Skip to main content
PMC home page
British Journal of Cancer logoLink to British Journal of Cancer
. 2011 May 10;105(1):162–169. doi: 10.1038/bjc.2011.172

Body mass index in early adulthood and colorectal cancer risk for carriers and non-carriers of germline mutations in DNA mismatch repair genes

A K Win 1, J G Dowty 1, D R English 1,2, P T Campbell 3,4, J P Young 5, I Winship 6, F A Macrae 7, L Lipton 8, S Parry 9,10, G P Young 11, D D Buchanan 5, M E Martínez 12,13, E T Jacobs 12,13, D J Ahnen 14, R W Haile 15, G Casey 15, J A Baron 16, N M Lindor 17, S N Thibodeau 17, P A Newcomb 4, J D Potter 4, L Le Marchand 18, S Gallinger 19, J L Hopper 1, M A Jenkins 1,*
PMCID: PMC3137400  PMID: 21559014

Abstract

Background:

Carriers of germline mutations in DNA mismatch repair (MMR) genes have a high risk of colorectal cancer (CRC), but the modifiers of this risk are not well established. We estimated an association between body mass index (BMI) in early adulthood and subsequent risk of CRC for carriers and, as a comparison, estimated the association for non-carriers.

Methods:

A weighted Cox regression was used to analyse height and weight at 20 years reported by 1324 carriers of MMR gene mutations (500 MLH1, 648 MSH2, 117 MSH6 and 59 PMS2) and 1219 non-carriers from the Colon Cancer Family Registry.

Results:

During 122 304 person-years of observation, we observed diagnoses of CRC for 659 carriers (50%) and 36 non-carriers (3%). For carriers, the risk of CRC increased by 30% for each 5 kg m–2 increment in BMI in early adulthood (hazard ratio, HR: 1.30; 95% confidence interval, CI: 1.08–1.58; P=0.01), and increased by 64% for non-carriers (HR: 1.64; 95% CI: 1.02–2.64; P=0.04) after adjusting for sex, country, cigarette smoking and alcohol drinking (and the MMR gene that was mutated in carriers). The difference in HRs for carriers and non-carriers was not statistically significant (P=0.50). For MLH1 and PMS2 (MutLα heterodimer) mutation carriers combined, the corresponding increase was 36% (HR: 1.36; 95% CI: 1.05–1.76; P=0.02). For MSH2 and MSH6 (MutSα heterodimer) mutation carriers combined, the HR was 1.26 (95% CI: 0.96–1.65; P=0.09). There was no significant difference between the HRs for MutLα and MutSα heterodimer carriers (P=0.56).

Conclusion:

Body mass index in early adulthood is positively associated with risk of CRC for MMR gene mutation carriers and non-carriers.

Keywords: body mass index, colorectal cancer, mismatch repair gene

Lynch syndrome, historically known as hereditary non-polyposis colorectal cancer (Jass, 2006), refers to colorectal and other cancers caused by germline mutations in DNA mismatch repair (MMR) genes MLH1, MSH2, MSH6 and PMS2 (Vasen et al, 1999). Approximately 1 in 3000 people in the general population carry a mutation in an MMR gene (Dunlop et al, 2000). These MMR gene mutation carriers are at substantially increased risk of colorectal cancer (CRC) with an estimated cumulative risk to age 70 years between 40% and 70% depending on the carrier's sex and the MMR gene that is mutated (Chen et al, 2006; Jenkins et al, 2006; Senter et al, 2008; Baglietto et al, 2010). Physical characteristics or environmental exposures of the mutation carriers could also modify their risks of developing CRC (Jenkins et al, 2007). Identifying modifiers of CRC risk for carriers of MMR gene mutations is important for understanding carcinogenesis. Identifying potentially protective or harmful and avoidable risk factors also creates opportunities for mutation carriers to reduce their risks of life-threatening diseases.

Previous studies have provided evidence for the existence of modifiers of CRC risk for MMR gene mutation carriers. For example, CRC risk for carriers is positively associated with alcohol consumption (Watson et al, 2004) and negatively associated with fruit consumption and dietary fibre intake (Diergaarde et al, 2007). For smoking, CRC risk for carriers is positively associated with current smoking (Diergaarde et al, 2007; Pande et al, 2010), but negatively or not associated with former smoking (Diergaarde et al, 2007; Pande et al, 2010). However, the association between body size and CRC risk for MMR gene mutation carriers is yet to be established.

One CRC risk factor for people in the general population is body size, a collective term for body mass index (BMI), waist circumference and height (Dai et al, 2007; Larsson and Wolk, 2007; Moghaddam et al, 2007; Renehan et al, 2008). Previous case–control studies of CRC cases with a family history (therefore, likely to be enriched for MMR gene mutation carriers) compared with population-based controls who were unselected for family history, have shown that current BMI is positively associated with CRC risk (Slattery et al, 2003; Campbell et al, 2007). However, studies that have examined the association between current BMI and the occurrence of tumours with microsatellite instability (MSI), a common characteristic of CRCs caused by MMR defects, have reported no statistically significant associations with this form of disease (Slattery et al, 2000; Campbell et al, 2010). One study that examined the association between BMI and colorectal adenoma risk for MMR gene mutation carriers observed a positive association for males but not females (Botma et al, 2010).

Identifying an association of BMI in early adulthood on subsequent cancer risk would be important for MMR gene mutation carriers who learn their mutation status as adults in their 20s, an increasingly common occurrence because of increased systematic testing in the population, with the consequent opportunity to reduce their risk of disease. In this study, we estimated an association between BMI at age 20 years and CRC risk for MMR gene mutation carriers. As a comparison, we also estimated the association for non-carriers to investigate whether BMI has a differential effect on risk for CRC by mutation status.

Materials and methods

Recruitment

Subjects comprised carriers and non-carriers of pathogenic mutations (see below) in the MMR genes MLH1, MSH2, MSH6 and PMS2, recruited and genetically characterised by the Colon Cancer Family Registry (Colon CFR). Study designs and recruitments for the Colon CFR can be found at http://epi.grants.cancer.gov/CFR/ and have been published in detail (Newcomb et al, 2007). Probands were either a recently diagnosed CRC case reported to a population-complete cancer registry or an attendee at a family cancer clinic. They were recruited between 1997 and 2007 and ascertained from family cancer clinics from Australia (Melbourne, Adelaide, Perth, Brisbane and Sydney), New Zealand (Auckland) and the United States (Mayo Clinic, Rochester, Minnesota and Cleveland) or from population-complete cancer registries in the United States (Puget Sound, Washington State; the State of Minnesota; Los Angeles, California; Arizona; Colorado; New Hampshire; North Carolina; and Hawaii), Australia (Victoria) and Canada (Ontario).

