Research Study the Swedish Mammography Cohort and The

Page 1 of 8

Research study

The Swedish mammography cohort and the cohort of Swedish men: study design and characteristics of two population- based longitudinal cohorts H Harris1,2*, N Håkansson1, C Olofsson1, O Stackelberg1, B Julin1, A Åkesson1, A Wolk1 Study Design

Abstract Cohort participants and 25% of Cohort Introduction Introduction of Swedish Men participants were cur- The Swedish Mammography Cohort The Swedish Mammography Cohort, rent smokers, and 23% and 15% were (SMC) and the Cohort of Swedish Men established in 1987, and the Cohort regular users of dietary supplements, (COSM) are two large population- of Swedish Men, established in 1997, respectively. based longitudinal cohorts established are two ongoing population-based Conclusion in Sweden. The SMC was established longitudinal cohorts designed to Recently, clinical sub-cohorts col- in 1987–1990 (together with a mam- study dietary exposures in relation to lecting more detailed information mography screening program) with chronic disease outcomes. This study including blood, urine, and adipose the aim to explore dietary exposures discusses these two cohorts. tissue collection, dual energy X-ray in relation to breast cancer risk and absorptiometry scan, and a cogni- nal manuscript. nal

Materials and methods has since been expanded to assess ﬁ Information on dietary intake includ- tive test phone interview have been risk of various health outcomes. The ing alcohol and dietary supplements established allowing for future in- COSM was established 10 years later, has been assessed by validated semi- depth analyses that examine health in 1997, with the aim to explore diet quantitative food frequency ques- outcomes in an aging population. and lifestyle in relation to chronic tionnaires at multiple time points. Questionnaires have also collected information including anthropomet- Table 1 Selected characterisƟ cs of parƟ cipants in the Swedish Mammography ric and physical activity measures Cohort (SMC) at baseline (1987) and follow-up (1997) and the Cohort of 1 across the life course, sleep habits, Swedish Men (COSM) at baseline (1997) urine and bowel habits, family his- Cohort/quesƟ onnaire year tory of selected diseases, stress and CharacterisƟ cs COSM SMC 1987 SMC 1997 social support. Annual linkage to 1997 multiple registries allows for up to date information on health status of N 61,433 38,984 45,906 all participants. Age (years) 52.9 (9.8) 62.0 (9.3) 60.3 (9.7) Results Married/living with partner (%) 76.3 — 82.4 The mean (±SD) age and body mass Body mass index (kg/m2) 24.7 (3.9) 25.0 (4.0) 25.8 (3.4) index at cohort enrolment were Height (cm) 164.1 (5.8) 164.3 (5.8) 177.2 (6.7) 52.9 (9.8) years and 24.7 (3.9) kg/ on of the manuscript, as well read and approved Ɵ m2 for the Swedish Mammography Waist circumference (cm) — 83.7 (10.8) 96.1 (11.7) Cohort and 60.3 (9.7) years and 25.8 Hip circumference (cm) — 102.7 (9.7) 101.9 (8.6) (3.4) kg/m2 for the Cohort of Swedish Men. At baseline, 76% of Swedish Postmenopausal (%) — 89.7 n/a Mammography Cohort participants Age at menopause — 49.6 (4.8) n/a and 82% of Cohort of Swedish Men Ever use of oral contracep ves (%) — 43.8 n/a participants were married. In 1997, none declared. ict of interests:

Smoking status (%) ﬂ 13% of Swedish Mammography on, design, and prepara Current — 13.1 25.0 Ɵ *Corresponding author

Email: [email protected] Past — 33.0 38.9 rules of disclosure. (AME) ethical Ethics Medical on for Ɵ 1 Unit of Nutritional Epidemiology, Institute Never — 53.9 36.2 of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden Regular user of dietary supplements (%) — 22.8 14.9 2 Obstetrics and Gynecology Epidemiology Average hours of sleep per night — 7.2 (1.1) 7.2 (1.1) Center, Brigham and Women’s Hospital, 1 Boston, Massachusetts, USA Data represent mean (standard devia on) unless otherwise indicated.

