Education level and survival after esophageal cancer surgery: a prospective population-based cohort study For peer review only

Transcription

1 Education level and survival after esophageal cancer surgery: a prospective population-based cohort study Journal: BMJ Open Manuscript ID: bmjopen-0-00 Article Type: Research Date Submitted by the Author: 0-Aug-0 Complete List of Authors: Brusselaers, Nele; Karolinska Institutet, Upper Gastrointestinal Research, Department of Molecular Medicine and Surgery Ljung, Rickard; Karolinska Institutet, Upper Gastrointestinal Research, Department of Molecular Medicine and Surgery Mattsson, Fredrik; Karolinska Institutet, Upper Gastrointestinal Research, Department of Molecular Medicine and Surgery Johar, Asif; Karolinska Institutet, Upper Gastrointestinal Research, Department of Molecular Medicine and Surgery Wikman, Anna; Karolinska Institutet, Upper Gastrointestinal Research, Department of Molecular Medicine and Surgery Lagergren, Pernilla; Karolinska Institutet, Upper Gastrointestinal Research, Department of Molecular Medicine and Surgery Lagergren, Jesper; Karolinska Institutet, Upper Gastrointestinal Research, Department of Molecular Medicine and Surgery; King's College London, Division of Cancer Studies <b>primary Subject Heading</b>: Gastroenterology and hepatology Secondary Subject Heading: Epidemiology, Oncology, Surgery Keywords: Gastrointestinal tumours < GASTROENTEROLOGY, Oesophageal disease < GASTROENTEROLOGY, Epidemiology < ONCOLOGY, Gastrointestinal tumours < ONCOLOGY BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

2 Page of BMJ Open Education level and survival after esophageal cancer surgery: a prospective population-based cohort study Authors: Nele BRUSSELAERS MD Phd, Rickard LJUNG MD PhD, Fredrik MATTSSON BSc, Asif JOHAR MSc, Anna WIKMAN PhD, Pernilla LAGERGREN PhD, Jesper LAGERGREN MD PhD,. Affiliations: Unit of Upper Gastrointestinal Research, Department of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden. Division of Cancer Studies, King s College London, London, United Kingdom. Corresponding author: Dr. Nele Brusselaers, Unit of Upper Gastrointestinal Research, Department of Molecular medicine and Surgery, Norra Stationsgatan, Level, Karolinska Institutet, SE- Stockholm, Sweden. Telephone: + (0) 0. Fax: + (0) 0. nele.brusselaers@ki.se Key words (MeSH terms): esophageal neoplasms; oesophagus; esophagectomy; education; educational status; socio-economic factors; health status disparities; Sweden. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

3 Page of Abstract (word count: ) Objectives: This study aimed to investigate if a higher education level is associated with improved long-term survival after esophagectomy for cancer. Design: A prospective, population-based cohort study Setting: Sweden - nationwide Participants: 0% of all esophageal and cardia cancer patients who underwent resection in Sweden in 00-00, were enrolled in this study (N=00; 0.% male) and followed-up until death or end of the study period (0). The study exposure was level of education, defined as compulsory ( years), moderate (0- years), or high ( years). Outcome measures: The main outcome measure was overall -year survival after esophagectomy. Cox regression was used to estimate associations between education level and mortality, expressed as hazard ratios (HRs) with % confidence intervals (CIs), with adjustment for sex, age, tumour stage, histological type, complications, comorbidities and annual surgeon volume. High education was the reference category. Results: Among 00 included patients, (.%) had compulsory education, (.%) moderate education, and (.%) had high education. The overall -year survival rate was.%,.%, and.% among patients with compulsory, moderate and high education, respectively. After adjustment for confounders, a slightly higher, yet not statistically significantly increased point HR was found among compulsory educated patients (HR.0, % CI 0.0-.). In patients with tumour stage IV, increased adjusted HRs were found for compulsory (HR., % CI.0-.) and moderately (HR., % CI.-.) educated patients. No statistically significant associations were found for the other tumour stages. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

4 Page of BMJ Open Conclusions: This study provides limited evidence of an association between lower education and worse long-term survival after esophagectomy for cancer. Article Summary Article focus - Level of education is one of the most important determinants of socio-economic differences. - Higher education and socio-economic status have been associated with better cancer prognosis for several other types of cancer. - This study evaluates the association between educational level and long-term survival after esophageal resection for cancer. Key messages - Only limited evidence for an association was found except for patients with advanced tumours. - A higher education level was significantly associated with improved survival in patients with oesophageal cancer and tumour stage. Strengths and limitations - Strength: nationwide population-based design with high participation rate and completeness of the follow-up, consequently reducing the risk for selection bias; and robust definitions of exposure and outcome reducing the risk of misclassification. - Weakness: limited statistical power to detect weak or moderate differences, particularly in the stratified analyses BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

