Does Dietary Fiber Protect Against Colorectal Cancer

1
Does Dietary Fiber Protect Against Colorectal
Cancer?

• Young-In Kim, MD, FRCP(C)
• Division of Gastroenterology
• Departments of Medicine Nutritional Sciences
• University of Toronto
• St. Michaels Hospital

2
Colorectal Cancer Overview

• Common cancer
• Significant morbidity mortality
• Largely preventable
• Well established precursor(s)
• Effective screening strategies available
• Potentially modifiable environmental risk factors

3
Colorectal CancerCanadian Cancer Statistics 2007
New Cases
Deaths
F
F
total
M
total
M
Life-time risk M 1/16 (6.3) F 1/18 (5.6)
159,900
72,700
82,700
77,200
38,400
34,300
23,300
19,900
12,400
10,900
11,000
8,900
22,500
5,400
22,300
50
5,300
170
22,300
4,300
22,300
4,300
20,800
11,400
9,400
8,700
4,700
4,000
4
(No Transcript)
5
CRC Risk Factors

• Age
• Family history of CRC/adenomas
• Colorectal adenomas
• Advanced Lesions villous gt 1cm multiple
  dysplasia
• Previous CRC
• IBD (UC, Crohns)
• Environmental factors

6
CRC Dietary Life Style Factors
Positive Association
Inverse Association
energy intake total fat saturated fat red
meat protein simple sugars alcohol smoking iron BM
I
fiber vegetables fruits antioxidant
vitamins calcium/vitamin D selenium folate fish
oil (?-3 FA) physical activities HRT/BCP ASA/NSAID
S
7
CRC Incidence Mortality Rates Canada, 1969-1997
M
incidence
Rates per 100,000
F
M
mortality
F
8
(No Transcript)
9
Does Diet Play a Role in the Development of CRC?
???
10
Diet Cancer

• 35 (10 70) of all cancers in US attributable
  to dietary factors
• 90 of CRC in US preventable by dietary
  modifications
• Doll Peto (J Natl Cancer Inst 1981)
• 1/3 of 500,000 cancer deaths in US each year due
  to dietary factors
• McGinnis MJ (JAMA 1993)

11
Cancer Nature vs. NurtureN Engl J Med 2000
343 78

• 44,788 pairs of twins in Scandinavian countires
• Cancer Site Attributable to Envt Factors
• Prostate 58
• Colorectum 65
• Breast 73

12
Diet CRC Epidemiological Evidence

• International variations
• High incidence areas
• westernized, industrialized, urbanized,
• economically well-off
• Within country variations
• Temporal variations
• Migration studies rapid changes

13
Diet CRC Evidence
Correlation
Case-control
Prospective
Intervention
Weight of Evidence
Degree of Association
14
WCRF/AICR 1997
15
Epidemiologic Studies Limitations

• Imprecise estimation of dietary intake
• Lack of qualitative data on dietary habit, food
  preparation cooking methods
• Incomplete control or correction of potential
  confounding factors
• Selection bias, appropriateness of control
• Study subjects - representative of the general
  population?
• Lack of variation in dietary intake within a
  population - Narrow range of dietary intake
• Short duration of f/u small sample size
• Genetic variability not accounted for

16
Diet CRC Intervention TrialsGold Standard?

• Ideal
• Many thousands of subjects
• Many years of f/u
• CRC end point
• 1 Prevention
• Possible
• 600 1200
• A few years of f/u
• Adenomas end point
• 2 Prevention

17
CRC Intervention Strategies
SEPB
High Risk Groups

• Familial CRC syndromes
• FAP, HNPCC
• Family history
• CRC, adenomas
• Personal history
• prior CRC, adenomas
• HNPCC-associated
• cancers
• IBD

• Adenomas
• ACF
• Mucins
• Mitotic index
• DNA aneuploidy
• Proliferation markers
• Molecular markers

Surrogate End Point Biomarkers
18
Intervention Studies Limitations

• Short duration of follow-up
• Uncertain relevant induction time
• Small number of subjects
• Biologically relevant dose?
• Optimal timing of intervention?
• Timing of intervention wrong
• Use of SEPB
• Adoption of treatment diet by controls
• Genetic variability not accounted for