Licensee OA Publishing London 2013. Creative Commons Attribution Licence (CC-BY) none declared. Con ng interests: Ɵ F : Harris H, Håkansson N, Olofsson C, Stackelberg O, Julin B, Åkesson A, et al. The Swedish mammography cohort and the cohort of Swedish men: study design and characteristics of two population-based Compe longitudinal cohorts. OA Epidemiology 2013 Oct 01;1(2):16. the concep to All authors contributed All authors abide by the Associa Page 2 of 8

Research study

SMC (Swedish Mammography Cohort) COSM (Cohort of Swedish Men) Source population (Born between 1914-1948, residing Source population (Born between 1918-1952, residing in Vastmanland and Uppsala counties) in Vastmanland and Orebro counties) 90,303 invited in 1987-1990 100,303 invited in 1997-1998

Respondents 66,651 (74% response rate) 1987 Diet and lifestyle questionnaire Baseline Cohort after exclusions 61,433

Contacted 56,030 SMC members 1997 Respondents alive and still residing in study Diet and lifestyle questionnaire 48,850 (49% response rate) area (70% response rate) Baseline Cohort after exclusions 1997 Cohort after exclusions 45,906 39,984 nal manuscript. nal ﬁ

Contacted 48,263 SMC members 2008 Contacted 37,861 COSM members alive and still residing in study Health questionnaire alive and still residing in study area area (63% response rate) (78% response rate)

Contacted 30,134 SMC members 2009 Contacted 29,068 COSM members who completed the 2008 who completed the 2008 Diet and lifestyle questionnaire questionnaire (84% response rate) questionnaire (90% response rate)

Figure 1: Description of recruitment and data collection for the Swedish Mammography Cohort (SMC; 1987–2009) and the Cohort of Swedish Men (COSM; 1997–2009). diseases. Since then, additional ques- All subjects gave full informed con- asked to return the completed questionnaire data have been collected sent to participate in this study. tionnaire in-person at their mammo- and sub-cohorts with biological speci- gram appointment. Seventy-four per mens have been established making SMC cent of women returned a completed these cohorts a valuable resource for questionnaire. After excluding those Recruitment the scientific community. The aim of with an incorrect or missing national this study was to discuss the study From March 1987 to December 1990, registration number, cancer diag- design and characteristics of these all women living in Uppsala County of nosis (except non-melanoma skin on of the manuscript, as well read and approved two population-based longitudinal central Sweden and who were born cancer) before baseline, implausible Ɵ cohorts. in 1914 through 1948 (n = 48,517) total energy intake (three standard and all women living in the adjacent deviations [SD] from the mean value Västmanland County (n = 41,786) for loge-transformed energy intake), Materials and methods who were born in 1917 through those outside the age-range of 38–76 The SMC and the COSM are coor- 1948 received an invitation by mail years, or other missing data, the final dinated by the Unit of Nutritional to participate in a population-based baseline cohort consisted of 61,433

Epidemiology within the Institute mammography screening program none declared. ict of interests: women. Characteristics of the SMC at ﬂ

of Environmental Medicine at the (Figure 1). The invitation included a on, design, and prepara

baseline are presented in Table 1. Ɵ Karolinska Institute, Stockholm, six-page questionnaire that sought 1 1997 SMC questionnaire rules of disclosure. (AME) ethical Ethics Medical on for

Sweden . This work conforms to the information on diet and alcohol Ɵ values laid down in the Declaration of intake (67-item Food Frequency In the fall of 1997, a second extended Helsinki (1964). The protocol of this Questionnaire [FFQ]), parity, age at questionnaire was sent to all SMC mem- study has been approved by the rele- first birth, family history of breast bers (n = 56,030) who were still alive vant ethical committee related to our cancer, weight, height, education and residing in the study area (Figure 1). institution in which it was performed. level and marital status. Women were The 1997 questionnaire collected