5 Page of Introduction Esophageal cancer is characterized by an increasing incidence in many Western populations, and the -year overall survival is less than % in Europe. Low socioeconomic position is a risk factor for both main histological types of esophageal cancer, i.e. adenocarcinoma and squamous cell carcinoma. The role of the socio-economic position in the development of esophageal carcinoma can only partly be explained by the main risk factors, i.e. gastroesophageal reflux and obesity for adenocarcinoma, and tobacco smoking and heavy alcohol intake for squamous cell carcinoma. A recent study indicated that the mortality rates for esophageal cancer were lower in patients with a higher education level. Education has been studied more extensively in relation to survival of other types of cancer, -0 showing a beneficial effect of higher education on survival that might be explained by differences in comorbidity burden, lifestyle factors, health awareness, adherence to treatment and health care seeking behaviour; factors which are also likely to influence timing of referral and tumour stage at diagnosis. -0 The impact of education on survival after esophageal cancer diagnosis has been examined in cohort studies, but no clear associations were found. Two studies have evaluated the influence of the socio-economic position on survival after esophagectomy for cancer, where one showed no long-term survival advantage of higher socio-economic position, while the other study showed a better short-term survival. Thus, the prognostic role of socio-economic position in esophageal cancer patients remains uncertain. A better insight into this association might improve pre-operative decision-making and information, and post-operative management, survival and health-related quality of life. The objective of this study was to clarify the influence of education level on the overall and disease-specific survival after curatively intended esophageal cancer surgery using a nationwide population-based design with long-term follow-up. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

6 Page of BMJ Open Patients and methods Study design and data sources This was a nationwide Swedish prospective, population-based cohort study. All surgically treated esophageal or cardia cancer patients operated in Sweden during the period April 00 and December 00 were eligible for the study. Follow-up for survival ended in August 0. Data sources In September 0, the following Swedish registers were linked: The Swedish Esophageal and Cardia Cancer database (SECC): SECC was used to identify the study cohort and assess data on clinical factors. This database is our all-encompassing, nationwide Swedish research database, including 0% of all esophageal and cardia cancer patients treated with surgery in Sweden during the inclusion period. The organization of the nationwide network of clinicians and the design of this study have been described in detail elsewhere. In brief, SECC contains prospectively collected and detailed information about tumour pathology and localization, tumour stage, surgical procedures and complications. The Swedish Patient Registry: The Patient Register comprises information on all in-hospital care and outpatient specialist care in Sweden, including codes for diagnoses and surgical procedures. It has a complete nationwide coverage of inpatient data since, and complete outpatient specialist care data since 00. The validity of esophageal cancer surgery in the Patient Registry has recently been assessed. Among patients with a code representing esophageal resection in the Patient Registry in -00, the positive predictive value was.%. Data on comorbidity at the time of surgery was collected from this registry. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

7 Page of The National Education Registry: This registry was established by Statistics Sweden in, and is annually updated with information on the highest formal education attained by each individual, from elementary to postgraduate level. Data on the study exposure, education level was collected from this registry. Swedish Causes of Death Registry: This registry contains information on date of death for all deceased Swedish residents since, and has a.% completeness of cause-specific death. Data on mortality was collected from this registry. Study exposure The study exposure was the highest attained education level (partially or fully completed) at the time of surgery, and was classified into three categories based on the Swedish National School Administration and Statistics Sweden: ) compulsory education or years: primary and lower secondary education (up to the age of years), ) moderate education or 0- years: upper secondary education (standard is years), ) high education or years: post-secondary education. Study outcomes The outcomes were: ) overall mortality up to years after esophagectomy for esophageal or cardia cancer; the main outcome measure, ) conditional mortality, defined as mortality within years of surgery after exclusion of the first 0 days (short-term mortality) after surgery, ) disease-specific mortality, representing all deaths with esophageal or cardia cancer as an underlying or contributing cause of death within years of surgery, ) conditional and disease-specific -year mortality, and ) short-term mortality (within 0 days of surgery), which was analysed for completeness. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

8 Page of BMJ Open Statistical analyses Cox regression was used to assess the association between level of education and mortality, expressed as hazard ratios (HRs) with % confidence intervals (% CIs). The patient group in the highest category of education was used as the reference group. Three regression models were used: ) a crude model without any adjustments, ) a second model adjusted for sex, age (<0, 0-, or years) and tumour stage (0-I, II, III, or IV); and ) a fully adjusted model further adjusted for histological type (adenocarcinoma or squamous cell carcinoma), number of major complications (0,, or >), number of comorbidities (0,, or >) and surgeon volume (< or esophagectomies per year). Tumour stage was categorized according to the TNM classification presented by the Union Internationale Contre le Cancer. 0 Pre-defined major medical and surgical complications which occurred within 0 days of surgery were extracted from the SECC database and included the following: anastomotic leakage, serious infections, respiratory insufficiency, cardiac failure, renal or liver failure, technical complications, damage to the recurrent laryngeal nerve or the thoracic duct, early anastomotic stricture, and others (embolus, deep venous thrombosis, rupture of the wound, intestinal obstruction, myocardial infarction, or stroke, all with a need for intervention). Data on comorbidity present at the time of surgery was extracted from the patient register, and included the following: diabetes mellitus, cardiovascular disease, pulmonary disease, hepatic or renal disease, renal failure and other cancer. Comorbidities within the same group (e.g. two different cardiovascular diseases) were counted only once. Surgeon volume was categorized into two equally sized groups based on the median number of operations per surgeon and year. To assess effect modification, stratified survival analyses were performed for tumour stage and histological type. The same three regression models were used in the stratified analyses, but without adjustment for the stratifying variable. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

