Fish oil supplements for the prevention and treatment of hypertension in adults
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To assess the effectiveness of fish oil supplements in reducing blood pressure in both normotensive and hypertensive populations.
Background
Many countries and major organisations recommend a diet that comprises two or three portions of oily fish per week (De Backer 2003, Kris‐Etherton 2002). This recommendation derives from the view that the consumption of long chain omega 3 fatty acids, particularly, eicospaentaenoic acid (EPA) and docosahexaenoic acid (DHA)protect against cardiovascular disease. Dietary sources of long‐chain omega‐3 fatty acids include herbs nuts and some plants but they are found in abundance in fish. In Western diets, the main source of EPA and DHA intake is fish, particularly oily fish such as mackerel, trout, salmon, herring and sardines. However, the reported benefits of fish oils has led to a massive and continuing growth in the sale of dietary supplements containing omega 3 fats. Fish oil supplements contain variable amounts of n‐3 fatty acids in the form of triglycerides, ethyl esters or free fatty acids (De Caterina 2011).
The protective effects of fish oils were first reported by Bang 1972 who observed that ischaemic heart disease was rare amongst Greenlandic Eskimos despite their high intake of saturated animal fats. The protective component was suggested to be the long‐chain n‐3 fatty acids consumed in very high amounts as a result of the regular intake of seal meat and whale blubber. (Bang 1976) Additional epidemiological work has found similar patterns of low cardiovascular disease in populations that consume a diet rich in seafood, in Japan, Norway, Holland and India (Bulliyya 1990, Kromhout 1985, Yano 1988). Not all epidemiological studies agree however and a meta‐analysis of cohort data has found no clear effect of long chain and shorter chain omega 3 fats (from both fish and vegetable oils) on cardiovascular events (Hooper 2006).
Description of the condition
Reducing blood pressure is a major factor in reducing the risk of cardiovascular events. About 7.6 million deaths (13.5% of all deaths) in the year 2001 were attributable to high blood pressure. (Lawes 2008) Public health strategies to shift the population BP distribution to a lower level are therefore of utmost importance in order to reduce the burden of cardiovascular disease. (MacMahon 2005) Relatively small reductions in blood pressure can result in significantly reduced rates of stroke and heart disease. Differences of only 5mmHg in DBP could potentially reduce risk of stroke by approximately one third and coronary heart disease by a fifth. (MacMahon 2005)
Description of the intervention
Polyunsaturated fatty acids and their sources n‐3 (or omega‐3) fatty acids are a family of naturally occurring PUFAs (polyunsaturated fatty acids). It is the position of the double bonds within their hydrocarbon chain that gives n‐3 fatty acids their name and also their physical and physiological properties. Most PUFAs have more than 18 carbon atoms and are consequently defined as long‐chain PUFAs. Important long‐chain omega‐3 fatty acids include DHA and EPA, both of which can be synthesised from short‐chain omega‐3 fatty acids such as a‐linolenic acid via other intermediates. However, as the rate of conversion is low in humans, long‐chain omega‐3 fatty acids must be obtained primarily through the diet. Fish oil supplements provide rich sources of EPA and DHA fatty acids and it is their effects that will be analysed in this review.