Research study

information on diet and alcohol intake Table 2 Comparison of the Swedish Mammography Cohort (SMC) and (96-item FFQ), dietary supplements, the Cohort of Swedish Men (COSM) with the Swedish populaƟ on in 1997 family history of cancer and myocardial for women aged 48–83 years and men aged 45–79 years, regarding age infarction, age at menarche, history of distribuƟ on, educaƟ onal level and body mass index1 oral contraceptive use, age at meno- Female Male pause, postmenopausal hormone use, PopulaƟ on, PopulaƟ on CharacterisƟ cs SMC 1997 COSM 1997 disease diagnoses, body weight across aged 48–83 aged 45–79 the life course (birth weight through years, 19972 years, 19972 weight at age 80 years), current waist Age groups, years and hip circumferences, medication use Total, n 38,984 1,633,520 45,906 1,594,952 and lifestyle factors including history of 45–49 — — 15.9 19.5 cigarette smoking, physical activity (at 48–54 28.5 27.6 — — ages 15, 30, 50 and current) and sleep 50–54 — — 18.8 20.8 habits. The response rate for the 1997 55–59 17.9 15.0 15.9 15.6 questionnaire was 70%. After exclusion 60–64 14.5 12.7 13.1 12.5 of those with an incorrect or missing 65–69 13.5 12.7 14.1 11.6 national registration number the final 70–74 12.4 12.7 12.4 10.8 1997 cohort consisted of 38,984 par- 75–79 10.1 12.1 9.9 9.2 ticipants. Characteristics of the SMC in 80–83 2.9 7.3 — — 1997 are presented in Table 1. The SMC EducaƟ on, ages nal manuscript. nal population is comparable to the general 48-74 years3 ﬁ Swedish population with regards to age Total, n 33,914 1,316,743 41,382 144,585 distribution, education level and body ≤12 years 78.9 78.7 82.3 77.1 mass index (BMI; see Table 2). >12 years 20.5 19.9 17.3 21.0 Body Mass Index 2008 and 2009 SMC (>25 kg/m2), questionnaires by age groups4 In 2008, a third questionnaire was sent 45–54 37.6 38.8 54.5 57.2 to all cohort members (n = 48,263) that 55–64 45.7 47.4 59.1 60.3 had completed the 1987 FFQ (response 65–74 49.7 52.0 59.8 57.0 rate 63%; Figure 1). The 2008 question- 75–84 43.0 42.3 47.5 43.0 naire collected information including 1All values are percentages unless otherwise indicated. Percentages may not add up due to missing values. 2 general health status, disease diagno- Data from Oﬃ cial Sta s cs Sweden (OSS). 3Educa onal level reported are for those ≤74 years of age since data from OSS were not available for ses, current body weight, height, waist older ages. and hip circumferences, dental health, 4Propor on of overweight (body mass index >25 kg/m2) people within each age group. medication use, sleep habits, urine and bowel habits, family history of selected participate in the study, along with a After excluding those with an incor- diseases, stress and social support. In self-administered questionnaire (Figure rect or missing national registration 2009, a fourth questionnaire collecting on of the manuscript, as well read and approved

1). The questionnaire, the same as used number, cancer diagnosis (except non- Ɵ information on diet and alcohol intake for the SMC in 1997, with the exception melanoma skin cancer) before baseline, (132-item FFQ), dietary supplements, of some sex-specific questions, collected death before baseline or other missing smoking, physical activity, and sun hab- information on diet and alcohol intake data the final baseline cohort consisted its was sent to women who completed (96-item FFQ), dietary supplements, of 45,906 men. Characteristics of the the 2008 questionnaire (84% response family history of selected diseases, dis- COSM at baseline are presented in Table rate). The complete questionnaires ease diagnoses, body weight across 1. The COSM population is comparable can be accessed at: http://ki.se/imm/ the life course, current waist and hip to the general Swedish population with none declared. ict of interests:

1 ﬂ nutrition-en. circumferences, medication use, lower regards to age distribution, education on, design, and prepara urinary tract symptoms (International level and BMI (Table 2). Ɵ COSM

Prostate Symptom Score [I-PSS]) and rules of disclosure. (AME) ethical Ethics Medical on for Ɵ Recruitment lifestyle factors including history of 2008 and 2009 COSM questionnaires In the fall of 1997, all men born in 1918 cigarette smoking and physical activ- through 1952 living in Västmanland ity (at ages 15, 30, 50 and current). A In 2008, a second questionnaire and Örebro counties in central Sweden total of 48,850 (49%) men returned was sent to all COSM members (n = (n = 100,303) received an invitation to a completed questionnaire via mail. 37,861) who were still alive. The