9 Page of Ethical considerations The study was approved by the Regional Ethical Review Board in Stockholm, Sweden. Informed consent was obtained from each patient before inclusion in the study. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

10 Page of BMJ Open Results Patients From the study cohort, patients were originally identified, which corresponded to 0% of all eligible patients in Sweden who underwent esophageal or cardia cancer surgery during the inclusion period. Of these, patients were excluded from the analyses because of missing values for education (N=), tumour stage (N=), or histology (N=), leaving 00 patients for final analysis. Some characteristics of the study patients are described in Table. Of all patients, (.%) had compulsory education, (.%) had moderate education, and (.%) had high education. Compared to the compulsory educated group, patients in the higher educated groups were younger, but the sex distribution was similar in all groups (approximately 0% male). Distribution of tumour stage was comparable among the three education groups, while the proportion of squamous cell carcinoma was slightly higher in the compulsory educated group compared with the highly educated group (% versus %). The group with high education had a lower frequency of comorbidity compared to the compulsory educated group (% versus %). The proportions of patients operated on by surgeons with high or low annual volume were similar between education groups. No major differences in the number of post-operative complications were found between educational levels. In total, (.%) patients died within years of surgery, of whom (.%) had esophageal or cardia cancer as an underlying or contributing cause of death. Education level and mortality All tumour stages The frequencies of overall -year mortality, conditional -year mortality and short-term mortality were highest in the compulsory educated group, closely followed by the moderately educated group, and lowest in patients with high education (Table ). The overall -year BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

11 Page 0 of survival rates were.%,.%, and.% among patients with compulsory, moderate and high education, respectively. Table presents the HRs for mortality after esophagectomy according to the education level of all included patients. The crude model showed increased overall point HRs in the compulsory (HR.0, % CI 0.-.) and moderately educated groups (HR., % CI 0.-.), compared with the reference group (high education), but without statistical significance. The point HRs became slightly attenuated after adjustment in the second model (HR., % CI 0.-.) and the third, fully adjusted model (HR.0, % CI 0.0-.). The HRs for mortality within years of surgery were similar when only disease-specific deaths were considered, as well as when the first 0 days were excluded from the analyses (Table ). Regarding short-term (0 days) mortality, no difference was identified between the compulsory educated and the highly educated in the fully adjusted model (HR 0., % CI 0.-.). Specific tumour stages The tumour stage-specific analyses addressing the overall -year mortality indicated some differences in survival between the education groups (Table ). In tumour stage I, the fully adjusted point HRs for compulsory and moderately educated patients were increased, compared to patients in the highest level of education, but no statistically significant differences were found (HR., % CI 0.-.0, and HR., % CI 0.-., respectively). In tumour stage IV, the fully adjusted HRs were almost -fold increased in compulsory and moderately educated patients, compared to highly educated patients (HR., % CI.0-., and HR., % CI.-.). 0 BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

12 Page of BMJ Open Specific histological types The crude -year overall survival analyses stratified for histological type indicated associations between compulsory education and survival in patients with adenocarcinoma, but the HRs were attenuated after full adjustments (HR., % CI 0.-. and HR., % CI 0.-.)(Table ). The adjusted point HRs for squamous cell carcinoma were also increased among compulsory and moderately educated groups, compared to highly educated, but no statistically significant differences were identified (HR., % CI and HR., % CI 0.-., respectively). BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

13 Page of Discussion This study does not provide much evidence for the hypothesis that education level influences the long-term survival after surgery for esophageal cancer, with the exception of worse survival in compulsory educated groups with advanced tumours (stage IV). Strengths of this study include the nationwide population-based design with high participation rate and completeness of the follow-up, consequently reducing the risk for selection bias. The results have also been adjusted for confounding by known prognostic factors. Moreover, the exposure and outcome measures were robust. A weakness is the limited statistical power to detect weak or moderate differences, particularly in the stratified analyses. Despite the inclusion of the vast majority of all patients operated in Sweden during the nearly -year inclusion period, who it was possible to follow-up for at least years, the low incidence of esophageal cancer in Sweden combined with a low resection rate (%) reduced the sample size. A post-hoc power analysis showed that if the HR would have been. or higher, the power for this study cohort would have been at least 0%. An adjusted HR of. (Table ) would e.g. require a sample size of,00 patients in the study cohort assuming similar conditions as those of the present study. Therefore, even larger studies are needed to statistically verify the level of potential associations indicated in the present study. Since the inclusion period was limited, differences in therapeutic management are unlikely to have influenced the survival. The multi-centre design, where several hospitals and surgeons were involved in the treatment of these patients, might have influenced the survival, but it is unlikely that the referral patterns would be different between education groups. Finally, confounding by other factors than those adjusted for could influence the results. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

