Menopausal Hormone Therapy Use and Risk of Invasive Colon Cancer

Transcription

1 Amerian Journal of Epidemiology ª The Author Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg Shool of Publi Health. All rights reserved. For permissions, please Vol. 171, No. 4 DOI: /aje/kwp434 Advane Aess publiation: January 11, 2010 Original Contribution Menopausal Hormone Therapy Use and Risk of Invasive Colon Caner The California Teahers Study Katherine DeLellis Henderson*, Lei Duan, Jane Sullivan-Halley, Huiyan Ma, Christina A. Clarke, Susan L. Neuhausen, Claire Templeman, and Leslie Bernstein * Correspondene to Dr. Katherine DeLellis Henderson, Division of Caner Etiology, Department of Population Sienes, Bekman Researh Institute, City of Hope National Medial Center,1500 East Duarte Road, Duarte, CA ( khenderson@oh.org). Initially submitted July 13, 2009; aepted for publiation Deember 2, Results from epidemiologi studies of hormone therapy use and olon aner risk are inonsistent. This question was investigated in the California Teahers Study ( ) among 56,864 perimenopausal or postmenopausal partiipants under 80 years of age with no prior oloretal aner by using Cox proportional hazards regression. Inident invasive olon aner was diagnosed among 442 partiipants. Baseline-reent hormone therapy users were at 36% lower risk for olon aner versus baseline-never users (baseline-reent users: relative risk (RR) ¼ 0.64, 95% onfidene interval (CI): 0.51, 0.80). Results did not differ by formulation. Estimated risk was lower among baseline-reent hormone therapy users with inreasing duration between 5 and 15 years of use (RR ¼ 0.49, 95% CI: 0.35, 0.68), but the trend did not persist in the longest duration group, more than 15 years of use (RR ¼ 0.69, 95% CI: 0.52, 0.92; P trend ¼ 0.60). Long-term rereational physial ativity, obesity, regular use of nonsteroidal antiinflammatory mediations, and daily alohol intake did not modify these effets; baseline-reent use was more strongly assoiated with olon aner risk among women with a family history of oloretal aner (P heterogeneity ¼ 0.04). Baseline-reent hormone therapy use was inversely assoiated with invasive olon aner risk among perimenopausal and postmenopausal women in the California Teahers Study. oloni neoplasms; hormone replaement therapy; lung neoplasms; parity; prospetive studies; reprodution; smoking Abbreviations: CI, onfidene interval; HT, hormone therapy; Q2000, third California Teahers Study questionnaire sent in 2000; RR, relative risk; SD, standard deviation. Previous studies have shown that menopausal hormone therapy is assoiated with dereased risk of olon aner (1). The Women s Health Initiative trials found that ontinuous ombined hormone therapy (2), but not unopposed estrogen therapy (3), was assoiated with dereased oloretal aner risk. Those assoiations have been repliated (4) and disputed (5, 6) by subsequent studies. Few studies have investigated a possible dose-response between inreasing duration of hormone therapy use and dereasing olon aner risk. The Nurses Health Study reported no duration dose response but did report a strong redution in risk among reent hormone therapy users, whih was attenuated with inreasing time sine last use (7). In the urrent analysis, the assoiation between olon aner and hormone therapy use overall, by formulation, by duration of use, and by time sine last use among women partiipating in the California Teahers Study was assessed. In addition, we investigated the extent to whih several fators previously hypothesized to be assoiated with olon aner risk, inluding physial ativity (8, 9), body mass index (10, 11), regular nonsteroidal antiinflammatory drug use (12), alium intake (13, 14), alium plus vitamin D intake (15), family history of oloretal aner (16), personal history of oloretal polyps (17), smoking history (18), and alohol intake (19), might at as effet modifiers of the hormone therapy and olon aner risk assoiation or 415

2 416 Henderson et al. might appear to interat with hormone therapy in the ontext of olon aner risk. MATERIALS AND METHODS The California Teahers Study is a prospetive ohort of urrent and former female publi shool teahers and administrators, who were members of the California State Teahers Retirement System in Cohort partiipants ompleted a questionnaire, mailed in 1995, providing detailed information on fators suh as hormone therapy use, personal medial history, reprodutive history, physial ativity, anthropometris, mediation use, diet, and family history of oloretal aner. The third California Teahers Study questionnaire, sent in 2000 (Q2000), updated information on menopausal status and hormone therapy use. A detailed desription of the California Teahers Study is available (20). Use of human subjet data was approved by the institutional review boards at eah ollaborating institution in aord with assuranes approved by the US Department of Health and Human Servies. The California Teahers Study ohort omprises 133,479 women. Exlusions, in sequene, were women who, at baseline, lived outside California (n ¼ 8,867), had a prior/ unknown history of oloretal aner (n ¼ 1,559), had limited partiipation in breast aner researh (n ¼ 18), were 80 years or older (n ¼ 5,532), were premenopausal (n ¼ 47,966), had unknown menopausal status (n ¼ 4,961), or had unknown hormone therapy or progestin-only use (n ¼ 7,712). The resulting ohort for the analysis of baseline data