19
Prevention
Treatment
Intraepithelial Neoplasia (ACF ? adenoma)
Symptomatic
Diagnosable
In Utero
Childhood
Early adulthood
0
5
10
15
20
25
30
Time (Years)
20
Intervention Studies

• Not an epidemiological gold standard
• Positive trials Good evidence that the dietary
  factors used are protective
• Null or detrimental effect Inconclusive

Byers T, Ca Cancer J Clin 1999
21
Fiber CRC
22
Fiber CRCCombined Analysis of 13 Case-Control
StudiesHowe GR, et al. J Natl Cancer Inst 1992
84 1887
Study R.R. 95 C.I. Argentina
0.07 0.02 - 0.26 Australia I 1.57 0.86 -
2.90 Australia II 0.21 0.13 -
0.35 Belgium 0.43 0.28 - 0.65 Canada
0.61 0.43 - 0.87 China 0.68 0.46 -
1.00 France 0.50 0.24 - 1.15 Greece
0.47 0.13 - 1.66 North American Chinese
0.41 0.27 - 0.63 Singapore 0.36 0.10 -
1.34 Spain 0.34 0.15 - 0.75 U.S.A. I
0.74 0.51 - 1.07 U.S.A. II 0.35 0.19 -
0.63 All studies combined 0.51 0.44 - 0.59
23
Fiber CRCCombined Analysis of 13 Case-Control
StudiesHowe GR, et al. J Natl Cancer Inst 1992
84 1887
Cases 5,287 Controls 10,470
plt0.001, trend
RR
lt10.1
31.2
Dietary Fiber Intake (g/d)
24
CRC FiberPooled-Analysis of Case-Control
Studies
Friedenreich CM, et al. Epidemiology 1994
25
CRC Mortality FiberSummary of Prospective
Studies
26
(No Transcript)
27
CRC FiberSummary of Prospective Studies (US)


28
Fiber CR Adenoma The PLCO Cancer Screening
Trial
OR

• The Prostate, Lung, Colorectal,
• Ovarian (PLCO) Cancer
• Screening Trial
• 43,611 participants in 10 US cities
• Flexible sigmoidoscopy
• 3591 w/ distal CR adenomas
• 10th 90th 12.6 36.4 g/day
• Most beneficial effects
• fiber from grain cereals
• fiber from fruits
• Lancet 2003 361 1491

P0.002
P0.0006
P0.03
15.4
19.8
24.3
30.6
12.6
Fiber intake (g/day)
29
CRC Fiber The EPIC Study
The European Prospective Investigation into
Cancer Nutrition (EPIC) N 519,978 10
European Countries 4.5 years f/u Total fiber, not
specific type of fiber Lancet 2003 361 1496
P-trend0.005
Fiber (g per day)
30
Why the differences?

• Different sources, types, amounts of dietary
  fiber in different populations studied
• NHS 9.8 g/d 24.9 g/d
• Cereal fiber (Q5 4.8 g/d)
• Homogenous cohort
• EPIC 12.6 g/d 33.1 g/d
• Cereal fiber (Q1 4.7 g/d)
• Several very different population groups w/
  different sources of fiber being preferentially
  consumed
• Different methods used for the analysis of
  dietary fiber

31
CRC FiberSummary of Intervention Studies (SEPB)
32
CRC FiberSummary of RCT I
Study
N
Duration
Outcome
Endpoint
Intervention
HiFib protective if gt 11.0 g/d
LoF(2.2 g/d) vit C, E HiF(22.5g/d) vit C, E
vs. placebo
De Cosse 1989
FAP 58
adenoma recurrence
4 yrs
20 fat calories 50 g fiber/d vs. placebo
Toronto 1994
adenoma 201
no effect
adenoma recurrence
2 yrs
2x2x2 lt25 fat calories 25 g wheat
bran/d ?-carotene (20 mg/d)
fiber alone no effect w/ LoFat protective gt10 mm
APPP 1995
adenoma 424
2 yrs 4 yrs
adenoma recurrence
33
(No Transcript)
34
(No Transcript)
35
CRC FiberSummary of RCT II
36
CRC Fiber Meta-Analysis