Probands were asked for permission to contact their relatives to seek their enrolment in the Colon CFR. For probands ascertained from family cancer clinics, there were pre-specified rules consistent across recruiting centres governing which relatives were to be approached for recruitment (Newcomb et al, 2007). For probands ascertained from population-complete cancer registries, first-degree relatives were recruited at all centres and in some centres, recruitment extended to more distant relatives. Written informed consent was obtained from all study participants, and the study protocol was approved at each Colon CFR centre.

Data collection

At recruitment, baseline information on demographics, personal characteristics, personal and family history of cancer, cancer screening history, history of polyps, polypectomy and other surgery were obtained from all participants. Participants were followed-up approximately 5 years after baseline to update all this information. The questionnaires are available from http://cfrisc.georgetown.edu/isc/ dd.questionnaires.do. Reported cancer diagnoses and age at diagnoses were confirmed, where possible, using pathology reports, medical records, cancer registry reports and/or death certificates. Blood samples and tumour tissues were collected for genetic testing. The present study was based on all available baseline and follow-up data.

Self-reported height and weight at approximately age 20 years were collected using standardised personal interviews (University of Southern California Consortium), telephone interviews (University of Southern California Consortium, Fred Hutchinson Cancer Research Centre and University of Melbourne) or mailed questionnaires (University of Hawaii, Cancer Care Ontario and Mayo Clinic).

Mutation screening and testing

Testing for MLH1, MSH2, MSH6 and PMS2 mutations was performed for all case probands ascertained from family cancer clinics and for all probands from population-based ascertainment who had a colorectal tumour displaying evidence of impaired MMR function as evidenced by either MSI, or by lack of MMR protein expression by immunohistochemistry. Mutation testing was performed by Sanger sequencing or denaturing high-pressure liquid chromatography, followed by confirmatory DNA sequencing. Large insertion and deletion mutations were detected by multiplex ligation-dependent probe amplification according to the manufacturer's instructions (MRC Holland, Amsterdam, The Netherlands) (Southey et al, 2005; Newcomb et al, 2007; Senter et al, 2008). All participants who donated a blood sample, and who were relatives of probands with a pathogenic mutation, underwent testing for the same mutation identified in the proband.

Study sample

For this study, we included male and female probands and their participating relatives who were confirmed carriers and confirmed non-carriers of an MMR gene mutation. In total, 1468 carriers and 1365 non-carriers were identified. Of these, 144 (10%) carriers and 146 (11%) non-carriers were excluded due to missing data on height or weight at age 20 years, leaving a total of 1324 carriers (252 from population-based sources) and 1219 non-carriers (109 from population-based sources) for the analysis. Excluded subjects did not differ in baseline characteristics from subjects who entered into the analyses (age, sex, country and gene mutated) (data not shown).

We were unable to analyse current BMI in this study, as it was not available 1–2 years prior to diagnosis or censored age for 833 subjects (636 carriers and 197 non-carriers), which is 33% of the total sample size. This was composed of 450 CRC cases (425 carriers and 25 non-carriers), 215 subjects (140 carriers and 75 non-carriers) with another cancer and 164 subjects with a prior polypectomy (70 carriers and 94 non-carriers) and 4 subjects (1 carrier and 3 non-carriers) who did not report current weight at last contact.

Definitions

A pathogenic mutation was defined to be a variant that was predicted to result in a stop codon, a frameshift mutation, a large insertion or deletion, or a missense mutation previously reported in the scientific literature to be pathogenic. Age at CRC was the age at first diagnosis of CRC. Height and weight were defined as self-reported height and weight at age 20 years, respectively. BMI at age 20 years was calculated as weight as reported at age 20 years in kilograms divided by height in metres squared (kg m–2). Body mass index was categorised as ⩽18.49 kg m–2 (underweight), 18.50–24.99 kg m–2 (normal weight), 25.00–29.99 kg m–2 (overweight) or ⩾30.00 kg m–2 (obese) according to World Health Organization criteria (World Health Organization, 1998). T and N staging of CRC was categorised according to the American Joint Committee on Cancer staging (Edge et al, 2009).

Statistical analysis

Mismatch repair gene mutation carriers and non-carriers were treated as a cohort from birth. A proportional Cox regression model with age as the time scale was used to estimate hazard ratios (HR) relating BMI to CRC risk for carriers and non-carriers. Time at risk started at birth and ended at age of diagnosis of cancer of any site, polypectomy, death or last contact, whichever occurred first. The rationale for censoring at diagnosis of any cancer was that the resultant treatment and surveillance might alter risk of subsequent cancers, including CRC and this would introduce bias if the first cancer was associated with body size. Similarly, as polypectomy may reduce CRC risk, censoring at time of polypectomy was also required to avoid potential bias. Of the total 1324 mutation carriers, observation time ended at age of diagnosis for 659 CRC cases, and was censored at age of polypectomy for 173 subjects, at age of diagnosis of another cancer for 144 subjects, at age of death for 4 subjects and at age of last contact for 344 subjects. Of the total 1219 non-carriers, time of observation ended at the age of diagnosis for 36 CRC cases, at age of polypectomy for 176 subjects, at age of diagnosis of another cancer for 101 subjects, at age of death for 14 subjects and at age of last contact for 892 subjects.

Since some subjects were ascertained because they were from multiple-case CRC families, and CRC cases were preferentially tested for MMR gene mutations, the selection of subjects was not random with respect to disease status. To adjust for this non-random ascertainment, we used the weighted cohort approach described by Antoniou et al (2005), which has been used for modifier studies of cancer risk for carriers of rare genetic mutations (e.g., Andrieu et al, 2006; Antoniou et al, 2008; Win et al, 2011). A simulation study of this approach (Antoniou et al, 2005) showed that allowing for non-random sampling of subjects by using rates from an external referent population removed bias when the external rates were correctly specified, and reduced bias if the external rates were not completely accurate.

Age-specific incidence rates of CRC for MMR gene mutation carriers were previously estimated as described in detail by Pande et al (2010). These age-specific incidence rates were used to calculate sampling fractions to weight the proportion of affected and unaffected carriers in each age stratum, so that the proportion of affected carriers in each age group equalled the population proportions (Supplementary Table S1).

Due to the rarity of mutations in the population, age-specific incidence rates for non-carriers were assumed to equal those for the general population. Data on age-specific (5-year intervals) population cancer incidence rates were obtained from the American Cancer Society Cancer Facts and Figures 2008 (American Cancer Society, 2008). Using these age-specific incidence rates, we calculated sampling fractions to weight the proportion of affected and unaffected non-carriers in each age stratum.

Body mass index was fitted as a continuous variable and as a categorical variable based on WHO BMI classification (World Health Organization, 1998). In order to determine whether a non-linear association with a continuous variable was a superior fit compared with a linear model, we fitted and compared fractional polynomial models. Models were compared using Wald's tests. The proportional hazards assumption was tested by examining the relationship between the scaled Schoenfeld residuals and survival time.