onsisted of 56,864 women (2,245 perimenopausal and 54,619 postmenopausal). Case asertainment and follow-up Inident invasive olon aners were identified through annual linkages with the California Caner Registry, whih reeives reports of over 99% of aner diagnoses ourring in California (21). Invasive olon aner was defined as an International Classifiation of Diseases for Onology, Seond Edition (ICD-O-2), topography ode C180 C189 or C260 (large intestine not otherwise speified). C180 C185 odes were onsidered proximal and C186 C189 and C260 odes were onsidered distal. Stage, proured from the California Caner Registry summary stage variable, was lassified as loalized, regional, and distant. A total of 442 eligible partiipants were diagnosed with invasive olon aner during follow-up, whih began on the baseline questionnaire date and ontinued until the diagnosis of olon aner or the first ourrene of a ensoring event (reloation outside California lasting more than 4 months (n ¼ 4,702), diagnosis of in situ olon aner (n ¼ 26), death (n ¼ 5,600), or end of follow-up, Deember 31, 2006 (n ¼ 46,094)). Exposure assessment Women reporting ongoing menstrual periods who had never used hormones for menopausal symptoms were onsidered premenopausal. Women were lassified as perimenopausal if periods had stopped within the last 6 months and they were not urrently pregnant, and as postmenopausal if they met any of the following riteria: 1) periods stopped more than 6 months ago, 2) bilateral oophoretomy, 3) age 56 years or older at baseline and not already lassified as premenopausal or perimenopausal, 4) started using hormone therapy for menopausal symptoms before periods stopped, and/or 5) hysteretomy before age 56 years but aged 56 years or more at baseline. The age riterion was based on previous work indiating that approximately 90% of women of age 56 years or older were biologially postmenopausal (22). The prinipal hormone therapy item on the baseline questionnaire asked, Have you ever taken estrogen for symptoms of menopause (the hange of life) or for other reasons? Response ategories were no, yes, and I am urrently taking estrogens, and yes, but I am no longer taking estrogens. Subsequent questions inluded type of hormone used, ages of first and last use, and total years of use for eah type of hormone. Variables were reated to haraterize the pattern of hormone therapy use over time with respet to formulation, duration of use, and years sine last use. Hormone therapy was ategorized in several ways, first as never or ever hormone therapy user. The ever hormone therapy users inluded partiipants who had only ever used a single formulation (ever hormone therapy user, estrogen therapy only or ever hormone therapy user, estrogenplus-progestin therapy only) and who had used more than one formulation over their lifetimes (ever hormone therapy user, mixed formulations). Analyses assessed the effets by various ategorizations of hormone therapy, and sensitivity analyses tested for differenes in the effets when analyses were limited to women who had used only one formulation, either estrogen alone or ombined estrogen-plus-progestin therapy. Women who reported no hormone therapy use, past hormone therapy use, and urrent hormone therapy use on the baseline questionnaire will be referred to as baselinenever hormone therapy (HT) users, baseline-former HT users, and baseline-reent HT users, respetively. In addition, for assessment of the possible effet of hanges in hormone therapy use status between baseline and Q2000, a time-dependent hormone therapy use variable was reated by using information from both questionnaires. Baseline values were used until Q2000, at whih point the Q2000 value was used. If no Q2000 value was available, the baseline value was retained. A more simple, ombined baseline/ Q2000 hormone therapy use variable was also reated: baseline-never/q2000-never, baseline-never/q2000-former, baseline-never/q2000-reent, baseline-former/q2000- former, baseline-former/q2000-reent, baseline-reent/ Q2000-former, or baseline-reent/q2000-reent. Statistial analysis Multivariable Cox proportional hazards regression methods were used to assess the assoiations of hormone therapy use by various ategorizations with invasive olon aner risk, using ages at the start and the end of follow-up (in days) to define time on study. Models were adjusted for rae/ethniity, body mass index (kg/m 2 ), and physial

3 Menopausal Hormone Therapy and Invasive Colon Caner 417 ativity, and they were stratified by age at baseline (in single years of age). Hazard rate ratios, presented as relative risks with 95% onfidene intervals, were estimated. For ordinal variables, we tested for linear trend in the log e (relative risk) aross exposure ategories. To assess the proportional hazards assumption using baseline hormone therapy status (baseline-never HT user vs. baseline-ever HT user), we first visually examined whether Kaplan-Meier survival urves had parallel lines (23). We also plotted saled Shoenfeld residuals by time to test for a zero slope and tested the null hypothesis of no orrelation between the residuals and time on study (24). No evidene for a violation of the proportional hazards assumption was apparent. We examined the assoiation between invasive olon aner and hormone therapy use by stage of disease at diagnosis (loalized (n ¼ 159), regional (n ¼ 181), or distant (n ¼ 90)) and by loation of disease (proximal (n ¼ 321) vs. distal (n ¼ 121) olon aner). Sensitivity analyses to test for exlusion of perimenopausal women and to test for differenes in effets among women whose hormone therapy use was limited to one formulation did not differ markedly from those presented. We further assessed possible effet modifiation by physial ativity, body mass index, regular nonsteroidal antiinflammatory drug use, alium intake, alium plus vitamin D intake ombined, first-degree family history of oloretal aner, personal history of oloretal polyps, smoking history, and alohol intake, aording to the homogeneity of trends in hormone therapy use aross ategories of eah modifier. Additional analyses testing for statistial interation between hormone therapy use and these fators in whih baseline-never HT users in the lowest level of eah fator were the referene group did not differ measurably from the results presented herein. Additional analyses to determine whether inlusion of updated exposure information influened the results inluded women eligible for analysis at baseline, who were postmenopausal at baseline and postmenopausal at Q2000, and who provided omplete hormone therapy information at Q2000. Cox regression models were used to test the potential effet of hormone therapy, inorporating hormone therapy information from both the baseline questionnaire and Q2000. Adjusted relative risks of invasive olon aner were omputed for the time-dependent hormone therapy variable desribed above. Additional models examined exlusion of partiipants with missing Q2000 values or use of the simpler variable. Covariate lassifiations used in the analyses are presented in Table 1. All statistial analyses were performed by using SAS, version 9.1, software (SAS Institute, In., Cary, North Carolina). P values were not adjusted for multiple omparisons. RESULTS Nearly 76% of the 56,864 perimenopausal and postmenopausal women inluded in the analyses of baseline data reported ever using hormone therapy; 15% were baselineformer users and 61% were baseline-reent users (Table 1). Of the 34,433 baseline-reent HT users, 44% were using unopposed estrogen and 56% were using ombined estrogen plus progestin; 44% of the 442 women diagnosed with invasive olon aner were baseline-reent HT users. Baseline-reent HT users were younger than baseline-former or baseline-never HT users, with mean ages of 58.9 (standard deviation (SD) ¼ 8.5), 65.0 (SD ¼ 8.8), and 62.7 (SD ¼ 9.2) years, respetively (Table 1). Any use of hormone therapy, ompared with baselinenever use, was assoiated with a 28% dereased risk of inident invasive olon aner (relative risk (RR) ever HT use ¼ 0.72, 95% onfidene interval (CI): 0.58, 0.88) (Table 2). This relative risk did not differ when any hormone therapy use was ategorized further as estrogen therapy use only, estrogen-plus-progestin therapy use, or mixed hormone therapy use (Table 2). The redution in olon aner risk was restrited to baseline-reent HT users (RR baseline-reent HT use ¼ 0.64, 95% CI: 0.51, 0.80). Redutions in risk for baseline-reent users of unopposed estrogen and for baseline-reent users of estrogen-plus-progestin therapy were similar (RR unopposed estrogen ¼ 0.59, 95% CI: 0.45, 0.77 and RR estrogen þ progestin ¼ 0.71, 95% CI: 0.54, 0.93). Redutions in risk for baseline-reent users of hormone therapy were similar for distal versus proximal olon aners (RR distal ¼ 0.57, 95% CI: 0.37, 0.87 and RR proximal ¼ 0.67, 95% CI: 0.52, 0.87) (data not shown). The risk dereased among baseline-reent HT users with inreasing duration of use through 5 15 years of use (RR <5 years use ¼ 0.76, 95% CI: 0.54, 1.08 and RR 5 15 years use ¼ 0.49, 95% CI: 0.35, 0.68), but the trend did not persist in the longest duration group (RR >15 years use ¼ 0.69, 95% CI: 0.52, 0.92; P trend ¼ 0.06). The variable years sine last hormone therapy use was assoiated with olon aner risk among baseline-former HT users. Table 3 demonstrates that the assoiation between hormone therapy and invasive olon aner risk appeared to be stronger for tumors diagnosed at the regional and distant stages. For regional-stage tumors, the assoiation was stronger for partiipants who had ever used a progestin (RR ever HT user, estrogen therapy only ¼ 0.84, 95% CI: 0.59, 1.20; RR ever HT user, estrogen þ progestin therapy only ¼ 0.49, 95% CI: 0.24, 0.99; and RR ever HT user, mixed formulations ¼ 0.65, 95% CI: 0.43, 0.98). For distant-stage tumors, the magnitude of the assoiation did not differ learly (RR ever HT user, estrogen therapy only ¼ 0.50, 95% CI: 0.30, 0.84; RR ever HT user, estrogen þ progestin therapy only ¼ 0.80, 95% CI: 0.38, 1.67; and RR ever HT user, mixed formulations ¼ 0.42, 95% CI: 0.23, 0.76). The effet was statistially signifiant among baseline-reent HT users only, for regionalstage tumors (RR baseline-reent HT user ¼ 0.68, 95% CI: 0.48, 0.96), and distant-stage tumors (RR baseline-reent HT user ¼ 0.33, 95% CI: 0.20, 0.56) (Table 3). First-degree family history of oloretal aner may modify the assoiation between hormone therapy and olonanerrisk(p heterogeneity ¼ 0.04) (Table 4). Further, olon aner risk was lower among baseline-reent HT users with a positive family history of oloretal aner (RR ¼ 0.45, 95% CI: 0.26, 0.78) than among baselinereent HT users with no family history (RR ¼ 0.71, 95% CI: 0.56, 0.90) (data not shown). No other statistially signifiant effet modifiation or interation was apparent in these data.

