American Journal of Epidemiology The Author 2011.

. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Vol. 173, No. 10 DOI: 10.1093/aje/kwq513 Advance Access publication: March 18, 2011

Original Contribution Long-Term Sedentary Work and the Risk of Subsite-specic Colorectal Cancer

Terry Boyle*, Lin Fritschi, Jane Heyworth, and Fiona Bull

* Correspondence to Terry Boyle, Western Australian Institute for Medical Research, University of Western Australia, B Block, Queen Elizabeth II Medical Centre, Hospital Avenue, Nedlands, Western Australia 6009, Australia (e-mail: tboyle@waimr.uwa.edu.au).

Initially submitted September 27, 2010; accepted for publication December 22, 2010.
Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

Research suggests that sedentary behavior may increase the risk of some chronic diseases. The aims of the study were to examine whether sedentary work is associated with colorectal cancer and to determine whether the association differs by subsite. A total of 918 cases and 1,021 controls participated in a population-based case-control study of colorectal cancer in Western Australia in 20052007. Data were collected on lifestyle, physical activity, and lifetime job history. The estimated effects of sedentary work on the risk of cancers of the proximal colon, distal colon, and rectum were analyzed by using multinomial logistic regression. Compared with participants who did not spend any time in sedentary work, participants who spent 10 or more years in sedentary work had almost twice the risk of distal colon cancer (adjusted odds ratio 1.94, 95% condence interval: 1.28, 2.93) and a 44% increased risk of rectal cancer (adjusted odds ratio 1.44, 95% condence interval: 0.96, 2.18). This association was independent of recreational physical activity and was seen even among the most recreationally active participants. Sedentary work was not associated with the risk of proximal colon cancer. These results suggest that long-term sedentary work may increase the risk of distal colon cancer and rectal cancer. colorectal neoplasms; motor activity; occupations; sedentary lifestyle

Abbreviations: AOR, adjusted odds ratio; CI, condence interval; MET, metabolic equivalent.

Research suggests that sedentary behavior may increase the risk of some chronic diseases (1). Sedentary behavior is characterized by prolonged sitting and other activities requiring very low energy expenditure (<1.5 metabolic equivalents (METs)) (13) and is considered to be an independent exposure, distinct from low levels of physical activity (1). Research on the more well-recognized risk factor of physical activity has shown that a low level of work-related activity is associated with an increased risk of colorectal cancer (4). However, these studies have mostly compared light-intensity occupations (or combined sedentary and light-intensity occupations) with physically active work. Physical activity and sedentary behavior (often called too much sitting) may act through different mechanisms (1) and, as such, it is not clear if a low level of physical activity, a high level of sedentary behavior, or a combination of both is behind the associations seen. Previous studies have generally not adjusted for recreational physical activity and also have not investigated whether recreational physical
1183

activity modied the effect of sedentary work on colorectal cancer risk. Additionally, few of these studies have investigated the effect of long-term sedentary work on the risk of colorectal cancer; most have examined only occupational activity at 1 point in time. Another unresolved issue is whether sedentary work has a different effect on distal and proximal colon cancer risks. It has been suggested that distal and proximal colon cancers may in fact be 2 different diseases: There are morphological, molecular, and epidemiologic differences between distal and proximal colon cancers, and research suggests that they may have different environmental and genetic risk factors (5). Different effects by tumor subsite may have screening implications (6) and may provide further clues into the etiology of colon cancer. We conducted a case-control study to investigate these issues. The aims of the study were as follows: to examine whether sedentary work was associated with colorectal cancer; to determine whether the association differed by
Am J Epidemiol. 2011;173(10):11831191

1184 Boyle et al.

Table 1. Selected Characteristics of the Participants, by Outcome, the Western Australian Bowel Health Study, 20052007
Cases Proximal Colon Cancer No. % Distal Colon Cancer No. % Rectal Cancer No. % All Colorectal Cancer No. % Controls No. %

Sex Male Female Age, years 4049 5059 6069 7079 Socioeconomic status 1 (most disadvantaged) 2 3 4 5 (least disadvantaged) Smoking, pack-years 00.9 119.9 20 Body mass index at age 20 years, kg/m2a <25 2529.9 30 Body mass index at age 40 years, kg/m2a <25 2529.9 30 143 95 27 54.0 35.8 10.2 123 89 40 48.8 35.3 15.9 158 92 44 53.7 31.3 15.0 425 276 111 52.3 34.0 13.7 538 296 87 58.4 32.1 9.4 225 39 4 84.0 14.6 1.5 182 60 5 73.7 24.3 2.0 223 56 12 76.6 19.2 4.1 630 155 21 78.2 19.2 2.6 767 136 21 83.0 14.7 2.3 132 62 90 46.5 21.8 31.7 116 66 86 43.3 24.6 32.1 135 79 104 42.4 24.8 32.7 383 207 280 44.0 23.8 32.2 493 236 267 49.5 23.7 26.8 12 48 106 118 44 51 63 67 59 4.2 16.9 37.3 41.6 15.5 18.0 22.2 23.6 20.8 15 80 94 79 54 49 41 64 60 5.6 29.8 35.1 29.5 20.2 18.3 15.3 23.9 22.4 18 83 120 97 57 78 70 60 53 5.7 26.1 37.7 30.5 17.9 24.5 22.0 18.9 16.7 45 211 320 294 155 178 174 191 172 5.2 24.2 36.8 33.8 17.8 20.5 20.0 22.0 19.8 70 219 355 352 168 192 231 195 210 7.0 22.0 35.6 35.3 155 129 54.6 45.4 165 103 61.6 38.4 216 102 67.9 32.1 536 334 61.6 38.4 594 402 59.6 40.4

Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

16.9 19.3 23.2 19.6 21.1

Table continues

subsite; and to investigate whether recreational physical activity modied the effect of sedentary work on colorectal cancer risk.
MATERIALS AND METHODS The Western Australian Bowel Health Study

A case-control study called the Western Australian Bowel Health Study (WABOHS) took place in Western Australia between 2005 and 2007. Details of this study have been described previously (7). Briey, all participants were aged between 40 and 79 years and resided in Western Australia. Incident cases of colorectal cancer occurring between June 2005 and August 2007 were recruited through the Western Australian Cancer Registry. With the exception of nonme-

lanoma skin cancer, it is mandatory to register all cases of cancer in Australia. Carcinomas occurring in the hepatic exure, cecum, ascending colon, and transverse colon were classied as proximal colon cancers; carcinomas occurring in the splenic exure, descending colon, and sigmoid colon were classied as distal colon cancers; and carcinomas occurring in the rectosigmoid junction and rectum were classied as rectal cancers. Controls were randomly selected from the Western Australian electoral roll, which is considered to be virtually complete as registration on the electoral roll is compulsory in Australia. Information on lifestyle, diet, occupation, and medication use was collected by using self-administered questionnaires from 918 histopathologically conrmed cases (59.5% response fraction) and 1,021 age- and sex-matched controls (46.5% response fraction). Forty-eight cases and 25 controls were excluded because of
Am J Epidemiol. 2011;173(10):11831191

Sedentary Work and Colorectal Cancer Risk

1185

Table 1. Continued
Cases Proximal Colon Cancer No. % Distal Colon Cancer No. % Rectal Cancer No. % All Colorectal Cancer No. % Controls No. %

Diabetes and/or high blood sugar Neither High blood sugar Diabetes Energy intake from food 10 years ago, kJ/day 06,301 6,3028,074 8,07510,406 10,407 Alcohol consumption 10 years ago, g/day 01.1292 1.12939.514 9.51527.198 27.199 Total lifetime recreational physical activity 0 (least active) 1 (most active) Vigorous lifetime recreational physical activity 0 (least active) 1 (most active)
a

225 19 40

79.2 6.7 14.1

211 16 41

78.7 6.0 15.3

259 14 45

81.4 4.4 14.2

695 49 126

79.9 5.6 14.5

851 52 93

85.4 5.2 9.3

70 61 83 70

24.6 21.5 29.2 24.6

60 64 64 80

22.4 23.9 23.9 29.8

55 77 94 92

17.3 24.2 29.6 28.9

185 202 241 242

21.3 23.2 27.7 27.8

245 250 252 249

24.6 25.1 25.3 25.0

Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

71 69 69 75

25.0 24.3 24.3 26.4

56 66 59 87

20.9 24.6 22.0 32.5

72 69 76 101

22.6 21.7 23.9 31.8

199 204 204 263

22.9 23.4 23.4 30.2

248 249 248 251

24.9 25.0 24.9 25.2

112 172

39.4 60.6

112 156

41.8 58.2

137 181

43.1 56.9

361 509

41.5 58.5

387 609

38.9 61.1

197 87

69.4 30.4

209 59

78.0 22.0

245 73

77.0 23.0

651 219

74.8 25.2

708 288

71.1 28.9

Total case and control numbers do not add up because of missing data.

missing data on 1 or more covariates, leaving 996 controls and 870 cases in this analysis. Ethics approval for the Western Australian Bowel Health Study was obtained from the University of Western Australia Human Research Ethics Committee and the Condentiality of Health Information Committee within the Western Australian Department of Health, and written informed consent was obtained from all participants.
Exposure measurement

Participants were asked to complete their lifetime occupational history, from their rst job to retirement. For each job that a participant held, he/she was required to record the job title and duties, company, location, age started and stopped, and whether the job was full time, part time, casual, or seasonal. Occupational physical activity was calculated by classifying each job (based on job title and duties) that a participant held into one of 5 categories of Physical Demands Strength Rating, according to the US Department of Labors Dictionary of Occupational Titles (8). The 5 categories are 1) sedentary (e.g., bookkeepers, computing
Am J Epidemiol. 2011;173(10):11831191

professionals); 2) light (e.g., teachers, hairdressers, housewives); 3) medium (e.g., mechanics, police ofcers, nurses); 4) heavy (e.g., plumbers, farmers); and 5) very heavy (e.g., miners, re ghters). The heavy and very heavy categories were combined as there were few participants in the very heavy category. Jobs classied as sedentary by the Dictionary of Occupational Titles involve sitting most of the time, while light jobs involve activity of a light intensity (8). The number of years that a participant spent in each level of occupational activity was calculated. Part time, casual, and seasonal occupations were considered to be equivalent to 0.5, 0.25, and 0.25 years of full-time work, respectively. A lifetime occupational activity variable was created by assigning participants to the occupational activity category in which they spent the most time. As more than half of the participants were in the light category, this was used as the reference level in the model that assessed the effect of lifetime occupational activity on colorectal cancer. To further quantify the effect of sedentary work on colorectal cancer, we categorized the number of years spent in

