Table 3.

Results from linear regression models within the 2 sections of the omega-3-PTB relationship^a.

	Change in the PTB rate (cases per 100 live births)	95% CI	p value
< 550 mg/day LC Omega-3 PUFA (n = 157)
Model 1: Unadjusted
LC Omega-3 PUFAb	− 3.09	− 4.47, − 1.71	2.1 × 10–5
Model 2: Income adjusted
LC Omega-3 PUFAb	− 1.69	− 3.06, − 0.32	1.7 × 10–2
Country Incomec	− 1.15	− 1.58, − 0.72	4.0 × 10–7
Model 3: Income and MDD adjusted
LC Omega-3 PUFAb	− 1.51	− 2.88, − 0.15	3.2 × 10–2
Country incomec	− 1.21	− 1.64, − 0.79	9.7 × 10–8
MDD prevalenced	0.36	0.03, 0.68	3.4 × 10–2
≥ 550 mg/day LC Omega-3 PUFA (n = 25)
Model 1: Unadjusted
LC Omega-3 PUFAb	− 0.41	− 1.71, 0.90	5.5 × 10–1
Model 2: Income adjusted
LC Omega-3 PUFAb	0.44	− 0.59, 1.46	4.1 × 10–1
Country incomec	− 2.13	− 3.05, − 1.22	1.5 × 10–4
Model 3: Income and MDD adjusted
LC Omega-3 PUFA b	0.33	− 0.72, 1.38	5.5 × 10–1
Country income c	− 2.02	− 2.97, − 1.07	4.3 × 10–4
MDD prevalence d	0.45	− 0.45, 1.34	3.4 × 10–1

^aLC Omega-3 PUFA = seafood based intake + (plant based intake × 0.15) Two countries were excluded as outliers: Afghanistan (MDD prevalence outlier: 22.5 cases per 100 people; over 9 SD from the global mean) and the Maldives (LC Omega-3 PUFA outlier: 3918 mg/day; over 9 SD from the global mean).

^bChange in the PTB rate (cases per 100 live births) associated with a 1 SD increase in LC Omega-3 PUFA intake (380 mg/day).

^cChange in the PTB rate (cases per 100 live births) associated with a 1 unit increase in GNI (a rank variable with four levels; higher rank corresponds to higher income).

^dChange in the PTB rate (cases per 100 live births) associated with a 1 unit increase of MDD (cases per 100 people).