14 Page of BMJ Open An interesting question is why differences in education in esophageal cancer seems to be a less strong prognostic factor compared to other cancer types. -0 A British study compared the deprivation gap, or the percentage difference in -year survival between the most affluent and most deprived patients in different cancer types, and showed major differences for cancer of the larynx (%), colon (-%), rectum (-%), prostate (%), bladder (%), and breast (%), but not for the oesophagus (in men.%, in women 0.%), stomach (.%), pancreas (%), or ovary (%). These results, supported by the findings of the present study, may suggest that the possible impact of the socio-economic position might be restricted by the high overall mortality rates in esophageal cancer. The reasons for the differences in survival between compulsory and highly educated patients with advanced tumour (stage IV) are unclear, and we can only speculate. Highly educated patients with advanced disease might receive (and possibly also request) a more aggressive pre- and post-operative treatment, which might result in a longer average survival period that would influence the HR for overall mortality. These findings are supported by a recent study, which showed an increased survival for patients with higher income levels, and a correlation between higher income and receiving curative treatment (including surgery and radiotherapy). Even a small increase in the length of survival in our group of patients with high education could be responsible for these significant results. Yet, although the prognosis is poor for patients with tumour stage, these differences are not purely due to differences in short-term survival, since.% of these patients survive at least years after surgery. The clinical implications of this study are limited; since the overall results suggest that the impact of education on survival after esophagectomy could be low or even absent. Yet, it could be discussed that more effort should be placed on the lower educated patients, as they BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

15 Page of could have more risk factors for mortality such as increased comorbidity, as shown in the present study. Nevertheless, continuous efforts are needed to guarantee optimal treatment for all patients, including patients with suspected low compliance to treatment, follow-up and lifestyle recommendations. To conclude, this population-based, nationwide and prospective cohort study provides limited evidence of education being a prognostic factor after esophagectomy for cancer. However, the increased mortality in lower educated groups with tumour stage IV warrants further attention. In addition, the generally increased point risk estimates combined with a limited statistical power indicate a need for studies with very large sample sizes that can allow for detailed stratification of the analyses. Funding statement: This work was supported by the Swedish Research Council (SIMSAM) and Swedish Cancer Society. The funding sources had no role in the design and conduct of the study, collection, management, analysis and interpretation of the data, or preparation review or approval of the manuscript. Competing interest statement: No competing interests. Author contributions: All authors contributed equally to the study design and study protocol. PL and JL were responsible for the data collection. FM and NB performed the statistical analyses and interpreted the results. The writing and revisions of the manuscript have been performed by NB. Critical evaluation of the manuscript was performed equally by all authors. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

16 Page of BMJ Open References. Lagergren J, Lagergren P. Oesophageal cancer. BMJ 00;:c0.. Sant M, Allemani C, Santaquilani M, Knijn A, Marchesi F, Capocaccia R. EUROCARE-. Survival of cancer patients diagnosed in -. Results and commentary. Eur J Cancer 00;():-.. Gossage JA, Forshaw MJ, Khan AA, Mak V, Moller H, Mason RC. The effect of economic deprivation on oesophageal and gastric cancer in a UK cancer network. Int J Clin Pract 00;():-.. Jansson C, Johansson AL, Nyren O, Lagergren J. Socioeconomic factors and risk of esophageal adenocarcinoma: a nationwide Swedish case-control study. Cancer Epidemiol Biomarkers Prev 00;():-.. Jemal A, Simard EP, Xu J, Ma J, Anderson RN. Selected cancers with increasing mortality rates by educational attainment in states in the United States, -00. Cancer Causes Control 0.. Cavalli-Bjorkman N, Lambe M, Eaker S, Sandin F, Glimelius B. Differences according to educational level in the management and survival of colorectal cancer in Sweden. Eur J Cancer 0;():-0.. Eloranta S, Lambert PC, Cavalli-Bjorkman N, Andersson TM, Glimelius B, Dickman PW. Does socioeconomic status influence the prospect of cure from colon cancer--a population-based study in Sweden Eur J Cancer 00;():-.. Sorbye H, Pfeiffer P, Cavalli-Bjorkman N, Qvortrup C, Holsen MH, Wentzel-Larsen T, et al. Clinical trial enrollment, patient characteristics, and survival differences in prospectively registered metastatic colorectal cancer patients. Cancer 00;(0):-. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