To control for potential confounders present at the age at which BMI was reported (i.e., 20 years old), we adjusted for (i) sex, (ii) country of recruitment, (iii) cigarette smoking at age 20 (never, ever) and (iv) alcohol drinking at age 20 (ever, never) in both carriers and non-carriers, and further adjusted for (v) specific MMR gene that was mutated in carriers. Interactions between (i) mutation carrier status and BMI, (ii) sex and BMI and (iii) the specific MMR gene mutated and BMI were tested by adding an interaction term in a multivariable model. We did not adjust for current BMI because of the well-known bias introduced by adjusting for variables that are in the causal pathway or partly caused by the exposure of interest. The associations between BMI and CRC risk were estimated separately for carriers of mutation in each MMR gene, and carriers of mutations in the MutLα heterodimer (MLH1 and PMS2) and MutSα heterodimer (MSH2 and MSH6) genes.

Two sensitivity analyses were conducted: (i) to determine whether unverified self-reported CRC cases or cases reported by relatives influenced the associations with BMI, we performed analyses after excluding these cases and (ii) to determine whether censoring at polypectomy influenced the associations with BMI, we performed analyses of colorectal neoplasia that is CRC and colorectal polyps combined, as the outcome.

We applied the Huber–White robust variance estimation by clustering on family membership to allow for any correlation of risk between family members (Rogers, 1993; Williams, 2000). All statistical tests were two sided and, following convention, statistical significance for testing a predetermined null hypothesis was set at P<0.05. All statistical analyses were performed using Stata 10.0 (StataCorp, 2007).

Results

The study comprised 1324 carriers (737 females) of MMR gene mutations (500 in MLH1, 648 in MSH2, 117 in MSH6 and 59 in PMS2) from 498 families contributing a total of 58 868 person-years, of which 659 (50%) were diagnosed with CRC, and 1219 non-carriers (712 females) from 287 families contributing 63 436 person-years, of which 36 (3%) were diagnosed with CRC. Of all participants, 839 carriers and 983 non-carriers (72%) were recruited in Australia or New Zealand, 367 carriers and 199 non-carriers (22%) in the United States and 118 carriers and 37 non-carriers (6%) in Canada. Baseline characteristics of the study subjects are summarised in Table 1.

Table 1. Baseline characteristics of study subjects (MMR gene mutation carriers and non-carriers).

  Mutation carriers
Non-carriers
  CRC-affected (N=659) number (%) CRC-unaffected (N=665) number (%) CRC-affected (N=36) number (%) CRC-unaffected (N=1183) number (%)
Sex
 Female 332 (50) 405 (61) 19 (53) 693 (59)
         
Age (year)a, mean (s.d.) 44.0 (11.1) 44.9 (13.7) 56.8 (14.8) 51.9 (15.6)
         
MMR gene mutated
 MLH1 273 (41) 227 (34)    
 MSH2 308 (47) 340 (51)    
 MSH6 40 (6) 77 (12)    
 PMS2 38 (6) 21 (3)    
         
Country of recruitment
 Australia or New Zealand 355 (54) 484 (72) 16 (44) 967 (81)
 USA 216 (33) 151 (23) 15 (42) 184 (16)
 Canada 88 (13) 30 (5) 5 (14) 32 (3)
         
Source of recruitment
 Clinic based 473 (72) 599 (90) 23 (64) 1087 (92)
 Population based 186 (28) 66 (10) 13 (36) 96 (8)
         
Cigarette smoking b
 Never 319 (49) 363 (55) 15 (44) 652 (55)
 Ever 338 (51) 299 (45) 19 (56) 527 (45)
 Unknown 2 3 2 4
         
Alcohol consumption c
 Never 233 (36) 193 (31) 16 (44) 417 (36)
 Ever 408 (64) 439 (69) 20 (56) 732 (64)
 Unknown 18 33 0 34
         
Body size, mean (s.d.)
 Height (cm) 170.1 (10.1) 169.8 (9.7) 169.7 (9.5) 169.4 (9.9)
 Weight at age 20 (kg) 66.9 (15.5) 65.4 (14.6) 65.8 (13.0) 64.6 (13.8)
 BMI at age 20 (kg m–2)d 23.0 (4.0) 22.5 (3.7) 22.7 (3.3) 22.4 (3.7)
Open in a new tab

Abbreviations: BMI=body mass index; CRC=colorectal cancer; MMR=mismatch repair.

a

Age at first diagnosis of CRC for affected subjects; age at first polypectomy or diagnosis of another cancer or last contact for unaffected subjects (whichever came first).

b

Cigarette smoking status at age 20 years; cigarette smoking was defined as ever smoking one cigarette per day for 3 months or longer.

c

Alcohol consumption status at age 20 years; alcohol beverages include beer, wine, cider, spirits, mixed drinks or cocktails.

d

Calculated from self-reported height and weight at age 20 years.

Of the 659 CRC cases in carriers, 595 (90%) were confirmed by pathology report or review, SEER or other cancer registries, or hospital or clinical records and 64 (10%) were unconfirmed reports by self or relatives (Table 2). The mean age at CRC diagnosis was 44 (s.d.: 11) years for carriers and 57 (s.d.: 15) years for non-carriers. The mean reported BMI at age 20 years was 22.7 (s.d.: 3.9) kg m–2 for carriers and 22.4 (s.d.: 3.7) kg m–2 for non-carriers. There were no significant differences in mean BMI or the distributions of mutated MMR genes by country (data not shown).

Table 2. Source of colorectal cancer diagnoses verifications and tumour characteristics.

  Carriers number (%) Non-carriers number (%)
Source of verification
 Pathology review/report 569 (86) 25 (69)
 SEER/other cancer registry 17 (3) 0 (0)
 Hospital or clinic record 9 (1) 1 (3)
 Self/relative 64 (10) 10 (28)
     
Tumour site
 Righta 356 (61.1) 15 (58)
 Leftb 119 (20.4) 6 (23)
 Rectosigmoid 30 (5.1) 2 (8)
 Rectum 78 (13.4) 3 (11)
 Unknown 76 10
     
Histology grade
 Well differentiated 46 (10) 1 (5)
 Moderately differentiated 279 (62) 14 (74)
 Poorly differentiated 124 (28) 4 (21)
 Unknown 210 17
     
AJCC stage c
 I 122 (28) 4 (2.5)
 II (A, B, C) 214 (50) 9 (53)
 III (A, B, C) 94 (22) 4 (23.5)
 Unknown 229 19
     
Synchronous
 Yes 44 (12) 2 (11)
 No 313 (88) 17 (89)
 Unknown 302 17
Open in a new tab

Abbreviation: AJCC=American Joint Committee on Cancer.

a

Right colon included caecum, ascending colon, hepatic flexure and transverse colon.

b

Left colon included splenic flexure, descending colon and sigmoid colon.

c

Metastasis status was unavailable; and any of this stage could be stage IV.