How the intervention might work
The precise mechanisms by which fish oils might have a beneficial effect on blood pressure and consequently upon cardiovascular health are unclear and a number of potential processes have been described. (Calder 2004) The suggested mechanisms include:
-
Reduce susceptibility of the heart to ventricular arrhythmia
-
Decreased platelet aggregation
-
Vasodilatation
-
Anti‐inflammatory effects
-
Improvements in endothelial function
-
Plaque stabilization
-
Reduce adhesion molecule expression
-
Reduce platelet‐derived growth factor
-
Anti‐inflammatory
-
-
Reduced atherosclerosis
-
Up‐regulated adiponectin synthesis
-
Reduced collagen deposition
Why it is important to do this review
Intervention studies exploring the effects of fish oil, either in a diet rich in oily fish or in supplements, upon blood pressure have been carried out in both normotensive and hypertensive populations. The findings have been inconsistent in both groups. Three major meta‐analyses (Dickinson 2006, Geleijnse 2002, Morris 1993) of RCTs have evaluated the effect of fish oil on blood pressure. The different inclusion criteria and publication dates of these reviews account for some of the discrepancies in both the studies included and review results (Table 1). Nevertheless, all 3 meta‐analyses found small, statistically significant reductions in both systolic blood pressure (SBP) (2‐4 mmHg) and diastolic blood pressure (2‐2.51mmHg) among individuals with hypertension. Dickinson 2006 limited inclusion criteria to trials among hypertensive populations with at least eight weeks follow‐up, accounting for the smaller number of studies in this review. Morris 1993 and Geleijnse 2002 also included studies in normotensive populations. These two earlier reviews (Geleijnse 2002, Morris 1993) found a similar but non‐significant effect, reflecting perhaps the reduced scope for reduction of BP in normotensive participants. The reviews differ in their conclusions about the impact of dose on blood pressure and the influence of publication bias. Multivariate regression analysis in the meta‐analysis of Morris 1993 found a dose response in hypertensive patients with greater decreases in blood pressure with increasing doses of fish oil. In contrast Geleijnse 2002 found no dose response relationship of fish oil intake (g/day) with change in BP. The inclusion criteria also differ: Geleijnse 2002 only included trials of at least two weeks duration while Morris 1993 did not limit inclusion by intervention duration. Whilst there appears to be a consistent finding of a small blood pressure lowering effect of fish oil, there remains some uncertainty about the benefits in normotensive populations, the dose needed and the best composition of fish oil. The quality of trial methods may also influence results; in a sensitivity analysis excluding non‐blinded trials the size of the effect was reduced in both SBP and DBP by 18% and 11% respectively. (Geleijnse 2002)
This review will expand an earlier published meta‐analysis (Dickinson 2006) by widening the inclusion criteria to include trials among normotensive participants (n=13 additional trials). It will also update the analyses of hypertensive patients, as three new trials have been published (Cazzola 2007, Damsgaard 2008, Wang 2008) which have not been included in previous published reviews. Additional analyses will be been carried out to explore possible dose response effects of fish oils. In contrast to earlier reviews which explored dose response to fish oil only, we will explore the effects of different dosages of EPA and DHA in the fish oil.
Table One.
| Study | Reduction in SBP (95% CI) | Reduction in DBP (95% CI) | Included studies | Included patients | Inclusion criteria |
| ‐3.4 (‐2.5 to 10.0) | ‐2.0 (‐1 to 6.0) | 9 | 361 | Included fish oil supplements and fish diets with interventions of any duration | |
| ‐3.65 (‐5.7 to ‐1.6) | ‐2.5 (‐3.7 to ‐1.3) | 23 | 760 | Interventions of 2 or more weeks duration. | |
| ‐2.3 (‐4.3 to ‐0.2) | ‐2.2 (‐4.0 to ‐0.4) | 8 | 375 | Included hyptertensive patients (SBP ?140 and/or DBP ?85) only Interventions of 8 or more weeks duration |
Objectives
To assess the effectiveness of fish oil supplements in reducing blood pressure in both normotensive and hypertensive populations.
Methods
Criteria for considering studies for this review
Types of studies
Randomized controlled trials, including cross over trials where there is at least a 4 week wash out period between treatments.
Types of participants
Adults (aged 18 years and over) with normal or raised blood pressure. Trials with pregnant women or patients with secondary hypertension or renal disease will be excluded.
Types of interventions
Fish oil supplements administered for a minimum of 4 weeks. Trials where hypertensive medication is changed during the study or where participants are advised to increase the consumption of oily fish only will be excluded.
Comparison: placebo.
Types of outcome measures
Primary outcomes
Means and standard deviations of systolic and diastolic blood pressure at the end of the trial or the mean and standard deviation of the change in blood pressure between randomization and the end of the trial. For crossover trials, mean and standard error of the change in blood pressure (if reported) otherwise the mean and standard deviation of within‐patient differences in blood pressure at the end of treatment and control periods.
Secondary outcomes
Number of participants withdrawing from treatment (as an indication of tolerability of treatment).
Search methods for identification of studies
Electronic searches
The Database of Abstracts of Reviews of Effectiveness (DARE) and the Cochrane Database of Systematic Reviews will be searched for related reviews.
The Cochrane Central Register of Controlled Trials (CENTRAL), the Hypertension Group Specialised Register, MEDLINE (1946‐), EMBASE (1974‐), and the International Clinical Trials Registry Platform (WHO‐ICTRP)
Electronic databases will be searched using a strategy combining the Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE: sensitivity‐maximizing version (2008 revision) with selected MeSH terms and free text terms relating to fish oil, blood pressure and hypertension. No language restrictions will be used. The MEDLINE search strategy (Appendix 1) will be translated into the other databases using the appropriate controlled vocabulary as applicable.
