Study Profile of the Tohoku Medical Megabank Community-Based Cohort Study

1On the afternoon of March 11, 2011, the Great East Japan Earthquake (GEJE) occurred in a large area within eastern Japan, and the following tsunami devastated coastal areas of northern Japan. Consequently, 15,897 individuals were confirmed dead, and 2,534 were reported missing in Japan. In Miyagi and Iwate Prefectures, 14,216 deaths and 2,334 missing people were reported. Thus, Miyagi and Iwate Prefectures could be considered most severely impacted by GEJE, with the most damage.

2There is a pressing concern regarding increased neuropsychiatric disorders, such as posttraumatic stress disorder (PTSD), cigarette and alcohol consumption caused by psychological stress, and high blood pressure (BP) and accompanying cardiovascular diseases (CVD) due to disaster-related stress in the areas where people were greatly affected by the GEJE. Additionally, some patients with high BP, diabetes, or other diseases may have had to discontinue treatment owing to psychosocial adverse effects. If such conditions are not addressed, the frequency of cerebrovascular diseases, CVD, cancer, or suicide might subsequently increase. Thus, it is necessary to clarify the associations between the multidimensional factors and the risk of short-term onset diseases (mental distress, hypertension, and infectious diseases), as well as to identify high-risk subjects and provide support to the individuals and the community.

3In addition to environmental and lifestyle factors, genetic factors might affect the disease onset or disease progression. It has been clarified that genetic factors contribute to the development of risk factors, such as elevated body mass index (BMI), high BP, and common diseases, such as stroke, CVD, cancer, and mental disorders. Furthermore, lifestyle habits, such as alcohol consumption, are modified by genetic factors,, so there is a need for studies that investigate the impact of genetic factors on relation between lifestyle and diseases (ie, gene-environment interactions). Thus, we decided to ask participants to provide their genetic information.

4To overcome the above-mentioned issues, we established the Tohoku Medical Megabank (TMM) Project, which is conducted by Tohoku University Tohoku Medical Megabank Organization (ToMMo) and Iwate Medical University Iwate Tohoku Medical Megabank Organization (IMM). We decided to conduct the TMM Project in Miyagi and Iwate Prefectures and started two prospective cohort studies in Miyagi and Iwate Prefectures: a population-based adult cohort study, the TMM Community-Based Cohort Study (TMM CommCohort Study), which recruited more than 80,000 participants, and a birth and three-generation cohort study, the TMM Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study), which recruited more than 70,000 participants, including fetuses and their parents, siblings, grandparents, and extended family members.

5The objective of TMM CommCohort Study is to assess the long-term impact of the GEJE on disaster victims. We will conduct long-term follow-up analyses to investigate the cause-specific mortality and incidences of cancer, cerebrovascular diseases, and heart diseases as the main outcomes, as well as the risk factors of these diseases. Additionally, we also assess the gene-environment interactions on the incidence of the above-mentioned major diseases. The associations between genetic risk scores and the main outcomes—particularly the absolute risk assessment—and the associations between genetic risk scores combined with environmental factors and the main outcomes—particularly the development of new disease risk charts, such as the Framingham risk score, a risk prediction model developed by the national cancer center,; NIPPON DATA risk chart; and Suita score—through the assessment of absolute risks will be conducted.

6TMM CommCohort Study is a community-based prospective cohort study conducted from 2013 in the areas affected by the GEJE. Inclusion criteria were persons aged 20 years old or over who were registered in the basic resident register of all municipalities in Miyagi Prefecture and Iwate Prefecture at the time of enrollment. Despite the above criteria, we did not ask participants aged 75 years and older at the site of municipality based health check-up. Individuals were excluded from the TMM CommCohort Study if they did not consent to participate in the study and/or if they were not able to complete the study questionnaires.