Table 3 shows that carriers who were obese (⩾30 kg m–2) had an approximately two-fold (HR: 2.35; 95% confidence interval, CI: 1.30–4.23) increased risk of CRC compared with carriers who had normal BMI (18.5–24.99 kg m–2). Colorectal cancer risk for carriers was estimated to be 30% higher for each increment of 5 kg m–2 in BMI (adjusted HR: 1.30; 95% CI: 1.08–1.58; P=0.01). The corresponding HR for female carriers was 1.28 (95% CI: 0.98–1.69; P=0.08) and for male carriers was 1.33 (95% CI: 1.01–1.76; P=0.04); there was no statistically significant evidence that these HRs were different between men and women (P=0.97).

Table 3. Hazard ratios for association between BMI at age 20 years and CRC risk for MMR gene mutation carriers and non-carriers.

  Total number Total person-years CRC number (%) Univariable analysis
Multivariable analysis
        HR (95% CI)a P HR (95% CI)a,b P
Carriers
 BMI (per 5 kg m–2) 1324 58 868 659 (50) 1.33 (1.11–1.58) 0.002 1.30 (1.08–1.58) 0.007
               
 WHO classification
  Underweight 121 5353 54 (45) 0.91 (0.60–1.38)   1.04 (0.66–1.64)  
  Normal 909 41 014 450 (50) 1.00 (referent)   1.00 (referent)  
  Overweight 233 9971 118 (51) 1.17 (0.83–1.63) 0.01 1.12 (0.78–1.62) 0.06
  Obese 61 2530 37 (61) 2.40 (1.35–4.27)   2.35 (1.30–4.23)  
               
Non-carriers
 BMI (per 5 kg m–2) 1219 63 436 36 (3) 1.86 (1.24–2.77) 0.002 1.64 (1.02–2.64) 0.04
               
 WHO classification
  Underweight 113 6157 3 (3) 0.81 (0.23–2.88)   1.01 (0.28–3.73)  
  Normal 904 47 641 25 (3) 1.00 (referent)   1.00 (referent)  
  Overweight 158 7694 6 (4) 3.28 (1.09–9.84) 0.01 3.62 (1.05–12.55) 0.05
  Obese 44 1944 2 (5) 4.98 (1.17–21.23)   3.00 (0.60–14.97)  
Open in a new tab

Abbreviations: BMI=body mass index; CI=confidence interval; CRC=colorectal cancer; HR=hazard ratio; MMR=mismatch repair.

a

With robust variance estimation for familial correlation in risk.

b

Adjusted for sex, country, cigarette smoking and alcohol drinking in both carriers and non-carriers; and further adjusted for specific MMR gene mutated in carriers.

The HR per 5 kg m–2 in BMI was 1.20 (95% CI: 0.99–1.44; P=0.06) for right-sided colon cancers; 1.03 (95% CI: 0.84–1.27; P=0.76) for left-sided colon and rectosigmoid cancers; and 1.31 (95% CI: 0.84–2.05; P=0.24) for rectal cancers. When a fractional polynomial model was fitted, there was no evidence that the (log) HR was non-linear (P=0.72).

Table 4 presents separate results for carriers of mutations in each MMR gene. There was a positive association between BMI and CRC risk for MLH1 and PMS2 (MutLα heterodimer) mutation carriers combined (HR: 1.36; 95% CI: 1.05–1.76; P=0.02). The HR for MSH2 and MSH6 (MutSα heterodimer) mutation carriers combined was 1.26 (95% CI: 0.96–1.65; P=0.09), and not statistically different from the HR for MutLα heterodimer mutation carriers (P=0.56).

Table 4. HR for association between BMI (per 5 kg m–2) at age 20 years and CRC risk separately for carriers of mutation in each MMR gene.

MMR gene mutated Total number Total person-years CRC cases number (%) Median age of diagnosis (range) HR (95% CI)a P
MLH1 500 21 449 273 (55) 41 (17–73) 1.36 (1.04–1.77) 0.03
MSH2 648 28 560 308 (48) 44 (17–91) 1.28 (0.96–1.70) 0.09
MSH6 117 5870 40 (34) 50 (22–72) 0.84 (0.38–1.80) 0.62
PMS2 59 2989 38 (64) 49 (30–74) 1.52 (0.48–4.85) 0.48
MLH1 and PMS2 (MutLα heterodimer) 559 24 438 311 (56) 43 (17–74) 1.36 (1.05–1.76) 0.02
MSH2 and MSH6 (MutSα heterodimer) 765 34 430 348 (45) 45 (17–91) 1.26 (0.96–1.65) 0.09
Open in a new tab

Abbreviations: BMI=body mass index; CI=confidence interval; CRC=colorectal cancer; HR=hazard ratio; MMR=mismatch repair.

a

adjusted for sex, country, cigarette smoking and alcohol drinking with robust variance estimation for familial correlation in risk.

For non-carriers of MMR gene mutations, CRC risk was estimated to be 64% higher for each increment of 5 kg m–2 in BMI (adjusted HR: 1.64; 95% CI: 1.02–2.64; P=0.04). The corresponding HR for female carriers was 1.40 (95% CI: 0.67–2.91; P=0.37) and for male carriers was 1.82 (95% CI: 0.93–3.53; P=0.08); the difference between sexes was not statistically significant (P=0.32). When a fractional polynomial model was fitted, there was no evidence that the (log) HR was non-linear (P=0.88).

The HRs for carriers and non-carriers were not statistically different (interaction between mutation status and BMI; P=0.50).

Sensitivity analyses showed that (i) after excluding unverified self-reported cases or cases reported by relatives, the adjusted HR for each 5 kg m–2 increment in BMI was 1.37 (95% CI: 1.13–1.66; P=0.002) for carriers and 1.65 (95% CI: 0.97–2.80; P=0.06) for non-carriers and (ii) defining the outcome as colorectal neoplasia (either CRC or polyp), the adjusted HR for each 5 kg m–2 increment in BMI was 1.20 (95% CI: 1.00–1.44; P=0.05) for carriers and 1.27 (95% CI: 0.96–1.66; P=0.09) for non-carriers.

Discussion

We have shown that greater BMI at age 20 years is associated with a higher risk of CRC in later life for both carriers and non-carriers of pathogenic germline mutations in MMR genes. For MMR mutation carriers, we found no statistically significant differences between the strengths of association for men and women, and for carriers of mutations in MutLα or MutSα heterodimer genes. These results suggest that BMI in early adulthood is a potential modifier of CRC risk in later life for MMR gene mutation carriers as a whole.