• 5 RCTs
• N 4349
• Adenoma recurrence
• 2 4 years of intervention
• No protective effect associated with increased
  dietary fiber

Cochrane Database Syst Rev 2002
37

• The Japanese eat very little fat and suffer fewer
  heart attacks than the British or Americans
• On the other hand, the French eat a lot of fat
  and also suffer fewer heart attacks than the
  British or Americans
• The Japanese drink very little red wine and
  suffer fewer heart attacks than the British or
  Americans
• The Italians drink excessive amounts of red wine,
  and also suffer fewer heart attacks than the
  British or Americans
• Conclusion Eat and drink what you like. Its
  speaking English that kills you.

38
Fiber Doesnt Work.
Does it?

• Short duration of follow-up
• Uncertain relevant induction time
• Fiber works but slowly?
• Small number of subjects
• Biologically relevant dose?
• Optimal timing of intervention?
• Timing of intervention wrong
• Use of SEPB
• Predisposed subjects 2 prevention
• Genetic variability not accounted for

39
Does Dietary Fiber Protect Against CRC?

• Probably Yes
• Definitive answers to questions about dietary
  fiber and CRC are probably beyond the reach of
  both observational epidemiologic studies and
  randomized controlled trials
• Prevention strategy Overall diet in conjunction
  w/ other lifestyle factors

40
WCRF/AICR 1997
41
Diet CRC New Paradigm Portfolio or Totality of
Evidence
Correlation
Case-control
Prospective
Animal
Intervention
Mechanisms
Evidence
Gene-Nutrient Gene-Gene
Genetic Variability
42
(No Transcript)
43
Relationship between Diet CRC
Overall diet
High correlation
Particular foods
Individual Nutrients
Low correlation
44
CRC, Diet Lifestyle FactorsUnifying Hypothesis
Diet Exercise Obesity ETOH
? Insulin IGF-1 other hormones
? TG glucose
CRC
Insulin Resistance

• ? TG
• ? glucose
• ? VLDL
• ? HDL
• Insulin
• IGFs
• ? C-peptide

in subjects with CRC polyps c/w controls
McKeown-Eyssen, Cancer Epidemiol Biomarkers Prev
1994 3 687
45
Adherence to the AICR Cancer Prevention
Recommendations Morbidity Mortality

• The Iowa Womens Health Study
• 29,564 women 55 69 yrs at baseline
• 1986 1998
• 14 ? 9 recommendations
• Smoking

Cancer Epidemiol Biomarkers Prev 2004 13 1114
46
AICR Recommendations
47
Adherence to the AICR Recommendations Morbidity
Mortality
Cancer Epidemiol Biomarkers Prev 2004 13 1114
48
Adherence to the AICR Recommendations Population
Attributable Risk

• The proportion of cancer incidence, cancer
  mortality, and CVD mortality that theoretically
  would have been avoidable if the entire cohort
  had followed 6 9 recommendations
• ( never smoked)
• 22 for cancer incidence (31)
• 11 for cancer mortality (30)
• 4 for CVD mortality (22)

Cancer Epidemiol Biomarkers Prev 2004 13 1114
49
Diet CRC Summary

• What works Calcium
• What doesnt work Antioxidants
• What might work Vegetables, Fruits, Fiber, ?Fat,
  ?Red Meat, Vitamin D
• Dual effects? Folate, Selenium
• Cocktail?
• Chemopreventive agents

50
What Should We Recommend to Prevent CRC?
51
(No Transcript)
52
CRC Diet Recommendations

• Eat each of the 5 food groups daily
• Meat, dairy products, grain, fruits, vegetables
• Reduce total fat intake to lt25-30
  saturated fat to lt10 of total calories
• Eat red meat infrequently
  substitute with chicken fish
• Eat 7-8 servings of fresh vegetables fruits
• Eat more fiber-rich foods
• gt30-35 g fiber/day all sources types of fiber
  whole-grain cereals, grains, legumes
• Other dietary lifestyle modifications
• ETOH smoking exercise weight

53
Oh, Yes. Screening.