (selection bias), allocation sequence concealment (selection bias), blinding of participants (performance bias), blinding of outcome assessors (detection bias), incomplete outcome data (attrition bias), selective outcome reporting (reporting bias) and other potential sources of bias.
Measures of treatment effect
The treatment effect will be defined as the mean change in SBP and DBP in mmHg (a continuous variable) for each hour over a 24‐hour period. For example, if a trial used 24‐hour ambulatory BP monitoring at different points in time after week 3, we will use the mean of all the measurements.
Unit of analysis issues
Our approach for assessing statistical heterogeneity (described below) is designed to avoid unit of analysis errors.
Dealing with missing data
We will attempt to contact the authors of selected articles via email or telephone in order to obtain any missing data. These attempts to retrieve information and the authors’ replies or lack thereof will be recorded in the review.
Standard deviation (SD) of change in BP at each hour is often not included in published reports. In the event that this information cannot be obtained from authors, SDs will be imputed according to the hierarchy below, which is a modified version of the strategy described by Sekhon 2008.
-
SD of the change in average daily BP from the same trial.
-
SD of the average BP at each hour at the end of treatment from the same trial.
-
SD of the average daily BP at the end of treatment from the same trial.
-
Weighted mean SD of the change in average daily BP from at least 3 other trials using the same drug and dose regimen.
-
Weighted mean SD of average daily BP at the end of treatment from at least 3 other trials using the same drug and dose regimen.
-
Weighted mean SD of change in average daily BP from other trials using the same drug.
-
Weighted mean SD of average daily BP at the end of treatment from other trials using the same drug.
-
Weighted mean SD of change in average daily BP from all included trials (any drug and dose).
-
Weighted mean SD of average daily BP at the end of treatment from all included trials (any drug and dose).
SD imputed according to steps 1 and 3 to 9 of the hierarchy will be used at each hour.
Assessment of heterogeneity
We will not use the Review Manager software's built‐in test for heterogeneity of treatment effects to test for differences in BP lowering effect at different hours because of correlated errors introduced by repeated observations on the same participants. Instead, we will use linear regression models that compensate for the correlated observations. We will use restricted maximum likelihood estimation (REML) and likelihood ratio tests to choose an appropriate correlation structure. Correlation structures that we will test with REML will include independent, unstructured, and autogressive.
Assessment of reporting biases
Publication bias will be assessed using funnel plots, as outlined in the Cochrane Handbook Chapter 10 (Sterne 2011).
Data synthesis
Randomized parallel group trials with placebo or no intervention as a control, and baseline controlled trials will be analyzed separately as subgroups.
For randomized parallel group controlled trials, data that will be entered in RevMan include: 1) mean change from baseline in BP or heart rate, 2) SD of change in BP or heart rate, and 3) the number of participants in both the treatment and control groups to obtain mean difference (MD) with 95% confidence interval (CI).
For baseline controlled trials, the MD from baseline measurement in BP or heart rate plus its standard error will be entered and pooled as generic inverse variance data to obtain MD with 95% CI.
RevMan cannot be used for the analyses of total effects across time points because of correlated errors introduced by repeated observations on the same patients. Instead, the observations in the SBP and the DBP analyses will be analyzed using linear regression models that will compensate for the correlated observations. The linear regressions will be done 1000 times each for the SBP and DBP data. This analysis will also be adjusted for differences in baseline BP levels between trials.
A single iteration of the process will involve generating values randomly selected from normal distribution defined by the reported MD and respective 95% CI for each observation. The generated values will then be inputted into a linear regression to obtain a total estimated MD. This process will be repeated 1000 times to obtain a distribution of total MDs.
An iterative process is needed because estimating only one SBP or DBP linear model with the reported MD for each hour and study will not account for the variation around each individual study‐hour observation (i.e. the SD). Kernel density estimation will be used to identify a normal density function for the 1000 values, and then the mean, upper 95%, and lower 95% confidence limits will be extracted from the density. These analyses will be completed using PROC MIXED in SAS versions 9.4 (SAS Institute Inc., Cary, NC).
To account for correlated observations, variance‐covariance matrices in each linear regression will be assumed to be compound symmetric. The generated values from each iteration will also be used to conduct analyses of variance (ANOVA). F‐tests will be computed for each iteration. The assumption will be made that observations across hours are likely to be homogeneous if the F‐tests rarely exceed the critical F values of 1.564 (SBP) or 1.559 (DBP).
Subgroup analysis and investigation of heterogeneity
ARBs will be divided according to dose as done in other reviews; see low dose and high dose subgroups in Wright 1999. Placebo or no treatment controls (and baseline controlled trials) will also be analyzed as subgroups.
Heterogeneity will be assessed using linear regression models that compensate for the correlated observations.
Sensitivity analysis
If possible, depending on the number of eligible studies identified, we will conduct sensitivity analyses according to patient characteristics (e.g. age, gender, ethnicity) and baseline BP.