Research study

2008 questionnaire collected infor- the COSM are linked annually to mul- is estimated to be nearly 100% mation including general health tiple registries to ensure up to date complete18. The Prescribed Drug status, disease diagnoses, current information on all participants health Registry contains information about weight, height, waist and hip cir- status (Table 3). There are two main drugs dispensed under prescription cumferences, dental health, medica- types of registries in Sweden, Health at pharmacies (including dispensed tion use, sleep habits, urine (I-PSS) Data Registries and Quality Registries amount, dosage, substance, brand and bowel habits, family history of (national and regional)14,15. name and formulation) and has been selected diseases, stress and social available for linkage since 200519. support (78% response rate). In Health data registries The NBHW is also responsible for 2009, a third questionnaire collect- The Health Data Registries are the the Cause of Death Registry (estab- ing information on diet and alcohol responsibility of the Department lished in 1961), which includes intake (132-item FFQ), dietary sup- of Statistics, Monitoring and information about dates and causes plements, smoking, physical activity, Evaluation (previously called of death. It is has been estimated and sun habits was sent to men who the Epidemiological Centre) at that 93% of all deaths in Sweden completed the 2008 questionnaire the National Board of Health and are reported to the registry within (90% response rate). The complete Welfare (NBHW) and cover the entire 10 days and 100% are reported questionnaires can be accessed at: Swedish population16. They include within 30 days13. These Health Data http://ki.se/imm/nutrition-en1. the National Patient Registry which Registries are mandated by law and encompasses the National Inpatient thus are an excellent source of medi- Questionnaire validation and Registry (IPR; also called the Hospital cal information. In addition to the nal manuscript. nal reproducibility Discharge Registry) and the National Health Data Registries described ﬁ Intake of nutrients and total energy Outpatient Registry (OPR); the above, the NBHW also has a Care were calculated from the food fre- Cancer Registry; and the Prescribed for the Elderly Registry and Persons 14. The National IPR and quency questionnaires using food Drug Registry with Impairments Registry, both OPR contain information about dates composition values obtained from established in 2007. of admission and discharge, hospitals the Swedish National Administration Besides the registries available or clinics providing care, procedures, Database2. The reproducibility and through the NBHW, the government validity of the FFQs used by the SMC and main and secondary diagnoses agency Statistics Sweden has infor- and COSM have been assessed for (coded with international classifi- mation available regarding socio- 17. National foods, nutrients, dietary supple- cation of disease codes) economic status and demographics coverage of the IPR has been 100% ments, glycaemic index and glycae- including level of education, income, since 1987, while coverage of the mic load by comparison with multiple employment/unemployment, early OPR was ~80% in 2007. As of 2011, 24-hour recall interviews and/or diet retirement due to disability, mari- 99% of all somatic and psychiatric records3–7. In addition, FFQ-based tal status, regional migration, immi- estimates of total antioxidant capac- hospital discharges were registered grations and emigrations. Statistics ity (TAC) have been validated with in the IPR and a primary diagnosis Sweden is also responsible for the plasma TAC8, and fatty acid dietary was listed for 99% of all discharges. Multi-Generation Registry, through intake has been validated with sub- A review of the validity of the IPR by which family linkage (parents, sib- 9 Ludvigsson et al. found positive pre- lings, children and cousins) and joint

cutaneous adipose tissue . Exposure on of the manuscript, as well read and approved to environmental pollutants via foods dictive values (PPVs) of IPR diagno- linkage with the NBHW may provide Ɵ has also been validated, including ses of 85–95% for most diagnoses, opportunities to study family history dietary polychlorinated biphenyls with PPVs for specific diagnoses of of diseases. The Multi-Generation (PCB) exposure in relation to PCB 98–100% for myocardial infarction, Registry contains information on concentrations in serum10 and die- 82–97% for heart failure, 99% for individuals born from 1932 and tary cadmium exposure in relation to stroke/transient ischemic attack, onwards. cadmium urine concentrations11. The 87–96% for rheumatoid arthritis,

Quality registries none declared. ict of interests:

95–98% for hip fractures and 99% ﬂ physical activity questionnaire data 17

for acute pancreatitis . The National Quality Registries on, design, and prepara have also been validated using 7-day Ɵ The Cancer Registry (established activity records and accelerometers12. (approximately n = 73 receiving gov- rules of disclosure. (AME) ethical Ethics Medical on for

in 1958) contains basic informa- ernmental funding as of June 2013) Ɵ tion on all individuals diagnosed and regional quality registries con- Follow-up of the cohorts through with cancer in Sweden including site tain more detailed clinical informa- Registry Linkages of tumour, histological type (from tion then the Health Data Registries Using each participant’s Swedish per- 1993), and stage (from 2004). The described above. These registries con- sonal identity number13, the SMC and Swedish National Cancer Registry tain individualised data concerning

