Is there a sex effect in colon cancer ? Disease characteristics , management , and outcomes in routine clinical practice

Introduction The incidence of colon cancer varies by sex. Whether women and men show differences in extent of disease, treatment, and outcomes is not well described. We used a large population-based cohort to evaluate sex differences in colon cancer. Methods Using the Ontario Cancer Registry, all cases of colon cancer treated with surgery in Ontario during 2002–2008 were identified. Electronic records of treatment identified use of surgery and adjuvant chemotherapy. Pathology reports for a random 25% sample of all cases were obtained, and disease characteristics, treatment, and outcomes in women and men were compared. A Cox proportional hazards model was used to identify factors associated with overall (os) and cancer-specific survival (css). Results The study population included 7249 patients who underwent resection of colon cancer; 49% (n = 3556) were women. Stage of disease and histologic grade did not vary by sex. Compared with men, women were more likely to have right-sided disease (55% vs. 44%, p ≤ 0.001). Surgical procedure and lymph node yield did not differ by sex. Adjuvant chemotherapy was delivered to 18% of patients with stage ii and 64% of patients with stage iii disease; when adjusted for patientand disease-related factors, use of adjuvant chemotherapy was similar for women and men [relative risk: 0.99; 95% confidence interval (ci): 0.94 to 1.03]. Adjusted analyses demonstrated that os [hazard ratio (hr): 0.80; 95% ci: 0.75 to 0.86] and css (hr: 0.82; 95% ci: 0.76 to 0.90) were superior for women compared with men. Conclusions Long-term survival after colon cancer is significantly better for women than for men, which is not explained by any substantial differences in extent of disease or treatment delivered.


INTRODUCTION
Colorectal cancer (crc) is the 3rd most common cancer in Canada, with an estimated incidence of 25,100 cases in 2015.After lung cancer, crc is the 2nd most common cause of cancer death, with an estimated 9300 deaths annually in Canada 1 .Variation by sex in the incidence and outcomes of cancer is well described 2 .Male sex has been associated with a higher crc incidence 3 and worse prognosis [4][5][6] .The reasons for those associations are not well described, but might be related to differences in tumour location 3,7,8 , stage at presentation 9 , and estrogen exposure [10][11][12] .Women are older at presentation 10,13 , present with lower-stage disease 13 , and are more likely to undergo curative surgical resection 13 .Furthermore, existing studies suggest that, compared with men, women with colon cancer experience better cancer-specific survival 3,9,13,14 .
Although those results are intriguing, the literature on this topic is limited to institution-based studies with small sample sizes that describe cases from the 1980s and 1990s.To our knowledge, no large population-based studies have evaluated the sex effect in crc in the contemporary era.To address this important gap in the literature, we undertook a population-based study to explore sex differences in disease characteristics, management, and outcomes for patients with colon cancer in Ontario.

METHODS Study Design and Population
This population-based retrospective cohort study considered patients treated for colon cancer in the Canadian province of Ontario.Ontario has a population of approximately 13.5 million people and a single-payer universal health insurance program.The study population included a 25% random sample of all patients who underwent resection of primary colon cancer in Ontario during 2002-2008.Using the Ontario Cancer Registry (ocr), we identified all incident cases of colon cancer in Ontario diagnosed during 2000-2008.The ocr does not capture stage of disease for all patients; we therefore obtained surgical pathology reports for a random sample of 25% of the identified patients.Patients were randomly selected using a computer-generated random sample algorithm.Reports were not available for patients undergoing surgery in 2005; the study cohort was therefore restricted to patients who underwent surgery during 2002-2004 and 2006-2008.The study was approved by the Research Ethics Board of Queen's University.

Data Sources and Linkage
The ocr is a passive population-based cancer registry that captures diagnostic and demographic information for at least 98% of all incident cases of cancer in the province of Ontario 15 .The ocr also provides information about vital status and cause of death.Records of hospitalization from the Canadian Institute for Health Information provided information about surgical procedures; those records are known to have a very high level of completeness for crc surgery 16 .Provincial physician billing records from the Ontario Health Insurance Plan, treatment records from regional cancer centres, and provincial records of chemotherapy delivery were used to identify chemotherapy use.A team of trained data abstractors reviewed the pathology reports and entered information about extent of disease into an electronic database.Patients with evidence of metastatic disease (that is, liver, omentum, peritoneum, ovary)-based on the pathology report from the primary tumour resection-were classified as stage iv.The fore going datasets were linked using unique encoded identifiers and were analyzed at the Institute for Clinical Evaluative Sciences.