4 418 Henderson et al. Table 1. Distribution of Baseline Charateristis by Baseline Status of Hormone Therapy Use Among 56,864 Perimenopausal and Postmenopausal Partiipants in the California Teahers Study, Charateristis Analyses of assoiation between hormone therapy use and olon aner risk, inorporating hormone therapy information from Q2000, produed essentially the same results as those presented. For example, using information provided at baseline and Q2000 and treating the baseline-never/q2000-never HT use group (n ¼ 6,878) as the referent, we found that being a reent user at both time points (n ¼ 18,023) was assoiated with a 40% redution in olon aner risk (RR ¼ 0.60, 95% CI: 0.44, 0.83) (data not shown). Similarly, the time-dependent hormone therapy exposure variable supported the above-stated results in that reent hormone therapy users were at dereased risk for olon aner (RR ¼ 0.60, 95% CI: 0.44, 0.81) (data not shown). Total No. Status of HT Use at Baseline (Estrogen, Estrogen 1 Progestin, or User of Mixed Formulations) Baseline-Never DISCUSSION Baseline-Former Baseline-Reent No. of partiipants 56,864 13,778 8,653 34,433 No. of invasive olon aner ases Mean age at baseline, years (SD) 62.7 (9.2) 65.0 (8.8) 58.9 (8.5) Menopausal status, % a Perimenopausal 2, Postmenopausal, natural menopause 29, Postmenopausal, bilateral oophoretomy 8, Postmenopausal, other reason 16, Rae/ethniity, % a White 50, Afrian Amerian 1, Other b 4, Long-term rereational physial ativity, % a, Low (<0.5 hour/week of any ativity) 8, Intermediate 28, High (3 hours/week of any ativity) 18, Body mass index, kg/m 2a <25 30, , , Regular NSAID use a,d Low 49, High 5, Calium intake, mg a,e , > , Calium, mg, þ vitamin D, IU, intake a,f Low 44, High 7, Table ontinues Our results support the hypothesis that hormone therapy use is assoiated with dereased risk of invasive olon aner. Baseline-reent HT use in the form of unopposed estrogen or ombined estrogen-plus-progestin therapy was assoiated with a 36% derease in olon aner risk. This risk redution was similar for unopposed estrogen users and for estrogen-plus-progestin users. Baseline-former use of hormone therapy was not assoiated with olon aner risk. Results of previous observational epidemiologi studies investigating the assoiation between hormone therapy use and olon aner risk have been inonsistent. A 1999 metaanalysis haraterized the broad heterogeneity in findings among 18 ohort and ase-ontrol studies addressing the

5 Menopausal Hormone Therapy and Invasive Colon Caner 419 Table 1. Continued Charateristis assoiation between hormone therapy use and olon and retal aner separately, reporting a summary age-adjusted relative risk for the assoiation between ever hormone therapy use and olon aner risk of 0.80 (95% CI: 0.74, 0.86) (1). The majority of hormone therapy presribed during the ourse of these studies was unopposed estrogen. Only 3 studies provided results for ombined estrogen-plus-progestin therapy: 2 found statistially nonsignifiant risk redutions (25, 26) and 1 found no effet (27). That baseline-reent HT use, but not baseline-former HT use, is assoiated with lower olon aner risk has been reported previously (7). In the Women s Health Initiative trials (28), ontinuous ombined estrogen-plus-progestin therapy was assoiated with dereased oloretal aner risk (2), whereas unopposed estrogen therapy was not (3). It is possible that the inlusion of retal aner in the outome may have attenuated the estrogen result, but a 1999 meta-analysis reported a summary age-adjusted relative risk for the hormone therapy retal aner assoiation of 0.81 (95% CI: 0.72, 0.92). Total No. Status of HT Use at Baseline (Estrogen, Estrogen 1 Progestin, or User of Mixed Formulations) Baseline-Never Baseline-Former Baseline-Reent Family history of oloretal aner a,g No 49, Yes 6, Personal history of oloretal polyps a,h No 51, Yes 3, Smoking history a,i Never smoker 33, Former smoker 19, Current smoker 3, Alohol intake, g/day a,j 0 17, <20 31, , Abbreviations: HT, hormone therapy; NSAID, nonsteroidal antiinflammatory drug; SD, standard deviation. a All perentages, exept those for rae/ethniity and menopausal status, are age standardized by 5-year age ategories to the age distribution of the analytial ohort; the numbers of missing and unknown partiipants are not shown in the table. b Other inludes Hispani, Asian, Native Amerian, mixed, or none reported. Long-term rereational physial ativity ombines strenuous and moderate ativity and is defined as low, intermediate, or high average weekly hours of long-term rereational physial ativity. d Regular NSAID use ombines information on aspirin and ibuprofen use. High NSAID use was defined as use 4 times per week for more than 4 years; low/no regular NSAID use inluded all others. e Calium intake (mg) from diet and supplements was ut at the median value for partiipants in this analysis, mg/day. f Calium intake plus vitamin D intake (from diet and supplements) was dihotomized as the highest quartile for both alium and vitamin D versus all others. g A positive family history of oloretal aner was defined as olon or retal aner in at least 1 first-degree relative (mother, father, sister, or brother). h Self-reported personal history of olon or retal polyps (not aner) was ategorized as no, yes, or unknown. i Smoking history was ategorized as never, former, or urrent smoker. j Alohol intake was ategorized as 0 g/day, <20 g/day, or 20 g/day. Three studies have been published on hormone