Research study

Table 3 Selected registries available for linkage to the Swedish Mammography Cohort (SMC) and the Cohort of Swedish Men (COSM) Name of registry Type of registry Year established/coverage Selected variables Na onal Pa ent Health Data Registry 100% coverage of in-pa ents Dates of admission and Registry (Na onal since 1987 discharge Inpa ent Registry and 80% coverage of out-pa ents in Hospital and clinic codes Na onal Outpa ent 2007 Procedures Registry) Main and secondary diagnoses Cancer Registry Health Data Registry Established in 1958 Site of tumour Es mated to be nearly 100% Histological type (from 1993) complete Stage (from 2004) Prescribed Drug Health Data Registry Available for linkage since 2005 Drugs dispensed under Registry prescrip on Dispensed amount Dosage Substance Brand name Formula on nal manuscript. nal ﬁ Cause of Death — Established in 1961 Date of death Registry 100% complete Cause(s) of death Na onal Prostate Na onal Quality Registry Established in 1996 TNM stage Cancer Registry Gleason grade PSA values Treatment Prostate volume (from 2007) Type of prostatectomy (from 2007) Type of radiotherapy (from 2007) Na onal Breast Cancer Na onal Quality Registry Established in 2007* Tumour characteris cs Registry (i.e. receptor status) TNM stage Grade Treatment Type of surgery Long- and short-term complica ons on of the manuscript, as well read and approved Wai ng mes for diagnosis Ɵ and treatment Na onal Diabetes Na onal Quality Registry Established in 1996 Diabetes type Registry Treatment Weight and height Abdominal circumference (from 2007) none declared. ict of interests: HbA1c ﬂ on, design, and prepara

Blood lipids (from 2002) Ɵ Blood pressure rules of disclosure. (AME) ethical Ethics Medical on for S-crea nine (from 2002) Ɵ Micro- and macro- albuminuria Physical ac vity (from 2007) *Prior to 2007 regional breast cancer quality registries allowed for coverage.

Research study

patient problems, medical interven- treatment characteristics since 1994 Sub-cohorts 20 tions and outcomes with the vision in Uppsala County and since 1997 in SMC-Clinical of the qualities registries to provide Västmanland County. Recruitment 1: From 2003 through data needed for health care quality Through these Health Data 2009, 8,311 women under age 85 and improvement14. Quality registries15 Registry and Quality Registry link- living in the town of Uppsala received that can be linked to the SMC and the ages, diagnosis data (with data a diet, lifestyle and physical activity COSM include those for cardiovascular through 2011) for participants in the questionnaire with an invitation to par- disease (including Swedeheart, Heart SMC and COSM have been obtained ticipate in a health examination. Sixty- Failure Registry [RiksSvikt], Vascular including cardiovascular disease five per cent of invited participants Registry in Sweden [Swedvasc], (CVD; n = 31,509; 8,004 MI; 9,321 completed the questionnaire and 61% National Qualify Registry for Stroke stroke; 6,719 heart failure), type II participated in the health examination [Riks-Stroke]), National Diabetes diabetes (n = 15,200), cancer (n = which included weight, height, waist, Registry (NDR), cancer-specific qual- 20,319; 3,608 breast; 3,969 prostate), hip, and blood pressure measurements, ity registries (i.e. National Prostate fractures (n = 20,027; 5,917 hip frac- and a dual-energy X-ray absorptiom- Cancer Registry, National Breast tures), cataract extracts (n = 19,855) etry scan (n = 5,022). Blood, urine and Cancer Registry, Swedish Gyn- and deaths (n = 28,726; 8,441 cancer; adipose tissue were also collected at Oncology Registry, Swedish Colon 12,203 CVD; Table 4). the health exam. Venous blood samples Cancer Registry), Swedish Rheumatoid Arthritis Registry, Swedish Renal Registry and the Swedish National nal manuscript. nal