Exposures and Outcomes
Indicators of the socioeconomic status of the community in which a patient resided at diagnosis were linked as described previously 17 .Quintiles for median household income were based on the household income distribution for the full province of Ontario, with Q1 representing the communities in which the poorest 20% of the Ontario population reside.Geographic regions reflect the catchment areas for Ontario's regional cancer centres 17 .Comorbidity was classified using the Charlson index (modified for administrative data) and was based on all non-cancer diagnoses recorded during any hospital admission within the 5 years preceding the crc surgery 18 .
Adjuvant chemotherapy (actx) was defined as chemotherapy initiated within 16 weeks after surgery.Based on the pathology report, disease resection laterality was defined as right-sided (cecum, ascending colon, hepatic flexure, transverse colon) or left-sided (splenic flexure, descending colon, sigmoid colon, rectosigmoid colon).
Overall (os) and cancer-specific survival (css) were determined from the date of primary tumour resection.To account for the potential of cause-of-death miscoding, css included death from any cancer.Complete information about vital status in the ocr was available up to 31 December 2012; cause of death was available up to 31 December 2010.

Statistical Analysis
Comparisons of proportions between study groups were made using the chi-square test.Survival was determined using the Kaplan-Meier method.The associations of patient-, disease-, and treatment-related factors with os and css were evaluated using Cox proportional hazards regression models.A priori, we undertook subgroup analyses to explore whether the management and outcomes of stage iii colon cancer varied by patient sex.To explore any effect modification by actx on the association of sex with survival, we tested for interactions between actx and sex in the Cox proportional hazards regression models.Results were considered statistically significant at p < 0.05.All analyses were performed using the SAS software application (version 9.3: SAS Institute, Cary, NC, U.S.A.).

Study Population
The linked administrative datasets identified 25,613 potentially eligible patients who underwent resection of colon cancer during 2002-2008 (Figure 1).Surgical pathology reports were reviewed for 7519 randomly selected patients.Age, sex, comorbidity, and survival for the randomly selected patients did not differ significantly from those for the 18,094 patients not selected (Table i and Figure 2).
Of the 7519 randomly selected patients, 270 (4%) were excluded, as shown in Figure 1.Accordingly, the study population included 7249 patients.Table ii shows the characteristics of the study population.Median age in the cohort was 71 years, and 51% were men.More women than men were more than 80 years of age (29% vs. 18%, p < 0.001).Overall stage distribution was 18% stage i, 36% stage ii, 39% stage iii, and 6% stage iv.

Extent of Disease
Women were more likely than men to have right-sided disease (55% vs. 44%, p < 0.001; Table ii).Histologic grade and presence of lymphovascular invasion did not differ by sex.Women were slightly more likely than men to have T4 tumours (22% vs. 18%, p < 0.001).Nodal status and overall stage of disease did not differ substantially between women and men.

Treatment Delivery
Surgical approach (open vs. laparoscopic) and lymph node yield did not differ by sex ( Adjuvant chemotherapy was delivered to 18% of patients with stage ii disease and to 64% of patients with stage iii disease.Median time from surgery to initiation of actx was 7 weeks and did not differ by sex.In univariate analysis, women were slightly less likely than men to receive actx for stage ii and stage iii disease (16% vs. 19%, p=0.051, and 61% vs. 67%, p < 0.001, respectively).However, those differences did not persist when results were adjusted for age, comorbidity, socioeconomic status, and stage of disease [relative risk: 0.99; 95% confidence interval (ci): 0.94 to 1.03; p = 0.562].

