Table 4.

The association between PPI or H2RA use and the risk of gastric cancer in the UK Biobank.

	Users		Non-users		Unadjusted HR (95% CI)	Adjusteda HR (95% CI)
	Gastric cancer, n	Person-years	Gastric cancer, n	Person-years
PPI user vs. non-user
Main analysis (starting follow-up at 1 y)	44	208,807	206	1,949,341	1.53 (1.10, 2.12)	1.28 (0.86, 1.90)
Male only	29	94,195	153	896,467	1.43 (0.96, 2.13)	1.14 (0.70, 1.87)
Female only	15	114,611	53	1,052,874	1.94 (1.09, 3.47)	1.73 (0.86, 3.45)
Adenocarcinoma	37	208,807	175	1,949,341	1.52 (1.07, 2.18)	1.18 (0.76, 1.83)
Gastric cardia	15	208,807	86	1,949,341	1.26 (0.72, 2.18)	0.81 (0.40, 1.64)
Gastric non-cardia	14	208,807	51	1,949,341	1.93 (1.06, 3.50)	1.44 (0.68, 3.06)
Main additionally adjusting for H2RAb	44	208,807	206	1,949,341	1.53 (1.10, 2.12)	1.26 (0.84, 1.88)
Main removing adjustment for GORD, oesophagitis and peptic ulcerc	44	208,807	206	1,949,341	1.53 (1.10, 2.12)	1.41 (1.00, 1.98)
Main additionally adjusting for year of cohort entryd	44	208,807	206	1,949,341	1.53 (1.10, 2.12)	1.28 (0.86, 1.90)
Starting follow-up at 2 y	30	162,955	170	1,525,464	1.28 (0.86, 1.89)	1.15 (0.73, 1.82)
Starting follow-up at 3 y	22	117,731	122	1,105,366	1.28 (0.81, 2.02)	1.12 (0.65, 1.92)
Omeprazole user vs. non-user
Main analysis (starting follow-up at 1 y)	25	122,860	225	2,035,288	1.43 (0.95, 2.17)	1.17 (0.74, 1.85)
Lansoprazole user vs. non-user
Main analysis (starting follow-up at 1 y)	16	73,848	234	2,084,299	1.49 (0.90, 2.48)	1.21 (0.71, 2.08)
H2RA user vs. non-user
Main analysis (starting follow-up at 1 y)	4	38,517	246	2,119,632	0.80 (0.30, 2.15)	0.49 (0.16, 1.56)

^aAdjusted for age at baseline, sex, socioeconomic status, alcohol, smoking, BMI, comorbidities at baseline (including diabetes, GORD, oesophagitis and peptic ulcer) and other medication uses at baseline (statins aand aspirin).

^bAdditionally adjusted for H2RA.

^cRemoving the GORD, oesophagitis and peptic ulcer adjustment from the main model.

^dAdditionally adjusted for year of cohort entry.