Soy and colorectal cancer.

Total soy foods: Data about total soy products in relation to cancer risk was provided by 4 asian cohorts, including a total of 2,381 cases. A significant protective effect against total colorectal cancer risk was found in 1 cohort including female 321 cases (32). And a significanct protective effect against colon cancer risk was found in one other cohort among among 102 female cases only (28). In the third cohort, a significant protective effect against colon cancer risk was found among men (350 cases), but not women (25). No association with total colorectal cancer risk was found in the fourth cohort (31).
Tofu: Data about tofu was provided by 5 cohorts.
No associations of high vs low consumption were found for any end point.
Miso (soybean paste) soup: Data about miso soup was provided by 4 cohorts.
No significant associations of high vs low consumption were found for any end point.

Conclusion: Significant protective associations of total soy foods against cancer risk were found in 3 cohorts (35% of all cases), but results were inconsistently found across cancer sites and gender. No associations were found with tofu or miso soup. Suggestive evidence for a protective effect of total soy foods against cancer risk was found when all end points were combined (average RR = 0.92). Results could not be attributed to any specific level of consumption. No evidence was found for an association between tofu, or soybean paste soup and colorectal, colon, or rectal cancer.

| More information: back to literature review.|

Prospective studies of total soy and colorectal cancer:

Author

Cohort name

Subjects

Years of follow-up

Cases

End point

Consumption of

Relative Risk (RR)

Adjustments

32) Yang G (2009)

The Shanghai Women's Health Study.

68,412 women aged 40-70.

6.4
(1998-2005)

321

Colorectal cancer incidence

Soy foods (including soy milk, tofu, fried tofu, dried or pressed tofu, green fresh soy beans, dry soy beans, soy sprouts, and other soy products)

RR = 0.67 (0.49-0.90 ;P = 0.008) for the highest vs lowest tertile of consumption.

Amount specific data (g/d):
≤ 12.8: RR = 1.
12.9-21.0: RR = 0.86 (0.66-1.13).
> 21.0: RR = 0.67 (0.49-0.89).

**Stratified by age (g/d):**
< 52 y (79 cases)	≥ 52 y (242 cases)
≤ 12.8: RR = 1. 12.9-21.0: RR = 0.94 (0.54-1.61). > 21.0: RR = 0.93 (0.51-1.70; P = 0.83).	≤ 12.8: RR = 1. 12.9-21.0: RR = 0.85 (0.63-1.16). > 21.0: RR = 0.62 (0.44-0.87; P = 0.006).

**Stratified by menopausal status (g/d):**
Premenopausal (76? cases)	Postmenopausal (273? cases)
≤ 12.8: RR = 1. 12.9-21.0: RR = 0.84 (0.39-1.79). > 21.0: RR = 1.28 (0.60-2.70; P = 0.44).	≤ 12.8: RR = 1. 12.9-21.0: RR = 0.87 (0.65-1.16). > 21.0: RR = 0.60 (0.43-0.83; P = 0.002).

**Stratified by anatomic site (g/d):**
Colon (195 cases)	Rectum (126 cases)
≤ 12.8: RR = 1. 12.9-21.0: RR = 0.98 (0.69-1.40). > 21.0: RR = 0.76 (0.52-1.13; P = 0.15).	≤ 12.8: RR = 1. 12.9-21.0: RR = 0.73 (0.48-1.12). > 21.0: RR = 0.55 (0.34-0.90; P = 0.02).

The soy and colorectal cancer association was not modified by traditional risk factors for colorectal cancer, including BMI, physical activity, and consumption of red meat, fruit, and vegetables (data not shown).

Analyses excluding individuals who had ever smoked cigarettes, consumed alcoholic beverages, or regularly used nonsteroidal antiinflammatory drugs showed no material changes in the risk estimates. RRs for the highest compared with the lowest tertiles of soy food intake ranged from 0.65 to 0.66.

Age, education, household income, physical activity, BMI, menopausal status, family history of colorectal cancer, total calorie intake, and average intakes of fruit, vegetables, red meat, nonsoy calcium, nonsoy fiber, and nonsoy folic acid. Stratified by birth year.

Additional adjustments for cigarette smoking, alcohol consumption, aspirin or other nonsteroidal antiinflammatory drug use, prior history of colorectal polyps and chronic ulcerative colitis, and intakes of fat and nonsoy protein did not appreciably alter the results; these variables were therefore not included in the final model.

31) Butler LM (2008)

The Singapore Chinese Health Study.

61,321 men and women aged 45-74.