17 Page of Bentley R, Kavanagh AM, Subramanian SV, Turrell G. Area disadvantage, individual socio-economic position, and premature cancer mortality in Australia to 000: a multilevel analysis. Cancer Causes Control 00;():-. 0. Frederiksen BL, Osler M, Harling H, Ladelund S, Jorgensen T. Do patient characteristics, disease, or treatment explain social inequality in survival from colorectal cancer? Soc Sci Med 00;():0-.. Thrift AP, Nagle CM, Fahey PP, Smithers BM, Watson DI, Whiteman DC. Predictors of survival among patients diagnosed with adenocarcinoma of the esophagus and gastroesophageal junction. Cancer Causes Control 0;():-.. Morgan MA, Lewis WG, Chan DS, Burrows S, Stephens MR, Roberts SA, et al. Influence of socio-economic deprivation on outcomes for patients diagnosed with oesophageal cancer. Scand J Gastroenterol 00;(0):0-.. Rouvelas I, Zeng W, Lindblad M, Viklund P, Ye W, Lagergren J. Survival after surgery for oesophageal cancer: a population-based study. Lancet Oncol 00;():-0.. Leigh Y, Seagroatt V, Goldacre M, McCulloch P. Impact of socio-economic deprivation on death rates after surgery for upper gastrointestinal tract cancer. Br J Cancer 00;():0-.. Lagergren J, Bergstrom R, Lindgren A, Nyren O. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med ;0():-.. Viklund P, Lindblad M, Lu M, Ye W, Johansson J, Lagergren J. Risk factors for complications after esophageal cancer resection: a prospective population-based study in Sweden. Ann Surg 00;():0-.. Lagergren K, Derogar M. Validation of oesophageal cancer surgery data in the Swedish Patient Registry. Acta Oncol 0;():-. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

18 Page of BMJ Open Statistics Sweden: Yearbook of Educational Statistics (Utbildningsstatistisk). In: centralbyran S, editor. Sweden, 00:-.. Johansson LA, Westerling R. Comparing Swedish hospital discharge records with death certificates: implications for mortality statistics. Int J Epidemiol 000;(): Sobin L. Classification of malignant tumours. th ed, 00.. Coupland VH, Lagergren J, Luchtenborg M, Jack RH, Allum W, Holmberg L, et al. Hospital volume, proportion resected and mortality from oesophageal and gastric cancer: a population-based study in England, Gut 0.. Derogar M, Sadr-Azodi O, Johar A, Lagergren P, Lagergren J. Hospital and surgeon volume in relation to survival after esophageal cancer surgery in a population-based study. J Clin Oncology.. Coleman MP, Rachet B, Woods LM, Mitry E, Riga M, Cooper N, et al. Trends and socioeconomic inequalities in cancer survival in England and Wales up to 00. Br J Cancer 00;0():-.. Chen MF, Yang YH, Lai CH, Chen PC, Chen WC. Outcome of Patients with Esophageal Cancer: A Nationwide Analysis. Ann Surg Oncol 0. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

19 Page of Table : Demographic, treatment, and tumour characteristics and mortality after oesophagectomy for cancer, categorized by education level. Level of education a Total years 0- years years Number (%) Number (%) Number (%) Number (%) Total (.%) (.%) (.%) 00 (00.0) Age <0 years (.) (.) (.) (.0) 0- years (.) (.) (.) (.) years (.) (.0) (.) (0.) Sex Male (.0) (.0) 0 (.) (0.) Female (.0) 0 (.0) (.) (.) Tumour stage I (.) (.) (.) (.) II (.) (0.) (.) (.) III (.0) (.) (.) (0.) IV (.) (.) (.0) (.) Histology Adenocarcinoma (.) (.) 0 (.0) (.) Squamous cell carcinoma (.) (.) (.0) (.) Comorbidity None (.) (.) 0 (.) (.0) One (.) (.) (.) 0 (.) More than one (.) (.) (.) (.) Surgical volume Low (< operations/year) (.) (0.0) (.) (.) High ( operations/year) (.) (0.0) (.) (.) Post-operative complications None (.) (.) (.) (.) One 0 (.) (.) (.) (.) More than one (.) (.) (.) 0 (.) Mortality Within 0 days (.) (.0) (.) (.0) Overall within years (.) 0 (.) (.) (.) Conditional within years b 0 (.) (.) (.) 0 (.0) a Level of education: years: post-secondary education; 0- years: upper secondary education; years: compulsory education; b Conditional mortality: excluding first 0 days after surgery. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

20 Page of BMJ Open Table : Multi-variable Cox regression models analysing the association between education level and mortality after oesophagectomy for cancer, expressed as hazard ratios (HR) with % confidence intervals (CI). Level of education a years (Reference) 0- years HR (% CI) years HR (% CI) Overall -year mortality Model. (0.-.).0 (0.-.) Model.0 (0.-.).0 (0.0-.) Model.0 (0.0-.). (0.-.) Conditional overall -year mortality b Model. (0.-.). (0.-.) Model.0 (0.-.).0 (0.-.0) Model.0 (0.-.). (0.-.) Disease-specific -year mortality c Model. (0.-.). (0.-.) Model.0 (0.-.). (0.-.) Model.0 (0.0-.). (0.-.) Conditional b and disease-specific c -year mortality Model. (0.-.). (0.-.) Model.0 (0.-.). (0.-.) Model.0 (0.-.). (0.-.) Overall 0-days mortality Model. (0.-.). (0.-.) Model.0 (0.0-.).00 (0.-.0) Model. (0.-.) 0. (0.-.) Values are expressed as hazard ratios. a Level of education: years: post-secondary education; 0- years: upper secondary education; years: compulsory education. b Conditional mortality: excluding first 0 days after surgery. c Disease-specific mortality: oesophageal or cardia cancer as underlying or contributing cause. Model : unadjusted. Model : adjusted for sex, age, tumour stage. Model : adjusted for sex, age, tumour stage, histology, major complications, comorbidity and surgeon volume. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