7The TMM CommCohort Study was approved by Ethical Committee of ToMMo (the first approval: 2012-4-617 and latest approval: 2018-4-087) and the Medical Ethics Committee of Iwate Medical University (HG H25-2). The study protocol was compiled with inputs from the Genome Cohort Collaboration Promotion Working Group, which consisted of leaders of the genome cohort in Japan. All participants provided informed consent to the TMM Project to have their data from the specific health check-ups and on medical expenditure utilized, or several other official sources of information for the sake of this study.

8In the TMM CommCohort Study, we recruited potential participants through two major approaches (Figure 1). The Type 1 survey collected basic information, such as blood and urine, questionnaire, and municipal health check-up data (eg, height, weight, and blood pressure). The Type 2 survey was conducted at the TMM facility without going through the municipal health check-ups with detailed measurement in addition to measurement list of the Type 1 survey.

9The first type of survey was performed in specific municipal health check-up sites. We visited the sites and ask participants to join our cohort. All participants living in Japan were eligible to undergo this screening, which is provided by an insurer. These check-ups focused on visceral fat obesity, which is used to screen for individuals with risk factors of CVD, such as hypertension, diabetes, dyslipidemia, and metabolic syndrome. Participants who attend the health check-up conducted by the municipality were mainly covered members of the National Health Insurance (NHI).

10In the health check-up settings, researchers explained the TMM Project overview and cohort study contents to participants. The briefings were to be performed mainly by staff, instructors, and Genome Medical Research Coordinators (GMRCs) who belonged to the ToMMo and IMM. While participants were waiting for health checkups and before blood was drawn, informed consent was obtained with face-to-face individual interviews by trained GMRC, and each participant checked all the items on the written form and signed. When they consented to join our cohort, additional blood was drawn from participants and they were handed questionnaires, which they would complete in their house and would send back to our research center.

11When there was a limitation with the setting of municipal specific health check-ups (the size of the venue), we conducted the Type 1 Additional survey. The survey was conducted in ToMMo only and was conducted on different dates from those of the specific health check-ups in the municipality; the municipality and the ToMMo arranged for the survey to be conducted in a selected place at a specified date. The way for obtaining informed consent was similar to the Type 1 survey.

12The Type 2 survey was performed in an assessment center, the Community Support Center in ToMMo and Satellite in IMM. All eligible participants (ie, residents in Miyagi and Iwate prefecture aged 20 years and over) could participate in the Type 2 survey, and participants were administered several detailed measurements.

13TMM Project conducted advertisements for enrollment using mass media, such as newspaper fliers, webpages, television and radio, distribution of advertisement leaflets to public institutions, and advertisement at the specific health check-up venues. For those who wished to participate in the study, the assessment centers office arranged appointments for them to visit on specified dates. The TMM Project outline and cohort study details were explained to examinees at the survey (briefing). Informed consent was obtained in writing from each participant by the GMRC. This survey collected information on physiological measurements, in addition to those collected for the Type 1 and Type 1 Additional survey.

14The municipalities that cooperated with the Type 1 or Type 1 Additional survey are provided in Figure 2. In Figure 2, we also describe the location of assessment centers that performed the Type 2 surveys.

15For all of the surveys using different recruitment methods, six tubes of blood were collected. At the Type 1 survey, participants were asked to provide extra blood in addition to that required for the specific health check-ups.

16The common blood collection amount in ToMMo and IMM were 34 mL. The measured blood consisted of 7 mL in an EDTA2Na blood tube, 5 mL in a sodium heparin blood tube, and two serum blood tubes containing 9 mL of blood, completing blood count tests (2 mL), and for a blood sugar or HbA1c test (2 mL).

17For urine collection, at the Type 1 survey, urine samples that were leftover following measurement for the specific health checkups were stored in urine centrifuge tubes. For the Type 1 Additional surveys or the Type 2 survey in ToMMo, participants were requested to collect urine at home on the morning of the survey; thereafter, they were requested to bring the urine to the sites or assessment centers. For Type 2 surveys, a spot urine sample was collected at each of the satellites in IMM. Details of blood and urine tests conducted for study purposes are provided in Table 1.

