Reference
Complete the bibliographic reference for the article according to AJE format.

Slattery, M., et al. Associations between BMI, energy intake, energy expenditure, VDR genotype and colon and rectal cancers. Cancer Causes and Control. 2004; 15: 863-872.

Category of HuGE information
Specify the types of information (from the list below) available in the article:

1. Prevalence of gene variant
2. Gene-disease association
3. Gene-environment interaction
4. Gene-gene interaction
5. Genetic test evaluation/monitoring

2. Gene-disease association
3. Gene-environment interaction

The main purpose for conducting the study was to determine if the associations between BMI, physical activity, and energy intake, components of energy balance, and colon and rectal cancer are altered by VDR genotype.

Gene(s)
Identification of the following:

1. Gene name
2. Chromosome location
3. Gene product/function
4. Alleles
5. OMIM #
6. GDPInfo link

1. Gene name: VDR (Vitamin D receptor)
2. Chromosome location: 12q12-q14
3. Gene function: Nuclear receptor involved in the regulation of many physiological processes, including cell growth and differentiation and metabolic homeostasis.
4. Alleles: BsmI (B or b), poly-A (S or L), and Fok1 (F or f)
5. OMIM #: 601769
6. Go to GDPInfo Genes A-Z result

• BMI
• Energy Intake
• Energy Expenditure

Health outcome(s)
Identification of the major health outcome(s) studied

• Colon Cancer
• Rectal Cancer

1. Case-control
2. Cohort 
3. Cross-sectional
4. Descriptive or case series
5. Clinical trial
6. Population screening

1. Disease case definition
2. Exclusion criteria
3. Gender
4. Race/ethnicity
5. Age
6. Time period
7. Geographic location
8. Number of participants

1. Disease case definition: First primary colon cancer diagnosis between October 1, 1991 to September 30, 1994
2. Exclusion criteria: No previous history of colorectal cancer and no known familial adenomatous polyposis, ulcerative colitis or Crohn's disease
3. Gender: male and female
4. Race/ethnicity: Caucasian, Hispanic, African-American
5. Age: 30 to 79 at time of diagnosis
6. Time period: Interviewedfrom February 1991 and May 1994
7. Geographic location: Northern California and the state of Utah
8. Number of participants: 1346 (80.8%)

1. Disease case definition: First primary tumor in the rectosigmoid junction or rectum between May 1997 to May 2001
2. Exclusion criteria: No previous history of colorectal cancer and no known familial adenomatous polyposis, ulcerative colitis or Crohn's disease
3. Gender: male and female
4. Race/ethnicity: Caucasian, Hispanic, African-American, Asian, Native American
5. Age: 30 to 79 at time of diagnosis
6. Time period: Interviewed from October 1997 and January 2002
7. Geographic location: Northern California and the state of Utah
8. Number of participants: 952 (71.6%)

1. Control selection criteria
2. Matching variables
3. Exclusion criteria
4. Gender
5. Race/ethnicity
6. Age
7. Time period
8. Geographic location
9. Number of participants

1. Control selection criteria: Randomly selected from membership lists at the KPMCP. Utah controls 65 and older were randomly selected from HCFA lists and controls younger than 65 selected from random digit-dialing and driver's license lists.
2. Matching variables: Sex and five-year age groups
3. Exclusion criteria: Not specified
4. Gender: Male and female
5. Race/ethnicity: Caucasian, Hispanic, African-American
6. Age: 30 to 79
7. Time period: Interviewed from February 1991 and May 1994
8. Geographic location: Northern California and the state of Utah
9. Number of participants: 1544 (73.2%)

1. Control selection criteria: Randomly selected from membership lists at the Kaiser Permanente Medical Care Program of Northern California (KPMCP). Utah controls 65 and older were randomly selected from HCFA (Medicare) lists and controls younger than 65 selected from random digit-dialing and driver's license lists.
2. Matching variables: Sex and five-year age groups
3. Exclusion criteria: Not specified
4. Gender: Male and female
5. Race/ethnicity: Caucasian, Hispanic, African-American, Asian, Native American
6. Age: 30 to 79
7. Time period: Interviewed from October 1997 and January 2002
8. Geographic location: Northern California and the state of Utah
9. Number of participants: 1205 (68.8%)

Cohort definition
For study designs 2, 3, and 6, define the following if available:

1. Cohort selection criteria
2. Exclusion criteria
3. Gender
4. Race/ethnicity
5. Age
6. Time period
7. Geographic location
8. Number of participants (% of total eligible)

1. Environmental factor
2. Exposure assessment
3. Exposure definition
4. Number of participants with exposure data (% of total eligible)

1. Environmental factor: BMI
2. Exposure assessment: BMI was calculated as weight/height ² and used as an estimate of obesity. Height and weight measured at time of interview and weight was reported for two and five years prior to referent date. Weight from five years prior was used to calculate BMI only if weight from two years prior was missing.
3. Exposure definition: BMI <25 (reference), BMI 25-29, BMI >= 30
4. Number of participants with exposure data: not specified

1. Environmental factor: Energy Intake
2. Exposure assessment:Dietary intake was obtained using as adaptation of the CARDIA diet history. Using the Minnesota Nutrition Coding Center (NCC) nutrient database they converted dietary intake data into nutrient data.
3. Exposure definition:High, intermediate, and low energy intake.
4. Number of participants with exposure data: not specified