Searching other resources
a) Reference lists of all papers and relevant reviews identified
b) Authors of relevant papers will be contacted regarding any further published or unpublished work
c) Authors of trials reporting incomplete information will be contacted to provide the missing information
Data collection and analysis
Selection of studies
Studies will be selected using an agreed inclusion criteria. All titles and abstracts of citations will be screened by one reviewer. All rejected citations will be checked by a second reviewer. All papers that appear to meet the inclusion criteria at screening or where there is uncertainty regarding inclusion status will be retrieved as a full paper copy. Full paper copies will again be screened to assess if they meet the inclusion criteria by one reviewer and those rejected screened by a second reviewer. The reason for excluding papers will be documented and summarised in the flow chart depicting the search for papers.
Data extraction and management
Data on study risk of bias, characteristics of participants, interventions and relevant outcomes will be independently abstracted by two reviewers. Differences will be resolved by discussion. A data extraction form will be designed in collaboration with clinical experts and piloted.
Assessment of risk of bias in included studies
Risk of bias will be assessed by examining the adequacy of randomization, concealment of allocation, blinding, and loss to follow‐up. Additional aspects of study design will also be assessed in the included cross‐over trials including the duration of the wash out period, the use of paired analysis and reporting of dropouts from both treatment periods.
Measures of treatment effect
For randomised studies of parallel design we will extract the numbers of participants randomised to each study arm, and either a) the mean and standard deviation of SBP and DBP at baseline and end of study in each arm, or b) mean and standard deviation of change in SBP and DBP in each arm, dependent on which was reported. For crossover trials, we will extract the number of participants along with a) mean difference in SBP and DBP together with its standard error if reported, or if not b) the mean and standard deviation of within‐patient differences. Where multiple doses of fish oil were used in the same trial, these arms will be combined.
All analyses will be performed using RevMan version 5.
Unit of analysis issues
We will pool results of parallel trials using the final SBP and DBP (and standard deviations) in the weighted mean difference facility in RevMan. For cross‐over trials, we will extract the number of participants and the mean difference (and standard deviation) between final blood pressure for fish oil and control interventions for both SBP and DBP. Where no standard deviations for the treatment effect are reported in the cross‐over trials, they will be imputed where possible by applying the same within‐person correlation observed in one other cross‐over trial. Data from crossover and parallel trials will then be combined in the same meta‐analysis (also using the generic inverse variance approach), if they did not show heterogeneity.
Dealing with missing data
Authors will be contracted for missing endpoint data. Missing standard deviations will be imputed.
Assessment of heterogeneity
Studies with normotensive participants (defined as SBP<140mmHg and DBP <90mmHg) and hypertensive participants (SBP≥140mmHg or DBP≥90mmHg) will be pooled separately. Meta‐analyses will be performed by the inverse‐variance method. Heterogeneity between trials will be explored using the I2 statistic, which estimates the percentage of variation between trials that cannot be ascribed to sampling variation.
Assessment of reporting biases
Funnel plots will be used to assess for evidence of publication bias.
Data synthesis
Meta‐regression will be used to test a dose response effect on blood pressures, with individual study effects weighted by the inverse of the estimated variance. The meta‐regression will analyse both normotensive and hypertensive trials together but with an interaction term to allow the dose‐response to differ between the two populations. Five components will be assessed (total fish oil, EPA, DHA, EPA+DHA and EPA/DHA), each in two ways: firstly the amount in g/day, and secondly taking into account exposure‐weeks (i.e. amount in g/day multiplied by length of treatment).
If a trial used two or more doses of fish oil, each arm will be included as a separate dose in the analysis. To ensure control patients are used only once, control groups will be divided into one subgroup per active arm, each subgroup having the same mean and standard deviation.
Subgroup analysis and investigation of heterogeneity
Normotensive and hypertensive participants will be analysed in separate subgroups and not pooled. Subgroup analysis will be performed to compare the effects of the intervention in cross‐over trials compared to parallel controlled trials. A subgroup analysis will also be used to analyse the effects of shorter term interventions i.e. those where the supplement is taken for less that 8 weeks and to compare these against those trials where the intervention duration is 8 weeks or longer.
Sensitivity analysis
Sensitivity analsysis will be undertaken to explore the effects of including trials with a high risk of bias in the meta‐analyses.
