|Home | About | Journals | Submit | Contact Us | Français|
To determine whether dietary supplementation with α linolenic acid (ALA) can modify established and emerging cardiovascular risk markers.
Systematic review and meta‐analysis of randomised controlled trials identified by a search of Medline, Embase, Cochrane Controlled Trials Register (CENTRAL), and the metaRegister of Controlled Trials (mRCT).
All human studies were reviewed.
Changes in concentrations of total cholesterol, low density lipoprotein (LDL) cholesterol, high density lipoprotein (HDL) cholesterol, very low density lipoprotein (VLDL) cholesterol, triglyceride, fibrinogen, and fasting plasma glucose, and changes in body mass index, weight, and systolic and diastolic blood pressure.
14 studies with minimum treatment duration of four weeks were reviewed. ALA had a significant effect on three of the 32 outcomes examined in these studies. Concentrations of fibrinogen (0.17 μmol/l, 95% confidence interval (CI) −0.30 to −0.04, p = 0.01) and fasting plasma glucose (0.20 mmol/l, 95% CI −0.30 to −0.10, p < 0.01) were reduced. There was a small but clinically unimportant decrease in HDL (0.01 mmol/l, 95% CI −0.02 to 0.00, p < 0.01). Treatment with ALA did not significantly modify total cholesterol, triglycerides, weight, body mass index, LDL, diastolic blood pressure, systolic blood pressure, VLDL, and apolipoprotein B.
Although ALA supplementation may cause small decreases in fibrinogen concentrations and fasting plasma glucose, most cardiovascular risk markers do not appear to be affected. Further trials are needed, but dietary supplementation with ALA to reduce cardiovascular disease cannot be recommended.
The cardiovascular benefits of fish oil are now well established,1, but it is unclear whether α linolenic acid (ALA) confers similar benefits. ALA is a plant ω‐3 fatty acid, precursor of docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), the two main ω‐3 polyunsaturated fatty acids found in fish oils.2 However, unlike fish oils, ALA is inexpensive to produce and is more palatable than cod liver oil.
Clinical trials of dietary supplementation such as the Lyon diet heart study, in which ALA was a component, suggest that ALA may confer cardiovascular benefits.3 This has led to calls for trials specifically evaluating the effect of substituting oils containing ALA.
We therefore systematically reviewed randomised controlled trials to investigate the impact of ALA on cardiovascular risk markers.
We searched Medline, Embase, and the Cochrane Controlled Trials Register (CENTRAL) databases for published studies and the metaRegister of Controlled Trials (mRCT) for unpublished studies by using the search terms linolenic acid, plant oils, flax, linseed, canola, rapeseed, perilla, juglans, pumpkin, and purslane with a standard search filter to identify randomised controlled trials. We identified additional studies by searching references cited in identified primary studies. We restricted our search to studies of humans and included articles in languages other than English.
Studies were included if they had a control or comparison arm and had either a randomised crossover design (with a washout interval of 4 weeks) or a parallel group design (with 4 weeks of intervention). The units of measurement were converted to the common unit suggested by SI notation. Where crossover studies provided independent data for each intervention period we used data for only the first intervention period. Criteria for assessment of trial quality were the method of randomisation, blinding or objective measurements, loss to follow up, and systematic difference in care between intervention groups. When there was more than one control or ALA comparison group, we pooled results of the similar groups. We analysed subgroups stratified by the ALA dose used and type of control, and by comparing included and excluded trials.
Outcomes were changes in total cholesterol, low density lipoprotein (LDL) cholesterol, high density lipoprotein (HDL) cholesterol, very low density lipoprotein (VLDL) cholesterol, triglyceride, fibrinogen, fasting plasma glucose, body mass index, weight, and systolic and diastolic blood pressures. All data were analysed with Review Manager (version 4.2.3; Update Software, Oxford, UK).
For each trial we calculated the changes in the means between the beginning and the end of each intervention and estimated the standard deviation of the treatment effect. If the standard deviations of change were not provided, we derived them from the 95% confidence intervals or the standard error.
We used a fixed effect meta‐analysis model to calculate overall results. When a significant heterogeneity was observed, we used a random effect model instead. Heterogeneity was determined by χ2 (p < 0.10).
From 2566 references identified (1800 in Medline, 1955 in Embase, and 931 in CENTRAL) we reviewed 46 published clinical studies and one unpublished trial reporting the effects of ALA on cardiovascular risk markers. We excluded 31 of the 47 studies because they were not placebo controlled, provided insufficient information, or had a treatment period of < 4 weeks. We identified 28 outcomes in 16 published papers, reporting 14 studies. Twelve studies, involving 744 subjects, had outcomes in common and were included in the quantitative meta‐analysis. Table 11 shows selected characteristics of the included trials.4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19
Six studies measured body weight and three4,5,6 reported systolic and diastolic blood pressures. Three reported the body mass index4,6,7 and three5,8,9 reported weight. Comparisons between ALA and control groups were not significant (p > 0.05) for either body weight or blood pressure (table 22).).
