Fish consumption and risk of gastrointestinal cancers: A meta-analysis of cohort studies

Table 1 Summary characteristics of studies included in the meta-analysis

Ref.	Year	Country	Follow-up period	Study subjects	Number of cases	Cancer site	Fish consumption	Relative risk (95%CI)	Adjustments
Willett et al[17]	1990	United States	6 yr	88751 F	150 F	Colon	< 1/mo	1	Age
							1-3/mo	1.29 (0.70-2.40)
							1/wk	0.92 (0.49-1.72)
							2-4/wk	0.75 (0.35-1.58)
							≥ 5/wk	1.06 (0.36-3.12)
Bostick et al[18]	1994	United States	6 yr	35215 F	212 F	Colon	< 1 time/wk	1	Age, energy intake,
							1 time/wk	0.73 (0.50-1.07)	height, parity, vitamin E,
							1.5 time/wk	0.96 (0.65-1.42)	vitamin E*age interaction,
							2-2.5 time/wk	0.83 (0.54-1.28)	vitamin A
							> 2.5 time/wk	0.76 (0.49-1.19)
Giovannucci et al[19]	1994	United States		47949 M	205 M	Colon	8.4 g/d	1	Age,
							20.9 g/d	0.85 (0.54-1.33)	total energy intake
							31.0 g/d	1.05 (0.68-1.61)
							47.8 g/d	0.80 (0.51-1.26)
							83.4 g/d	1.06 (0.70-1.60)
Kato et al[20]	1997	United States	7.1 yr	14727 F	100 F	Colorectal	Q1	1	Age, total energy intake,
							Q2	1.01 (0.62-1.67)	education, place of
							Q3	0.65 (0.37-1.13)	residence
							Q4	0.49 (0.27-0.89)
Hsing et al[21]	1998	United States	20 yr	17633 M	145 M	Colorectal	< 0.8 time/mo	1	Age, calories, smoking,
							0.8-1.6 time/mo	1.1 (0.7-1.9)	alcohol intake,
							1.7-4.0 time/mo	1.2 (0.7-2.0)	total energy
							> 4 times/mo	1.5 (0.9-2.6)
Knekt et al[22]	1999	Finland	24 yr	9985	73	Colorectal	Q1	1	Age, energy intake,
							Q2	1.07 (0.53-2.17)	gender, municipality,
							Q3	1.60 (0.81-3.16)	smoking
							Q4	1.11 (0.55-2.28)
Pietinen et al[5]	1999	Finland	8 yr	27111 M	185 M	Colorectal	13 g/d	1	Age, education, smoking,
							26 g/d	1.1 (0.7-1.6)	BMI, alcohol, physical
							40 g/d	0.8 (0.5-1.3)	activity, calcium intake
							68 g/d	0.9 (0.6-1.4)
Tiemersma et al[23]	2002	The Netherlands	8.5 yr	537	102	Colorectal	0-1 time/mo	1	Age, gender, center, total
							1-4 time/mo	1.1 (0.7-1.9)	energy intake, alcohol,
							> 4 times/mo	0.7 (0.4-1.3)	height
English et al[24]	2004	Australia	9 yr	37112	451	Colorectal	< 1 times/wk	1	Age, energy intake,
							1.0-1.4 times/wk	0.9 (0.7-1.2)	country of birth,
							1.5-2.4 times/wk	0.9 (0.7-1.1)	gender, fat,
							≥ 2.5 times/wk	0.9 (0.7-1.2)	cereal intake
Kojima et al[25]	2004	Japan	9.9 yr	107824	457	Colorectal	0-2 times/wk	1	Age, family history, BMI,
							3-4 times/wk	0.88 (0.65-1.12)	smoking, physical activity,
							every day	0.96 (0.71-1.16)	education, alcohol intake,
									region
Sanjoaquin et al[26]	2004	United Kingdom	17 yr	10998	95	Colorectal	0 times/wk	1	Age, gender,
							0-1 time/wk	1.21 (0.71-2.06)	smoking, alcohol
							≥ 1 time/wk	1.17 (0.71-1.92)
Larsson et al[27]	2005	Sweden	13.9 yr	61433 F	733 F	Colorectal	< 0.5 servings/wk	1	Age, energy, education,