Table 4 Incident diagnoses of select health condiƟ ons and deaths in the ﬁ Cataract Registry. Swedish Mammography Cohort (SMC; 1987–2011) and the Cohort of Swedish Linkage to these quality reg- Men (COSM; 1998–2011) istries provides detailed clinical data on a variety of conditions. For Cohort example, the NDR (established in SMC (n = 61,433) COSM (n = 45,906) 1996) contains a wealth of infor- Cardiovascular disease (CVD) 19,638 11,871 mation on diabetes care including diabetes type, diabetes treat- Myocardial Infarc on 4,171 3,833 ment, body weight, height, HbA1c, Heart failure 3,762 2,957 blood pressure, micro- and macro- Stroke 5,488 3,833 albuminuria, blood lipids (from Cancer 11,532 8,787 2002), s-creatinine (from 2002), abdominal circumference (from Breast 3,597 11 2007) and physical activity (from Prostate — 3,969 2007). The National Prostate Cancer Colorectal 1,439 1,009 Registry (established in 1996) includes information on tumour- Lung 892 638 node-metastasis (TNM) stage, Endometrial 915 — Gleason grade, prostate-specific Ovarian 477 — on of the manuscript, as well read and approved antigen values and treatment. Ɵ Additional data have been collected Pancrea c 308 165 since 2007 including prostate Type II diabetes1 7,980 7,223 volume, type of prostatectomy, and Rheumatoid arthri s 839 521 type of radiotherapy. The National Fractures 14,767 5,260 Breast Cancer Registry (established in 2007) includes information on Hip fractures 4,547 1,370 none declared. ict of interests: tumour characteristics (i.e. hormone Cataract extrac ons 13,113 6,742 ﬂ

receptor status), TNM stage, grade, on, design, and prepara Kidney stones 851 1,297 Ɵ treatment, type of surgery, long- and

Abdominal aor c aneurysm 353 1,005 rules of disclosure. (AME) ethical Ethics Medical on for short-term complications and wait- Ɵ ing-times for diagnosis and treat- Total deaths 17,181 11,545 ment. Prior to 2007, regional breast CVD deaths 7,036 5,167 cancer quality registries allowed Cancer deaths 5,167 3,274 for coverage and have provided the SMC with data on tumour and 1Includes prevalent diagnoses reported on the 1997 ques onnaire.

Research study

were collected after a 12-hour over- been invited to participate. Data col- myocardial infarctions among women night fast. Samples were immediately lection for this clinical sub-cohort is in our study population30. In addi- centrifuged and separated in a dark ongoing and, to date, participation in tion to these etiological studies, our room, and stored at −80°C. the cognitive phone interview por- cohorts have been the first to assess A subset of these women com- tion is over 83%. the validity of dietary questionnaire pleted an additional FFQ and physi- for estimates of glycaemic load7, total cal activity questionnaire between Results anti-oxidative capacity of consumed 8 10 2004 and 2006 (n = 303) and wore The mean (±SD) age and BMI at foods and PCB exposure and iden- a physical activity monitor and com- cohort enrolment were 52.9 (9.8) tified the saturated fatty acids penta- pleted a 7-day activity record (n = years and 24.7 (3.9) kg/m2 for the decanoic acid (15:0) and heptadeca- 12 151) . A second subset of women SMC and 60.3 (9.7) years and 25.8 noic acid (17:0) as valid biomarkers 9 (n = 116), completed a questionnaire (3.4) kg/m2 for the COSM. At baseline, of long-term dairy fat intake . on sun exposure in conjunction with 76% of SMC participants and 82% of a blood sample collected during the COSM participants were married. In Conclusion winter months (January-March) and 1997, 13% of SMC participants and The SMC and the COSM are unique 100 of these women completed the 25% of COSM participants were cur- epidemiologic resources with their same sun questionnaire and donated rent smokers, and 23 and 15% were population-based design, high a blood sample during the summer regular users of dietary supplements, response rates, and multiple meas- 21 (August and September) . respectively (Table 1). ures over a 20+ year period (1987– A second health examination, bio-