9.8
(1993-98 to 2005)

961

Colorectal cancer incidence

Soy foods (not defined)

HR = 0.95 (0.78-1.16; P = 1.0) for the highest vs lowest quartile of consumption.

Age, sex, dialect group, interview year, diabetes, smoking history, BMI, alcohol intake, education, any weekly physical activity, first-degree relative diagnosed with colorectal cancer, and total daily energy intake.

28) Oba S (2007)

The Takayama Study.

30,221 Japanese subjects (13,894 men and 16,327 women) aged ≥ 35.

8
(1993-2000)

111 men, and 102 women

Colon cancer incidence

Soy products (tofu, miso, soybeans, natto, soymilk, okara, dried tofu, fried tofu, deep fried tofu, and fried tofu with minced vegetables/seaweed)


Men:	Women:
HR = 1.24 (0.77-2.00; P = 0.37) for the highest vs lowest tertile of consumption. Amount specific data (g/day): 49.22: HR = 1. 85.52: HR = 1.12 (0.69-1.84). 141.09: HR = 1.24 (0.77-2.00).	HR = 0.56 (0.34-0.92; P = 0.02) for the highest vs lowest tertile of consumption. Amount specific data (g/day): 46.29: HR = 1. 79.60: HR = 0.84 (0.53-1.33). 128.03: HR = 0.56 (0.34-0.92).

Age, height, alcohol intake, smoking status, BMI, physical activity, coffee intake, and use of hormone replacement therapy (women only).

25) Akhter M (2008)

The Japan Public Health Center (JPHC) Study.

83,063 Japanese subjects (39,069 men and 43,994 women) ages 45-74.

Mean 7.6
(1995-98 to 2004)

See variables

Colorectal cancer incidence

Soy food (tofu for miso soup, tofu for other dishes, yushidofu [predrained tofu], koyadofu or shimitofu [freeze-dried tofu], and soymilk)

**HRs for soy food among men. Stratified by cancer site:**
Colon (350 cases)	Proximal colon (145 cases)	Distal colon (189 cases)	Rectum (178 cases)
HR = 0.77 (0.55-1.06; P = 0.05) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.99 (0.74-1.33). Q3: HR = 0.78 (0.57-1.07). Q4: HR = 0.77 (0.55-1.06).	HR = 0.51 (0.30-0.87; P = 0.009) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.91 (0.58-1.44). Q3: HR = 0.76 (0.47-1.22). Q4: HR = 0.51 (0.30-0.87).	HR = 1.01 (0.65-1.55; P = 0.77) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 1.03 (0.68-1.54). Q3: HR = 0.80 (0.52-1.25). Q4: HR = 1.01 (0.65-1.55).	HR = 1.20 (0.74-1.95; P = 0.47) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 1.26 (0.79-2.00). Q3: HR = 1.40 (0.88-2.22). Q4: HR = 1.20 (0.74-1.95).

**HRs for soy food among women. Stratified by cancer site:**
Colon (259 cases)	Proximal colon (146 cases)	Distal colon (97 cases)	Rectum (99 cases)
HR = 1.11 (0.77-1.60; P = 0.52) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.91 (0.63-1.33). Q3: HR = 0.93 (0.64-1.34). Q4: HR = 1.11 (0.77-1.60).	HR = 1.16 (0.69-1.97; P = 0.72) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 1.25 (0.75-2.09). Q3: HR = 1.19 (0.71-2.00). Q4: HR = 1.16 (0.69-1.97).	HR = 1.08 (0.61-1.92; P = 0.73) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.72 (0.40-1.31). Q3: HR = 0.67 (0.37-1.24). Q4: HR = 1.08 (0.61-1.91).	HR = 0.90 (0.50-1.62; P = 0.82) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.78 (0.44-1.39). Q3: HR = 0.87 (0.49-1.54). Q4: HR = 0.90 (0.50-1.62).

No change was seen in the results for soy food consumption and colorectal cancer in premenopausal and postmenopausal women (data not shown).

Age; public health center area; history of diabetes mellitus; BMI; leisure time physical activity; cigarette smoking; alcohol drinking; and intake of vitamin D, dairy products, meat, vegetable, fruit, and fish; and menopausal status and current use of female hormones (women onl).

Prospective studies of tofu and colorectal cancer:

Author

Cohort name

Subjects

Years of follow-up

Cases

End point

Consumption of

Relative Risk (RR)

Adjustments

23) Lin J. (2005)

The Women's Health Study.

36,976 women aged > or = 45 years.
(USA)

10

223?