1186 Boyle et al.

sedentary work as none, more than zero but less than 10 years, and 10 or more years. The latter 2 categories were determined from the distribution among controls that led to 2 groups of roughly equal size.
Statistical analysis

Polytomous logistic regression models were used to estimate the odds ratio for the risk of proximal colon cancer, distal colon cancer, and rectal cancer. Postestimation commands were used to test whether the effect of each exposure differed signicantly by subsite (9). As this was an imputed data set (see below), the command MIM was used to generate parameter estimates (10). All analyses were performed by using STATA, version 11.1, statistical software (StataCorp LP, College Station, Texas). Sex and age group (5 year) were included in all models, as the control distribution was matched to the case distribution on these variables. Energy intake 10 years ago, cigarette use, alcohol intake 10 years ago, diabetes, socioeconomic status, body mass index at ages 20 and 40 years, and lifetime recreational physical activity were included in the nal analyses as they are established risk factors for colorectal cancer and were considered to be potential confounders (11). The model assessing the effect of years in sedentary work additionally adjusted for combined years in heavy and very heavy occupational activity. Energy intake (kJ/day from food) and alcohol intake (g/day) were categorized on the basis of the control quartiles. Cigarette use was categorized into 3 groups (0, 119, 20 pack-years). Diabetes was classied as having diabetes, high blood sugar level only, or neither. Socioeconomic status was classied by using the Index of Relative Socio-economic Disadvantage from the Socio-Economic Indexes for Areas (12) and was categorized into 5 groups. Body mass index was classied as under or normal weight (<25 kg/m2), overweight (2529.9 kg/m2), and obese (30 kg/m2). Body mass index at ages 20 years and 40 years was imputed by using the multiple imputation command ICE (13), as there were 8% missing data for these variables. Information on recreational physical activity performed over the adult lifetime was collected by using a questionnaire based upon others that have been shown to be reliable (1416). Participants were asked to record the recreational physical activity that they performed regularly (>10 times) during 3 periods of their life (up to the date of their participation in the study): 1934 years; 3550 years; and 51 years or more. For each age period, a MET-hours per week total was calculated. This total was then categorized as no physical activity (i.e., 0 MET-hours per week), and the remaining active participants were split into 3 groups of roughly equal size (0.111.9, 1229.9, and 30 MET-hours per week). A lifetime recreational physical activity score was created by summing the category (0, 1, 2, or 3) from each age period. The total score was then split into 2 groups (04, 59 for participants aged 51 years or older; 02, 36 for participants aged <51 years). This was also done for vigorous activities (activities with a MET value of 6 or more) only. To determine whether recreational physical activity modied the effect of sedentary work on colorectal cancer risk, we

included an interaction term in the full model. Recreational physical activity was considered to be a potential effect modier and is an established risk factor for colorectal cancer. We also examined whether vigorous recreational physical activity or body mass index at age 40 years modied the effect of sedentary work on colorectal cancer risk. Postestimation commands were used to generate odds ratios and condence intervals for the models that included an interaction term. Chi-square tests were used to assess the differences between outcomes for the selected characteristics in Table 1. All P values were 2 sided. The controls were selected from the electoral roll, but we did not exclude the cases who were not on the electoral roll. We therefore conducted a sensitivity analysis, in which we repeated all analyses without cases who were not on the electoral roll (n 54) and compared the effect sizes with those from the original models. There was no statistically signicant interaction by sex in any of the analyses, and the associations observed were similar in both sexes, so all results are presented for males and females combined.
RESULTS

Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

The demographic and various dietary and lifestyle characteristics of the participants by cancer status are shown in Table 1. There was a higher proportion of females among the proximal colon cancer cases than the distal colon or rectal cancers. The age distribution among participants with proximal colon cancer was signicantly different from that of participants with distal colon cancer or rectal cancer, with 79% of proximal colon cancer aged over 60 years compared with 65% of distal colon cancer cases and 68% of rectal cancer cases. Participants from more disadvantaged areas were more likely to have rectal cancer than distal colon cancer. Compared with all cases, controls were less likely to be 20 pack-year smokers and were also less likely to have diabetes. Proximal colon cancer cases were less likely than distal colon or rectal cancer cases to be overweight or obese at 20 years of age. Compared with participants who spent the most time in a light job, participants who spent the most time in a sedentary occupation had 2 times the risk of distal colon cancer (adjusted odds ratio (AOR) 2.07, 95% condence interval (CI): 1.25, 3.44) (Table 2). Participants who spent 10 or more years in sedentary work had almost twice the risk of distal colon cancer (AOR 1.94, 95% CI: 1.28, 2.93) and nearly 1.5 times the risk of rectal cancer (AOR 1.44, 95% CI: 0.96, 2.18) compared with participants who never held a sedentary occupation (Table 2). When the interaction term between total recreational physical activity and sedentary work was included in the model (Table 3), it was not statistically signicant (P 0.29). The risk of distal colon cancer and rectal cancer increased with increasing duration of sedentary work among both the most recreationally active and least recreationally active participants (Table 3). The risk of proximal colon cancer was not linked with sedentary work in the whole population (Table 2), the most recreationally active participants, or the least recreationally active participants (Table 3). Neither vigorous recreational physical
Am J Epidemiol. 2011;173(10):11831191

Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

Am J Epidemiol. 2011;173(10):11831191

Table 2. Association Among Lifetime Occupational Activity, Years in Sedentary Work, and Site-specic Colorectal Cancer Risk, the Western Australian Bowel Health Study, 20052007
Proximal Colon Cancer No. ORa 95% CI AORb 95% CI No. ORa Distal Colon Cancer 95% CI AORb 95% CI No. ORa Rectal Cancer 95% CI AORb 95% CI

Controls, no.

Lifetime occupational activity Sedentary Light Medium Heavy/very heavy Years in sedentary work 0 >0<10 10 years Ptrend
a

57 563 264 112 805 95 96

18 178 68 20 232 24 28

1.13 1.00 0.84 0.56 1.00 0.93 1.14

0.64, 2.01 Referent 0.60, 1.17 0.33, 0.96 Referent 0.58, 1.49 0.72, 1.79 0.70

1.12 1.00 0.81 0.55 1.00 0.89 1.11c

0.63, 2.00 Referent 0.58, 1.15 0.32, 0.95 Referent 0.55, 1.44 0.69, 1.76 0. 84
d

31 146 68 23 196 27 45

2.07 1.00 0.99 0.83 1.00 1.15 1.87

1.26, 3.38 Referent 0.70, 1.40 0.50, 1.38 Referent 0.73, 1.82 1.26, 2.78 0.003

2.07 1.00 0.92 0.75 1.00 1.21 1.94c

1.25, 3.44 Referent 0.64, 1.31 0.44, 1.27 Referent 0.76, 1.95 1.28, 2.93 0.002d

23 161 97 37 251 26 41

1.27 1.00 1.12 0.98 1.00 0.89 1.29

0.75, 2.16 Referent 0.82, 1.54 0.63, 1.52 Referent 0.56, 1.41 0.87, 1.93 0.33

1.34 1.00 1.03 0.88 1.00 0.93 1.44

0.78, 2.30 Referent 0.75, 1.42 0.56, 1.38 Referent 0.58, 1.49 0.96, 2.18

Sedentary Work and Colorectal Cancer Risk

0.14

Abbreviations: AOR, adjusted odds ratio; CI, condence interval; OR, odds ratio. Adjusted for age group and sex. b Adjusted for age group, sex, lifetime recreational physical activity level, cigarette smoking (pack-year tertiles), diabetes, educational level, energy intake from food, alcohol intake, body mass index at age 20 years, body mass index at age 40 years, and socioeconomic status. The AOR in the years in sedentary work analysis is additionally adjusted for years in a heavy or very heavy occupation. c P < 0.05 for the difference between the proximal colon and the distal colon. d P < 0.05 for the difference between the proximal colon trend and the distal colon trend.

1187

1188 Boyle et al.

Table 3. Effects of Lifetime Recreational Physical Activity and Body Mass Index on the Association Between Years in Sedentary Work and Sitespecic Colorectal Cancer Risk, the Western Australian Bowel Health Study, 20052007
Proximal Colon Cancer by Lifetime Recreational Physical Activity Groups 0 and 1 (Less Active) AORa 95% CI Groups 2 and 3 (More Active) AORa 95% CI Distal Colon Cancer by Lifetime Recreational Physical Activity Groups 0 and 1 (Less Active) AORa 95% CI Groups 2 and 3 (More Active) AORa 95% CI Rectal Cancer by Lifetime Recreational Physical Activity Groups 0 and 1 (Less Active) AORa 95% CI Groups 2 and 3 (More Active) AORa 95% CI

Years in Sedentary Work

0 >0<10 10
a

1.13 0.56 0.68

0.83, 1.53 0.21, 1.50 0.29, 1.58

1.00 1.13 1.49

Referent 0.64, 1.97 0.86, 2.60

1.25 1.35 1.42

0.90, 1.73 0.63, 2.91 0.71, 2.82

1.00 1.30 2.60

Referent 0.72, 2.35 1.57, 4.31

1.20 1.22 1.49

0.89, 1.62 0.57, 2.59 0.79, 2.79

1.00 0.88 1.57

Referent 0.49, 1.59 0.93, 2.65

Abbreviations: AOR, adjusted odds ratio; CI, condence interval. Adjusted for age group, sex, lifetime recreational physical activity level, cigarette smoking (pack-year tertiles), diabetes, educational level, energy intake from food, alcohol intake, body mass index at age 20 years, body mass index at age 40 years, socioeconomic status, and years in a heavy or very heavy occupation.

Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

activity nor body mass index at age 40 years modied the association between years in sedentary work and colorectal cancer (data not shown). Participants who spent the most time in jobs requiring heavy or very heavy activity had a signicantly reduced risk of proximal colon cancer (AOR 0.56, 95% CI: 0.32, 0.96) compared with participants who spent the most time in a light job (Table 2). There were no meaningful changes to the effect estimates in the above analyses when the cases that were not on the electoral roll were excluded.
DISCUSSION

In this study, we found that participants who spent the most time in sedentary work had a risk of distal colon cancer that was 2 times higher than those who spent the most time in a job requiring light activity. Similarly, participants who spent 10 or more years in sedentary work had almost twice the risk of distal colon cancer and almost 1.5 times the risk of rectal cancer, of those who did not do any sedentary work. Proximal colon cancer risk was not associated with sedentary work, although the risk was signicantly reduced among participants who spent the most time in jobs requiring heavy or very heavy activity. Previous research that has investigated the association between long-term exposure to a low level of work-related activity and colorectal cancer has generally found an increased risk (1725), although these studies have in the main compared light-intensity work with more physically active work. The 4 studies that have investigated the effect of time spent sitting, or a proxy measure, on colorectal cancer risk have all revealed an increased risk (2629). Howard et al. (27) found that participants who spent 9 or more hours per day sitting had a 25% greater risk of colon cancer than those who spent less than 3 hours per day sitting. Whittemore et al. (29) found that, compared with ChineseAmerican men who spent less than 5 hours sitting, those who spent 59 hours sitting had 2.4 times the risk of colorectal cancer, and participants who spent 10 or more hours sitting had 3.9 times the risk of colorectal cancer. Colbert et al. (26) found that participants who reported that their

occupation in the past year involved mainly sitting had a 67% greater risk of colon cancer and a 41% increased risk of rectal cancer than those who reported walking quite a lot in their occupation, although neither of these associations reached statistical signicance, and the authors did not nd an association between sedentary behavior in leisure time and colon or rectal cancer risk. Finally, Steindorf et al. (28) found that those who spent more than 2 hours a day watching television more than doubled their risk of colorectal cancer compared with those who spent less than 1.14 hours a day watching television. The positive association found between sedentary work and colorectal cancer in this study was independent of recreational physical activity and was seen among the most recreationally active participants. Indeed, even a high level of vigorous recreational physical activity did not modify the effect of sedentary work. This nding ts with the Active Couch Potato phenomenon, in which high amounts of sedentary behavior and physical activity coexist, and lends support to the notion that sedentary behavior affects disease risk independently of physical activity (1). There are several plausible biologic mechanisms through which sedentary work and sedentary behavior in general may increase the risk of colorectal cancer (refer to reference 30 for a review of the evidence and proposed mechanisms linking sedentary behavior and cancer). Sedentary behavior has been shown to increase blood glucose levels and to decrease insulin sensitivity, independently of physical activity (31). Increased blood glucose and decreased insulin resistance are both thought to promote colorectal cancer carcinogenesis (32, 33). Sedentary behavior has also been linked to an increased risk of diabetes and obesity (again independently of exercise level) (34), both of which are established risk factors for colorectal cancer (11). Bed-rest studies and experiments on mice provide some evidence that sedentary behavior is associated with markers of inammation and mitochondrial function (30). Other proposed mechanisms through which sedentary behavior may increase the risk of colorectal cancer include increasing levels of proinammatory factors, decreasing levels of antiinammatory factors, and decreasing levels of vitamin D (30). Although several of these mechanisms are the same as those proposed
Am J Epidemiol. 2011;173(10):11831191

Sedentary Work and Colorectal Cancer Risk

1189

for the inverse relation between physical activity and colorectal cancer (35), many of them appear to act independently of physical activity. It has also been proposed that the loss of local contractile activity caused by sitting results in unique gene expressions that may be involved in disease etiology (36). Sitting has been shown to suppress the regulation of skeletal muscle lipoprotein lipase, which plays an important role in lipid metabolism and may have a positive effect on diet-induced adiposity and insulin resistance (36). Our nding that sedentary work is associated with the risk of distal colon cancer but not proximal colon cancer adds to the evidence suggesting that lifestyle factors may play a larger role in distal colon carcinogenesis than in proximal colon carcinogenesis (37). Fruit and vegetable intake (38), calcium intake (39), and meat consumption (40) all appear to be more strongly associated with distal colon cancer than proximal colon cancer, although the literature concerning physical activity and distal and proximal colon cancer is inconsistent (41). There is no clear mechanistic explanation, however, for why sedentary behavior would increase the risk of distal colon and rectal cancers but not proximal colon cancer. Obesity is one possibility: An increased body mass index may have a greater effect on distal colon cancer (37), and there is evidence that obesity has a greater effect on microsatellite-stable tumors (42), which are more likely to occur in the distal colon than the proximal colon (5). There is some evidence that vitamin D has a stronger protective effect on distal colon and rectal cancers than proximal colon cancer (43); however, there is little evidence that insulin affects proximal and distal colon cancers differently (4447), and antiinammatories may have a more pronounced protective effect on proximal tumors than distal tumors (48, 49). Further mechanistic evidence is needed to understand why the association between sedentary behavior (and lifestyle factors in general) and colon cancer may differ by subsite. The ndings of this study have occupational health implications, especially given that advances in technology have led to increasing amounts of sedentary behavior at work and in other settings. Findings from the National Health and Nutrition Examination Survey suggest that the average American adult spends 55% (7.7 hours) of his/her waking time sedentary (50). This percentage increases to more than 60% (8 hours) among adults aged 60 years or more. Sedentary work is unlikely to be eliminated from modern life. However, strategies to minimize the amount of time spent sitting, especially long bouts of sitting, have been suggested (1). For example, it has been suggested that new occupational regulations could be introduced to break up prolonged sedentary time by taking regular nonsitting breaks (1), which have been shown to have a benecial effect on metabolic biomarkers (51). A recent systematic review, however, found that there is a lack of evidence concerning the effectiveness of workplace interventions for reducing sitting and called for more research to develop effective programs that specically target sitting (52). It has also been suggested that future guidelines on physical activity and health should include recommendations regarding sedentary behavior (53); this will become especially important if, as we found in our study, sedentary behavior has a detrimental effect on disease risk even among individAm J Epidemiol. 2011;173(10):11831191

