Table 4.

Human studies of MSG which were referenced by ICHD-III beta but were not complied with the criteria for the systematic review

First author	Vehicle for MSG administration	Protocol	Number of subjects	Incidence of headache	Statistical difference
Gore M (1980) [16]	150 ml cold tap water	After an overnight fast, subjects ingested, on different days, 1.5, 3, 6 g MSG and three paired placebo materials, the order of each pair being randomized. (There was no indication about the content "three paired placebo materials")	30 men, 25 women	[Total number of positive responses by 3 doses] MSG: 8 episodes (7 subjects), Placebo: 2 episodes (2 subjects)	No difference (Fisher test) Statistical analysis in each symptom was not performed in the paper.
Kenny RA (1972) [17]	150 ml tomato juice	(Phase 1) 2 h after breakfast, 5 g MSG on1 day and 0.8 g NaCl on the other 2 days. Breakfast type: (1)no breakfast (2)liquids (milk, coffee, juice) or instant breakfast (3)largely of carbohydrate(cereal, toast, etc.) (4)containing protein (eggs, ham, etc.)	(Phase 1) 77 subjects	[Number of positive responses] MSG : 4, Placebo : 2	No difference (Fisher test) Statistical analysis was not performed in the paper.
	150 ml tomato juice (J) or water (W)	(Phase 2) JP1 : 0.8 g NaCl, WP: 0.7 g NaCl, WM5 : 5 g MSG + 0.2 g NaCl, JM1 : 1 g MSG, JM2 : 2 g MSG, JM3 : 3 g MSG, JM4 : 4 g MSG, JP2 : 0.8 g NaCl	(Phase 2) 22 out of 25 subjects, who reacted only to MSG on phase 1 study	[Number of positive responses / subjects] JM1 : 1 /15, JM2 : 2/13, JM3 : 4/13, JM4 : 5/14, WM5 : 8/17	Since the incidence of placebo group (JP1, WP, JP2) is not shown, statistical analysis cannot be done.