Our study was based on 1324 carriers and 1219 non-carriers of a confirmed pathogenic mutation in any of four MMR genes, making it the largest study to date of this issue. To overcome bias when subjects are selected on the basis of phenotype, we used a weighted cohort approach (Antoniou et al, 2005), which has been successfully utilised for studies of modifiers of cancer risk of rare genetic mutations (e.g., Andrieu et al, 2006; Antoniou et al, 2008; Win et al, 2011). Familial correlation in the risk of CRC was also accounted for using a robust variance estimation to derive appropriate measure of HR estimate imprecision (Rogers, 1993; Williams, 2000). Another strength was the standardised and uniformly high-quality testing for MMR gene mutations by the Colon CFR (Newcomb et al, 2007).

Potential limitations of this study include the use of self-reported weight and height and the often long period between age 20 years and the interview, meaning that participants often had to recall weight many years in the past. However, other studies have shown that measures of weight taken at age 20–30 years are well correlated with recalled values at age 50–70 years, with correlation coefficients ranging between 0.73 and 0.95 (Stevens et al, 1990; Perry et al, 1995; Troy et al, 1995; Norgan and Cameron, 2000; Tamakoshi et al, 2003). There is a possibility of response bias, where CRC-affected subjects recall their weight at age 20 years differently than unaffected subjects. It is also possible that the recall of weight at age 20 years may have depended on the methods of data collection, which varied by recruitment centre, and, therefore, may have resulted in decreased precision of reported weight data. Since cases with poor survival were less likely to be included in this analysis (as they were unable to provide a blood sample for genetic testing and complete a questionnaire), there is a possibility of survival and selection bias if age at onset and survival of cases were related to BMI and/or mutation status. We acknowledge that exposures such as physical activity and meat consumption that we have not adjusted for in this analysis may be confounders for the observed associations.

One previous study investigated the role of BMI in colorectal neoplasms for MMR gene mutation carriers, but the focus of their study was colorectal adenomas (Botma et al, 2010). They reported that MMR gene mutation carriers who are overweight or obese (current BMI ⩾25 kg m–2) had higher risk of adenomas compared with those with normal BMI (<25 kg m–2) for men (HR: 8.72; 95% CI: 2.06–36.96), but not for women (HR: 0.75; 95% CI: 0.19–3.07). A case–control study of CRC cases who met the revised Bethesda or Amsterdam criteria for Lynch syndrome and population controls observed both increased BMI at age 20 years and current BMI were associated with increased CRC risk for men, but not for women (Campbell et al, 2007). In our study, we did not observe statistical evidence for a difference between men and women for associations of BMI with CRC risk. Consistent with our results, Slattery et al (2003) found that obese people with a family history of CRC had approximately two-fold increased risk of CRC compared with those with normal BMI for both sexes combined (odds ratio: 1.65, 95% CI: 1.01–2.73).

There have been two population-based case–control studies examining the association between BMI and CRC with and without MSI, a phenotype that indicates loss of MMR function somatically, but not necessarily due to inherited mutations. Approximately 50% of CRCs with MSI diagnosed before age 50 years arise in MMR gene mutation carriers and 12% of cases diagnosed at age ⩾50 years are MMR gene mutation carriers (Hampel et al, 2008). These studies found positive associations between current BMI and microsatellite stable tumours, but no association with MSI-high tumours (Slattery et al, 2000; Campbell et al, 2010). It is difficult to compare these findings with our results because MSI tumours include CRCs caused not by an inherited MMR gene mutation but by methylation of MLH1 (Herman et al, 1998; Poynter et al, 2008), and, therefore, these previous studies are probably not generalisable to MMR gene mutation carriers. Using the population-based cases from the Colon CFR (the same source of data for this analysis), Campbell et al (2010) found some suggestion that obese MMR gene mutation carriers, compared with those with normal BMI, had an increased risk of CRC (odds ratio: 3.96; 95% CI: 0.59–26.48) when MSI-high was further stratified by known MMR gene mutation status. As evidenced by the broad CIs, however, the study had insufficient numbers (only 85 carriers) to adequately investigate the association between BMI and CRC due to MMR gene mutations.

We observed a statistically significant association between BMI in early adulthood and CRC risk for MLH1 and PMS2 mutation carriers, but not for MSH2 and MSH6 mutation carriers. However, the associations were not statistically different from each other (P=0.51). The proteins produced by MSH2 and MSH6 from the MutSα heterodimer are responsible for detecting DNA mismatches occurring during replication, whereas the proteins produced by MLH1 and PMS2 from the MutLα heterodimer are responsible for cleavage (Kadyrov et al, 2006; Hewish et al, 2010). The association we observed with MLH1 and PMS2 mutations is consistent with obesity affecting CRC risk by inducing a second hit in the remaining wild-type allele of MLH1 or PMS2. It is also consistent with obesity being a tumour promoter that is more relevant to tumours induced by MMR defects due to mutations in the MutLα heterodimer.

In this study, as all non-carriers are relatives of mutation carriers and, therefore, in general matched for familial factors that may be risk factors for CRC, the two groups are comparable. Given that we observed similar associations of BMI with CRC risk for carriers and non-carriers, we can surmise that having a mutation in an MMR gene does not alter the degree of increased risk imposed by BMI in early adulthood. In contrast, we have shown, using the same analytical design and data source, that non-carrier women were 74% more likely to develop endometrial cancer for every 5 kg m–2 increase in BMI at age 20 (HR: 1.74; 95% CI: 1.27–2.37; P<0.001), while carriers were not more likely to develop endometrial cancer with increased BMI (Win et al, 2011). Given MMR mutation carriers are rare in the general population (Dunlop et al, 2000), the vast majority of the general population are non-carriers and, therefore, our finding for non-carriers will be generalisable to the general population. Our findings for non-carriers are consistent, in terms of direction, with previously published studies that observed a positive association between BMI in early adulthood and subsequent CRC risk for the general population (Le Marchand et al, 1992; Lee and Paffenbarger, 1992; Must et al, 1992; Russo et al, 1998).

The biologic mechanisms underlying an association between body size in early adulthood and increased risk of CRC are not well understood. One theory suggests that obesity, particularly central obesity, increases insulin resistance and hyperinsulinaemia (Giovannucci, 1995; Kahn and Flier, 2000) via alterations in the signalling of endogenous hormones, particularly not only insulin and insulin-like growth factors, but also steroid hormones, and possibly, adipocyte-derived factors such as leptin and adiponectin (Gunter and Leitzmann, 2006; John et al, 2006). High circulating concentrations of insulin and C-peptide, as well as type 2 diabetes, are associated with increased risk of CRC (Kaaks et al, 2000; Larsson et al, 2005). It is of interest that MMR defects frequently result in mutations of the insulin-like growth factor receptor (IGFR), which involves this same insulin signalling system (Duval and Hamelin, 2002). It is not known if the effect of obesity in MMR gene mutation carriers is related to IGFR mutational events.