18For the Type 1 surveys, after participants provided informed consent, they were handed questionnaires for self-administration. Staff asked participants to complete and return them in person within the 2 weeks following survey administration. For the Type 2 surveys, following the organization of appointments for the participants to visit the assessment centers, participants were sent questionnaires via mail.

19The following question items were established on the basis of the question items from the Japan Multi-Institutional Collaborative Cohort Study (J-MICC Study), a cohort study consisting of tens of thousands of the general population, and based on the results of the Japan Public Health Center-based Prospective Study (JPHC Study). The J-MICC study intended to investigate the influence of both living circumstances and genetic contributors as risk factors of cancer and other lifestyle-related diseases. Such information would be useful in gaining information to establish life-style related disease prevention methods suitable for individual constitutions, or tailor-made prevention. The JPHC Study was conducted as a joint study by a designated group to establish evidence on the benefit of health maintenance and improvements including cancer prevention based on multipurpose cohort studies. Additionally, we have included items from the survey performed in the GEJE-affected area, such as K6 and Lubben Social Network Scale-6 (LSNS-6), since our cohort assessed victims of GEJE. Moreover, the questionnaire included unique items regarding the genetic disposition and birthplace of family members. A detailed list of the items included in the questionnaires is shown in Table 2. In IMM (Iwate), the participants were also asked to complete additional questionnaires.

20In order to perform a more detailed assessment of the physiological health status of the participants, several additional physiological measurements were conducted. Physiological measurements take about 2 hours. In ToMMo, ophthalmic examinations, hearing acuity, respiratory function tests, estimated central BP, calcaneal ultrasound bone mineral density, muscular strength, oral examinations, and additional questions were asked using tablet personalized computers.

21Additionally, for participants who agreed to have measurements taken at home, we asked them to measure home blood pressure and step counts for 2 weeks. In IMM, ophthalmic examinations, calcaneal ultrasound bone mineral density, visceral fat, electrocardiography, pulse wave velocity, carotid echography, and flow mediated dilation tests were performed. We describe the list of physiological measurements taken and the devices used at the assessment centers in Table 3A and Table 3B. As shown in Figure 1, some participants who were recruited in the Type 1 survey visited the assessment centers later and received additional testing.

22We returned the results of blood, urine, questionnaire, and other measurements to participants within 2 to 6 months. If they showed extraordinary data, which might affect the health immediately, we returned the data as soon as possible. We also provided mental health follow-up when the participants showed high risk for depression or post-traumatic stress response.

23We performed several ways to ascertain follow-up information of our participants. In brief, the ways were mail survey; information on medical expenditure and health check-up; medical record review; public statistics, such as basic resident registration and vital statistics; and the following secondary surveys.

24We sent surveys that collected information on self-reported onset of diseases; change in psychological distress, assessed by K6; and treatment status of hypertension, diabetes, and hyperlipidemia via postal mail. Additionally, the surveys included questions regarding the onset of the following diseases: cerebrovascular diseases, myocardial infarction (MI), angina pectoris, chronic obstructive pulmonary disease (COPD), endoscopic treatment and/or surgery owing to gastric ulcers or duodenal ulcers, influenza, PTSD, bronchial asthma, atopic dermatitis, and glaucoma.

25To ascertain information on the specific health check-ups and medical expenditures, we asked the municipality to provide data on blood and urine tests conducted during the specific health check-ups for participants covered by the NHI. The NHI covers almost all medical treatment required for beneficiaries, including diagnostic tests, medication, surgery, supplies and materials, physician and other staffing costs, and most dental treatment. NHI data are reported separately by inpatient and outpatient files. When a beneficiary is withdrawn from the NHI owing to death or emigration, the corresponding reasons and data are entered on the NHI withdrawal history files. Both NHI claims and withdrawal history files were linked with our baseline survey data files using the beneficiary ID number. We also asked the municipality to provide information on medical expenditure withdrawn from the NHI. This information enabled us to clarify the participants who were admitted to hospital.