							0.5-< 1.0 servings/wk	0.94 (0.72-1.22)	BMI, alcohol, saturated
							1.0-< 2.0 servings/wk	1.21 (0.94-1.55)	fat, calcium, fruits and vegetables, whole-grain
							≥ 2 servings/wk	1.08 (0.81-1.43)	foods, red meat, poultry
Lüchtenborg et al[28]	2005	The Netherlands	5 yr	2948	588	Colorectal	0 g/d	1	Age, energy intake,
							4.6 g/d	1.13 (0.84-1.41)	gender, family history
							15.2 g/d	0.86(0.65-1.06)	of colorectal cancer,
							29.4 g/d	1.00 (0.74-1.27)	smoking, BMI
Norat et al[29]	2005	10 European	4.8 yr	478040	1329	Colorectal	< 10 g/d	1	Age, energy intake,
		countries					10-20 g/d	0.88 (0.74-1.06)	gender, height, weight,
							20-40 g/d	0.86 (0.72-1.02)	occupational physical
							≥ 80 g/d	0.69 (0.54-0.88)	dietary fiber, alcohol, center
Engeset et al[30]	2007	Norway	8 yr	63914 F	254 F	Colon	< 70.8 g/d	1	Age, daily intake of
							70.8-117 g/d	0.93 (0.66-1.31)	energy, smoking, fish
							> 117 g/d	1.28 (0.90-1.81)	liver, fruit and vegetables,
									fiber, fats, sauces
Hall et al[31]	2008	United States	22 yr	21406 M	500 M	Colorectal	< 1 time/wk	1	Age, smoking, BMI,
							1-< 2 time/wk	0.88 (0.65-1.20)	multivitamin use, history
							2-< 5 time/wk	0.82 (0.61-1.10)	of diabetes, random
							≥ 5 times/wk	0.63 (0.42-0.95)	assignment to aspirin or
									placebo, vigorous exercise,
									alcohol, red meat intake
Lee et al[32]	2009	China	7.4 yr	73224 F	394 F	Colorectal	< 20 g/d	1	Age, education, income,
							< 33 g/d	1.2 (0.9-1.5)	survey season, tea
							< 49 g/d	1.2 (0.8-1.6)	consumption, NSAID
							< 74 g/d	1.5 (1.1-1.9)	use, energy intake,
							≥ 74 g/d	1.3 (0.9-1.9)	fiber intake
Sugawara et al[33]	2009	Japan	8 yr	39498	566	Colorectal	0-26.2 g/d	1	Age, BMI, family history
							26.3-53.3 g/d	1.04 (0.79-1.39)	of cancer, history of
							53.4-96.3 g/d	1.11 (0.81-1.53)	stroke, hypertension,
							≥ 96.4 g/d	1.07 (0.78-1.46)	myocardial infarction
									and diabetes mellitus,
									education, marital
									status, job status,
									smoking, alcohol, time
									spent walking, total
									calories, fruit and
									vegetables
Spencer et al[34]	2010	United Kingdom		2575	579	Colorectal	< 1 g/d	1	Age, height, weight,
							1 < 15 g/d	0.89 (0.71-1.08)	smoking, energy,
							15 < 30 g/d	1.10 (0.90-1.30)	alcohol, dietary fiber
							≥ 30 g/d	0.78 (0.62-0.95)
Daniel et al[35]	2011	United States	9 yr	492186	6979	Colorectal	3.6 g/1000 kcal	1	Red meat intake, age, sex,
							7.0 g/1000 kcal	0.97 (0.90-1.04)	education, marital status,
							9.9 g/1000 kcal	0.92 (0.85-0.99)	family history of cancer,
							13.4 g/1000 kcal	0.93 (0.86-1.00)	race, BMI, smoking status,
							21.4 g/1000 kcal	0.95 (0.88-1.03)	frequency of vigorous physical activity,
									MHT in women, intake of alcohol,
									fruit, vegetables, and total energy
Nomura et al[36]	1990	United States	19 yr	7990 M	150	Gastric	≤ 1 time/wk	1	Age
							2-4 times/wk	1.4 (1.0-1.9)