1997–2008/2009) for the SMC manuscript. nal ﬁ logical sample collection and diet and Discussion and 10+ year period for the COSM lifestyle questionnaire, is planned for Research from the SMC and the (1997–2008/2009). This allows for this clinical sub-cohort in November COSM has resulted in over 200 sci- the examination of the associations 2013. entific peer-reviewed publications. between long-term diet and lifestyle Recruitment 2: From March 2012 Important findings in the area of factors and many different health out- to March 2013, 1,030 women in the nutrition and cancer include protec- comes. At present, very few prospec- SMC who were over age 85 and liv- tive associations observed between tive cohorts in Europe or the United ing in Uppsala County received an dietary folate intake and risk of ovar- States have repeated measurements invitation to participate in a validated ian cancer23 and pancreatic cancer24; and very few are representative of 22 cognitive test phone interview , com- between magnesium intake and colo- the whole population. In addition, plete a FFQ, and receive a home health rectal cancer risk25; between vitamin the ability to link these cohorts to the examination (n = 627, 404 and 309 B6 intake and risk of colorectal can- extensive registry data available in participants, respectively). The home cer26; and between consumption of Sweden is a strength that allows for health examination included weight, fatty fish and kidney cancer risk5. The near complete assessment of multi- height, waist, hip and blood pressure SMC was the first to report a statisti- ple health conditions. The recent and measurements, blood, urine, adipose cally significant interaction between ongoing data collection in the clinical tissue and faecal sample collection, alcohol consumption and hormone sub-cohorts will allow for future in- a physical function test and the Mini replacement therapy for the risk of depth analyses that examine health Mental State Examination. oestrogen/progesterone-positive outcomes in an aging population. on of the manuscript, as well read and approved breast cancer, indicating that alcohol Ɵ COSM-Clinical consumption is more harmful dur- Acknowledgements From May 2010 through June 2013, ing hormone substitution27, and has This work was supported by over 3,000 men in the COSM who observed an association between the Swedish Research Council/ were over age 75 and living in long-term dietary cadmium exposure Committee for Infrastructure, the Västerås have received an invitation (a proposed endocrine disruptor) Karolinska Institutet’s Strategic to participate in a validated cognitive and endometrial and breast cancer Funds, and the Karolinska Institutet’s

11,28 none declared. ict of interests: phone interview22, complete an FFQ incidence . Distinguished Professor Award ﬂ on, design, and prepara and a one-day food diary, and par- Beyond individual foods and nutri- (A.W.). Ɵ ticipate in a health examination. The ents we have also reported that the rules of disclosure. (AME) ethical Ethics Medical on for health examination includes the same metabolic syndrome was associated References Ɵ data collection as in the SMC-Clinical with an increased risk of age-related 1. KarolinskaInstitute.http://ki.se/imm/ recruitment 2 described above. In cataracts29 and that a combina- nutrition-en [Accessed January 20, 2014]. addition, the wives of these men who tion of healthy dietary and lifestyle 2. Bergström L, Kylberg E, Hagman U, are enrolled in the SMC have also behaviours could prevent 77% of Erikson H, Bruce Å. The food composition