Outcomes
Table iv shows patient outcomes overall and for the stage iii subgroup.Length of stay (median: 8 days) and postoperative mortality at 30 and 90 days did not differ by sex.For the cohort overall and for those with stage iii disease, long-term survival (os and css) was superior for women compared with men.
Tables v and vi show factors associated with os and css.Adjusted overall os [hazard ratio (hr): 0.80; 95% ci: 0.75 to 0.86] and css (hr: 0.82; 95% ci: 0.76 to 0.90) were superior for women compared with men.Similar results were observed for patients with stage iii disease (os hr: 0.83; 95% ci: 0.75 to 0.92; css hr: 0.86; 95% ci: 0.77 to 0.97).Other factors associated with long-term survival included age, comorbidity, extent of disease, lymphovascular invasion, and grade.Adjuvant chemotherapy was associated with improved survival in the overall cohort and in the stage iii subgroup.In the stage iii subgroup, a test for interaction between actx and sex suggested no differential effect of actx on survival by sex (os p = 0.400, css p = 0.939).
a Quintile 1 represents patients from the poorest communities in Ontario.Data were not available for 98 patients (0.4%).CCI = Charlson comorbidity index.

DISCUSSION
This population-based study evaluated the extent to which disease characteristics, management, and outcomes in colon cancer vary by sex.Several important findings emerged.First, although women were more likely than men to have right-sided disease, no substantial differences in overall disease stage or grade were evident.Second, general aspects of management, including type of surgery, lymph node yield, and use of actx did not vary by sex.Third, short-term postoperative outcomes in women and men were comparable.Finally, despite minimal observed differences in extent of disease or management, our results demonstrated a substantial sex difference in long-term survival.
Several earlier studies have investigated sex differences in disease characteristics and treatment in crc.McArdle and colleagues 13 evaluated sex differences in 3200 patients with crc treated at 11 institutions in Scotland during 1991-1994.In their analysis, the authors found that women were older, more likely to have right-sided tumours, and more likely to be diagnosed emergently.They also found  that, compared with men, women were less likely to have metastatic disease at the time of diagnosis and more likely to undergo curative-intent surgery.Amri et al. 9 evaluated sex differences in colon cancer in 1071 patients treated with curative-intent surgery at Massachusetts General Hospital during 2004-2011.In that study, the authors found higher-grade disease, more advanced stage, and lower rates of radical resection for women than for men.Those two institutional studies reported differences in stage of cancer between women and men that were not observed in our large population-based study.Several other reports have consistently found that, compared with men, women are more likely to develop right-sided colon cancer 8,13,19 .Although the cause of this difference in laterality of disease is unclear, it might have implications in terms of disease stage at presentation and outcomes, with right-sided disease presenting at a later stage-often because of the lack of early symptoms and the greater challenge of diagnosis through fecal occult blood testing 13 .That association between laterality of disease and overall survival has previously been evaluated.Loupakis et al. 20 recently pooled data for 2027 patients enrolled in three clinical trials of first-line chemotherapy for metastatic crc.They found that right-sided tumours were independently associated with inferior survival.In a study using the Surveillance, Epidemiology, and End Results Program database in the United States, Weiss et al. 19 evaluated the prognostic significance of laterality in 53,801 patients with resected stages i-iii colon cancer.They similarly found that women were more likely to have right-sided disease; however, in their adjusted analysis, laterality across all cases was not associated with survival.In our analysis, we found that right-sided colon cancer was not associated with inferior survival.It is therefore possible that right-sided disease could indeed be associated with a worse prognosis, but that the effect could be offset by unmeasured biologic factors associated with female sex that confer improved survival.
The most striking finding of our study is the observation that female sex is independently associated with   a Because of missing data, 6 patients were excluded from the socioeconomic status and geographic region analyses, and because of missing or ambiguous data, 149 patients were excluded from the laterality analysis.b Quintile 1 represents patients from the poorest communities in Ontario.HR = hazard ratio; CI = confidence interval; CCI = Charlson comorbidity index.
improved survival in colon cancer-an observation that has emerged in several other institution-based studies.The eurocare ii study evaluated the prognostic role of sex for cancer survival in 1,188,469 patients from European registries treated during 1985-1989 2 .Overall, female sex was found to be independently associated with improved survival; however, there was no observed relationship between sex and outcomes in colon cancer specifically.