							≥ 5 times/wk	0.9 (0.5-1.8)
Ngoan et al[37]	2002	Japan	10.5 yr	13250	116	Gastric	≤ 2-4 times/mo	1	Sex, age, smoking
							2-4 times/wk	0.90 (0.40-2.20)	smoking, processed meat, liver,
							≥ 1 time/d	0.90 (0.30-2.10)	cooking or salad oil, suimono, pickled food
Sauvaget et al[38]	2005	Japan	20 yr	38576	1270	Gastric	> 2 times/wk	1	Sex, sex-specific age,
							2-4 times/wk	1.09 (0.96-1.23)	city, radiation dose,
							≥ 5 times/wk	1.16 (0.97-1.39)	sex-specific smoking habits,
									education level
Tokui et al[39]	2005	Japan	11 yr	110792	859	Gastric	≤ 1-2 times/mo	1	Age
							1-2 times/wk	0.85 (0.61-1.19)
							3-4 times/wk	0.90 (0.65-1.26)
							≥ 1 time/d	0.95 (0.68-1.33)
Larsson et al[40]	2006	Sweden	18 yr	61433	156	Gastric	< 1.2 servings/wk	1	Age, education,
							1.2-1.9 servings/wk	0.97 (0.64-1.46)	body mass index,
							≥ 5 servings/wk	1.14 (0.75-1.72)	intake of total energy,
									alcohol, fruits and vegetables
Zheng et al[41]	1993	United States	20 yr	17633 M	57	Pancreatic	Q1	1	Age, smoking index,
							Q2	1.2 (0.5-3.0)	alcohol index, total calories
							Q3	2.0 (0.8-4.7)
							Q4	1.4 (0.6-3.7)
Stolzenberg- Solomon et al[42]	2002	Finland	13 yr	27111 M	163 M	Pancreatic	≤ 17.9 g/d	1	Energy intake,
							> 17.9 and ≤ 27.7 g/d	1.22 (0.75-1.97)	age and years of smoking
							> 27.7 and ≤ 38.6 g/d	1.14 (0.70-1.86)
							> 38.6 and ≤ 55.8 g/d	1.07 (0.65-1.76)
							> 55.8 g/d	0.91 (0.54-1.52)
Michaud et al[43]	2003	United States	18 yr	88802 F	178 FM	Pancreatic	< 4/mo	1	pack-years of smoking, BMI,
							1/wk	1.42 (1.01-1.98)	history of diabetes mellitus,
							≥ 2/wk	1.30 (0.86-1.98)	caloric intake, height,
									physical activity, menopausal status
Nöthlings et al[44]	2005	United States	7 yr	190545	482	Pancreatic	1.1	1	Age, ethnicity,
							3.8	0.85 (0.70-1.03)	history of diabetes mellitus,
							6.4	0.84 (0.69-1.03)	familial history of pancreatic cancer,
							9.8	0.90 (0.74-1.10)	smoking status, energy intake
							17.3	0.91 (0.75-1.11)
Larsson et al[45]	2006	Sweden	17 yr	61433 F	172	Pancreatic	≤ 1.0 servings/wk	1	Age, education,
							1.1 < 1.5 servings/wk	0.88 (0.58-1.34)	BMI, smoking,
							1.5 < 2.0 servings/wk	1.52 (0.96-2.40)	intakes of total energy,
							≥ 2.0 servings/wk	1.22 (0.77-1.92)	alcohol, energy-adjusted folate
Lin et al[46]	2006	Japan	11 yr	110792	300	Pancreatic	0-2/mo	1	Age, area,
							1-4/wk	1.21 (0.60-1.81)	pack-years of smoking
							Almost every day	0.98 (0.43-1.53)
Heinen et al[47]	2009	Netherlands	13.3 yr	120852	350	Pancreatic	0	1	Gender, age, energy, smoking,
							0-10 g/d	1.22 (0.89-1.67)	alcohol, history of diabetes mellitus,
							10-20 g/d	1.02 (0.75-1.38)	history of hypertension, BMI,
							≥ 20 g/d	1.05 (0.75-1.47)	vegetables, fruit
Rohrmann et al[48]	2012	10 European	16 yr	477202	865	Pancreatic	0 to <10 g/d	1	Height, weight,