21 Page 0 of Table : Multi-variable Cox regression models analysing the association between education level and all-cause -year mortality after oesophagectomy for cancer, stratified by tumour stage, and expressed as hazard ratios (HR) with % confidence intervals (CI). Level of education a Total number (%) years (Reference) 0- years HR (% CI) years HR (% CI) Model Stage I. (0.-.0). (0.-.) (.) Stage II. (0.-.). (0.-.) (.) Stage III 0. (0.-.0).0 (0.-.) (0.) Stage IV. (0.-.). (0.-.) (.) Model Stage I. (0.-.0). (0.-.) (.) Stage II. (0.-.0) 0. (0.-.) (.) Stage III 0. (0.-.) 0. (0.-.) (0.) Stage IV. (0.-.).0 (0.-.0) (.) Model Stage I. (0.-.). (0.-.0) (.) Stage II 0. (0.-.) 0. (0.-.) (.) Stage III 0. (0.-.).0 (0.-.) (0.) Stage IV. (.-.). (.0-.) (.) Values are expressed as hazard ratios or as number of patients (%). a Level of education: years: post-secondary education; 0- years: upper secondary education; years: compulsory education. Model : unadjusted. Model : adjusted for sex and age. Model : adjusted for sex, age, histology, major complications, comorbidity and surgeon volume. 0 BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

22 Page of BMJ Open Table : Multi-variable Cox regression models analysing the association between education level and all-cause -year mortality after oesophagectomy for cancer, stratified by tumour histology, and expressed as hazard ratios (HR) with % confidence intervals (CI). Level of education a Total number (%) years (Reference) 0- years HR (% CI) years HR (% CI) Overall -year mortality / (.) 0/ (.) / (.) /00 (.) Model Adenocarcinoma.0 (0.-.0). (.00-.) (.) Squamous cell carcinoma.00 (0.-.) 0. (0.-.) (.) Model Adenocarcinoma.0 (0.-.).0 (0.-.) (.) Squamous cell carcinoma.0 (0.-.).0 (0.-.) (.) Model Adenocarcinoma. (0.-.). (0.-.) (.) Squamous cell carcinoma. (0.-.). (0.-.0) (.) Values are expressed as hazard ratios or as number of patients (%). a Level of education: years: post-secondary education; 0- years: upper secondary education; years: compulsory education. Model : unadjusted. Model : adjusted for sex and age. Model : adjusted for sex, age, tumour stage, major complications, comorbidity and surgeon volume. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

23 Page of STUDY PROTOCOL Project title: Level of education in relation to -year survival after oesophageal cancer surgery (SECC) Collaborators (in alphabetic order) Anna, Asif, Fredrik, Jesper (last author), Nele (first author), Pernilla, Rickard. Objectives To clarify the potential effects of the socio-economic factor educational level on the overall survival within years after oesophageal cancer surgery, using the SECC database linked to the LISA-registry Background Previous studies have shown that the socio-economic status might affect survival after a cancer diagnosis [-]. Possible explanations may include differences in comorbidity burden, life style, health awareness and health care seeking behaviour []. In some studies, marital status and patient s partner s level of education has influenced choice of, and adherence to, treatment. Timing (referral) and stage at diagnosis may also play a role [,, ]. Only little is known about the specific impact of socio-economic factors on the outcome after oesophageal cancer surgery. Studies have tried to explore determinants of short term survival (in-hospital or 0-days), survival after oesophageal cancer diagnosis, or tried to predict the success of chemotherapy [, -]. Several studies explored determinants for long-term survival after oesophageal cancer surgery [-], but most of these focused on clinical and tumor-related characteristics. Only a minority of these studies included determinants used to describe the socio-economic status of the patient (e.g. ethnicity, marital status, education status, smoking and deprived socio-economic status [,,, 0-]. Taken together, surprisingly little is known about the specific impact of education on survival after oesophageal cancer surgery. A better insight in the prognostic variables may improve pre-operative information, post-operative management, and consequently also survival and quality of life after surgery for oesophageal carcinoma. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

24 Page of BMJ Open Methods Data collection The following databases are linked and will be used: - SECC-database (Swedish Esophageal and Cardia Cancer database): an allencompassing, nationwide research database that includes almost all esophageal and cardia cancer patients in Sweden treated with surgery. SECC contains detailed patient file-based information about tumor pathology and localization, TNM stage, surgical procedures, techniques and complications, and additionally contains repeated health-related quality of life (HRQL) assessments and clinical (prognostic) factors such as BMI. The SECC register was initiated in April 00 and since then until December 00 about 0% of all surgically treated esophageal cancer patients are included. The almost complete national coverage and the detailed prospective data collection and objective manual review of each case ensure a good validity. The patients are all followed up for at least years for survival (up to 0). - LISA-registry (Longitudinell integrationsdatabas för sjukförsäkrings- och arbetsmarknadsstudier): includes socio-economic variables, e.g. educational level, occupation, marital status and income. - The Causes of Death Registry: includes dates and underlying causes of all deaths among persons residing in the country, independently of where they died (including abroad). Only a small number of deaths (0.%) do not have a recorded cause of death. This enables us to study disease-specific mortality. - Patient Registry: will be used to collect data on comorbidities based on discharge diagnoses, surgical procedures, and hospitalisation dates. The Swedish Patient Registry was % complete in and 00% in and onwards. The Patient Registry provides us with possibilities to better adjust the results for comorbidities. Study design This prospective population-based cohort study will be based on the SECC database (N=), of patients who underwent oesophageal resection in April 00 to December 00 and followed up for years. The specific impact of education on long-term survival after surgery will be evaluated. The data from the available registries will be used for each cohort member, including comorbidities, operations, cancer, date of death, cause of death, etc. Study exposures - Socio-economic factors: education = defined by the number of years of education at time of surgery; divided in groups (based on data from 0 and onwards education register-lisa): BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

