Table 4.

Meta‐regression results for CHD deaths^a

Variable assessed	P value
LCn3 dose	0.94
ALA dose	0.93
Omega‐6 dose	0.66
Total PUFA dose	0.64
Duration, months	0.41
Primary or secondary CVD prevention	0.63
Food or capsule	0.78
Risk of bias	0.89
Duration + Primary or secondary prevention + PUFA dose	0.73 0.90 0.76

ALA: alpha‐linolenic acid; CVD: cardiovascular disease; LCn3: long‐chain omega‐3 fatty acids; PUFA: poly‐unsaturated fatty acids.

^aRandom‐effects meta‐regression exploring effects of LCn3 dose, ALA dose, omega‐6 dose, total PUFA dose, study duration, primary or secondary prevention, food or capsule intervention, and summary risk of bias (low or moderate to high) on CHD mortality. We ran the meta‐regression using all included trials that reported this outcome in this review, and its sister reviews (update of Hooper 2018, and Abdelhamid 2018). For each variable the P value presented represents probability that the relationship was due to chance (as we had limited power we assumed a true relationship when P < 0.10). Meta‐regression was of each variable singly, plus a multivariate meta‐regression of the 3 single variables with lowest P values. See methods for further information.