		countries					10 to < 20 g/d	1.13 (0.90-1.41)	physical activity index,
							20 to <40 g/d	1.20 (0.97-1.50)	cigarette smoking, education,
							≥ 40 g/d	1.16 (0.92-1.47)	history of diabetes, total energy intake
Kinjo et al[50]	1998	Japan	15 yr	220272	440	Esophageal	1-3 times/mo or less	1	Age, prefecture,
							1-3 times/wk or less	0.9 (0.7-1.1)	occupation, sex
							4 times/wk or more	1.1 (0.9-1.3)
Kjaerheim et al[49]	1998	Norway	24 yr	10960 M	71 M	Esophageal	< monthly	1	Age, smoking level,
							1-5 times/mo	0.73 (0.18-1.28)	frequency of alcohol consumption,
							≥ 6 times/mo	0.96 (0.15-1.77)
Kurozawa et al[51]	2004	Japan	12 yr	110792	401	Liver	≤ 1-2 times/wk	1	Age, gender
							3-4 times/wk	0.34 (0.13-0.55)
							Almost every day	0.46 (0.16-0.76)
Sawada et al[52]	2012	Japan	11.2 yr	90296	398	Liver	35.0 g/d	1	Age, area, sex, smoking status,
							60.6 g/d	0.83 (0.59-1.17)	alcohol frequency, body mass index,
							82.8 g/d	0.84 (0.59-1.20)	past history of diabetes mellitus,
							109.9 g/d	0.75 (0.51-1.11)	and intake of coffee, soy foods, vegetables,
							160.6 g/d	0.64 (0.41-1.02)	vegetable oil, protein, and iron

Table 2 Summary relative risk for various cancer sites or different geographical regions and incremental estimates for 20-g/d increment of fish consumption

Cancer sitesregions		CorrespondingRR (95%CI) for cancer	Heterogeneity test			RR for 20 g/dIncrement of fish
			Q value	P value	I2(%)
	Total	0.93 (0.88-0.98)	80.14	0.000	67.6	0.98 (0.96-1.01)
	Total male	0.95 (0.87-1.02)	13.97	0.082	42.8
	Total female	0.96 (0.89-1.03)	27.60	0.001	71.0
	Low consumption	0.94 (0.89-0.99)	50.29	0.003	48.3
	High consumption	0.91 (0.84-0.97)	70.27	0.000	63.0
Colorectum	Total	0.93 (0.87-0.99)	53.85	0.000	64.7	0.99 (0.97-1.01)
	Low consumption	0.95 (0.91-0.98)	13.12	0.832	0.0
	High consumption	0.91 (0.82-0.99)	55.50	0.000	65.8
Colon	Total	0.95 (0.91-0.98)	10.53	0.160	33.5
	Low consumption	0.97 (0.92-1.02)	3.15	0.871	0.0
	High consumption	0.90 (0.81-0.99)	12.65	0.081	44.7
Rectum	Total	0.85 (0.75-0.95)	16.66	0.020	58.0
	Low consumption	0.86 (0.80-0.93)	3.32	0.853	0.0
	High consumption	0.85 (0.70-0.99)	17.53	0.014	60.1
Esophagus	Total	0.91 (0.83-0.99)	2.43	0.297	17.6
	Low consumption	0.90 (0.79-1.02)	0.43	0.807	0.0
	High consumption	0.95 (0.73-1.17)	4.71	0.095	57.5
Stomach	Total	1.04 (0.97-1.10)	5.94	0.430	0.0	1.03 (1.00-1.05)
	Low consumption	1.02 (0.94-1.11)	6.75	0.344	11.1
	High consumption	1.06 (0.96-1.17)	3.06	0.802	0.0
Liver	Total	0.71 (0.38-1.03)	29.04	0.000	93.1	0.89 (0.84-0.94)
	Low consumption	0.73 (0.34-1.13)	23.75	0.000	91.6
	High consumption	0.71 (0.48-0.95)	6.60	0.037	69.7
Pancreas	Total	1.07 (0.96-1.17)	19.49	0.012	59.0	1.02 (0.96-1.08)
	Low consumption	1.05 (0.93-1.17)	15.38	0.052	48.0
	High consumption	1.04 (0.96-1.11)	8.73	0.366	8.4
Asia		0.98 (0.81-1.15)	49.49	0.000	87.9
Europe		0.94 (0.86-1.01)	16.86	0.051	46.6
North America		0.88 (0.81-0.97)	19.73	0.011	59.5