Colorectal cancer risk

Soybeans, or
Tofu

Intakes of individual vegetable items were found not appreciably to be associated with colorectal cancer risk (data not shown).

age, randomized treatment assignment, BMI, family history of colorectal cancer, history of colon polyps, physical activity, smoking status, baseline aspirin use, red meat intake, alcohol consumption, total energy intake, menopausal status, baseline post-menopausal HT use, folate intake and multivitamin use. Glycemic load in the multivariate model did not change the overall results.

22) Iso H (2007)

The JACC Study.

44,118 men, and 60,624 women.
(Japan)

Not defined.

207 men, and
202 women.

Colon cancer mortality

Tofu (soybean curd)


Men:	Women:
HR = 1.12 (0.78-1.62) for the highest vs lowest tertile of consumption. Amount specific data: < 3/w: HR = 1. 3-4/w: HR = 1.16 (0.81-1.64). ≥ 5/w: HR = 1.12 (0.78-1.62).	HR = 0.87 (0.60-1.26) for the highest vs lowest tertile of consumption. Amount specific data: < 3/w: HR = 1. 3-4/w: HR = 1.22 (0.86-1.72). ≥ 5/w: HR = 0.87 (0.60-1.26).

Age and study area.

22) Iso H (2007)

The JACC Study.

44,118 men, and 60,624 women.
(Japan)

Not defined.

157 men, and
82 women.

Rectal cancer mortality

Tofu (soybean curd)


Men:	Women:
HR = 1.21 (0.81-1.83) for the highest vs lowest tertile of consumption. Amount specific data: < 3/w: HR = 1. 3-4/w: HR = 0.96 (0.64-1.46). ≥ 5/w: HR = 1.21 (0.81-1.83).	HR = 1.00 (0.60-1.68) for the highest vs lowest tertile of consumption. Amount specific data: < 3/w: HR = 1. 3-4/w: HR = 0.43 (0.23-0.81; P = < 0.01). ≥ 5/w: HR = 1.00 (0.60-1.68).

Age and study area.

22) Kojima M (2004)

The Japan Collaborative Cohort Study (JACC).

107,824 (45,181 men and 62,643 women) aged 40-79.

9.9.
(1988-90 to 1999)

253 colon (123 men, and 130 women), and 156 rectal (106 men, and 50 women)

Colorectal cancer mortality

Tofu

**Colon cancer:**
Men	Women
HR = 1.13 (0.72-1.76; P = 0.46) for the highest vs lowest tertile of consumption Amount specific data (times of consumption): 0-2 per week: HR = 1. 3-4 per week: HR = 1.01 (0.65-1.58). Every day: HR = 1.13 (0.72-1.76).	HR = 0.75 (0.48-1.18; P = 0.35) for the highest vs lowest tertile of consumption Amount specific data (times of consumption): 0-2 per week: HR = 1. 3-4 per week: HR = 1.17 (0.78-1.77). Every day: HR = 0.75 (0.48-1.18).

**Rectal cancer:**
Men	Women
HR = 0.97 (0.60-1.55; P = 0.53) for the highest vs lowest tertile of consumption Amount specific data (times of consumption): 0-2 per week: HR = 1. 3-4 per week: HR = 0.81 (0.51-1.31). Every day: HR = 0.97 (0.60-1.55).	HR = 0.87 (0.45-1.67; P = 0.55) for the highest vs lowest tertile of consumption Amount specific data (times of consumption): 0-2 per week: HR = 1. 3-4 per week: HR = 0.49 (0.23-1.03). Every day: HR = 0.87 (0.45-1.67).

Age, family history of colorectal cancer, BMI, alcohol, smoking status, walking time, and education. Stratified by regions of enrollment.

21) Khan MM. (2004)

No cohort name.

1,524 men and 1,634 women aged ≥ 40 living in Hokkaido, Japan.

1984-2002

15 men, 14 women?

Colorectal cancer mortality

Tofu/soybean curd


Men:	Women:
Men: RR = 1.5 (0.2-11.2; No P-value).	RR = 0.9 (0.1-6.9; No P-value).

RRs are for consumption ≥ several times/wk vs ≤ several times/month.

Men: age and smoking. Women: age, health status, health education, health screening + smoking.

6) Michels KB. (2000)

The Nurses' Health Study
&
The Health Professionals' Follow-up Study.

88,764 women aged 34-59 and 47,325 men aged 40-75.
(USA)

Women: 16
(1980-1996).