Our results suggest that BMI in early adulthood is a potential modifier of CRC risk for carriers of an MMR gene mutation. Currently, the only methods available for MMR gene mutation carriers to reduce their risk for CRC are screening by colonoscopy (de Jong et al, 2006), prophylactic subtotal colectomy (Smith and Rodriguez-Bigas, 2009) and aspirin chemoprevention (Burn et al, 2009). Avoiding obesity in early adulthood might also reduce the risk of CRC for MMR gene mutation carriers.

Acknowledgments

This work was supported by the National Cancer Institute, National Institutes of Health under RFA #CA-95-011 and through cooperative agreements with members of the Colon CFR and Principal Investigators. Collaborating centres include Australasian Colorectal Cancer Family Registry (U01 CA097735), the USC Familial Colorectal Neoplasia Collaborative Group (U01 CA074799), Mayo Clinic Cooperative Family Registry for Colon Cancer Studies (U01 CA074800), Ontario Registry for Studies of Familial Colorectal Cancer (U01 CA074783), Seattle Colorectal Cancer Family Registry (U01 CA074794) and University of Hawaii Colorectal Cancer Family Registry (U01 CA074806). This work was also supported by a grant from NHMRC Australia (ID 400160). The content of this manuscript neither necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centres in the CFRs, nor does mention of trade names, commercial products or organisations implies endorsement by the US Government or the CFR. We had full responsibility for the design of the study, the collection of the data, the analysis and interpretation of the data, the decision to submit the manuscript for publication and the writing of the manuscript. We thank Dr Elizabeth Williamson from Murdoch Childrens Research Institute, Melbourne, Australia for statistical advice. We thank all study participants of the Colon Cancer Family Registry and staff for their contributions to this project.

Footnotes

Supplementary Information accompanies the paper on British Journal of Cancer website (http://www.nature.com/bjc)

The authors declare no conflict of interest.

Supplementary Material

Supplementary Information
Click here for additional data file. (41.5KB, doc)

