Table 2.
Summary of findings for effects of omega-6 on primary outcomes. High compared with low omega-6 fats for prevention or treatment of diabetes
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect (95% CI) | No of participants (studies) | Certainty of the evidence (GRADE)† | Comments | |
|---|---|---|---|---|---|---|
| Risk with low omega-6 fats | Risk with high omega-6 fats | |||||
| Diagnosis of type 2 diabetes mellitus | 1 per 1000 | 1 (0 to 12) per 1000 | RR 1.52 (0.19 to 12.05) | 2087 (2 RCTs) | ⨁◯◯◯ VERY LOWa,b,c | Effect of omega-6 fats on diagnoses of type 2 diabetes mellitus is unclear as quality of evidence is very low. Downgraded once for risk of bias and twice for imprecision |
| Diagnosis of impaired glucose tolerance | 0 per 1000 | 0 (0 to 0) per 1000 | Not estimable | (0 RCTs) | - | No RCTs assessed diagnosis of impaired glucose tolerance |
| Glycated haemoglobin (HbA1c, %) | Mean HbA1c 7.9% | Mean HbA1c 0.0% (1.01% lower to 1.01% higher) | - | 64 (2 RCTs) | ⨁⨁◯◯ LOWd,e | Omega 6 fat may have little or no effect on glycated haemoglobin. Downgraded once each for imprecision and risk of bias |
| Plasma glucose, fasting (mmol/L) | Mean plasma glucose 7.05 mmol/L | Mean plasma glucose 0.09 lower (0.39 lower to 0.20 higher) | - | 134 (3 RCTs) | ⨁◯◯◯ VERY LOWf,g,h | Effect of omega-6 fats on plasma glucose is unclear as quality of evidence is very low. Downgraded once each for imprecision, risk of bias, and publication bias |
| Insulin, fasting (pmol/L) | Mean insulin 55.4 pmol/L | Mean insulin 14.7 higher (19.8 lower to 49.2 higher) | - | 124 (3 RCTs) | ⨁◯◯◯ VERY LOWi,j,k | Effect of omega-6 on fasting insulin is unclear as data are of very low quality. Downgraded once each for imprecision, inconsistency, and risk of bias |
| HOMA-IR | Mean HOMA-IR 2.4 | Mean HOMA-IR 1.5 (0.59 to 2.41) higher | - | 60 (1 RCT) | ⨁◯◯◯ VERY LOWl,m | Effect of increasing omega-6 on HOMA-IR is unclear as data are of very low quality. Downgraded once for risk of bias and twice for indirectness |
HOMA-IR=homoeostatic model assessment for insulin resistance; RCT=randomised controlled trial; RR=risk ratio.
Patient or population: people with and without diabetes; setting: these are long term trials, so participants live in the community; intervention: high omega-6 fats; comparison: low omega-6 fats.
Risk in intervention group (and its 95% CI) is based on assumed risk in comparison group and relative effect of intervention (and its 95% CI). Note that GRADE describes risk and 95% CI without using negative numbers; for example, GRADE states “0.02% lower (0.07 lower to 0.04 higher),” which would normally be described as “–0.02% (−0.07% to 0.04%).”
High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.
a. Risk of bias: effect size did not differ with fixed effects analysis, but neither included study was at low risk of bias from compliance, at low summary risk of bias, or at low risk of bias from allocation concealment. Downgraded once.
b. Indirectness: both studies were carried out in the UK, in men, and more than 25 years ago. Not downgraded.
c. Imprecision: 2087 randomised participants, but only 3 diagnoses reported; very limited data available. Downgraded twice.
d. Risk of bias: effect size did not differ with fixed effects analysis, but neither included study was at low summary risk of bias; 1 was at low risk of bias from allocation concealment, and 1 was at low risk of bias from compliance problems (both suggesting no effect of omega-6 on HbA1c). Downgraded once.
e. Imprecision: data included from 252 participants in 3 RCTs; confidence intervals do not exclude important harm. Downgraded once.
f. Risk of bias: effect size did not differ with fixed effects analysis; none of the 3 included trials was at low risk from allocation concealment or at low summary risk of bias; 1 trial was at low risk of bias from compliance, suggesting a small (but not statistically significant) fall in glucose with more omega-6. Downgraded once.
g. Imprecision: 134 included participants from 3 trials; 95% CI include both important benefits and harms. Downgraded once.
h. Publication bias: we are aware of a further 3 trials that measured glucose but did not report it in a way that could be included in meta-analysis. Downgraded once.
i. Risk of bias: small non-statistically significant increase in insulin with omega-6 seen also in fixed effects analysis; no trials were at low summary risk of bias or at low risk of bias from allocation concealment; 1 trial at low risk of bias from compliance suggested no effect of omega-6 on insulin. Downgraded once.
j. Inconsistency: I2=82%. Downgraded once.
k. Imprecision: data from 124 participants (2 RCTs); 95% CI included important harms and benefits. Downgraded once.
l. Risk of bias: the single study was not at low risk from compliance, allocation concealment, or summary risk of bias. Downgraded once.
m. Indirectness: small single study only. Downgraded twice.