Endometriosis and Recurrent Pregnancy Loss as New Risk Factors for Venous Thromboembolism during Pregnancy and Post-Partum: The JECS Birth Cohort

► venous thromboembolism ► birth cohort study ► pregnancy ► post-partum ► recurrent pregnancy loss

Background Since venous thromboembolism (VTE) is one of the causes of maternal mortality, several guidelines recommend prophylaxis using low molecular weight heparin for women in high-risk groups. The number of large population-based studies examining predictors for VTE has been limited, and there has been no study based on a Japanese population.

Our objective was to examine VTE risk factor during the pregnancy and post-partum period.

A nationwide birth cohort study known as the ‘Japan Environ- ment and Children’s Study (JECS)’ was conducted by the Ministry of the Environment. The subjects consisted of 103,070 pregnancies recruited by the JECS between January 2011 and March 2014. Pregnant women completed the questionnaires during the first and second/ third trimester. Their medical records were transcribed by physicians or research coordi- nators at registration, just after delivery and at 1 month after delivery.

The frequency of VTE was 7.5 per 10, 000 pregnancies (77 of 103,070) during the pregnancy and post-partum period. After the adjustment of multiple covariates for each factor, endometriosis and recurrent pregnancy loss (RPL) were identi fied as novel independent risk factors for VTE. Adjusted odds ratios were as follows: 2.70 (95% confidence interval, 1.21 –6.00) for endometriosis and 6.13 (2.48 –15.16) for RPL. Threatened abortion, threatened pre-term birth, pre-term birth and caesarean section were ascertained to be risk factors for VTE.

Careful attention should be given to novel predictors, such as endometriosis and a history of RPL, to prevent VTE during the pregnancy and post-partum period. received October 11, 2018 accepted after revision December 17, 2018 DOI https://doi.org/ 10.1055/s-0039-1677733. ISSN 0340-6245. © 2019 Georg Thieme Verlag KG Stuttgart · New York Coagulation and Fibrinolysis THIEME 606 Published online: 2019-02-05

It is well known that pregnancy and puerperium is asso- ciated with venous thromboembolism (VTE). The frequency of post-partum VTE was reported to be 7.2 per 10,000 deliveries in an English cohort registered between 1997 and 2014 and 7.9 in a Swedish cohort registered between 2005 and 2011. 1 Since VTE is one of the causes of maternal mortality, several guidelines recommend prophylaxis using low molecular weight heparin (LMWH) for women in the high-risk group. 2,3 The strongest risk factor is a previous episode.2,3 The recurrence rate of withholding heparin was 2.4% in 125 pregnant women with previous VTE, 0% in 44 women without thrombophilia and 5.9% in 51 women with thrombophilia and/or previous episode of idiopathic throm- bosis. 4 Pregnant women who have had two or more episodes, one episode plus thrombophilia associated with anti-phos- pholipid antibodies, protein C (PC), protein S (PS) or an anti- thrombin (AT) de ficiency, and one episode plus a history of VTE in a first-degree relative are all included in the highest risk group. 2,3 The second highest risk group includes women with an isolated episode during bed rest, dehydration and surgery, thrombophilia or with a complication such as heart disease, lung disease, systemic lupus erythematosus (SLE), cancer, in flammatory bowel disease or nephrotic syn- drome. 2,3,5 The third highest group includes women with an age 3 35 years, a body mass index (BMI) of > 30 kg/m2,a smoker, parity 3 3, a systemic infection, gross varicose veins, paraplegia, pre-eclampsia, hyperemesis, ovarian hypersti- mulation syndrome, multiple pregnancies, use of arti ficial reproductive technology, caesarean section and post-partum haemorrhage. 2,3,5 Accordingly, LMWH is recommended for pregnant women with three or more of these risk factors. However, both the Royal College of Obstetricians and Gynaecologists (RCOG) guidelines and the American College of Chest Physicians (ACCP) evidence-based clinical practice guidelines have recommended that LMWH use should be decided upon after discussion between two or more doctors in the case of the second and third highest risk groups since these guidelines were based on expert opinion due to limited evidence concerning the effect of thromboprophylaxis during pregnancy. 2,3,6 The Japan Society of Obstetrics and Gynecology (JSOG) drew up guidelines in 2014 that conformed to the RCOG and ACCP guidelines due to a lack of evidence in Japan. 7 Before April 2014, thromboprophylaxis was performed according to the previous guidelines, under which an episode of VTE, thrombophilia, age 3 35 years, a caesarean section and obesity were considered to be risk factors for Japanese women. 8 For the highest risk group, low dose unfractionated heparin was recommended for women with a past episode of VTE or thrombophilia who underwent a caesarean section during the post-partum period. Unfractionated heparin or intermittent pneumatic compression was recommended for older obese women after caesarean section or for women with an episode of VTE or thrombophilia after vaginal delivery in the high-risk group. There has been no study of VTE risk factors in Japanese women during the pregnancy and post-partum period. Therefore, it was considered neces- sary to examine these factors because the frequency of VTE is smaller in Japanese populations than in Caucasian popula- tion and the frequency of thrombophilia depends on race. 9 We have conducted a nationwide population-based birth cohort study known as the ‘Japan Environment and Chil- dren’s Study (JECS) ’, a project planned by the Ministry of the Environment, Government of Japan. 10–14 The study subjects consisted of 104,102 registered children or foetuses recruited during the first 3 years of the JECS, and the babies are now being followed up for 13 years mainly to examine the influence of the environment on the foetus. This is the first birth cohort used to examine the frequency of VTE and to determine VTE risk factors during the preg- nancy and post-partum period in a Japanese population.