therapy and olon aner risk sine the Women s Health Initiative publiation. Newomb et al. (4) reported an inverse assoiation between oloretal aner risk and urrent estrogenplus-progestin formulations in a large ase-ontrol study but no assoiation with use of unopposed estrogen therapy or former estrogen-plus-progestin therapy. In ontrast, Campbell et al. (5) reported lower risk of oloretal aner among ever users of hormone therapy in a ase-ontrol study, with no statistially signifiant differene in the main effet by formulation. However, Campbell et al. did report a differential effet of urrent versus past use omparable to our results, but only among unopposed estrogen users. Finally, a report from the Breast Caner Detetion Demonstration Projet follow-up study has provided results similar to ours with respet to formulation and reeny of use (baseline-reent vs. baseline-former) (6). The apparent attenuation of the duration dose response among baseline-reent HT users after 15 years of use has

6 420 Henderson et al. Table 2. Adjusted a Risks and 95% Confidene s for the Assoiation Between Baseline Status of Menopausal Hormone Therapy Use and Inident Invasive Colon Caner Among Perimenopausal and Postmenopausal Partiipants in the California Teahers Study, Status of HT Use at Baseline (Estrogen, Estrogen 1 Progestin, or User of Mixed Formulations) not been reported previously. Most prior studies have ategorized duration of hormone therapy use as <5 years or 5 years (4, 5, 7, 27, 29). The report from the Breast Caner Detetion Demonstration Projet follow-up study showed a dose response with inreasing duration up to 10 years of use (6). In the California Teahers Study, with restrition to women who had used only one hormone therapy formulation, this attenuation in risk for the longest duration (>15 years) was stronger among reent users of ombined estrogen-plus-progestin therapy than among reent users of unopposed estrogen (data not shown). Further analytial strategies aimed at testing the Total No. Person-Years No. of Cases 95% Risk a Confidene Ever HT use, at baseline Ever HT user (former and 43, , , 0.88 reent HT users) Ever HT use, at baseline b Ever HT user, 16, , , 0.90 estrogen therapy only Ever HT user, estrogen 5,324 53, , 1.06 þ progestin therapy only Ever HT user, mixed formulations 21, , , 0.93 Former or reent HT use, at baseline b Former HT user 8,653 84, , 1.15 Reent HT user 34, , , 0.80 Type of HT used at baseline b Former HT user 8,653 84, , 1.14 Reent estrogen therapy user 15, , , 0.77 Reent estrogen 19, , , 0.93 þ progestin therapy user Duration of HT use, at baseline b Ever HT user, <5 years duration 19, , , 1.10 Ever HT user, 5 15 years duration 13, , , 0.76 Ever HT user, >15 years duration 8,097 79, , 0.96 P trend Duration of HT use, at baseline by former/ reent use b Former HT user <5 years duration 5,788 56, , years duration 1,770 17, , 1.39 >15 years duration 730 6, , 1.73 P trend Table ontinues onsisteny of the attenuation of the duration doseresponse effet among the long-term baseline-reent HT users, suh as exlusion of partiipants who were perimenopausal, those with early menopause (<43 years), or those who ended follow-up on Deember 31, 2001, whih tests for mislassifiation of the longest duration baselinereent HT due to undeteted essation of use, did not alter the results in the California Teahers Study. No statistially signifiant assoiation was evident with years sine last hormone therapy use among baseline-former HT users. These results agree with some (30, 31), but not all (26), previous studies.

7 Menopausal Hormone Therapy and Invasive Colon Caner 421 Table 2. Continued Status of HT Use at Baseline 95% No. of (Estrogen, Estrogen 1 Progestin, Total No. Person-Years Cases Risk a Confidene or User of Mixed Formulations) Reent HT user <5 years duration 13, , , years duration 11, , , 0.68 >15 years duration 7,367 72, , 0.92 P trend 0.60 Duration of HT use, at baseline by formulation b Ever HT user, estrogen therapy only <5 years duration 6,231 61, , years duration 4,489 45, , 0.91 >15 years duration 5,203 50, , 0.94 P trend 0.08 Ever HT user, estrogen þ progestin therapy only <5 years duration 2,387 23, , years duration 1,873 18, , 1.12 >15 years duration 701 6, , 1.79 P trend 0.64 Ever HT user, mixed formulations <5 years duration 10, , , years duration 7,179 73, , 0.79 >15 years duration 2,193 21, , 1.25 P trend 0.58 Years sine last HT use for former HT users, at baseline b Former HT user, 5 years sine last use 3,873 38, , 1.24 Former HT user, >5 years sine last use 4,747 45, , 1.23 Abbreviation: HT, hormone therapy. a Adjusted for rae (as shown), body mass index (ontinuous measure), and physial ativity (low, intermediate, high) and stratified by age at ohort entry (ontinuous measure in years). b The number of missing is not shown in the table. The trend effet was estimated by using the ontinuous variable of HT duration (in years) among users. Few previous epidemiologi studies have addressed the possible assoiation between hormone therapy use and olon aner risk with respet to stage of disease at diagnosis. Nurses Health Study investigators reported similar assoiations with hormone therapy for higher and lower stage olon aners (7). Results from the Women s Health Initiative estrogen-plus-progestin trial indiated a statistially signifiant dereased risk for less advaned aners and no statistially signifiant effet for regional or metastati disease; in addition, olon aners diagnosed in the