25 Page of i. Low ( y in total; compulsory education), ii. Middle (0-y), iii. High (y or more; at least y of university college or university). The changes in the educational system shouldn t have influenced these subcategories. If necessary, the middle and high education groups will be combined into one group. Study outcome ) Overall mortality up to years after oesophagectomy for oesophageal or cardia cancer. ) Overall mortality up to years after oesophagectomy for oesophageal or cardia cancer, after exclusion of the first 0 days after surgery. ) Disease-specific mortality up to years after oesophagectomy for oesophageal or cardia cancer. (Possibly not report results in paper, only mention in text) ) Disease-specific mortality up to years after oesophagectomy for oesophageal or cardia cancer, after exclusion of the first 0 days after surgery. (Possibly not report results in paper, only mention in text) ) Potential outcome: 0-days overall mortality after oesophagectomy for oesophageal or cardia cancer (possible power problem) The last surgical resections were performed in 00, so all patients have sufficiently long follow-up. Possible confounders () Tumour stage: based on TNM classification (categorised in groups: 0-I; II and III, IV). Stage III=TN is largest group ±0%), () Comorbidities: combined in groups: none one more than one; or groups if insufficient statistical power; possible comorbidities: diabetes, cardiovascular disease, pulmonary disease, liver disease, renal failure, other cancer. Data should be complete (not reported = none), but patients can have more than co-morbidity. In former SECC studies hypertension and other were also reported, but will not be included in this binary variable. () Age (at surgery): groups: <0y, 0-y, >y or continuous. () Sex (adjustment might lead to power problem; only 0% female) () Histological type (adenocarcinoma - % or squamous cell carcinoma %) adjustment in the regression model + stratification. () Complications (surgical or medical; combined in groups: none one more than one; or groups if insufficient statistical power. In SECC database: types of surgical complications and 0 medical) (Not in 0 days survival) () Surgery volume: equal sized groups BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

26 Page of BMJ Open Statistical analyses The adjusted associations between education and survival after oesophageal cancer surgery will be analysed by means of multivariable Cox regression analyses (hazard ratio s) with % confidence intervals. Kaplan-Meier method will be used to present mortality, and comparison between survival curves in various exposure groups will be tested with the log-rank test. Interaction analysis or separate stratified analysis will be conducted for different tumour stages and histological type (possibly not presented in study). Three models will be used: - Model : crude model - Model : age, sex, TNM, - Model : all variables (except for complications in short term/0d mortality) Time plan What to do When to do it ) Study protocol October-November 0 ) Statistical analyses October-November 0 ) Manuscript writing October-December 0 BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

