juice results being less than expected (Kirkhus et al., 2012). This is could be due to various nutrients effecting absorption from the fruit. Unlike Dawczynski et al. (2013), Laidlaw et al. (2017) and Casanova et al. (2017) the total n-3 index was not tested, which has been shown to offer improvement towards risk factors of CVD. The larger sample size should show significant effects and as this doesn’t, future research should consider higher dosage through food intake such as in older studies like Marckmann et al. (1997).
A similar review from Bowen et al. (2016) shows trials published <6yrs fails to see any significant effects of n-3, the authors suggest this could be due to low dose CVD also seen in Zeng et al. (2017), Laidlaw et al. (2017) Dawczynski et al. (2013). The literature suggests notably that effects were not seen significantly in n-3 1g/d (Kirkhus et al, 2012). A further recommendation for reviews would be focusing on g/d dose (>2g/d) which provided smaller difference from the controls. Mozaffarian & Wu (2011), Burr et al., (1989) and the GISSI-Prevenzione Investigators (GISSI-Prevenzione, 1999) all show positive trend n-3 effects on cardiovascular risk in long term trials with statically significant group sizes. This would suggest a long follow up would increase results and validity. We see in Dawczynski et al. (2013) we see a notable difference with 0.8 and 3g/d in yoghurt. Although the p values and CI encourage a wide reliability, further research into this area would create better approaches in the n-3 guidelines given to people at risk of CVD.
In addition, a recent review from Balk et al. (2016) included 147 articles of RCT and prospective cohorts, found insufficient or low strength evidence between n-3FA in food or supplementation and major adverse CVD affects. It concedes with this review on high strength of evidence that n-3 oils increase levels of HDL-C and LDL-C (Balk et al., 2016). N-3 effects on CVD remain varied, Balk et al. (2016) advises that further trials may produce similar effective and ineffective results. This is important because it clarifies similar review problems of how the evidence in this study is so varied. Other reviews have assessed follow up period as showing the same results as 4 weeks-1 year in circulation of nutrients, with plasma levels changing in 6 weeks in Sanders & Hinds (1992). That the studies selected will produce similar results and external validity should still be relatively high from this (Mozaffarian et al., 2012; Thorngren et al., 1986).
Another way of improving future trials would be having a longer trial period with no n-3FA consumed, for longer than 3 months beforehand, Din et al. (2004), suggested a high intake of fish before a study can influence results, by unchanging levels significantly. The studies in this review state <1 per week for fatty fish intake, as an exclusion criteria. We must assume this criterion is followed and cannot check retrospectively, which could explain varied results between groups (Mann & Truswell, 2002). Dosage is an important aspect of the effectiveness of results on biomarkers, Sjoberg et al. (2010) research on dosage in heart rate showed significant changes in biomarkers on 6g/d n-3FA (0.34 p<0.02) associating the greater dosage to benefit individuals in a short follow up. This evidence would suggest high validity and reliability due to small margin of error and sample size. Breslow (2006) also suggests that clinical studies should determine what component has more effect on CVD risk, EPA, DHA and dosage of each. Laidlaw et al. (2017) does provide evidence in favour of EPA being a high in reducing risk factors. Additionally, Wang et al. (2006) reviewed showed a reduction in secondary factors of CVD from n-3 oil, whereas fish intake showed a greater strength of evidence for reduction of primary factors. This agrees with this review in suggesting the biggest differences and strongest evidence through methodological and statistical analysis is shown in consumption of fish oil as a food rather than supplement. Limitations Study specific limitations have been detailed in Table 1, collectively studies suffered from similar limitations. Compliance in all studies, based on participants self-administering at home. For researchers accurate measurement may be unreliable due to compliance being assumed on what is show to them at face value. Problems of self-reporting measures create recall error including arbitrary responses or socially acceptable ones will impact baseline and follow up results (Babbie, 2012). Observation blood levels are assessed in all studies but were unable to provide strong evidence on any lipid levels. Questionnaire assessment in all the studies relating to food and health status retrospectively, indicates recall error due to memory leading to arbitrary or social desirably towards answers, identifying questions and potential impacts should be considered when interpreting results. It is impossible to completely remove bias and these limitations apply a general caution towards the internal validity of the studies reviewed here. Sampling bias was not randomly achieved, most participants were selected and screened for disorders with relation to CVD. All participants were recruited from one location or type of clinic, leading to including external validity errors. Small group sizes, meet recommendation of sample size but with larger sample you would see more accurate results, similarly the differences in time length of the studies and g/d creates the varied results, external validity of these studies may be hard to generalise to the population with larger studies due to large CI. Review Limitations Review limitations include reporting bias, looking at English available texts in the past 5 years. Studies included looked at healthy participants with some including studies of participants with high blood pressure. The review didn't include studies of patients that had or were recovering from some form of CVD, therefore an association between aftercare treatment and n-3 was not established, restricting generalisability of our review to prevention methods.