References

  1. American Cancer Society (2008) Cancer Facts & Figures 2008. American Cancer Society: Atlanta [Google Scholar]
  2. Andrieu N, Easton DF, Chang-Claude J, Rookus MA, Brohet R, Cardis E, Antoniou AC, Wagner T, Simard J, Evans G, Peock S, Fricker J-P, Nogues C, Van’t Veer L, Van Leeuwen FE, Goldgar DE (2006) Effect of chest X-rays on the risk of breast cancer among BRCA1/2 mutation carriers in the international BRCA1/2 carrier cohort study: a report from the EMBRACE, GENEPSO, GEO-HEBON, and IBCCS Collaborators’ Group. J Clin Oncol 24(21): 3361–3366 [DOI] [PubMed] [Google Scholar]
  3. Antoniou AC, Goldgar DE, Andrieu N, Chang-Claude J, Brohet R, Rookus MA, Easton DF (2005) A weighted cohort approach for analysing factors modifying disease risks in carriers of high-risk susceptibility genes. Genet Epidemiol 29(1): 1–11 [DOI] [PubMed] [Google Scholar]
  4. Antoniou AC, Spurdle AB, Sinilnikova OM, Healey S, Pooley KA, Schmutzler RK, Versmold B, Engel C, Meindl A, Arnold N, Hofmann W, Sutter C, Niederacher D, Deissler H, Caldes T, Kämpjärvi K, Nevanlinna H, Simard J, Beesley J, Chen X, Neuhausen SL, Rebbeck TR, Wagner T, Lynch HT, Isaacs C, Weitzel J, Ganz PA, Daly MB, Tomlinson G, Olopade OI, Blum JL, Couch FJ, Peterlongo P, Manoukian S, Barile M, Radice P, Szabo CI, Pereira LHM, Greene MH, Rennert G, Lejbkowicz F, Barnett-Griness O, Andrulis IL, Ozcelik H, Gerdes A-M, Caligo MA, Laitman Y, Kaufman B, Milgrom R, Friedman E, Domchek SM, Nathanson KL, Osorio A, Llort G, Milne RL, Benítez J, Hamann U, Hogervorst FBL, Manders P, Ligtenberg MJL, van den Ouweland AMW, Peock S, Cook M, Platte R, Evans DG, Eeles R, Pichert G, Chu C, Eccles D, Davidson R, Douglas F, Godwin AK, Barjhoux L, Mazoyer S, Sobol H, Bourdon V, Eisinger F, Chompret A, Capoulade C, Bressac-de Paillerets B, Lenoir GM, Gauthier-Villars M, Houdayer C, Stoppa-Lyonnet D, Chenevix-Trench G, Easton DF (2008) Common breast cancer-predisposition alleles are associated with breast cancer risk in BRCA1 and BRCA2 mutation carriers. Am J Hum Genet 82(4): 937–948 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Baglietto L, Lindor NM, Dowty JG, White DM, Wagner A, Gomez Garcia EB, Vriends AHJT, Dutch Lynch Syndrome Study G, Cartwright NR, Barnetson RA, Farrington SM, Tenesa A, Hampel H, Buchanan D, Arnold S, Young J, Walsh MD, Jass J, Macrae F, Antill Y, Winship IM, Giles GG, Goldblatt J, Parry S, Suthers G, Leggett B, Butz M, Aronson M, Poynter JN, Baron JA, Le Marchand L, Haile R, Gallinger S, Hopper JL, Potter J, de la Chapelle A, Vasen HF, Dunlop MG, Thibodeau SN, Jenkins MA (2010) Risks of lynch syndrome cancers for MSH6 mutation carriers. J Natl Cancer Inst 102(3): 193–201 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Botma A, Nagengast FM, Braem MGM, Hendriks JCM, Kleibeuker JH, Vasen HFA, Kampman E (2010) Body mass index increases risk of colorectal adenomas in men with Lynch syndrome: the GEOLynch cohort study. J Clin Oncol 28(28): 4346–4353 [DOI] [PubMed] [Google Scholar]
  7. Burn J, Gerdes A-M, Mecklin J-P, Macrae F, Moeslein G, Olschwang S, Bisgaard M-L, Ramesar R, Eccles D, Maher ER, Bertario L, Lindblom A, Evans G, Lubinski J, Morrison PJ, Ho JWC, Vasen HFA, Lynch HT, Mathers JC, Bishop DT, the CAPP2 Consortium (2009) Aspirin prevents cancer in Lynch Syndrome: results of long term follow-up in the CAPP2 trial. the 59th Annual Meeting of The American Society of Human Genetics; 23 October; Honolulu, Hawaii
  8. Campbell PT, Cotterchio M, Dicks E, Parfrey P, Gallinger S, McLaughlin JR (2007) Excess body weight and colorectal cancer risk in canada: associations in subgroups of clinically defined familial risk of cancer. Cancer Epidemiol Biomarkers Prev 16(9): 1735–1744 [DOI] [PubMed] [Google Scholar]
  9. Campbell PT, Jacobs ET, Ulrich CM, Figueiredo JC, Poynter JN, McLaughlin JR, Haile RW, Jacobs EJ, Newcomb PA, Potter JD, Le Marchand L, Green RC, Parfrey P, Younghusband HB, Cotterchio M, Gallinger S, Jenkins MA, Hopper JL, Baron JA, Thibodeau SN, Lindor NM, Limburg PJ, Martinez ME, for the Colon Cancer Family Registry (2010) Case-control study of overweight, obesity, and colorectal cancer risk, overall and by tumor microsatellite instability status. J Natl Cancer Inst 102(6): 391–400 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chen S, Wang W, Lee S, Nafa K, Lee J, Romans K, Watson P, Gruber SB, Euhus D, Kinzler KW (2006) Prediction of germline mutations and cancer risk in the Lynch syndrome. JAMA 296(12): 1479. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dai Z, Xu Y-C, Niu L (2007) Obesity and colorectal cancer risk: a meta-analysis of cohort studies. World J Gastroenterol 13(31): 4199–4206 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. de Jong A, Nagengast F, Kleibeuker J, van de Meeberg P, van Wijk H, Cats A, Griffioen G, Vasen H (2006) What is the appropriate screening protocol in Lynch syndrome? Fam Cancer 5(4): 373–378 [DOI] [PubMed] [Google Scholar]
  13. Diergaarde B, Braam H, Vasen HF, Nagengast FM, van Muijen GN, Kok FJ, Kampman E (2007) Environmental factors and colorectal tumor risk in individuals with hereditary nonpolyposis colorectal cancer. Clin Gastroenterol Hepatol 5(6): 736–742.e1 [DOI] [PubMed] [Google Scholar]
  14. Dunlop MG, Farrington SM, Nicholl I, Aaltonen L, Petersen G, Porteous M, Carothers A (2000) Population carrier frequency of hMSH2 and hMLH1 mutations. Br J Cancer 83(12): 1643–1645 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Duval A, Hamelin R (2002) Mutations at coding repeat sequences in mismatch repair-deficient human cancers. Cancer Res 62(9): 2447–2454 [PubMed] [Google Scholar]
  16. Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A (eds) (2009) The AJCC Cancer Staging Manual. Springer: New York, NY [Google Scholar]
  17. Giovannucci E (1995) Insulin and colon cancer. Cancer Causes Control 6(2): 164–179 [DOI] [PubMed] [Google Scholar]
  18. Gunter MJ, Leitzmann MF (2006) Obesity and colorectal cancer: epidemiology, mechanisms and candidate genes. J Nutr Biochem 17(3): 145–156 [DOI] [PubMed] [Google Scholar]
  19. Hampel H, Frankel WL, Martin E, Arnold M, Khanduja K, Kuebler P, Clendenning M, Sotamaa K, Prior T, Westman JA, Panescu J, Fix D, Lockman J, LaJeunesse J, Comeras I, de la Chapelle A (2008) Feasibility of screening for lynch syndrome among patients with colorectal cancer. J Clin Oncol 26(35): 5783–5788 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Herman JG, Umar A, Polyak K, Graff JR, Ahuja N, Issa JPJ, Markowitz S, Willson JKV, Hamilton SR, Kinzler KW (1998) Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proc Natl Acad Sci USA 95(12): 6870. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hewish M, Lord CJ, Martin SA, Cunningham D, Ashworth A (2010) Mismatch repair deficient colorectal cancer in the era of personalized treatment. Nat Rev Clin Oncol 7: 197–208 [DOI] [PubMed] [Google Scholar]
  22. Jass JR (2006) Hereditary non-polyposis colorectal cancer: the rise and fall of a confusing term. World J Gastroenterol 12(31): 4943–4950 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Jenkins M, Southey M, Giles G, Hopper J (2007) Rationale for, and approach to, studying modifiers of risk in persons with a genetic predisposition to colorectal cancer. Curr Oncol Rep 9(3): 202–207 [DOI] [PubMed] [Google Scholar]
  24. Jenkins MA, Baglietto L, Dowty JG, Vliet CMV, Smith L, Mead LJ, Macrae FA, John DJBS, Jass JR, Giles GG, Hopper JL, Southey MC (2006) Cancer risks for mismatch repair gene mutation carriers: a population-based early onset case-family study. Clin Gastroenterol Hepatol 4(4): 489–498 [DOI] [PubMed] [Google Scholar]