estrogenplus-progestin arm tended be of more advaned stage ompared with those in the plaebo arm (32). Reent Women s Health Initiative analyses have not shown lower oloretal mortality in the estrogen-plus-progestin arm than in the plaebo arm (33). In ontrast, the urrent California Teahers Study results suggest that the assoiation between hormone therapy and invasive olon aner is stronger in nonloal disease; however, it is possible that this differene may be explained by a high level of sreening among hormone therapy users in the California Teahers Study, whih might inrease the number of deteted loal ases. The California Teahers Study does not have data on olon sreening speifially, preluding evaluation of an assoiation between high soioeonomi status and overdiagnosis of olon aner. Of the plethora of California Teahers Study data items olleted on the baseline questionnaire, several have been hypothesized to be assoiated with risk of olon or oloretal aner. These inlude dereased physial ativity (8, 9), body mass index (10, 11), nonsteroidal antiinflammatory drug use (12), alium intake (13, 14), alium plus vitamin D intake (15), family history of oloretal

8 422 Henderson et al. Table 3. Adjusted a Risks and 95% Confidene s for the Assoiation Between Baseline Status of Menopausal Hormone Therapy Use and Inident Invasive Colon Caner, by Stage of Disease at Diagnosis, b Among Perimenopausal and Postmenopausal Partiipants in the California Teahers Study, Status of HT Use at Baseline No. of Cases aner (16), personal history of oloretal polyps (17), smoking history (18), and alohol intake (19). Suh fators were hypothesized to modify the assoiation between hormone therapy and olon aner risk or perhaps to interat with hormone therapy in the ontext of olon aner risk. In this investigation, the most lear modifiation was a greater risk redution assoiated with baseline-reent HT use among partiipants with a positive first-degree family history of oloretal aner. Although the main effet between positive family history and inreased olon aner risk has been established by previous studies (16), only one ase-ontrol study has investigated the possible modifiation of the hormone therapy olon aner risk assoiation by family history of olon aner (34); no evidene for suh an effet was deteted (17, 18, 35). A likely mehanism linking hormone therapy use with dereased olon aner risk is regulation of apoptosis genes, perhaps through regulation by the estrogen reeptor-b. Estrogen reeptor-b is the dominant estrogen reeptor in the human olon (36), and 17b-estradiol indues apoptosis in olon aner ells through gene expression regulation mediated by estrogen reeptor-b (37, 38). Our study has several strengths. Its prospetive design prevents the differential mislassifiation of exposure due to errors in reall, beause exposures were measured before diagnosis. The large number of inident olon aners provides substantial statistial power to detet assoiations of modest size. Linkage with the California Caner Registry provides virtually omplete ase asertainment and aess to detailed information about tumor stage and loation. Loalized Regional Distant Risk 95% Confidene No. of Cases Risk 95% Confidene No. of Cases Risk 95% Confidene Ever HT use, at baseline Never HT user Referent Referent Referent Ever HT user, estrogen , , , 0.84 therapy only Ever HT user, estrogen , , , 1.67 þ progestin therapy only Ever HT user, mixed , , , 0.76 formulations Former or reent HT use, at baseline Never HT user Referent Referent Referent Former HT user , , , 1.58 Reent HT user , , , 0.56 Abbreviation: HT, hormone therapy. a Adjusted for rae (as shown), body mass index (ontinuous measure), and physial ativity (low, intermediate, high) and stratified by age at ohort entry (ontinuous measure in years). b Partiipants with missing stage (n ¼ 12) are not shown in the table. A onsideration for the analyses herein is that lassifiation of reent hormone therapy users was defined by hormone therapy status prior to The widely publiized early stopping of the Women s Health Initiative ombined estrogen-plus-progestin trial due to evidene that overall risks exeeded the benefits of the treatment (2) in 2002 inreases the likelihood that baseline-reent HT users stopped use shortly after baseline. Results of a sensitivity analysis in whih follow-up was ended on Deember 31, 2001, did not differ measurably from those presented herein. Other limitations inlude the lak of information in the California Teahers Study on dose for eah formulation, whih preluded examination of this detail, and possible residual onfounding, whih annot be ompletely ruled out, even in light of the information on possible onfounders obtained by the California Teahers Study. The California Teahers Study ohort has a high baseline prevalene of hormone therapy use (20), onsistent with the partiipants above-average aess to health insurane and health are and with levels of hormone therapy use in omparable ohorts (39). As desribed above, sensitivity analyses were performed to test the stability of our results to several suh issues and found results onsistent with an overall redution in risk for baseline-reent HT users. Baseline-reent HT use was assoiated with dereased risk of invasive olon aner. Similar results were found for unopposed estrogen and estrogen-plus-progestin therapy. This assoiation was modified by family history of oloretal aner. These results highlight a need to understand the organ-speifi effets of hormone therapy.