Study Design and Participants Pregnant women were recruited by the JECS between January 2011 and March 2014. Eligibility criteria for expectant mothers were as follows: that they (1) resided at the time of recruit- ment in any of the study areas selected by 15 regional JECS centres located countrywide; (2) had an expected delivery date after August 1, 2011; and (3) were capable of compre- hending the Japanese language and completing the self-admi- nistered questionnaire. 10–14 The sample size has been calculated in the JECS protocol by the Ministry of the Environ- ment.15 In principle, pregnant women completed the ques- tionnaires during the first (MT1) and second/third trimester (MT2). Their medical records were transcribed by physicians or research coordinators at registration (DrT1), just after delivery (Dr0m) and at 1 month after delivery (Dr1m). This study was based on the jecs-ag-20160424 dataset, which includes 104,102 registered children (foetuses and embryos), and was released restrictively to all concerned in June 2016. The second and third children of multiple preg- nancies were excluded and these numbered 1,003 (0.96%). Twenty-nine participants (0.03%) withdrew their consent completely. Finally, 103,070 pregnancies were included in the main analysis. The mean (standard deviation [SD]) age at registration was 30.7 (5.1) years. The mean (SD) gestational weeks at registration was 14.0 (5.7) weeks. The JECS popula- tion has been recognized as representative of the pregnant women in Japan. 11 The JECS protocol was reviewed and approved by the Ministry of the Environment ’s Institutional Review Board on Epidemiological Studies and by the Ethics Committees of all participating institutions. Written informed consent was obtained from all participating women. Data Collection The first questionnaire (MT1) included socio-demographic characteristics, medical histories, the details of all previous pregnancies and exercise habits. The socio-economic status was assessed by the education level and annual household income in the second question- naire (MT2). The MT2 included lifestyle details. Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al. 607 The first medical record transcript (DrT1) included mater- nal age, gestational weeks at registration, maternal body weight, height, conception and details of all previous preg- nancies (vaginal delivery/caesarean delivery/miscarriage/ induced abortion/stillbirth). The Dr0m included maternal age, gestational weeks at miscarriage and delivery, single/multiple pregnancies, live birth/stillbirth, miscarriage/induced abortion, male/female, birth weight, vaginal/caesarean delivery, pregnancy compli- cations and perinatal outcome. VTE was described in the Dr0m and in the third medical record transcription (Dr1m). Exposures and Covariates Potential exposures were compiled from medical and obste- trical histories as well as from information on gynaecological diseases, obstetric complications and lifestyle. Potential covariates were maternal age at registration (categorized as < 20, 20–29, 30–39, 3 40 years), BMI (cate- gorized as < 18.5, 18.5 –25.0, 3 25.0), the presence/absence of in vitro fertilization and embryo transfer (IVF-ET), smok- ing status and income level (categorize as < 200, 200–< 400, 400–< 600, 600 –< 800, 800 –< 1,000 JPY /C2 10,000 [1 US $ ¼ 114.66 JPY, November 13, 2018]). Statistical Analysis Details of potential exposures are listed in ►Supplementary Table S1 (available in the online version). Frequencies for discrete data between VTE and each factor were counted with proportion. To compare their respective distributions, a Fisher ’s exact test was performed. Crude odds ratios (ORs) for all exposures were calculated and multiple logistic regression analyses were further conducted to determine what kind of exposures was predictive of VTE after controlling for maternal age, BMI, the presence/absence of IVF-ET, smoking status and income level. In case the observed number of each factor and VTE was less than 6, adjusted ORs were not calculated because of potential biases depending on the number of cases in the less frequencies. All calculations were conducted using SPSS version 23 and 24 (IBM Corp., Japan), and a p-value of < 0.05 was regarded as statistically signi ficant.

The frequency of VTE was 7.5 per 10,000 pregnancies (77 of 103,070) during the pregnancy and post-partum period. The presence/absence of VTE according to each factor is shown in ►Tables 1 and 2. Age, BMI, the presence of IVF-ET, smoking status and income level were significantly associated with VTE. After the adjustment of multiple covariates for each factor, endometriosis and recurrent pregnancy loss (RPL) were judged to be novel independent predictors for VTE ( ►Table 3 ). Adjusted ORs were as follows: 2.70 (95% con- fidence interval [CI], 1.21 –6.00) for endometriosis and 6.13 (2.48–15.16) for RPL. There was no statistically signi ficant risk related to any of the listed medical histories or lifestyle other than those mentioned above. Regarding pregnancy complications and outcomes, oligo- hydramnios might be a novel independent predictor for VTE though adjusted ORs were not shown since the number was less than 6 ( ►Table 4 ). Threatened abortion, threatened pre- term birth, pre-term birth at < 37 weeks and caesarean section remained signi ficantly associated with VTE. Adjusted ORs were as follows: 3.61 (95% CI, 2.16 –6.02) for threatened abortion, 2.98 (1.83 –4.86) for threatened pre-term birth, 2.64 (1.30 –5.36) for pre-term birth both at < 37 weeks and 2.19 (1.32 –3.63) for caesarean section. There was no association with hyperemesis or mild hypertensive disorders of pregnancy (HDP) with VTE. The association with a history of placenta praevia, adenomyosis and PCOS, pre-term birth < 34 weeks’ gestation, severe HDP, placenta praevia, abruptio placenta, multiple pregnancy and VTE were unclear in this study because the sample size was relatively small. There were no cases of VTE in patients with a history of stroke, myocardial infarction, congenital heart disease, SLE, Crohn ’s disease, ulcerative colitis, nephrotic syndrome or various forms of cancer.