Research study

database KOST: the National Food 11. Julin B, Wolk A, Bergkvist L, Bottai M, 22. Gatz M, Reynolds C, John R, Administration’s Information System Åkesson A. Dietary cadmium expo- Johansson B, Mortimer J, Pedersen N. for nutritive values of food. Vår Fö da. sure and risk of postmenopausal breast Telephone screening to identify poten- 1991;43:439–47. cancer: a population-based prospec- tial dementia cases in a population-based 3. Messerer M, Johansson S-E, Wolk A. tive cohort study. Cancer Res. 2012 sample of older adults. Int J Psychogeriatr. The validity of questionnaire-based Mar;72(6):1459–66. 2002 Sep;14(3):273–89. micronutrient intake estimates is 12. Orsini N, Bellocco R, Bottai M, 23. Larsson SC, Giovannucci E, increased by including dietary supple- Hagströmer M, Sjöström M, Pagano M, Wolk A. Dietary folate intake and inci- ment use in Swedish men. J Nutr. 2004 et al. Validity of self-reported total dence of ovarian cancer: the Swedish Jul;134(7):1800–05. physical activity questionnaire among Mammography Cohort. J Natl Cancer Inst. 4. Messerer M, Wolk A. Sensitivity and older women. Eur J Epidemiol. 2008 2004 Mar;96(5):396–402. specificity of self-reported use of die- Oct;23(10):661–67. 24. Larsson SC, Håkansson N, tary supplements. Eur J Clin Nutr. 2004 13. Ludvigsson J, Otterblad-Olausson P, Giovannucci E, Wolk A. Folate intake and Dec;58(12):1669–71. Pettersson B, Ekbom A. The Swedish per- pancreatic cancer incidence: a prospec- 5. Wolk A, Larsson SC, Johansson J-E, sonal identity number: possibilities and tive study of Swedish women and men. J Ekman P. Long-term fatty fish con- pitfalls in healthcare and medical research. Natl Cancer Inst. 2006 Mar;98(6):407–13. sumption and renal cell carcinoma Eur J Epidemiol. 2009;24(11):659–67. 25. Larsson S, Bergkvist L, Wolk A. incidence in women. JAMA. 2006 14. National Healthcare Quality Registries Magnesium intake in relation to risk of Sep;296(11):1371–76. in Sweden. In: Regions SAoLAa, ed. colorectal cancer in women. JAMA. 2005 6. Wolk A, Ljung H, Vessby B, Hunter D, Stockholm, 2007. Jan;293(1):86–9. Willett W. Effect of additional questions 15. Garpenby P, Carlsson P. The role of 26. Larsson SC, Giovannucci E, Wolk A. about fat on the validity of fat estimates National Quality Registers in the Swedish Vitamin B6 intake, alcohol consump- nal manuscript. nal from a food frequency questionnaire. health service. Health Policy. 1994 tion, and colorectal cancer: a longitudi- ﬁ Study Group of MRS SWEA. Eur J Clin Nutr. Sep;29(3):183–95. nal population-based cohort of women. 1998 Mar;52(3):186–92. 16. http://www.socialstyrelsen.se/statistics Gastroenterol. 2005 Jun;128(7):1830–37. 7. Levitan EB, Westgren CW, Liu S, 17. Ludvigsson JF, Andersson E, Ekbom A, 27. Suzuki R, Ye W, Rylander-Rudqvist T, Wolk A. Reproducibility and validity of Feychting M, Kim JL, Reuterwall C, et al. Saji S, Colditz GA, Wolk A. Alcohol and post- dietary glycemic index, dietary glycemic External review and validation of the menopausal breast cancer risk defined load, and total carbohydrate intake in Swedish national inpatient register. BMC by estrogen and progesterone receptor 141 Swedish men. Am J Clin Nutr. 2007 Public Health. 2011 Jun;11(1):450. status: a prospective cohort study. J Natl Feb;85(2):548–53. 18. Mattsson B, Wallgren A. Completeness Cancer Inst. 2005 Nov;97(21):1601–08. 8. Rautiainen S, Serafini M, of the Swedish Cancer Register. Non- 28. Åkesson A, Julin B, Wolk A. Long- Morgenstern R, Prior RL, Wolk A. The notified cancer cases recorded on death term dietary cadmium intake and post- validity and reproducibility of food- certificates in 1978. Acta Radiol Oncol. menopausal endometrial cancer inci- frequency questionnaire-based total 1984;23:305–13. dence: a population-based prospec- antioxidant capacity estimates in 19. Wettermark B, Hammar N, Fored CM, tive cohort study. Cancer Res. 2008 Swedish women. Am J Clin Nutr. 2008 Leimanis A, Otterblad Olausson P, Aug;68(15):6435–41. May;87(5):1247–53. Bergman U, et al. The new Swedish 29. Lindblad BE, Håkansson N, 9. Wolk A, Vessby B, Ljung H, Barrefors P. Prescribed Drug Register: opportunities Philipson B, Wolk A. Metabolic syn- Evaluation of a biological marker of for pharmacoepidemiological research drome components in relation to risk of dairy fat intake. Am J Clin Nutr. 1998 and experience from the first six months. cataract extraction: a prospective cohort Aug;68(2):291–95. Pharmacoepidemiol Drug Saf. 2007 study of women. Ophthalmol. 2008 10. Bergkvist C, Akesson A, Glynn A, Jul;16(7):726–35. Oct;115(10):1687–92.

Michaëlsson K, Rantakokko P, 20. http://www.kvalitetsregister.se/ 30. Åkesson A, Weismayer C, Newby PK, on of the manuscript, as well read and approved Ɵ Kiviranta H, et al. Validation of ques- 21. Burgaz A, Åkesson A, Michaëlsson K, Wolk A. Combined effect of low-risk tionnaire-based long-term dietary expo- Wolk A. 25-hydroxyvitamin D accumula- dietary and lifestyle behaviors in pri- sure to polychlorinated biphenyls using tion during summer in elderly women at mary prevention of myocardial infarc- biomarkers. Mol Nutr Food Res. 2012 latitude 60 degrees N. J Inter Med. 2009 tion in women. Arch Intern Med. 2007 Nov;56(11):1748–54. Nov;266(5):476–83. Oct;167(19):2122–27. none declared. ict of interests: ﬂ on, design, and prepara Ɵ rules of disclosure. (AME) ethical Ethics Medical on for Ɵ