9 Table 4. Adjusted a Risks and 95% Confidene s of Invasive Colon Caner, by Baseline Status of Menopausal Hormone Therapy Use, for Seleted Potential Effet Modifiers, Among Perimenopausal and Postmenopausal Partiipants in the California Teahers Study, No. of Cases Baseline Status of HT Use b Baseline- Never Baseline- Former Baseline- Reent Risk a 95% Confidene Risk a 95% Confidene Risk a 95% Confidene Long-term rereational physial ativity Low Referent , , Intermediate/high Referent , , Body mass index, kg/m 2 < Referent , , Referent , , Regular NSAID use d Low Referent , , 0.85 < High Referent , , Calium intake, mg e Referent , , > Referent , , Calium, mg, þ vitamin D, IU, intake a,f Low Referent , , 0.83 < High Referent , , Family history of oloretal aner g No Referent , , Yes Referent , , Personal history of oloretal polyps h No Referent , , 0.80 < Yes Referent , , Smoking history i Never smoker Referent , , Ever smoker Referent , , Alohol intake, g/day j Referent , , > Referent , , P trend P homogeneity Abbreviations: HT, hormone therapy; NSAID, nonsteroidal antiinflammatory drug. a Adjusted for rae (as shown), body mass index (ontinuous measure), and physial ativity (low, intermediate, high) and stratified by age at ohort entry (ontinuous measure in years). b HT use refers to use of any of the following: estrogen therapy only, estrogen þ progestin therapy only, and mixed formulations. Long-term rereational physial ativity ombines strenuous and moderate ativity and is defined as low, intermediate, or high average weekly hours of long-term rereational physial ativity. d Regular NSAID use ombines information on aspirin and ibuprofen use. High NSAID use was defined as use 4 times per week for more than 4 years; low/no regular NSAID use inluded all others. e Calium intake (mg) from diet and supplements was ut at the median value for partiipants in this analysis, mg/day. f Calium intake plus vitamin D intake (from diet and supplements) was dihotomized as highest quartile for both alium and vitamin D versus all others. g A positive family history of oloretal aner was defined as olon or retal aner in at least 1 first-degree relative (mother, father, sister or brother). h Self-reported personal history of olon or retal polyps (not aner) was ategorized as no, yes, or unknown. i For analyses of effet modifiation, smoking was ategorized as never smoker or ever smoker. j For analyses of effet modifiation, alohol intake was ategorized as 0 g/day or >0 g/day. Menopausal Hormone Therapy and Invasive Colon Caner 423

10 424 Henderson et al. ACKNOWLEDGMENTS Author affiliations: Division of Caner Etiology, Department of Population Sienes, Bekman Researh Institute, City of Hope National Medial Center, Duarte, California (Katherine DeLellis Henderson, Lei Duan, Jane Sullivan- Halley, Huiyan Ma, Leslie Bernstein); Northern California Caner Center, Fremont, California (Christina A. Clarke); Department of Epidemiology, Shool of Mediine, University of California-Irvine, Irvine, California (Susan L. Neuhausen); and Kek Shool of Mediine, University of Southern California, Los Angeles, California (Claire Templeman). This work was supported by the National Institutes of Health (grant CA 77398). The olletion of aner inidene data used in this study was supported by the California Department of Health Servies as part of the statewide aner reporting program mandated by California Health and Safety Code Setion ; the National Caner Institute s Surveillane, Epidemiology, and End Results Program under ontrat N01-PC awarded to the Northern California Caner Center, ontrat N01-PC awarded to the University of Southern California, and ontrat N02-PC awarded to the Publi Health Institute; and the Centers for Disease Control and Prevention s National Program of Caner Registries, under agreement U55/CCR awarded to the Publi Health Institute. Conflit of interest: none delared. REFERENCES 1. Grodstein F, Newomb PA, Stampfer MJ. Postmenopausal hormone therapy and the risk of oloretal aner: a review and meta-analysis. Am J Med. 1999;106(5): Rossouw JE, Anderson GL, Prentie RL, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: prinipal results from the Women s Health Initiative randomized ontrolled trial. JAMA. 2002;288(3): Anderson GL, Limaher M, Assaf AR, et al. Effets of onjugated equine estrogen in postmenopausal women with hysteretomy: the Women s Health Initiative randomized ontrolled trial. JAMA. 2004;291(14): Newomb PA, Zheng Y, Chia VM, et al. Estrogen plus progestin use, mirosatellite instability, and the risk of oloretal aner in women. Caner Res. 2007;67(15): Campbell PT, Newomb P, Gallinger S, et al. Exogenous hormones and oloretal aner risk in Canada: assoiations stratified by linially defined familial risk of aner. Caner Causes Control. 2007;18(7): Johnson JR, Laey JV Jr, Lazovih D, et al. Menopausal hormone therapy and risk of oloretal aner. Caner Epidemiol Biomarkers Prev. 2009;18(1): Grodstein F, Martinez ME, Platz EA, et al. Postmenopausal hormone use and risk for oloretal aner and adenoma. Ann Intern Med. 1998;128(9): Mai PL, Sullivan-Halley J, Ursin G, et al. Physial ativity and olon aner risk among women in the California Teahers Study. Caner Epidemiol Biomarkers Prev. 2007;16(3): Samad AK, Taylor RS, Marshall T, et al. A meta-analysis of the assoiation of physial ativity with redued risk of oloretal aner. Coloretal Dis. 2005;7(3): MaInnis RJ, English DR, Hopper JL, et al. Body size and omposition and olon aner risk in women. Int J Caner. 