Table 3.

Multivariate logistic regression of the odds ratios (95% confidence interval) for the incidence of depressive symptoms by change in seaweed intakes at 3-year follow-up (n = 500)

	Change in seaweed intakes			p for trend
	Decreased	Unchanged	Increased
Range, g/1000 kcal/day	–46.1, −2.0	−2.0, 2.0	2.0, 39.5
No. of participants	152	193	155
No. of participants with depressive symptoms	21	14	11
Model 1a	1.00 (ref)	0.42 (0.18–0.97)	0.39 (0.16–0.94)	0.043
Model 2b	1.00 (ref)	0.32 (0.13–0.78)	0.33 (0.13–0.82)	0.025
Model 3c	1.00 (ref)	0.33 (0.13–0.83)	0.32 (0.13–0.81)	0.022
Model 4d	1.00 (ref)	0.32 (0.13–0.81)	0.34 (0.13–0.88)	0.032

^aAdjusted for baseline Self-rating Depression Scale, baseline seaweed intakes, sex, log age, and log body mass index

^bSame as Model 1 + occupation (desk work), education level (≥college), and living status (alone)

^cSame as Model 2 + smoking status (current, former, never), alcohol use frequency (every day, sometimes, never), breakfast skipping (≥3 times/week), physical activity (0, 0.1–22.9, ≥ 23 metabolic equivalent hours/week), and metabolic syndrome

^dSame as Model 3 + change in intakes of protein, green leafy vegetables, and fish