Eleven studies4,5,6,7,8,9,10,11,12,13,14 with a total of 790 subjects reported changes in total cholesterol and eight studies4,5,6,7,8,9,13,14 reporting triglyceride concentrations enrolled a total of 629 subjects. As significant heterogeneity was evident (p < 0.05) a random effect model was used. The pooled mean differences were −0.01 mmol/l for total cholesterol and 0.01 mmol/l for triglycerides (table 22).). HDL and LDL cholesterol4,5,6,7,8,9,10,12,14,16 were studied in a total of 661 and 680 patients, respectively. We also identified two studies that measured VLDL cholesterol.6,14 Heterogeneity between the studies for both HDL cholesterol and VLDL cholesterol were not significant (p > 0.10), although heterogeneity was significant (χ2 = 16.38, p = 0.06) for LDL cholesterol. The pooled mean difference in HDL concentrations was −0.01 mmol/l (p < 0.01). The effect sizes of LDL and VLDL cholesterol were not significant (table 22).).
We identified two studies assessing the effect of ALA on fasting plasma glucose.14,15 A fixed effects analysis showed a significant (p < 0.01) reduction in the mean difference of 0.20 mmol/l (−0.30, −0.10 mmol/l). The effect of ALA on fibrinogen was examined in three studies4,10,15 with 382 subjects. The pooled mean difference in fibrinogen was a significant (p = 0.01) decrease of 0.17 μmol/l (−0.30, −0.04 µmol/l) (fig 11).). Heterogeneity between the studies was not significant for both outcomes (p > 0.10).
Several changes in plasma markers were reported in only one of the identified studies. The following markers of inflammation were identified in only one study: tumour necrosis factor α, interleukin 6, C reactive protein, cell adhesion molecule 1, vascular cell adhesion molecule 1,7,17 and thrombogenic factors such as factor VII, factor XII, von Willebrand factor, thromboxane, Imax, platelet aggregation velocity, plasminogen activator inhibitor 1, tissue plasminogen activator, and D dimer.4,5,10,14,18
Apolipoproteins A and A IV, fatty acids, apolipoprotein B, and Lp(a) lipoprotein were also reported.5,14,19 Although the effect of some markers was significant over time, only vascular cell adhesion molecule 117 was significantly different between treatments.
Subgroup analyses did not show significant differences when analysed by type of placebo or by the dose used of ALA (above and below 5 g/day). Funnel plots for selected outcomes did not provide evidence of publication bias in favour of trials with positive outcomes.
This systematic review provides the most reliable assessment yet of whether ALA is associated with established and emerging risk markers for coronary heart disease. Our systematic review indicates that ALA significantly affects fibrinogen and fasting plasma glucose concentrations, decreasing fibrinogen concentrations by 0.17 μmol/l and fasting glucose by 0.20 mmol/l. No other statistically or clinically significant findings were evident in the quantitatively evaluated cardiovascular risk markers.
A limitation of the meta‐analysis was that most trials were small; they did not describe the method of randomisation and not all of them were blinded. For some potential risk markers we were unable to identify two or more studies to allow pooling. We were unable to obtain data from unpublished studies and did not attempt to obtain patient level data. Although we did not adjust statistically for multiple comparisons, the two clinically important differences we observed were highly significant. The subgroup analysis showed no significant difference by either type or dose of placebo; the small dose of olive oil used as a placebo (table 11)) is therefore unlikely to have masked any important differences.
On the basis of estimates from a meta‐analysis of observational studies, a 2.9 μmol/l reduction in fibrinogen concentration would lead to a relative risk reduction of 80% in coronary heart disease.20 Therefore, a reduction of 0.17 μmol/l attributable to ALA would be expected to lead to a reduction of 6% in coronary heart disease. This is a much smaller reduction than that observed in the Lyon diet heart study, in which patients were randomly assigned to a Mediterranean diet and margarine high in ALA. Fibrinogen is therefore unlikely to mediate a clinically important effect of ALA on cardiovascular risk.
ALA is a metabolic precursor of DHA and EPA and any risk reduction may be mediated through conversion to this fatty acid. However, the metabolic overall conversion rate is low2 and varies between the sexes, being higher in women.21 Our review suggests that the impact of ALA on decreased cardiovascular risk is unlikely to be mediated through conversion to DHA or EPA, since we noted no changes consistent with increased concentrations of these fatty acids.
Although supplementation with ALA may lead to a small decrease in fibrinogen concentrations and fasting plasma glucose, most established cardiovascular risk factors or emerging risk markers do not appear to be affected. Further trials are needed but on the basis of this meta‐analysis, dietary supplementation with ALA to reduce cardiovascular disease cannot be recommended.
The authors have no conflicts of interest to declare. EW was supported by Fundação Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). AF is supported by an NHS R&D career scientist award.
ALA - α linolenic acid
CENTRAL - Cochrane Controlled Trials Register
DHA - docosahexaenoic acid
EPA - eicosapentaenoic acid
HDL - high density lipoprotein
LDL - low density lipoprotein
mRCT - metaRegister of Controlled Trials
VLDL - very low density lipoprotein
Ethical approval was not required for this study.