uals who are physically active in their leisure time. Some countries, including Australia (54), Canada (55), and the United Kingdom, have or are developing sedentary behavior recommendations for early years (<5 years of age), children, and/or adults. This study had several limitations. Selection bias is a possible explanation for the results, especially given the low response fraction among the controls (46.5%). Basing occupational activity level on job title is not an ideal assessment and may have led to nondifferential exposure misclassication. Recreational activity was determined by selfreport of activities; however, this approach to measuring physical activity has been found to be reliable and valid (56). We had data on only sedentary behavior in 1 domain (occupation) and were not able to take into account sedentary behavior during leisure time, at home, or during transport. This study also had several strengths. It was a population-based study with histopathologically conrmed cases, and we were able to control for many potential confounders including lifetime recreational physical activity. We were also able to investigate whether recreational physical activity modied the effect of sedentary work on colorectal cancer risk. Aside from the increased risk of distal colon cancer and rectal cancer seen among those who worked in a sedentary occupation for 10 or more years in this study, an increasing level of sedentary behavior has been linked to an increased risk of several other chronic diseases, as well as increased mortality (57, 58). Sedentary behavior appears to be a novel and important risk factor for many chronic diseases (59).

Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

ACKNOWLEDGMENTS

Author afliations: School of Population Health, the University of Western Australia, Perth, Australia (Terry Boyle, Jane Heyworth, Fiona Bull); and Western Australian Institute for Medical Research, the University of Western Australia, Perth, Australia (Terry Boyle, Lin Fritschi). This work was supported by the Australian National Health and Medical Research Council (grant 353568). T. B. is supported by an Australian Postgraduate Award from the University of Western Australia and a scholarship from the Lions Cancer Institute of Western Australia. L. F. is supported by an Australian National Health and Medical Research Council fellowship. The authors thank Barry Iacopetta, Cameron Platell, Kieran McCaul, David Crawford, Cassandra Clayforth, Jenny Landrigan, Jen Girschik, Clare Tran, Beatriz Cuesta-Briand, and Anna Timperio for their contributions to this study. Conict of interest: none declared.

REFERENCES 1. Owen N, Healy GN, Matthews CE, et al. Too much sitting: the population health science of sedentary behavior. Exerc Sport Sci Rev. 2010;38(3):105113.