In the present cohort, we found endometriosis and three or more pregnancy losses as novel risk factors for VTE. Endometriosis is characterized by endometrial-like tissue outside the uterine and is a major contributor to pelvic pain and infertility. 16 This is the first study that we know of to show that endometriosis is a VTE risk factor. In fact, there was one study in which no episode of deep vein thrombosis and VTE occurred after laparoscopic surgery in a group of 266 patients in which 21% had endometriosis. 17 Lesions and activated macrophages can secrete pro-in flammatory cyto- kines such as interleukin-1 β, interleukin-6, interleukin-8 and tumour necrosis factor α in the peritoneal cavity of patients with endometriosis. 16 These pro-in flammatory cytokines might trigger VTE during pregnancy. With regard to RPL, to our knowledge, this is the first study to show that a history of RPL has a direct predictive value for assessing VTE risk. Stillbirth was not included in eight women with both a history of RPL and VTE, though stillbirth is well-known to be a strong risk factor for post- partum VTE. 1 RPL and VTE are considered to be associated with acquired and inheritable thrombophilia such as AT deficiency, PC de ficiency, PS de ficiency and homozygous factor V (FV) Leiden. 18 Clinical criteria for anti-phospholipid syndrome (APS) include thrombosis, recurrent miscarriage, intrauterine foetal death and early-onset pre-term birth following pre-eclampsia or placental insuf ficiency.19 Patients with APS received combined therapy of low dose aspirin and heparin, and this applied to 162 women in the present cohort with APS. RPL was an independent predictor when APS was used as a covariate (not shown). Recently, 472 variants in 187 genes have been reported to be associated with RPL. 20 A meta-analysis revealed a signi ficant associa- tion between RPL and 21 variants, including loss due to thrombophilia with ORs of 0.51 to 2.37. Thrombophilia related to FV Leiden mutation, pro-thrombin mutation, Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al.608 Table 1 The presence/absence of venous thromboembolism according to each factor Factors N % Venous thromboembolism p-Valuea (Missing data) Absence Presence Medical histories Atopic dermatitis 98,766 95.8 Absence 83,178 60 0.207 (4,304) (4.2) Presence 15,512 16 Asthma 98,766 95.8 Absence 87,905 64 0.194 (4,304) (4.2) Presence 10,785 12 Collagen disease 98,766 95.8 Absence 98,560 76 /C0 (4,304) (4.2) Presence 130 0 Autoimmune disease 98,766 95.8 Absence 98,548 76 /C0 (4,304) (4.2) Presence 142 0 SLE 98,766 95.8 Absence 98,618 76 /C0 (4,304) (4.2) Presence 72 0 RA 98,766 95.8 Absence 98,483 76 /C0 (4,304) (4.2) Presence 207 0 IDDM 98,766 95.8 Absence 98,614 76 /C0 (4,304) (4.2) Presence 76 0 NIDDM 98,766 95.8 Absence 98,556 76 /C0 (4,304) (4.2) Presence 134 0 Gestational diabetes 98,766 95.8 Absence 97,950 75 0.436 (4,304) (4.2) Presence 740 1 Hyperthyroidism 98,766 95.8 Absence 97,635 76 /C0 (4,304) (4.2) Presence 1,055 0 Hypothyroidism 98,766 95.8 Absence 97,714 75 0.530 (4,304) (4.2) Presence 976 1 Anaemia 98,766 95.8 Absence 80,446 60 0.555 (4,304) (4.2) Presence 18,244 16 Hypertension 98,766 95.8 Absence 98,220 76 /C0 (4,304) (4.2) Presence 470 0 Hyperlipidaemia 98,766 95.8 Absence 98,206 76 /C0 (4,304) (4.2) Presence 484 0 Stroke 98,766 95.8 Absence 98,579 76 /C0 (4,304) (4.2) Presence 111 0 Myocardial infarction 98,766 95.8 Absence 98,628 76 /C0 (4,304) (4.2) Presence 62 0 Congenital heart disease 98,766 95.8 Absence 98,387 76 /C0 (4,304) (4.2) Presence 303 0 Kawasaki disease 98,766 95.8 Absence 98,271 76 /C0 (4,304) (4.2) Presence 419 0 Depression 98,766 95.8 Absence 95,701 72 0.295 (4,304) (4.2) Presence 2,989 4 Dysautonomia 98,766 95.8 Absence 95,044 72 0.366 (4,304) (4.2) Presence 3,646 4 Anxiety disorder 98,766 95.8 Absence 95,904 71 0.064 (4,304) (4.2) Presence 2,786 5 Gastritis 98,766 95.8 Absence 90,166 65 0.097 (4,304) (4.2) Presence 8,524 11 (Continued ) Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al. 609 Table 1 (Continued ) Factors N % Venous thromboembolism p-Valuea (Missing data) Absence Presence Gastric ulcer 98,766 95.8 Absence 96,998 73 0.142 (4,304) (4.2) Presence 1,692 3 Irritable colon 98,766 95.8 Absence 97,152 75 1.000 (4,304) (4.2) Presence 1,538 1 Crohn’s disease 98,766 95.8 Absence 98,655 76 /C0 (4,304) (4.2) Presence 35 0 Ulcerative colitis 98,766 95.8 Absence 98,468 76 /C0 (4,304) (4.2) Presence 222 0 Fatty liver 98,766 95.8 Absence 98,446 76 /C0 (4,304) (4.2) Presence 244 0 Chronic nephritis 98,766 