27 Page of References. Cavalli-Bjorkman, N., et al., Differences according to educational level in the management and survival of colorectal cancer in Sweden. Eur J Cancer, 0. (): p Cavalli-Bjorkman, N., et al., Lower treatment intensity and poorer survival in metastatic colorectal cancer patients who live alone. Br J Cancer, 0. 0(): p. -.. Eloranta, S., et al., Does socioeconomic status influence the prospect of cure from colon cancer--a populationbased study in Sweden Eur J Cancer, 00. (): p. -.. Morgan, M.A., et al., Influence of socio-economic deprivation on outcomes for patients diagnosed with oesophageal cancer. Scand J Gastroenterol, 00. (0): p Sorbye, H., et al., Clinical trial enrollment, patient characteristics, and survival differences in prospectively registered metastatic colorectal cancer patients. Cancer, 00. (0): p. -.. Bentley, R., et al., Area disadvantage, individual socio-economic position, and premature cancer mortality in Australia to 000: a multilevel analysis. Cancer Causes Control, 00. (): p. -.. Frederiksen, B.L., et al., Do patient characteristics, disease, or treatment explain social inequality in survival from colorectal cancer? Soc Sci Med, 00. (): p Morris, A.M., et al., Residual treatment disparities after oncology referral for rectal cancer. J Natl Cancer Inst, (0): p. -.. Bergquist, H., et al., Factors predicting survival in patients with advanced oesophageal cancer: a prospective multicentre evaluation. Aliment Pharmacol Ther, 00. (): p Kii, T., et al., Evaluation of prognostic factors of esophageal squamous cell carcinoma (stage II-III) after concurrent chemoradiotherapy using biopsy specimens. Jpn J Clin Oncol, 00. (): p. -.. Farrow, D.C. and T.L. Vaughan, Determinants of survival following the diagnosis of esophageal adenocarcinoma (United States). Cancer Causes Control,. (): p. -.. Sariego, J., et al., Prediction of outcome in "resectable" esophageal carcinoma. J Surg Oncol,. (): p. -.. Steyerberg, E.W., et al., Surgical mortality in patients with esophageal cancer: development and validation of a simple risk score. Journal of clinical oncology : official journal of the American Society of Clinical Oncology, 00. (): p. -.. Thrift, A.P., et al., Predictors of survival among patients diagnosed with adenocarcinoma of the esophagus and gastroesophageal junction. Cancer Causes Control, 0. (): p. -.. Lagarde, S.M., et al., Prognostic nomogram for patients undergoing oesophagectomy for adenocarcinoma of the oesophagus or gastro-oesophageal junction. Br J Surg, 00. (): p. -.. Mirinezhad, S.K., et al., Survival Rate and Prognostic Factors of Esophageal Cancer in East Azerbaijan Province, North-west of Iran. Asian Pac J Cancer Prev, 0. (): p. -.. Hosokawa, Y., et al., Clinicopathological features and prognostic factors of adenocarcinoma of the esophagogastric junction according to Siewert classification: experiences at a single institution in Japan. Ann Surg Oncol, 0. (): p. -.. Stassen, L.P., et al., Recurrence and survival after resection of adenocarcinoma of the gastric cardia. Rotterdam Esophageal Tumor Study Group. Dis Esophagus, 000. (): p. -.. Rouvelas, I., et al., Survival after surgery for oesophageal cancer: a population-based study. Lancet Oncol, 00. (): p Ra, J., et al., Postoperative mortality after esophagectomy for cancer: development of a preoperative risk prediction model. Ann Surg Oncol, 00. (): p. -.. Aghcheli, K., et al., Prognostic factors for esophageal squamous cell carcinoma--a population-based study in Golestan Province, Iran, a high incidence area. PLoS One, 0. (): p. e.. Bashash, M., et al., The prognostic effect of ethnicity for gastric and esophageal cancer: the population-based experience in British Columbia, Canada. BMC Cancer, 0. : p... Gossage, J.A., et al., The effect of economic deprivation on oesophageal and gastric cancer in a UK cancer network. Int J Clin Pract, 00. (): p. -. BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.

28 Page of BMJ Open Section/Topic Item # STROBE 00 (v) Statement Checklist of items that should be included in reports of cohort studies Recommendation Reported on page # Title and abstract (a) Indicate the study s design with a commonly used term in the title or the abstract, Introduction (b) Provide in the abstract an informative and balanced summary of what was done and what was found, Background/rationale Explain the scientific background and rationale for the investigation being reported Objectives State specific objectives, including any prespecified hypotheses Methods Study design Present key elements of study design early in the paper Setting Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data collection Participants (a) Give the eligibility criteria, and the sources and methods of selection of participants. Describe methods of follow-up - (b) For matched studies, give matching criteria and number of exposed and unexposed Variables Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if Data sources/ measurement applicable * For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe comparability of assessment methods if there is more than one group Bias Describe any efforts to address potential sources of bias -, Study size 0 Explain how the study size was arrived at -, Quantitative variables Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and why Statistical methods (a) Describe all statistical methods, including those used to control for confounding Results on January 0 by guest. Protected by copyright. (b) Describe any methods used to examine subgroups and interactions (c) Explain how missing data were addressed (d) If applicable, explain how loss to follow-up was addressed (e) Describe any sensitivity analyses BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from - n.a. - - n.a. n.a.

29 Page of Participants * (a) Report numbers of individuals at each stage of study eg numbers potentially eligible, examined for eligibility, confirmed eligible, included in the study, completing follow-up, and analysed (b) Give reasons for non-participation at each stage Descriptive data (c) Consider use of a flow diagram * (a) Give characteristics of study participants (eg demographic, clinical, social) and information on exposures and potential confounders (b) Indicate number of participants with missing data for each variable of interest n.a. +table (c) Summarise follow-up time (eg, average and total amount) Outcome data * Report numbers of outcome events or summary measures over time -0 + table Main results (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (eg, % confidence -, table - interval). Make clear which confounders were adjusted for and why they were included (b) Report category boundaries when continuous variables were categorized Whole manuscript (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period n.a. Other analyses Report other analyses done eg analyses of subgroups and interactions, and sensitivity analyses 0- Discussion Key results Summarise key results with reference to study objectives - Limitations Interpretation 0 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from - similar studies, and other relevant evidence Generalisability Discuss the generalisability (external validity) of the study results Other information Funding Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on which the present article is based *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and published examples of transparent reporting. The STROBE checklist is best used in conjunction with this article (freely available on the Web sites of PLoS Medicine at Annals of Internal Medicine at and Epidemiology at Information on the STROBE Initiative is available at BMJ Open: first published as 0./bmjopen-0-00 on December 0. Downloaded from on January 0 by guest. Protected by copyright.