2006;118(6): Pishon T, Lahmann PH, Boeing H, et al. Body size and risk of olon and retal aner in the European Prospetive Investigation into Caner and Nutrition (EPIC). J Natl Caner Inst. 2006;98(13): Chan AT, Giovannui EL, Meyerhardt JA, et al. Longterm use of aspirin and nonsteroidal anti-inflammatory drugs and risk of oloretal aner. JAMA. 2005;294(8): Flood A, Peters U, Chatterjee N, et al. Calium from diet and supplements is assoiated with redued risk of oloretal aner in a prospetive ohort of women. Caner Epidemiol Biomarkers Prev. 2005;14(1): MCullough ML, Robertson AS, Rodriguez C, et al. Calium, vitamin D, dairy produts, and risk of oloretal aner in the Caner Prevention Study II Nutrition Cohort (United States). Caner Causes Control. 2003;14(1): Ding EL, Mehta S, Fawzi WW, et al. Interation of estrogen therapy with alium and vitamin D supplementation on oloretal aner risk: reanalysis of Women s Health Initiative randomized trial. Int J Caner. 2008;122(8): Fuhs CS, Giovannui EL, Colditz GA, et al. A prospetive study of family history and the risk of oloretal aner. N Engl J Med. 1994;331(25): Giardiello FM, Brensinger JD, Petersen GM. AGA tehnial review on hereditary oloretal aner and geneti testing. Gastroenterology. 2001;121(1): Giovannui E. An updated review of the epidemiologial evidene that igarette smoking inreases risk of oloretal aner. Caner Epidemiol Biomarkers Prev. 2001;10(7): Moskal A, Norat T, Ferrari P, et al. Alohol intake and oloretal aner risk: a dose-response meta-analysis of published ohort studies. Int J Caner. 2007;120(3): Bernstein L, Allen M, Anton-Culver H, et al. High breast aner inidene rates among California teahers: results from the California Teahers Study (United States). Caner Causes Control. 2002;13(7): Kwong S, Perkin C, Morris C, et al. Caner in California: Saramento, CA: Caner Surveillane Setion, California Department of Health Servies; Weiss LK, Burkman RT, Cushing-Haugen KL, et al. Hormone replaement therapy regimens and breast aner risk. Obstet Gyneol. 2002;100(6): Therneau T. Extending the Cox Model. Tehnial Report Series No. 58. Rohester, MN: Department of Health Siene Researh, Mayo Clini; Shoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika. 1982;69(1): Persson I, Yuen J, Bergkvist L, et al. Caner inidene and mortality in women reeiving estrogen and estrogen-progestin replaement therapy long-term follow-up of a Swedish ohort. Int J Caner. 1996;67(3): Newomb PA, Storer BE. Postmenopausal hormone use and risk of large-bowel aner. J Natl Caner Inst. 1995;87(14): Troisi R, Shairer C, Chow WH, et al. A prospetive study of menopausal hormones and risk of oloretal aner (United States). Caner Causes Control. 1997;8(2):

11 Menopausal Hormone Therapy and Invasive Colon Caner Design of the Women s Health Initiative Clinial Trial and Observational Study. The Women s Health Initiative Study Group. Control Clin Trials. 1998;19(1): Folsom AR, Mink PJ, Sellers TA, et al. Hormonal replaement therapy and morbidity and mortality in a prospetive study of postmenopausal women. Am J Publi Health. 1995;85(8 pt 1): Hannaford P, Elliott A. Use of exogenous hormones by women and oloretal aner: evidene from the Royal College of General Pratitioners Oral Contraeption Study. Contraeption. 2005;71(2): Purdue MP, Mink PJ, Hartge P, et al. Hormone replaement therapy, reprodutive history, and oloretal adenomas: data from the Prostate, Lung, Coloretal and Ovarian (PLCO) Caner Sreening Trial (United States). Caner Causes Control. 2005;16(8): Chlebowski RT, Watawski-Wende J, Ritenbaugh C, et al. Estrogen plus progestin and oloretal aner in postmenopausal women. N Engl J Med. 2004;350(10): Prentie RL, Pettinger M, Beresford SA, et al. Coloretal aner in relation to postmenopausal estrogen and estrogen plus progestin in the Women s Health Initiative Clinial Trial and Observational Study. Caner Epidemiol Biomarkers Prev. 2009;18(5): Newomb PA, Taylor JO, Trentham-Dietz A. Interations of familial and hormonal risk fators for large bowel aner in women. Int J Epidemiol. 1999;28(4): Paskett ED, Reeves KW, Rohan TE, et al. Assoiation between igarette smoking and oloretal aner in the Women s Health Initiative. J Natl Caner Inst. 2007;99(22): Enmark E, Pelto-Huikko M, Grandien K, et al. Human estrogen reeptor beta-gene struture, hromosomal loalization, and expression pattern. J Clin Endorinol Metab. 1997; 82(12): Qiu Y, Waters CE, Lewis AE, et al. Oestrogen-indued apoptosis in olonoytes expressing oestrogen reeptor beta. J Endorinol. 2002;174(3): Qiu Y, Langman MJ, Eggo MC. Targets of 17beta-oestradiolindued apoptosis in olon aner ells: a mehanism for the protetive effets of hormone replaement therapy? J Endorinol. 2004;181(2): Jaobson BC, Moy B, Colditz GA, et al. Postmenopausal hormone use and symptoms of gastroesophageal reflux. Arh Intern Med. 2008;168(16):