95.8 Absence 98,349 76 /C0 (4,304) (4.2) Presence 341 0 Nephrotic syndrome 98,766 95.8 Absence 98,599 76 /C0 (4,304) (4.2) Presence 91 0 Breast cancer 98,766 95.8 Absence 98,635 76 /C0 (4,304) (4.2) Presence 55 0 Cervical cancer 98,766 95.8 Absence 97,883 75 0.464 (4,304) (4.2) Presence 807 1 Endometrial cancer 98,766 95.8 Absence 98,682 76 /C0 (4,304) (4.2) Presence 8 0 Gastric cancer 98,766 95.8 Absence 98,687 76 /C0 (4,304) (4.2) Presence 3 0 Colorectal cancer 98,766 95.8 Absence 98,679 76 /C0 (4,304) (4.2) Presence 11 0 Blood cancer 98,766 95.8 Absence 98,650 76 /C0 (4,304) (4.2) Presence 40 0 Other cancers 98,766 95.8 Absence 98,515 76 /C0 (4,304) (4.2) Presence 175 0 Pregnancy histories Pregnancy loss (total number) 99,687 96.7 Absence 76,885 50 0.012 (3,383) (3.3) Presence 22,725 27 Pregnancy loss (once) 94,645 91.8 Absence 76,885 50 0.343 (8,425) (8.2) Presence 17,695 15 Pregnancy loss (twice) 80,842 78.4 Absence 76,885 50 0.333 (22,228) (21.6) Presence 3,903 4 Pregnancy loss (three times or more) 78,070 75.7 Absence 76,885 50 < 0.001 (25,000) (24.3) Presence 1,127 8 Hypertensive disorders of pregnancy 98,766 95.8 Absence 96,791 76 /C0 (4,304) (4.2) Presence 1,899 0 Gestational diabetes 98,766 95.8 Absence 98,223 76 /C0 (4,304) (4.2) Presence 467 0 Abruptio placentae 98,766 95.8 Absence 98,455 75 0.166 (4,304) (4.2) Presence 235 1 Ectopic pregnancy 98,766 95.8 Absence 97,742 75 0.520 (4,304) (4.2) Presence 948 1 Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al.610 Table 1 (Continued ) Factors N % Venous thromboembolism p-Valuea (Missing data) Absence Presence Placenta praevia 98,766 95.8 Absence 98,216 74 0.052 (4,304) (4.2) Presence 474 2 Hydatidiform mole 98,766 95.8 Absence 98,262 76 /C0 (4,304) (4.2) Presence 428 0 Gynaecological disease Menstrual disorder 98,766 95.8 Absence 87,529 68 1.000 (4,304) (4.2) Presence 11,161 8 Endometriosis 98,766 95.8 Absence 95,088 69 0.021 (4,304) (4.2) Presence 3,602 7 Uterine fibroids 98,766 95.8 Absence 92,671 67 0.049 (4,304) (4.2) Presence 6,019 9 Adenomyosis 98,766 95.8 Absence 98,355 74 0.028 (4,304) (4.2) Presence 335 2 Uterine anomaly 98,766 95.8 Absence 98,408 76 /C0 (4,304) (4.2) Presence 282 0 Ovarian tumour 98,766 95.8 Absence 95,267 74 1.000 (4,304) (4.2) Presence 3,423 2 Polycystic ovarian syndrome 98,766 95.8 Absence 96,477 71 0.028 (4,304) (4.2) Presence 2,213 5 Lifestyle Strong exercise during pregnancy 96,753 93.9 No 93,448 72 0.325 (6,317) (6.1) Yes 3,229 4 Moderate exercise during pregnancy 96,091 93.2 No 72,207 64 0.049 (6,979) (6.8) Yes 23,809 11 Walking during pregnancy 95,424 92.6 No 27,018 25 0.298 (7,646) (7.4) Yes 68,333 48 Night shift work during pregnancy 97,035 94.1 No 88,910 70 1.000 (6,035) (5.9) Yes 8,049 6 Breakfast during pregnancy 97,064 94.2 Everyday 69,100 58 0.374 (6,006) (5.8) Not everyday 27,888 18 Working h/wk (MT1) 59,428 57.7 0 h 37,614 28 0.995 (43,642) (42.3) 0 –35 h 21,770 16 Working h/wk (MT1) 63,449 61.6 0 h 37,614 28 0.878 (39,621) (38.4) 36 –45 h 25,789 18 Working h/wk (MT1) 49,807 48.3 0 h 37,614 28 0.277 (53,263) (51.7) > 46 h 12,152 13 Working h/wk (MT2) 64,615 62.7 0 h 45,672 33 1.000 (38,455) (37.3) 0 –35 h 18,896 14 Working h/wk (MT2) 68,246 66.2 0 h 45,672 33 0.663 (34,824) (33.8) 36 –45 h 22,522 19 Working h/wk (MT2) 55,240 53.6 0 h 45,672 33 0.311 (47,830) (46.4) > 46 h 9,525 10 Abbreviations: IDDM, insulin-dependent diabetes mellitus; MT1, maternal questionnaires during the first trimester; MT2, maternal questionnaires during the second/third trimester; NIDDM, non-insulin-dependent diabetes mellitus; RA, rheumatoid arthritis; SLE, systemic lupus erythematosu s. aThe p-value was not calculated when expected frequency is less than 1. Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al. 611 Table 2 The presence/absence of venous thromboembolism according to each factor Factors N %V e n o u s thromboembolism p-Valued (Missing data) Absence Presence Threatened abortion 100,818 97.8 Absence 88,624 52 < 0.001 (2,252) (2.2) Presence 12,117 25 Threatened pre-term labour 100,818 97.8 Absence 80,909 47 < 0.001 (2,252) (2.2) Presence 19,832 30 Hyperemesis 97,070 94.2 Absence 16,745 11 0.648 (6,000) (5.8) Presence 80,249 65 Early miscarriage ( < 12 wk gestation) a 99,614 96.6 Absence 99,130 76 /C0 (3,456) (3.4) Presence 408 0 Stillbirth ( /C21 12 wk gestation) b 100,048 97.1 Absence 99,130 76 0.478 (3,022) (2.9) Presence 841 1 Pre-term birth < 37 wk gestation c 99,475 96.5 Absence 94,090 67 0.020 (3,595) (3.5) Presence 5,309 9 Pre-term birth at 34 –36 wk gestation c 98,196 95.3 Absence 