1190 Boyle et al.

2. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc. 2000;32(suppl 9):S498S504. 3. Pate RR, ONeill JR, Lobelo F. The evolving denition of "sedentary". Exerc Sport Sci Rev. 2008;36(4):173178. 4. Samad AK, Taylor RS, Marshall T, et al. A meta-analysis of the association of physical activity with reduced risk of colorectal cancer. Colorectal Dis. 2005;7(3):204213. 5. Iacopetta B. Are there two sides to colorectal cancer? Int J Cancer. 2002;101(5):403408. 6. Friedenreich CM, Neilson HK, Lynch BM. State of the epidemiological evidence on physical activity and cancer prevention. Eur J Cancer. 2010;46(14):25932604. 7. Iacopetta B, Heyworth J, Girschik J, et al. The MTHFR C677T and D DNMT3B C-149T polymorphisms confer different risks for right- and left-sided colorectal cancer. Int J Cancer. 2009;125(1):8490. 8. US Department of Labor. Dictionary of Occupational Titles. 4th ed. Washington, DC: US Government Printing Ofce; 1991. (http://www.oalj.dol.gov/LIBDOT.HTM). (Accessed July 26, 2009). 9. Long JS, Freese J. Regression Models for Categorical Dependent Variables Using Stata. College Station, TX: Stata Press; 2006. 10. Carlin JB, Galati JC, Royston P. A new framework for managing and analyzing multiple imputed data in Stata. Stata J. 2008;8(1):4967. 11. Potter JD, Hunter D. Colorectal cancer. In: Adami HO, Hunter D, Trichopoulos D, eds. Textbook of Cancer Epidemiology. Oxford, United Kingdom: Oxford University Press; 2008. 12. Australian Bureau of Statistics. Census of population and housing: socio-economic indexes for areas (SEIFA). Australiadata only, 2006. Canberra, Australia: Australian Bureau of Statistics; 2008. (http://www.abs.gov.au/ AUSSTATS/abs@.nsf/allprimarymainfeatures/ 8298F6AB51491A7CCA2570A400770380?opendocument). (Accessed August 26, 2009). 13. Royston P. Multiple imputation of missing values: update of ICE. Stata J. 2005;5(4):527536. 14. Chasan-Taber L, Erickson JB, McBride JW, et al. Reproducibility of a self-administered lifetime physical activity questionnaire among female college alumnae. Am J Epidemiol. 2002;155(3):282289. 15. Friedenreich CM, Courneya KS, Bryant HE. The lifetime total physical activity questionnaire: development and reliability. Med Sci Sports Exerc. 1998;30(2):266274. 16. Kriska AM, Sandler RB, Cauley JA, et al. The assessment of historical physical activity and its relation to adult bone parameters. Am J Epidemiol. 1988;127(5):10531063. 17. Arbman G, Axelson O, Fredriksson M, et al. Do occupational factors inuence the risk of colon and rectal cancer in different ways? Cancer. 1993;72(9):25432549. 18. Benito E, Obrador A, Stiggelbout A, et al. A population-based case-control study of colorectal cancer in Majorca. I. Dietary factors. Int J Cancer. 1990;45(1):6976. 19. Fredriksson M, Bengtsson NO, Hardell L, et al. Colon cancer, physical activity, and occupational exposures. A case-control study. Cancer. 1989;63(9):18381842. 20. Gerhardsson M, Floderus B, Norell SE. Physical activity and colon cancer risk. Int J Epidemiol. 1988;17(4):743746. 21. Longnecker MP, Gerhardsson le Verdier M, Frumkin H, et al. A case-control study of physical activity in relation to risk of cancer of the right colon and rectum in men. Int J Epidemiol. 1995;24(1):4250.

22. Moradi T, Gridley G, Bjo rk J, et al. Occupational physical activity and risk for cancer of the colon and rectum in Sweden among men and women by anatomic subsite. Eur J Cancer Prev. 2008;17(3):201208. 23. Peters RK, Garabrant DH, Yu MC, et al. A case-control study of occupational and dietary factors in colorectal cancer in young men by subsite. Cancer Res. 1989;49(19):54595468. 24. Tavani A, Braga C, La Vecchia C, et al. Physical activity and risk of cancers of the colon and rectum: an Italian case-control study. Br J Cancer. 1999;79(11-12):19121916. 25. Zhang Y, Cantor KP, Dosemeci M, et al. Occupational and leisure-time physical activity and risk of colon cancer by subsite. J Occup Environ Med. 2006;48(3):236243. 26. Colbert LH, Hartman TJ, Malila N, et al. Physical activity in relation to cancer of the colon and rectum in a cohort of male smokers. Cancer Epidemiol Biomarkers Prev. 2001;10(3): 265268. 27. Howard RA, Freedman DM, Park Y, et al. Physical activity, sedentary behavior, and the risk of colon and rectal cancer in the NIH-AARP Diet and Health Study. Cancer Causes Control. 2008;19(9):939953. 28. Steindorf K, Tobiasz-Adamczyk B, Popiela T, et al. Combined risk assessment of physical activity and dietary habits on the development of colorectal cancer. A hospital-based casecontrol study in Poland. Eur J Cancer Prev. 2000;9(5):309316. 29. Whittemore AS, Wu-Williams AH, Lee M, et al. Diet, physical activity, and colorectal cancer among Chinese in North America and China. J Natl Cancer Inst. 1990;82(11):915926. 30. Lynch BM. Sedentary behavior and cancer: a systematic review of the literature and proposed biological mechanisms. Cancer Epidemiol Biomarkers Prev. 2010;19(11):26912709. 31. Healy GN, Dunstan DW, Salmon J, et al. Objectively measured light-intensity physical activity is independently associated with 2-h plasma glucose. Diabetes Care. 2007;30(6): 13841389. 32. Giovannucci E. Insulin, insulin-like growth factors and colon cancer: a review of the evidence. J Nutr. 2001;131(suppl 11): 3109S3120S. 33. Giovannucci E. Metabolic syndrome, hyperinsulinemia, and colon cancer: a review. Am J Clin Nutr. 2007;86(3):836S842S. 34. Hu FB, Li TY, Colditz GA, et al. Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA. 2003;289(14):17851791. 35. McTiernan A. Mechanisms linking physical activity with cancer. Nat Rev Cancer. 2008;8(3):205211. 36. Hamilton MT, Hamilton DG, Zderic TW. Role of low energy expenditure and sitting in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease. Diabetes. 2007;56(11): 26552667. 37. Laake I, Thune I, Selmer R, et al. A prospective study of body mass index, weight change, and risk of cancer in the proximal and distal colon. Cancer Epidemiol Biomarkers Prev. 2010; 19(6):15111522. 38. Koushik A, Hunter DJ, Spiegelman D, et al. Fruits, vegetables, and colon cancer risk in a pooled analysis of 14 cohort studies. J Natl Cancer Inst. 2007;99(19):14711483. 39. Huncharek M, Muscat J, Kupelnick B. Colorectal cancer risk and dietary intake of calcium, vitamin D, and dairy products: a meta-analysis of 26,335 cases from 60 observational studies. Nutr Cancer. 2009;61(1):4769. 40. Larsson SC, Wolk A. Meat consumption and risk of colorectal cancer: a meta-analysis of prospective studies. Int J Cancer. 2006;119(11):26572664. 41. Friedenreich C, Norat T, Steindorf K, et al. Physical activity and risk of colon and rectal cancers: the European Prospective Am J Epidemiol. 2011;173(10):11831191

Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

Sedentary Work and Colorectal Cancer Risk

1191

42. 43. 44. 45.

46. 47.

48.

49.

50.

Investigation into Cancer and Nutrition. Cancer Epidemiol Biomarkers Prev. 2006;15(12):23982407. Ogino S, Chan AT, Fuchs CS, et al. Molecular pathological epidemiology of colorectal neoplasia: an emerging transdisciplinary and interdisciplinary eld. Gut. 2011;60(3):397411. Feskanich D, Ma J, Fuchs CS, et al. Plasma vitamin D metabolites and risk of colorectal cancer in women. Cancer Epidemiol Biomarkers Prev. 2004;13(9):15021508. Larsson SC, Orsini N, Wolk A. Diabetes mellitus and risk of colorectal cancer: a meta-analysis. J Natl Cancer Inst. 2005; 97(22):16791687. Limburg PJ, Stolzenberg-Solomon RZ, Vierkant RA, et al. Insulin, glucose, insulin resistance, and incident colorectal cancer in male smokers. Clin Gastroenterol Hepatol. 2006;4(12):15141521. Gunter MJ, Hoover DR, Yu H, et al. Insulin, insulin-like growth factor-I, endogenous estradiol, and risk of colorectal cancer in postmenopausal women. Cancer Res. 2008;68(1):329337. Jenab M, Riboli E, Cleveland RJ, et al. Serum C-peptide, IGFBP-1 and IGFBP-2 and risk of colon and rectal cancers in the European Prospective Investigation into Cancer and Nutrition. Int J Cancer. 2007;121(2):368376. Rothwell PM, Wilson M, Elwin CE, et al. Long-term effect of aspirin on colorectal cancer incidence and mortality: 20-year follow-up of ve randomised trials. Lancet. 2010;376(9754): 17411750. Mahipal A, Anderson KE, Limburg PJ, et al. Nonsteroidal anti-inammatory drugs and subsite-specic colorectal cancer incidence in the Iowa Womens Health Study. Cancer Epidemiol Biomarkers Prev. 2006;15(10):17851790. Matthews CE, Chen KY, Freedson PS, et al. Amount of time spent in sedentary behaviors in the United States, 20032004. Am J Epidemiol. 2008;167(7):875881.

51. Healy GN, Dunstan DW, Salmon J, et al. Breaks in sedentary time: benecial associations with metabolic risk. Diabetes Care. 2008;31(4):661666. 52. Chau JY, der Ploeg HP, van Uffelen JG, et al. Are workplace interventions to reduce sitting effective? A systematic review. Prev Med. 2010;51(5):352356. 53. Hamilton M, Healy G, Dunstan D, et al. Too little exercise and too much sitting: inactivity physiology and the need for new recommendations on sedentary behavior. Curr Cardiovasc Risk Rep. 2008;2(4):292298. 54. Australian Department of Health and Ageing. Get Up and Grow: Healthy Eating and Physical Activity Guidelines for Early Childhood (Director/Coordinator Book). Canberra, Australia: Department of Health and Ageing, Government of Australia; 2009. 55. Kesa niemi A, Riddoch CJ, Reeder B, et al. Advancing the Future of Physical Activity Guidelines in Canada: An Independent Expert Panel Interpretation of the Evidence. Ottawa, Ontario Canada: Canadian Society for Exercise Physiology; 2010. (http://www.csep.ca/English/view.asp?x724&id260). 56. van Poppel MN, Chinapaw MJ, Mokkink LB, et al. Physical activity questionnaires for adults: a systematic review of measurement properties. Sports Med. 2010;40(7): 565600. 57. Katzmarzyk PT, Church TS, Craig CL, et al. Sitting time and mortality from all causes, cardiovascular disease, and cancer. Med Sci Sports Exerc. 2009;41(5):9981005. 58. Patel AV, Bernstein L, Deka A, et al. Leisure time spent sitting in relation to total mortality in a prospective cohort of US adults. Am J Epidemiol. 2010;172(4):419429. 59. Owen N, Bauman A, Brown W. Too much sitting: a novel and important predictor of chronic disease risk? Br J Sports Med. 2009;43(2):8183.

Downloaded from http://aje.oxfordjournals.org/ by guest on November 12, 2012

Am J Epidemiol. 2011;173(10):11831191