The superior survival of women observed in our study is consistent with most studies to date that have considered sex effects in colon cancer.In a single-centre study of 894 patients in Germany, Wichmann et al. 14 reported significantly longer os (57.8 months vs. 52 months, p < 0.05) and disease-free survival (51.6 months vs. 46 months, p < 0.05) in women than in men.Similarly, in the previously described multi-institution study from Scotland, McArdle et al. 13 found that female sex was independently associated with improved css (hr: 0.84; 95% ci: 0.73 to 0.98).
Although our study strongly suggests improved survival for women, the results do not provide insight into the causal mechanisms of that finding.Potentially contributing factors include earlier detection in women, resulting in more timely treatment of less-advanced disease.However, if improved survival were driven by earlier detection, we would expect more women to present with lower-stage disease, which is not reflected in our cohort.Further contributing factors might include differences in confounding risk factors that result in more favourable outcomes in women than in men and that were not captured in our cohort, as well as culturally-mediated factors influencing illness behaviours in men and women, which might lead to differences in their use of available health care services.It is also notable that women are more likely than men to have right-sided colon cancer.Right-sided tumours are more likely to have microsatellite instability, which is known to be associated with improved outcomes in colon cancer 21 .Unfortunately, our dataset lacked details about microsatellite instability, and so we were unable to evaluate the extent to which that factor might explain the differential survival between women and men.
Another potential mechanism is the concept that immunologic differences between women and men might influence cancer outcomes.Disease progression in cancer patients is known to be influenced by the host inflammatory response and the immune system 22 , with an elevated systemic inflammatory response being associated with worse outcome independent of tumour stage 23,24 .It has been proposed that female sex hormones might exert a protective immunologic effect on the inflammatory response.In support of that hypothesis, crc progression has been linked to the presence of estrogen receptors in colonic mucosa 10 .Furthermore, several studies have shown that oral contraceptive use and hormone replacement therapy have a protective effect for both colon and rectal cancer 11,12 .Further supporting evidence comes from the Women's Health Initiative trial of estrogen plus progestin in postmenopausal women.Although that trial identified more risks than benefits for women in the hormone group, the use of estrogen plus progestin was associated with a significant decrease in the risk of crc 25 , and the authors concluded that the short-term use of estrogen plus progestin was associated with a decreased risk of crc.However, it is important to note that although invasive crcs in the hormone group were similar in histologic features and grade to those in the placebo group, women in the hormone replacement group were diagnosed with crc at a more advanced stage 25 .In a similar study by Cao et al. 26 , women receiving hormone replacement therapy had a 20% decreased risk of developing crc.They proposed that the presence of female sex steroid hormones might be associated with the expression of estrogen receptors on mesothelioma cells, which might induce cell growth arrest, thereby slowing disease progression.
If estrogen levels confer improved survival in patients with colon cancer, it might be expected that the differential survival between women and men would be restricted to younger patients.In a post-hoc exploratory analysis, we evaluated for potential interactions between age, sex, and survival.That analysis did not support the premise that the difference in survival between women and men is associated with age (p = 0.725 for os; p = 0.432 for css).Further exploratory analyses included a stratified Cox model for women less than 50 years of age and one for women 50 years of age and older.Contrary to our hypothesis, we found that the sex difference in survival remained significant for postmenopausal women and was no longer statistically significant for younger women.The latter finding might simply relate to limited statistical power, given that only 442 patients were less than 50 years of age.
To our knowledge, our study is the most contemporary and the only population-based one to examine differences in disease characteristics, treatment, and survival between women and men with colon cancer.Several study limitations merit comment.As with all observational studies, the results could be biased by unmeasured prognostic factors, and we were only partly able to control for known variables such as comorbidities.In addition, disease stage is based solely on the pathology report at the time of surgical resection of the primary tumour, with stage iv being defined as omental or liver metastases resected at the time of primary tumour resection.It is therefore possible that some patients classified as having stage i, ii, or iii disease might have been understaged if their metastatic disease was not resected at the time of primary colonic resection.However, the proportion of such cases is likely to be relatively low, and there is no reason to believe that it would differ by sex.Despite those limitations, the very large sample size and, particularly, the random sample of all registry cases of colon cancer treated with curative intent-and therefore unselected-is a major strength of the present study; referral and selection biases that affect traditional institution-based observational studies are thus minimized 27,28 .