94,090 67 0.539 (4,874) (4.7) Presence 4,035 4 Pre-term birth < 34 wk gestation c 95,436 92.6 Absence 94,090 67 0.003 (7,634) (7.4) Presence 1,274 5 Placenta praevia c 99,519 96.6 Absence 98,777 75 0.400 (3,551) (3.4) Presence 666 1 Abruptio placentae c 99,519 96.6 Absence 99,009 75 0.283 (3,551) (3.4) Presence 434 1 Adherent placenta c 99,519 96.6 Absence 99,209 76 /C0 (3,551) (3.4) Presence 234 0 Premature rupture c 99,519 96.6 Absence 90,648 68 0.545 (3,551) (3.4) Presence 8,795 8 Oligohydramnios c 99,519 96.6 Absence 98,169 72 0.017 (3,551) (3.4) Presence 1,274 4 Mild hypertensive disorders of pregnancy c 99,519 96.6 Absence 97,141 74 0.696 (3,551) (3.4) Presence 2,302 2 Severe hypertensive disorders of pregnancy c 99,519 96.6 Absence 98,469 72 0.007 (3,551) (3.4) Presence 974 4 Uterine infection 100,818 97.8 Absence 100,008 75 0.109 (2,252) (2.2) Presence 733 2 Caesarean section c 99,165 96.2 Vaginal 79,785 49 0.001 (3,905) (3.8) Caesarean 19,304 27 SFD ( < 10%)c 99,360 96.4 Absence 89,385 69 1.000 (3,710) (3.6) Presence 9,899 7 Multiple pregnancy 100,733 97.7 Single 99,670 74 0.040 (2,337) (2.3) Multiple 986 3 Abbreviation: SFD, small for dates. aStillbirth ( /C21 12 weeks ’ gestation) and arti ficial abortion were excluded from analyses. bEarly miscarriage ( < 12 weeks ’ gestation) and arti ficial abortion were excluded from analyses. cMiscarriage and arti ficial abortion were excluded from analyses. dThe p-value was not calculated when expected frequency is less than 1. Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al.612 MTHFR and ANXA5 single-nucleotide polymorphisms is reported to be associated with RPL. 20,21 Recent large cohort study proved that the rate of VTE within 42 days of an induced abortion was 3.0 per 10,000 women (hazard ratio, 0.16, 95% CI, 0·12 –0·22) when compared with women in the live birth cohort, whose VTE rate was 18.5 per 10,000 women. 22 The risk might be due to pathologies common to RPL and VTE. Table 3 An association of medical histories, gynaecological disease and lifestyle with venous thromboembolism Factors Crude ORs (95% CI) p-Value Adjusted ORs a (95% CI) p-Value Medical histories Atopic dermatitis 1.43 (0.82/C0 2.48) 0.204 1.58 (0.89/C0 2.82) 0.119 Asthma 1.53 (0.83/C0 2.83) 0.178 1.55 (0.81/C0 2.96) 0.188 Gestational diabetes 1.77 (0.25/C0 12.71) 0.573 /C0/C0 Hypothyroidism 1.34 (0.19/C0 9.61) 0.774 /C0/C0 Anaemia 1.18 (0.68/C0 2.04) 0.565 1.22 (0.68/C0 2.21) 0.506 Depression 1.78 (0.65/C0 4.87) 0.263 /C0/C0 Dysautonomia 1.45 (0.53/C0 3.97) 0.471 /C0/C0 Anxiety disorder 2.42 (0.98/C0 6.01) 0.056 /C0/C0 Gastritis 1.79 (0.94/C0 3.39) 0.074 1.80 (0.92/C0 3.53) 0.089 Gastric ulcer 2.36 (0.74/C0 7.48) 0.146 /C0/C0 Irritable colon 0.84 (0.12/C0 6.06) 0.865 /C0/C0 Cervical cancer 1.62 (0.23/C0 11.65) 0.633 /C0/C0 Pregnancy histories Pregnancy loss (total number) 1.83 (1.14/C0 2.92) 0.012 1.42 (0.84/C0 2.39) 0.186 Pregnancy loss (once) 1.30 (0.73/C0 2.32) 0.368 1.11 (0.59/C0 2.06) 0.755 Pregnancy loss (twice) 1.58 (0.57/C0 4.37) 0.382 /C0/C0 Pregnancy loss (three times or more) 10.92 (5.16/C0 23.08) < 0.001 6.13 (2.48/C0 15.16) < 0.001 Abruptio placentae 5.59 (0.77/C0 40.34) 0.088 /C0/C0 Ectopic pregnancy 1.38 (0.19/C0 9.90) 0.752 /C0/C0 Placenta praevia 5.60 (1.37/C0 22.88) 0.016 /C0/C0 Gynaecological disease Menstrual disorder 0.92 (0.44/C0 1.92) 0.829 0.93 (0.42/C0 2.03) 0.851 Endometriosis 2.68 (1.23/C0 5.83) 0.013 2.70 (1.21/C0 6.00) 0.015 Uterine fibroids 2.07 (1.03/C0 4.15) 0.041 1.73 (0.84/C0 3.58) 0.139 (Continued ) Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al. 613 The frequency of VTE was found to be 7.5 per 10,000 pregnancies in the present Japanese population registered between 2011 and 2014. It was similar to that in Caucasian population, 1 but VTE has been speculated to be less fre- quent in the Japanese population due to the fact that no FV Leiden and pro-thrombin mutations were found in the Japanese population. 9 The frequency of PS de ficiency was higher in the Japanese population because of a domestic mutation known as PS-Tokushima (K196E) which has a frequency of 1.8%. 23 According to the old guidelines, thromboprophylaxis with the use of unfractionated heparin was speculated to be common for post-partum women with a previous episode of VTE or thrombophilia or for older obese women post-partum after a caesarean section that took place between 2011 and 2014. 8 After April 2014, thrombopro- phylaxis during pregnancy was introduced into the JSOG guidelines. 7 Thus, the effect of thromboprophylaxis during pregnancy might be small in this analysis, but we should consider that some portion of patients with an unex- plained RPL may have received a combination of low d o s ea s p i r i na n du n f r a c t i o n a t e dh e p a r i ne v e nt h o u g ht h e r e was no recorded evidence. 24 Actually, 267 and 1,609 women in the two groups above received heparin and low dose aspirin in the present cohort. However, any indication of treatment after VTE, prophylaxis for VTE or prevention of APS or gestational weeks at the start of heparin therapy was not available. This is one of the