1. Environmental factor: Energy Expenditure
2. Exposure assessment:A detailed questionnaire was used to assess physical activity patterns during the referent period and 10 and 20 years prior to referent date.
3. Exposure definition:Indicator of long-term vigorous activity was used to estimate physical activity level as high, intermediate, and none.
4. Number of participants with exposure data: not specified

1. Gene
2. DNA source
3. Methodology
4. Number of participants genotyped (% of total eligible) 

1. Gene: VDR
2. DNA source: Not specified
3. Genotyping method: PCR-based amplification of r estriction fragment length analysis only. Standard PCR protocol was used with a PE9600 thermocycler. Amplification times and annealing temperatures varied depending on the assay. Electrophoresis of the products was carried out in a 6% denaturing polyacrylamide gel at 70 W for three hours. Repeat length was determined for those alleles of different sizes through sequencing.
4. Number of participants genotyped:
   • Cases: 1956 colon and rectal cancer cases were genotyped (85.1% of the total eligible)
   • Controls: 2174 colon and rectal controls were genotyped (79.1% of the total eligible) controls

They did not have genotype information for both BsmI and poly-A polymorphisms for all individuals. Because the BsmI and poly-A alleles are in linkage disequilibrium with each other they combined the data for these two polymorphisms and studied them together. They used the notation SS/BB, LL/bb, and all others were denoted as “mostly SL/Bb.”

Separate analysis was done for the Fok1 polymorphism.

Analysis
Comment on the analysis carried out by the author(s), e.g. matching, modeling or statistical tests used. Were the analyses appropriate?

The authors matched on sex and five-year age groups. Logistic models included several potential confounders: age at selection, sex, BMI, long-term vigorous physical activity, energy intake, dietary calcium, and dietary fiber. Because the population was 87-89% Caucasian they did not evaluate race-specific associations. However they still should have controlled for race as a potential confounder. Race, alcohol consumption, and tobacco use are all additional factors that are known risk factors for colorectal cancer that should have been considered as potential confounders.

VDR is involved in metabolic homeostasis, so the researchers thought it was reasonable to assume it might affect cancer risk when combined with different components of energy balance, such physical activity or diet.

The authors assessed the interactions between BMI, energy intake, and energy expenditure with the VDR genotype on both multiplicative and additive scales. The Wald ?² test was used to assess if changes in ORs holding VDR genotype constant while varying BMI, energy intake, and energy expenditure were statistically meaningful.

1. Prevalence of gene variant
2. Gene-disease association
3. Gene-environment interaction
4. Gene-gene interaction
5. Genetic test evaluation/monitoring

1. Prevalence of gene variant: The SS/BBVDR polymorphism was seen in 19.1% of colon cancer controls and 16.4% of rectal cancer controls. The ff VDR polymorphism was seen in 13.5% of colon cancer controls and 13.1% of rectal cancer controls. See Table 1
   
   
2. Gene-disease association: The authors assessed the SS and BB genotypes separately and found both were inversely associated with colon cancer (OR=0.79 and 0.84 respectively) compared to the LL and bb genotypes. The data for these calculations is not shown. Those individuals with the FF genotypes were at greater risk for colon cancer than those with the ff genotype (OR=1.28). In addition, I produced ORs for each genotype for rectal cancer and attributable risk fractions for colon and rectal cancer with respect to VDR genotype.
   See Table 2 | See Table 3
   
   
3. Gene-environment interaction: The SS/BBVDR genotypes were associated with a significant increase in colon cancer risk among obese individuals (OR=3.50), but not rectal cancer risk. The Fok1 VDR genotype was not significantly associated with BMI. See Table 4
   
   
4. Gene-environment interaction: The FF VDR genotype increased the risk of colon cancer for those with high physical activity (OR=2.30). The ff VDR genotype increased the risk of colon and rectal cancer (OR= 3.46 and 3.10, respectively) for those with low physical activity levels. The Bsm1/ployA VDR genotypes were not significantly associated with physical activity for colon cancer or rectal cancer at the 0.05 level.
   See Table 5
   
   
5. Gene-environment interaction: The Fok1 polymorphism increased the risk of colon cancer for those with the highest level of energy intake. The FF VDR genotype decreased the risk of rectal cancer for those with low energy intake compared to those with the ff genotype (OR=0.58). The Bsm1/poly A VDR genotypes were not significantly associated with energy intake for colon or rectal cancer.
   See Table 6

The authors assessed colorectal cancer risk associated with the VDR genotype by level of energy intake. The SS/BB VDR genotype reduced the risk of proximal tumors among those who had low energy consumption. The ff VDR genotype lowered the risk of distal tumors among those with high-energy intake.
See Table 7 | See Table 8

Slattery et al. concluded that the data provide some support for an interaction between obesity and the VDR genotype in risk associated with colon cancer. The associations of physical activity and energy intake with colon cancer were not as strong as the association of obesity with colon cancer. There may be some possible genetic regulation of energy balance as it relates to colon cancer. Energy intake may modify the effects of the VDR genotype on rectal cancer risk. (1)

One of the strengths of this study is the large sample size, but even with such a large sample size there is limited power to look at combinations of BMI, energy intake, and energy expenditure. The attributable fraction for the association found between obesity and the SS/BB genotype and the risk of colon cancer was only 3.5%. The attributable fraction for the association found between high level of energy expenditure and the FF genotype and risk of colon cancer was 7.8%. The attributable fraction for no physical activity and the ff genotype and risk of colon cancer was 4.0%. The inverse association found between low energy intake and the ff genotype and rectal cancer suggests a protective effect. The range of attributable risks and odds ratios led the authors to believe that VDR may be acting through multiple pathways to influence carcinogenesis.