  25. John BJ, Irukulla S, Abulafi AM, Kumar D, Mendall MA (2006) Systematic review: adipose tissue, obesity and gastrointestinal diseases. Aliment Pharmacol Ther 23(11): 1511–1523 [DOI] [PubMed] [Google Scholar]
  26. Kaaks R, Toniolo P, Akhmedkhanov A, Lukanova A, Biessy C, Dechaud H, Rinaldi S, Zeleniuch-Jacquotte A, Shore RE, Riboli E (2000) Serum C-peptide, insulin-like growth factor (IGF)-I, IGF-binding proteins, and colorectal cancer risk in women. J Natl Cancer Inst 92(19): 1592–1600 [DOI] [PubMed] [Google Scholar]
  27. Kadyrov F, Dzantiev L, Constantin N, Modrich P (2006) Endonucleolytic function of MutLα in human mismatch repair. Cell 126(2): 297–308 [DOI] [PubMed] [Google Scholar]
  28. Kahn BB, Flier JS (2000) Obesity and insulin resistance. J Clin Invest 106(4): 473–481 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Larsson SC, Orsini N, Wolk A (2005) Diabetes mellitus and risk of colorectal cancer: a meta-analysis. J Natl Cancer Inst 97(22): 1679–1687 [DOI] [PubMed] [Google Scholar]
  30. Larsson SC, Wolk A (2007) Obesity and colon and rectal cancer risk: a meta-analysis of prospective studies. Am J Clin Nutr 86(3): 556–565 [DOI] [PubMed] [Google Scholar]
  31. Le Marchand L, Wilkens LR, Mi MP (1992) Obesity in youth and middle age and risk of colorectal cancer in men. Cancer Causes Control 3(4): 349–354 [DOI] [PubMed] [Google Scholar]
  32. Lee IM, Paffenbarger Jr RS (1992) Quetelet's index and risk of colon cancer in college alumni. J Natl Cancer Inst 84(17): 1326–1331 [DOI] [PubMed] [Google Scholar]
  33. Moghaddam AA, Woodward M, Huxley R (2007) Obesity and risk of colorectal cancer: a meta-analysis of 31 studies with 70 000 events. Cancer Epidemiol Biomarkers Prev 16(12): 2533–2547 [DOI] [PubMed] [Google Scholar]
  34. Must A, Jacques PF, Dallal GE, Bajema CJ, Dietz WH (1992) Long-term morbidity and mortality of overweight adolescents. A follow-up of the Harvard Growth Study of 1922 to 1935. N Engl J Med 327(19): 1350–1355 [DOI] [PubMed] [Google Scholar]
  35. Newcomb PA, Baron J, Cotterchio M, Gallinger S, Grove J, Haile R, Hall D, Hopper JL, Jass J, Le Marchand L, Limburg P, Lindor N, Potter JD, Templeton AS, Thibodeau S, Seminara D, for the Colon Cancer Family Registry (2007) Colon Cancer Family Registry: an international resource for studies of the genetic epidemiology of colon cancer. Cancer Epidemiol Biomarkers Prev 16(11): 2331–2343 [DOI] [PubMed] [Google Scholar]
  36. Norgan NG, Cameron N (2000) The accuracy of body weight and height recall in middle-aged men. Int J Obes Relat Metab Disord 24(12): 1695. [DOI] [PubMed] [Google Scholar]
  37. Pande M, Lynch PM, Hopper JL, Jenkins MA, Gallinger S, Haile RW, LeMarchand L, Lindor NM, Campbell PT, Newcomb PA, Potter JD, Baron JA, Frazier ML, Amos CI (2010) Smoking and colorectal cancer in Lynch syndrome: results from the Colon Cancer Family Registry and The University of Texas M.D. Anderson Cancer Center. Clin Cancer Res 16(4): 1331–1339 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Perry GS, Byers TE, Mokdad AH, Serdula MK, Williamson DF (1995) The validity of self-reports of past body weights by US adults. Epidemiology 6(1): 61–66 [DOI] [PubMed] [Google Scholar]
  39. Poynter JN, Siegmund KD, Weisenberger DJ, Long TI, Thibodeau SN, Lindor N, Young J, Jenkins MA, Hopper JL, Baron JA (2008) Molecular characterization of MSI-H colorectal cancer by MLHI promoter methylation, immunohistochemistry, and mismatch repair germline mutation screening. Cancer Epidemiol Biomarkers Prev 17(11): 3208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Renehan AG, Tyson M, Egger M, Heller RF, Zwahlen M (2008) Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 371(9612): 569–578 [DOI] [PubMed] [Google Scholar]
  41. Rogers WH (1993) Regression standard errors in clustered samples. Stata Tech Bull 3(13): 19–23 [Google Scholar]
  42. Russo A, Franceschi S, La Vecchia C, Dal Maso L, Montella M, Conti E, Giacosa A, Falcini F, Negri E (1998) Body size and colorectal-cancer risk. Int J Cancer 78(2): 161–165 [DOI] [PubMed] [Google Scholar]
  43. Senter L, Clendenning M, Sotamaa K, Hampel H, Green J, Potter JD, Lindblom A, Lagerstedt K, Thibodeau SN, Lindor NM, Young J, Winship I, Dowty JG, White DM, Hopper JL, Baglietto L, Jenkins MA, de la Chapelle A (2008) The clinical phenotype of lynch syndrome due to germ-line PMS2 mutations. Gastroenterology 135(2): 419–428.e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Slattery ML, Curtin K, Anderson K, Ma K-N, Ballard L, Edwards S, Schaffer D, Potter J, Leppert M, Samowitz WS (2000) Associations between cigarette smoking, lifestyle factors, and microsatellite instability in colon tumors. J Natl Cancer Inst 92(22): 1831–1836 [DOI] [PubMed] [Google Scholar]
  45. Slattery ML, Levin TR, Ma K, Goldgar D, Holubkov R, Edwards S (2003) Family history and colorectal cancer: predictors of risk. Cancer Causes Control 14(9): 879–887 [DOI] [PubMed] [Google Scholar]
  46. Smith KD, Rodriguez-Bigas MA (2009) Role of surgery in familial adenomatous polyposis and hereditary nonpolyposis colorectal cancer (lynch syndrome). Surg Oncol Clin N Am 18(4): 705–715 [DOI] [PubMed] [Google Scholar]
  47. Southey MC, Jenkins MA, Mead L, Whitty J, Trivett M, Tesoriero AA, Smith LD, Jennings K, Grubb G, Royce SG, Walsh MD, Barker MA, Young JP, Jass JR, St John DJB, Macrae FA, Giles GG, Hopper JL (2005) Use of molecular tumor characteristics to prioritize mismatch repair gene testing in early-onset colorectal cancer. J Clin Oncol 23(27): 6524–6532 [DOI] [PubMed] [Google Scholar]
  48. StataCorp (2007) Stata Statistical Software: Release 10. StataCorp LP: College Station, TX [Google Scholar]
  49. Stevens J, Keil JE, Waid LR, Gazes PC (1990) Accuracy of current, 4-, and 28-year self-reported body weight in an elderly population. Am J Epidemiol 132(6): 1156. [DOI] [PubMed] [Google Scholar]
  50. Tamakoshi K, Yatsuya H, Kondo T, Hirano T, Hori Y, Yoshida T, Toyoshima H (2003) The accuracy of long-term recall of past body weight in Japanese adult men. Int J Obes 27(2): 247–252 [DOI] [PubMed] [Google Scholar]
  51. Troy LM, Hunter DJ, Manson JE, Colditz GA, Stampfer MJ, Willett WC (1995) The validity of recalled weight among younger women. Int J Obes Relat Metab Disord 19(8): 570. [PubMed] [Google Scholar]
  52. Vasen HFA, Watson P, Mecklin JP, Lynch HT (1999) New clinical criteria for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch syndrome) proposed by the International Collaborative Group on HNPCC. Gastroenterology 116(6): 1453–1456 [DOI] [PubMed] [Google Scholar]
  53. Watson P, Ashwathnarayan R, Lynch HT, Roy HK (2004) Tobacco use and increased colorectal cancer risk in patients with hereditary nonpolyposis colorectal cancer (lynch syndrome). Arch Intern Med 164(22): 2429–2431 [DOI] [PubMed] [Google Scholar]
  54. Williams RL (2000) A note on robust variance estimation for cluster-correlated data. Biometrics 56(2): 645–646 [DOI] [PubMed] [Google Scholar]
  55. Win AK, Dowty JG, Antill YC, English DR, Baron JA, Young JP, Giles GG, Southey MC, Winship IMD, Lipton L, Parry S, Thibodeau SN, Haile R, Gallinger S, Le Marchand L, Lindor NM, Newcomb PA, Hopper JL, Jenkins MA (2011) Body mass index in early adulthood and endometrial cancer risk for mismatch repair gene mutation carriers. Obstet Gynecol 117(4): 899–905 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. World Health Organization (1998) Report of a WHO Consultation on Obesity. Obesity: Preventing and Managing the Global Epidemic. WHO: Geneva Switzerland [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information
Click here for additional data file. (41.5KB, doc)

Articles from British Journal of Cancer are provided here courtesy of Cancer Research UK

RESOURCES