of this study. Associations with pre-term birth, HDP , uterine infection, multiple pregnancies and caesarean section with VTE were reported in the previous studies. 2,3,5 A recent risk prediction model showed that emergency caesarean delivery, stillbirth, varicose veins, pre-eclampsia and post-partum infection were the strongest predictors of post-partum VTE. 1 Ap r e - c o n c e p - tional history of VTE was reportedly associated with an Table 3 (Continued ) Factors Crude ORs (95% CI) p-Value Adjusted ORs a (95% CI) p-Value Adenomyosis 7.94 (1.94/C0 32.46) < 0.001 /C0/C0 Ovarian tumour 0.75 (0.19/C0 3.07) 0.691 /C0/C0 Polycystic ovarian syndrome 3.07 (1.24/C0 7.61) 0.015 /C0/C0 Lifestyle Strong exercise during pregnancy 1.61 (0.59/C0 4.40) 0.356 /C0/C0 Moderate exercise during pregnancy 0.52 (0.28/C0 0.99) 0.046 0.62 (0.33/C0 1.19) 0.154 Walking during pregnancy 0.76 (0.47/C0 1.23) 0.264 0.72 (0.43/C0 1.19) 0.199 Night shift work during pregnancy 0.95 (0.41/C0 2.18) 0.898 1.03 (0.44/C0 2.39) 0.954 Breakfast during pregnancy 0.77 (0.45/C0 1.31) 0.330 0.86 (0.48/C0 1.54) 0.607 Working h/wk (MT1) (0 h vs. 1 –35 h) 0.99 (0.53/C0 1.83) 0.968 0.82 (0.42/C0 1.60) 0.562 Working h/wk (MT1) (0 h vs. 36 –45 h) 0.94 (0.52/C0 1.70) 0.831 0.77 (0.40/C0 1.50) 0.443 Working h/wk (MT1) (0 h vs. 46 h) 1.44 (0.74/C0 2.78) 0.280 1.28 (0.62/C0 2.63) 0.500 Working h/wk (MT2) (0 h vs. 1 –35 h) 1.03 (0.55/C0 1.92) 0.937 0.82 (0.41/C0 1.64) 0.575 Working h/wk (MT2) (0 h vs. 36 –45 h) 1.17 (0.66/C0 2.05) 0.591 0.95 (0.50/C0 1.82) 0.884 Working h/wk (MT2) (0 h vs. 46 h) 1.45 (0.72/C0 2.95) 0.301 1.53 (0.72/C0 3.25) 0.274 Abbreviations: BMI, body mass index; CI, con fidence interval; IVF-ET, in vitro fertilization and embryo transfer; MT1, maternal questionnaires during the first trimester; MT2, maternal questionnaires during the second/ third trimester; OR, odds ratio; VTE, venous thromboembolism. aAdjusted for maternal age at registration, BMI, the presence/absence of IVF-ET, smoking and income. Only crude ORs were shown when the number of presence of both factor and VTE was less than 6. Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al.614 increased risk of pre-eclampsia, stillbirth and placental abrup- tion.25 These might also be induced by common genetic factors. An association with oligohydramnios was not confirmed by a logistic model because of the small sample size. The major

was that there was no distinction made between VTE during pregnancy and that occurring post-partum. Data on gestational weeks at the VTE occurrence were also not available. This might be a reason why hyperemesis were not associated with VTE in this study. Thus, we could not dismiss these associations. The majority of women were recruited at 14 weeks’ gestation, therefore, we should keep in mind that we did not cover all early miscarriages or hyperemesis. Endometriosis and a history of RPL were found to be novel risk factors for VTE. Endometriosis and RPL affect 6 to 10% and 4.2% of women of reproductive age, respectively. 16,26 Consequently, we should pay careful attention to novel predictors, such as endometriosis and RPL, to prevent VTE in pregnant women. Table 4 An association between pregnancy complication and venous thromboembolism Factors Crude ORs (95% CI) p-Value Adjusted ORs a (95% CI) p-Value Threatened abortion 3.52 (2.18/C0 5.67) < 0.001 3.61 (2.16/C0 6.02) < 0.001 Threatened pre-term labour 2.60 (1.65/C0 4.12) < 0.001 2.98 (1.83/C0 4.85) < 0.001 Hyperemesis 1.23 (0.65/C0 2.34) 0.521 1.21 (0.62/C0 2.38) 0.573 Stillbirth ( /C21 12 wk gestation) 1.55 (0.22/C0 11.17) 0.663 /C0/C0 Pre-term birth < 37 wk gestation b 2.38 (1.19/C0 4.78) 0.015 2.64 (1.30/C0 5.36) 0.007 Pre-term birth 34 –36 wk gestation b 1.39 (0.51/C0 3.82) 0.521 /C0/C0 Pre-term birth < 34 wk gestation b 5.51 (2.22/C0 13.70) < 0.001 /C0/C0 Placenta praevia b 1.98 (0.28/C0 14.24) 0.499 /C0/C0 Abruptio placentae b 3.04 (0.42/C0 21.93) 0.270 /C0/C0 Premature rupture b 1.21 (0.58/C0 2.52) 0.606 1.14 (0.52/C0 2.50) 0.744 Oligohydramnios b 4.28 (1.56/C0 11.73) 0.005 /C0/C0 Mild hypertensive disorders of pregnancy b 1.14 (0.28/C0 4.65) 0.854 /C0/C0 Severe hypertensive disorders of pregnancy b 5.62 (2.05/C0 15.40) < 0.001 /C0/C0 Uterine infection 3.64 (0.89/C0 14.85) 0.072 /C0/C0 Caesarean section b 2.28 (1.42/C0 3.64) < 0.001 2.19 (1.32/C0 3.63) < 0.001 SFD ( < 10%)b 0.92 (0.42/C0 1.99) 0.825 0.95 (0.41/C0 2.20) 0.899 Multiple pregnancy 4.10 (1.29/C0 13.02) 0.017 /C0/C0 Abbreviations: BMI, body mass index; CI, con fidence interval; IVF-ET, in vitro fertilization and em bryo transfer; OR, odds ratio; SFD, small for dates; VTE, venous thromboembolism. aAdjusted for maternal age at registration, BMI, the presence/absence of IVF-ET, smoking and income. Only crude ORs were shown when the number of presence of both factor and VTE was less than 6. bMiscarriage and arti ficial abortion were excluded from analyses. Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al. 615 What is known about this topic?  Venous thromboembolism (VTE) prophylaxis during pregnancy and post-partum is recommended for women in high-risk groups.  The strongest risk factor is a previous episode. Throm- bophilia, an age3 35 years, obesity, a smoker, parity3 3, a systemic infection, gross varicose veins, paraplegia, pre-eclampsia, hyperemesis, multiple pregnancies, cae- sarean section, stillbirth and post-partum haemorrhage were considered to be risk factors.  The number of large population-based studies exam- ining predictors for VTE has been limited, and there has been no study based on a Japanese population. What does this paper add?  Endometriosis and recurrent pregnancy loss (RPL) were identified as novel independent risk factors for VTE.  Threatened abortion, threatened pre-term birth, pre- term birth and caesarean section were ascertained to be risk factors for VTE. Authors’ Contributions The JECS group conducted the nationwide study project. M.S.O. designed the present study, analysed the data and wrote the first draft of the manuscript. T.E. organized the study team and was responsible for obtaining and analys- ing the data. M.K., a member of the JECS Steering Com- mittee, was responsible for data acquisition and supervision of the study. M.K. and T.O. took the initiative in the launch of the Aichi regional sub-cohort of JECS. T.M. analysed the data. Y.Y., T.O., Y.M., S.K. and S.S. were responsible for data acquisition. All authors interpreted the data, contributed to the writing of the manuscript and revised it critically for important intellectual content. Note Data sharing is not permitted by the JECS due to a government policy restricting the deposition of data containing personal information. See the reference for more details. 12 Funding This study was funded by the Ministry of the Environ- ment, Government of Japan. Conflict of Interest M.S.-O. received grants from the Japanese Ministry of Education, Science, and Technology for conducting stu- dies on the topic of RPL in the different patient population, and payment for lectures from Kaken Pharmaceutical Co. Ltd., Kissei Pharmaceutical Co., Aska Pharmaceutical Co. Ltd., Sekisui Medical Co. Ltd. and Siemens Japan. The remaining authors report no con flict of interest.

We would like to express our gratitude to all of the JECS participants and all obstetricians in the 15 areas. The findings and conclusions of this article are solely the responsibility of the authors and do not represent the official views of the Ministry of the Environment. The chief members of the Japan Environment and Chil- dren’s Study (JECS) as of 2017 are the National Institute for Environmental Studies, Tsukuba (principal investigator, Toshihiro Kawamoto); the National Center for Child Health and Development, Tokyo (Hirohisa Saito); Hokkaido Uni- versity, Sapporo (Reiko Kishi); Tohoku University, Sendai (Nobuo Yaegashi); Fukushima Medical University, Fukush- ima (Koichi Hashimoto); Chiba University, Chiba (Chisato Mori); Yokohama City University, Yokohama (Shuichi Ito); University of Yamanashi, Chuo (Zentaro Yamagata); Uni- versity of Toyama, Toyama (Hidekuni Inadera); Nagoya City University, Nagoya (Michihiro Kamijima); Kyoto University, Kyoto (Takeo Nakayama); Osaka University, Suita (Hiroyasu Iso); Hyogo College of Medicine, Nishinomiya (Masayuki Shima); Tottori University, Yonago (Yasuaki Hirooka); Kochi University, Nankoku (Narufumi Suganuma), University of Occupational and Environmental Health, Kitakyushu (Koi- chi Kusuhara); and Kumamoto University, Kumamoto (Takahiko Katoh).