Men: 10
(1986-1996)

1,181?
(937 colon cancer [368 men, 569 women], 244 rectal cancer [89 men, 155 women])

Colorectal cancer risk

Soybeans, or tofu

Intake of individual vegetables that constitute the composite items was not appreciably associated with colon or rectal cancer risk in women or men (No data data shown).

age, family history of colorectal cancer, sigmoidoscopy, height, body mass index, pack-years of smoking, alcohol intake, physical activity, (women: menopausal status, postmenopausal hormone use,) aspirin use, vitamin supplement intake (ever use of multivitamins or vitamins A, C, or E), total caloric intake, and red meat consumption.

Prospective studies of miso soup and colorectal cancer:

Author

Cohort name

Subjects

Years of follow-up

Cases

End point

Consumption of

Relative Risk (RR)

Adjustments

25) Akhter M (2008)

The Japan Public Health Center (JPHC) Study.

83,063 Japanese subjects (39,069 men and 43,994 women) ages 45-74.

Mean 7.6
(1995-98 to 2004)

See variables

Colorectal cancer incidence

Miso soup (fermented soybean paste)

**HRs for miso soup among men. Stratified by cancer site:**
Colon (350 cases)	Proximal colon (145 cases)	Distal colon (189 cases)	Rectum (178 cases)
HR = 0.85 (0.61-1.17; P = 0.27) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.98 (0.72-1.33). Q3: HR = 0.90 (0.66-1.24). Q4: HR = 0.85 (0.61-1.17).	HR = 0.72 (0.43-1.21; P = 0.20) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.96 (0.59-1.56). Q3: HR = 0.91 (0.56-1.47). Q4: HR = 0.72 (0.43-1.21).	HR = 0.92 (0.60-1.43; P = 0.73) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.98 (0.64-1.49). Q3: HR = 0.97 (0.63-1.49). Q4: HR = 0.92 (0.60-1.43).	HR = 0.78 (0.49-1.26; P = 0.27) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 1.19 (0.77-1.85). Q3: HR = 1.16 (0.75-1.80). Q4: HR = 0.78 (0.49-1.26).

**HRs for miso soup among women. Stratified by cancer site:**
Colon (259 cases)	Proximal colon (146 cases)	Distal colon (97 cases)	Rectum (99 cases)
HR = 0.88 (0.60-1.27; P = 0.60) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.95 (0.65-1.38). Q3: HR = 1.05 (0.73-1.50). Q4: HR = 0.88 (0.60-1.27).	HR = 0.90 (0.55-1.47; P = 0.74) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 0.90 (0.54-1.48). Q3: HR = 0.91 (0.56-1.49). Q4: HR = 0.90 (0.55-1.47).	HR = 0.83 (0.44-1.60; P = 0.84) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 1.04 (0.56-1.93). Q3: HR = 1.42 (0.80-2.54). Q4: HR = 0.83 (0.44-1.60).	HR = 1.62 (0.87-3.02; P = 0.13) for the highest vs lowest quartile of consumption. Amount specific data (Quartiles not defined): Q1: HR = 1. Q2: HR = 1.31 (0.70-2.46). Q3: HR = 1.44 (0.77-2.69). Q4: HR = 1.62 (0.87-3.02).

No change was seen in the results for miso soup consumption and colorectal cancer in premenopausal and postmenopausal women (data not shown).

Age; public health center area; history of diabetes mellitus; BMI; leisure time physical activity; cigarette smoking; alcohol drinking; and intake of vitamin D, dairy products, meat, vegetable, fruit, and fish; and menopausal status and current use of female hormones (women onl).

22) Iso H (2007)

The JACC Study.

42,696 men, and 58,494 women.
(Japan)

Not defined.

206 men, and
201 women.

Colon cancer mortality

Miso soup

**Men:**
Bowls at present	Bowls at 30 years old
HR = 0.87 (0.58-1.28) for the highest vs lowest tertile of consumption. Amount specific data (bowls/day): < 1: HR = 1. 1: HR = 0.79 (0.52-1.19). ≥ 2: HR = 0.87 (0.58-1.28).	HR = 0.81 (0.56-1.17) for the highest vs lowest tertile of consumption. Amount specific data (bowls/day): < 1: HR = 1. 1: HR = 0.88 (0.57-1.36). ≥ 2: HR = 0.81 (0.56-1.17).

**Women:**
Bowls at present	Bowls at 30 years old
HR = 0.84 (0.58-1.23) for the highest vs lowest tertile of consumption. Amount specific data (bowls/day): < 1: HR = 1. 1: HR = 0.71 (0.47-1.07). ≥ 2: HR = 0.84 (0.58-1.23).	HR = 0.82 (0.57-1.18) for the highest vs lowest tertile of consumption. Amount specific data (bowls/day): < 1: HR = 1. 1: HR = 0.69 (0.44-1.07). ≥ 2: HR = 0.82 (0.57-1.18).