26Information on the self-reported onset of diseases and hospital admissions were ascertained using data from the mail surveys or from the medical expenditure reports from the NHI. However, it is well known that self-reported onset of diseases has a lower positive predictive value (PPV). In the JPHC Study, the PPVs for self-reported cerebrovascular diseases and MI were 57% and 43%, respectively. Therefore, we confirmed our results by conducting medical record examinations.

27Conversely, the sensitivity of a survey regarding the development of cerebrovascular diseases and MI was 70% and 82%, respectively. Thus, they confirmed that self-reported onset of cerebrovascular diseases and MI was sensitive enough to screen for baseline cerebrovascular diseases and MI in Japanese cohort studies. In addition to chart reviews, we collected information on disease registration, such as cancer or CVD (currently Iwate Prefecture only).

28In principle, information on the address of the subjects was confirmed using the follow-up mail surveys sent by participants. If a participant did not return the follow-up mail survey or the TMM Project did not have the information on the new address of a participant, the participant’s information would be confirmed using the Basic Resident Register of each municipality. In viewing this information, we complied with the Basic Resident Registration Act. If a participant died, the TMM Project confirmed the death using the Basic Resident Register. After an application was submitted for outside use to confirm vital statistics, we applied to vital statistics to confirm cause of death.

29A secondary survey was conducted from 2017 (4 years after the baseline survey) to assess the changes in the health status of participants, including information about ophthalmic examination, hearing acuity, respiratory function tests, estimated central BP, calcaneal ultrasound bone mineral density, muscular strength, oral examinations, and other tests.

30In this paper, we described information on current smoking and psychological distress assessed using the Kessler 6-item psychological distress scale. We compared age-adjusted prevalence of current smoking and higher K6 according to sex, recruitment methods, and residential area.

31From May 2013, we identified 62,439 and 34,980 participants in the ToMMo and the IMM sites, respectively, to partake in the Type 1 surveys. In total 40,433 (65%) and 26,922 (77%) participants in the ToMMo and IMM provided informed consent to be enrolled. For the Type 1 Additional survey conducted only in the ToMMo, 664 participants provided informed consent to be enrolled. For the Type 2 surveys, 13,855 and 5,991 participants based in ToMMo or IMM provided informed consent to be enrolled, and 13,782 and 5,987 returned questionnaires, respectively (Table 4). More than 99% of participants provided blood and urine samples.

32Figure 3 shows the distribution of age groups in ToMMo and IMM. The mean age of participants who completed the Type 1 surveys, Type 1 Additional surveys, and the Type 2 surveys were 59.7, 50.5, and 57.4 years, respectively in ToMMo. Similarly, the mean age of participants who completed the Type 1 survey and Type 2 surveys in IMM were 61.8 and 54.7 years, respectively. Age- and sex-specific distributions of participants are provided in eTable 1.

33For the participants with Type 1 surveys, the prevalence of current smoking did not differ between coastal and inland areas (Figure 4). However, for volunteer-based recruitment, the prevalence of current smoking was higher in coastal areas than in inland areas. Except for IMM women, the prevalence of current smoking was higher in participants who completed the Type 1 surveys than those who completed the Type 2 surveys. Detailed information is provided in eTable 2.

34The participants with Type 1 survey were more likely to have psychological distress assessed using the K6 than those of the Type 2 survey, and women were more likely to have psychological distress than the men (Figure 5). Additionally, coastal residents were more likely to have higher degrees of psychological distress than inland residents regardless of sex (the age-adjusted rates in men ranged from 3.2–4.8% and 4.4–6.6% in the inland and the coastal areas, respectively, and those in women ranged from 3.1–7.4% and 5.4–7.5% in inland and coastal areas, respectively). Detailed information is provided in eTable 3. Participation rate according to area was as follows: inland IMM, 68.9%; coastal IMM, 79.5%; inland ToMMo, 65.6%; and coastal ToMMo, 63.3%.