1 Sultan AA, West J, Grainge MJ, et al. Development and validation of risk prediction model for venous thromboembolism in post- partum women: multinational cohort study. BMJ 2016;355:i6253 2 Bates SM, Greer IA, Middeldorp S, Veenstra DL, Prabulos AM, Vandvik PO. Venous thromboembolism, thrombophilia, antithrom- botic therapy, and pregnancy. American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (9th Edition). Chest 2012;141:691–736 3 Royal College of Obstet Gynecol. Reducing the risk of thrombosis and embolism during pregnancy and the puerperium. RCOG Green-top Guideline No. 37a;2015 4 Brill-Edwards P, Ginsberg JS, Gent M, et al. Recurrence of Clot in This Pregnancy Study Group. Safety of withholding antepartum heparin in women with a previous episode of venous throm- boembolism. N Engl J Med 2000;343:1439 –1444 5 Sultan AA, Tata LJ, West J, et al. Risk factors for first venous thromboembolism around pregnancy: a population-based cohort study from the United Kingdom. Blood 2013;121(19):3953 –3961 6 Bain E, Wilson A, Tooher R, Gates S, Davis LJ, Middleton P. Prophylaxis for venous thromboembolic disease in pregnancy and the early postnatal period. Cochrane Database Syst Rev 2014; 2(02):CD001689 7 Minakami H, Maeda T, Fujii T, et al. Guidelines for obstetrical practice in Japan: Japan Society of Obstetrics and Gynecology (JSOG) and Japan Association of Obstetricians and Gynecologists (JAOG) 2014 edition. J Obstet Gynaecol Res 2014;40(06):1469–1499 8 Editorial Committee on Japanese Guidelines for Prevention of Venous Thromboembolism. Japanese guidelines for prevention of venous thromboembolism. Tokyo, Japan: Medical Front Interna- tional Limited; 2004:47 –54 9 James AH, Jamison MG, Brancazio LR, Myers ER. Venous throm- boembolism during pregnancy and the postpartum period: inci- dence, risk factors, and mortality. Am J Obstet Gynecol 2006;194 (05):1311–1315 10 Kawamoto T, Nitta H, Murata K, et al; Working Group of the Epidemiological Research for Children ’s Environmental Health. Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al.616 Rationale and study design of the Japan Environment and Chil- dren’s Study (JECS). BMC Public Health 2014;14:25 11 Michikawa T, Nitta H, Nakayama SF, et al; Japan Environment and Children’s Study Group. Baseline pro file of participants in the Japan Environment and Children’s Study (JECS). J Epidemiol 2018; 28(02):99–104 12 Suzuki K, Shinohara R, Sato M, Otawa S, Yamagata Z. Association between maternal smoking during pregnancy and birth weight: an appropriately adjusted model from the Japan Environment and Children’s Study. J Epidemiol 2016;26(07):371 –377 13 Japan Environment and Children ’s Study. Tokyo: Ministry of the Environment. Available at: http://www.env.go.jp/en/chemi/hs/ jecs/ [updated August 12, 2016]. Accessed August 16, 2016 14 Ishitsuka K, Nakayama SF, Kishi R, et al. Japan Environment and Children’s Study: backgrounds, activities, and future directions in global perspectives. Environ Health Prev Med 2017;22(01):61 15 National Institute for Environmental Studies/National Centre for Japan Environment and Children ’s Study, Japan Environment and Children’s Study (JECS) Study Protocol (ver. 1.4). Available at: https://www.env.go.jp/chemi/ceh/outline/data/jecs-study_proto col_14_en.pdf. Accessed January 18, 2017 16 Giudice LC. Clinical practice. Endometriosis. N Engl J Med 2010; 362(25):2389–2398 17 Ageno W, Manfredi E, Dentali F, et al. The incidence of venous thromboembolism following gynecologic laparoscopy: a multi- center, prospective cohort study. J Thromb Haemost 2007;5(03): 503–506 18 Croles FN, Nasserinejad K, Duvekot JJ, Kruip MJ, Meijer K, Leebeek FW. Pregnancy, thrombophilia, and the risk of a first venous thrombosis: systematic review and Bayesian meta-analysis. BMJ 2017;359:j4452 19 Miyakis S, Lockshin MD, Atsumi T, et al. International consensus statement on an update of the classi fication criteria for de finite antiphospholipid syndrome (APS). J Thromb Haemost 2006;4 (02):295–306 20 Pereza N, Ostoji ć S, Kapović M, Peterlin B. Systematic review and meta-analysis of genetic association studies in idiopathic recur- rent spontaneous abortion. Fertil Steril 2017;107(01):150 –159 21 Hayashi Y, Sasaki H, Suzuki S, et al. Genotyping analyses for polymorphisms of ANXA5 gene in patients with recurrent preg- nancy loss. Fertil Steril 2013;100(04):1018 –1024 22 Liu N, Vigod SN, Farrugia MM, Urquia ML, Ray JG. Venous thromboembolism after induced abortion: a population-based, propensity-score-matched cohort study in Canada. Lancet Hae- matol 2018;5(07):e279 –e288 23 Kimura R, Honda S, Kawasaki T, et al. Protein S-K196E mutation as a genetic risk factor for deep vein thrombosis in Japanese patients. Blood 2006;107(04):1737 –1738 24 Kaandorp SP, Goddijn M, van der Post JAM, et al. Aspirin plus heparin or aspirin alone in women with recurrent miscarriage. N Engl J Med 2010;362(17):1586 –1596 25 Hansen AT, Schmidt M, Horváth-Puhó E, et al. Preconception venous thromboembolism and placenta-mediated pregnancy complications. J Thromb Haemost 2015;13(09):1635 –1641 26 Sugiura-Ogasawara M, Suzuki S, Ozaki Y, Katano K, Suzumori N, Kitaori T. Frequency of recurrent spontaneous abortion and its influence on further marital relationship and illness: the Okazaki Cohort Study in Japan. J Obstet Gynaecol Res 2013;39(01):126–131 Thrombosis and Haemostasis Vol. 119 No. 4/2019 New Risk Factors for VTE during Pregnancy and Post-Partum Sugiura-Ogasawara et al. 617