Age and study area.

22) Iso H (2007)

The JACC Study.

42,696 men, and 58,494 women.
(Japan)

Not defined.

154 men, and
86 women.

Rectal cancer mortality

Miso soup

**Men:**
Bowls at present	Bowls at 30 years old
HR = 0.75 (0.48-1.18) for the highest vs lowest tertile of consumption. Amount specific data (bowls/day): < 1: HR = 1. 1: HR = 0.72 (0.0.45-1.16). ≥ 2: HR = 0.75 (0.48-1.18).	HR = 0.81 (0.54-1.23) for the highest vs lowest tertile of consumption. Amount specific data (bowls/day): < 1: HR = 1. 1: HR = 0.56 (0.33-0.95; P = < 0.05). ≥ 2: HR = 0.81 (0.54-1.23).

**Women:**
Bowls at present	Bowls at 30 years old
HR = 1.02 (0.56-1.85) for the highest vs lowest tertile of consumption. Amount specific data (bowls/day): < 1: HR = 1. 1: HR = 0.80 (0.41-1.56). ≥ 2: HR = 1.02 (0.56-1.85).	HR = 0.93 (0.52-1.66) for the highest vs lowest tertile of consumption. Amount specific data (bowls/day): < 1: HR = 1. 1: HR = 0.92 (0.46-1.84). ≥ 2: HR = 0.93 (0.52-1.66).

Age and study area.

21) Khan MM. (2004)

No cohort name.

1,524 men and 1,634 women aged ≥ 40 living in Hokkaido, Japan.

1984-2002

15 men, 14 women?

Colorectal cancer mortality

Miso soup


Men:	Women:
Not estimated due to 0 case in either of the 2 groups.	Not estimated due to 0 case in either of the 2 groups.

RRs are for consumption ≥ several times/wk vs ≤ several times/month.

Men: age and smoking. Women: age, health status, health education, health screening + smoking.

5) Hirayama T. (1990)

No cohort name.

265,118 adults (122,261 men, 142,857 women) aged ≥ 40 from 6 prefectures in Japan.

17
(1966-1982)

552? (248 men, 304 women)

Colon cancer mortality

Soybean paste soup

RR = 1.13 (90% CI = 0.97-1.32; No P-value) for consumption ≥ 4 vs ≤ 3 times/wk.

Age and sex.

5) Hirayama T. (1990)

No cohort name.

265,118 adults (122,261 men, 142,857 women) aged ≥ 40 from 6 prefectures in Japan.

17
(1966-1982)

563? (316 men, 247 women)

Rectal cancer mortality

Soybean paste soup

RR = 1.04 (90% CI = 0.89-1.21; No P-value) for consumption ≥ 4 vs ≤ 3 times/wk.

Age and sex.

Prospective studies of other soy foods and colorectal cancer:

Author

Cohort name

Subjects

Years of follow-up

Cases

End point

Consumption of

Relative Risk (RR)

Adjustments

22) Iso H (2007)

The JACC Study.

40,964 men, and 55,969 women.
(Japan)

Not defined.

185 men, and
183 women.

Colon cancer mortality

Tukudani (soy sauce-preserved foods)


Men:	Women:
HR = 1.00 (0.66-1.50) for the highest vs lowest tertile of consumption. Amount specific data: < 1/w: HR = 1. 1-2/w: HR = 0.93 (0.63-1.35). ≥ 3/w: HR = 1.00 (0.66-1.50).	HR = 1.07 (0.72-1.57) for the highest vs lowest tertile of consumption. Amount specific data: < 1/w: HR = 1. 1-2/w: HR = 0.88 (0.60-1.30). ≥ 3/w: HR = 1.07 (0.72-1.57).

Age and study area.

22) Iso H (2007)

The JACC Study.

40,964 men, and 55,969 women.
(Japan)

Not defined.

146 men, and
70 women.

Rectal cancer mortality

Tukudani (soy sauce-preserved foods)


Men:	Women:
HR = 0.79 (0.48-1.29) for the highest vs lowest tertile of consumption. Amount specific data: < 1/w: HR = 1. 1-2/w: HR = 0.96 (0.63-1.46). ≥ 3/w: HR = 0.79 (0.48-1.29).	HR = 1.98 (1.05-3.73; P = < 0.05) for the highest vs lowest tertile of consumption. Amount specific data: < 1/w: HR = 1. 1-2/w: HR = 1.33 (0.68-2.58). ≥ 3/w: HR = 1.98 (1.05-3.73).

Age and study area.