CONCLUSIONS
The present population-based study demonstrates that long-term survival is substantially higher in women than in men with colon cancer, but it does not suggest an explanation, given a lack of any substantial differences in extent of disease or treatment.

ACKNOWLEDGMENTS
Parts of this material are based on data and information provided by Cancer Care Ontario; however, the analyses, conclusions, opinions, and statements expressed herein are those of the authors and not necessarily those of Cancer Care Ontario.
Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information (cihi); however, the analyses, conclusions, opinions, and statements expressed herein are those of the authors and not necessarily those of cihi.
This study was supported by the Institute for Clinical Evaluative Sciences (ices), which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (mohltc).The opinions, results, and conclusions reported in this paper are those of the authors and are independent from the funding sources.No endorsement by ices or mohltc is intended or should be inferred.
CMB is supported as a Canada Research Chair in Population Cancer Care.This work was also supported by the Canada Foundation for Innovation and the Canadian Institutes of Health Research.

CONFLICT OF INTEREST DISCLOSURES
We have read and understood Current Oncology's policy on disclosing conflicts of interest, and we declare that we have none.

AUTHOR AFFILIATIONS
*Division of Cancer Care and Epidemiology, Queen's University Cancer Research Institute, and the departments of † Oncology, ‡ Surgery, § Medicine, and || Public Health Sciences, Queen's University, Kingston, ON.

FIGURE 1
FIGURE 1 Identification of patients with resected colon cancer in Ontario, 2002-2008.a Exclusions were histology other than adenocarcinoma, rectal cancer, inconsistency between date of surgical resection and date of pathology report, and use of preoperative chemotherapy.

FIGURE 2
FIGURE 2 Cancer-specific survival curves for patients with and without randomly selected pathology reports treated by surgical resection for colon cancer in Ontario, 2002-2008.

TABLE I
Characteristics of 25,613 patients with and without randomly selected pathology reports treated with surgical resection for colon cancer inOntario, 2002-2008

TABLE II
Patient and pathologic characteristics in 7249 cases of resected colon cancer inOntario, 2002-2008 a Quintile 1 represents the communities in which the poorest 20% of the Ontario population resided.Data were not available for 21 patients.b T ≤ 1 = pTX, or pT0, or pTis or pT1.CCI = Charlson comorbidity index.

TABLE III
Surgical characteristics and chemotherapy delivery in 7249 patients with resected colon cancer inOntario, 2002-2008

TABLE IV
Outcomes in patients with resected colon cancer in Ontario, 2002-2008, by sex a Covariates were age, socioeconomic status, score on the Charlson comorbidity index, stage, lymphovascular invasion, grade, and adjuvant chemotherapy.b Covariates were age, socioeconomic status, score on the Charlson comorbidity index, T stage, N status, lymphovascular invasion, grade, and adjuvant chemotherapy.CI = confidence interval; OS = overall survival; CSS cancer-specific survival.e20 Current Oncology, Vol. 24, No. 1, February 2017 © 2017 Multimed Inc.

TABLE V
Variables associated with cancer-specific survival and overall survival in 7249 patients a with resected colon cancer in Ontario, 2002-2008 SEX EFFECT AND COLON CANCER, Quirt et al.
a Because of missing data, 21 patients were excluded from the socioeconomic status analysis and 378 patients were excluded from the laterality analysis.b Quintile 1 represents patients from the poorest communities in Ontario.HR = hazard ratio; CI = confidence interval; CCI = Charlson comorbidity index.e21 Current Oncology, Vol. 24, No. 1, February 2017 © 2017 Multimed Inc.

TABLE VI
Variables associated with cancer-specific survival (CSS) and overall survival (OS) in 2920 patients a with resected stage III colon cancer inOntario, 2002-2008