35The TMM CommCohort Study was a large scale prospective cohort study conducted in areas greatly impacted by the GEJE. Since we recruited participants from both coastal and inland areas, we could compare health status of participants between these areas. As shown in Figure 5, the prevalence of elevated K6 score was higher in coastal areas than that in inland areas. This difference may be because of housing damage and the death of relatives. Further analyses of this dataset will be performed in the future. We have a limitation in this topic. Some participants did not reside in Miyagi and Iwate Prefectures at the time of GEJE. This might cause inaccurate evaluation. However, the inclusion of these participants was unavoidable in a real-world setting. Additionally, this might not largely affect the main study findings.

36As shown in Table 4, Type 1 survey based on municipal specific health check-up site achieved relatively higher consent rate; around 70% of participants who visited health check-up participated in our study. Because we asked participants to provide their genomic information, it might be unrealistic to obtain the very high consent rate for this survey. We considered our consent rate was satisfactorily high. Of course, participants who undergo health check-ups have higher health consciousness than those who do not, and participants who agreed to participate in our study might have higher health consciousness. Thus, although we considered that our participants satisfactory represent the participants who underwent health check-ups in Miyagi and Iwate areas, we should take care that the prevalence obtained from our survey might underestimate the prevalence in all participants. However, this effect of health consciousness might be almost identical among all participants in this survey; we considered that the internal validity of our participants was preserved.

37Conversely, the Type 2 survey recruited participants who voluntarily visited our assessment center. Thus, these participants might not represent the general population who lived in Miyagi and Iwate Prefectures. However, the surveys provide a lot of detailed information on health status, as shown in Table 3A and Table 3B. This information might contribute to understanding the effect of disaster-related factors on health status. Furthermore, following up these detailed health conditions might clarify not only long-term effects of the natural disaster on health status but also general issues that affect atherosclerosis progression, worsening pulmonary function, and other diseases. As with the Type 1 survey, since all participants voluntarily participated in this survey, internal validity was preserved. However, these volunteers might have higher health consciousness. In fact, smoking status was different from the Type 1 and Type 2 surveys, as shown in Figure 4. Thus, there was a problem in simply analyzing the combination of recruitment methods; an analysis, such as a stratified analysis, should be required to handle these data simultaneously.

38We asked participants to provide their genomic information, and we have already analyzed whole genome and array information. As of March 2019, most of genome information has been analyzed in this cohort. Thus, our information can be analyzed together with genome information. We have already reported the several papers regarding genome-wide association analysis or gene-environment interaction for health status., Furthermore, we also analyzed information on metabolome. Thus, we could analyze the gene-metabolome association, and the metabolome-disease association. In addition, we considered that using metabolome information might clarify the pathway from gene to health status. We have also added epigenomic information to this cohort, and we conducted epigenome-wide association studies.

39In this project, we have obtained the agreement of all the participants to transfer their biological samples and data to researchers. Although only a part of our cohort information has already been distributed, several groups used our information as controls, and several collaborative studies were published.– We will expand the data size and will add the follow-up information, including mail surveys; information on medical expenditures and health check-up; medical record review; public statistics, such as basic resident registration and vital statistics; and secondary survey data. This expansion may contribute to the development of personalized medicine and healthcare.

40Collaborations are highly welcome. Please contact the corresponding authors (AH, KT) with any enquiries. We also established a system for data transfer to Japanese researchers. Information on data distribution is provided on the webpage (http://www.dist.megabank.tohoku.ac.jp/ currently in Japanese only).

41This cohort comprised a large sample size and contains information on the natural disaster, genome information, and metabolome information. This cohort also had several detailed measurements. Using this cohort enabled us to clarify the long-term effect of the natural disaster and also to establish personalized prevention based on genome, metabolome, and other omics information.