BMJ Open

A nationwide population -based cohort study of peripartum hysterectomy and arterial embolisation in : results from the Belgian Obstetric Surveillance System.

For peer review only Journal: BMJ Open

Manuscript ID bmjopen-2017-016208

Article Type: Research

Date Submitted by the Author: 14-Mar-2017

Complete List of Authors: Vandenberghe, Griet; Ghent University Hospital, Department of Obstetrics and Gynaecology Guisset, Marine; Leuven University Hospital, Department of Obstetrics and Gynaecology Janssens, Iris; Ghent University, Faculty of Medicine Van Leeuw, Virginie; Perinatal Epidemiology Centre (Centre d'Épidémiologie Périnatale, CEpiP); School of Public Health, Université Libre de Bruxelles (ULB) Roelens, Kristien; Ghent University Hospital, Department of Obstetrics and Gynaecology Hanssens, Myriam; University Hospital Leuven, Department of Obstetrics & Gynaecology Russo, Erika; Intercommunale de Santé Publique du Pays de Charleroi (ISSPC), Department of Obstetrics and Gynaecology Vankeirsbilck, Joachim; St Jan's Hospital, Department of Obstetrics and Gynaecology Englert, Yvon; Perinatal Epidemiology Center (Centre d'Épidémiologie Périnatale, CEpiP); Université Libre de Bruxelles, Faculty of Medicine, Research Laboratory on Human Reproduction Verstraelen, Hans; Ghent University Hospital, Department of Obstetrics and Gynaecology

Primary Subject Obstetrics and gynaecology Heading:

Secondary Subject Heading: Epidemiology

peripartum hysterectomy, arterial embolisation, Interventional radiology < Keywords: RADIOLOGY & IMAGING, maternal near miss, severe maternal morbidity

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 37 BMJ Open

1 2 3 1 TITLE PAGE 4 5 2  Title of the article: 6 7 3 A nationwide population-based cohort study of peripartum hysterectomy and arterial embolisation in 8 9 4 Belgium: results from the Belgian Obstetric Surveillance System. 10 5 11 12 6  Author and Co-Author’s name: 13

14 7 Vandenberghe G, Guisset M, Janssens I, Van Leeuw V, Roelens K, Hanssens M, Russo E, Van Keirsbilck J, 15 For peer review only 16 8 Englert Y, Verstraelen H. 17 9 18 19 10  Corresponding author: Griet Vandenberghe 20 21 11 Department of Obstetrics & Gynaecology 22 12 Ghent University Hospital 23 24 13 De Pintelaan 185, 9000 Ghent, Belgium 25 26 14 E-mail: [email protected] 27 28 15 Telephone number: +32 9 33 27 849 29 16 30 31 17  Co-author: Marine Guisset 32 33 18 Department of Obstetrics & Gynaecology 34 35 19 Leuven University Hospital 36 20 Leuven, Belgium 37 38 21 E-mail: [email protected] 39 40 22 41 23  Co-author: Iris Janssens 42 43 24 Faculty of Medicine 44 45 25 Ghent University 46 47 26 Ghent, Belgium 48 27 E-mail: [email protected] 49 50 28 51 52 29  Co-author: Virginie Van Leeuw 53 54 30 Perinatal Epidemiology Center (Centre d’Épidémiologie Périnatale, CEpiP) 55 31 School of Public Health, Université Libre de Bruxelles (ULB), , Belgium 56 57 32 Brussels, Belgium 58 59 33 E-mail: [email protected] 60 34 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 37

1 2 3 35  Co-author: Kristien Roelens 4 36 Department of Obstetrics & Gynaecology 5 6 37 Ghent University Hospital 7 8 38 Ghent, Belgium 9 10 39 E-mail: [email protected] 11 40 12 13 41  Co-author: Myriam Hanssens 14 15 42 Department Forof Obstetrics peer & Gynaecology review only 16 43 University Hospital Leuven 17 18 44 Leuven, Belgium 19 20 45 E-mail: [email protected] 21 22 46 23 47  Co-author: Erika Russo 24 25 48 Department of Obstetrics & Gynaecology 26 27 49 Intercommunale de Santé Publique du Pays de Charleroi (ISPPC) 28 29 50 Hôpital Civil Marie Curie 30 51 Charleroi, Belgium 31 32 52 E-mail: [email protected] 33 34 53 35 54  Co-author: Joachim Vankeirsbilck 36 37 55 Department of Obstetrics & Gynaecology 38 39 56 St Jan’s Hospital 40 41 57 , Belgium 42 58 E-mail: [email protected] 43 44 59 45 46 60  Co-author: Yvon Englert 47 48 61 Perinatal Epidemiology Center (Centre d’Épidémiologie Périnatale, CEpiP) 49 62 Research Laboratory on Human Reproduction, Faculty of Medicine, Université Libre de Bruxelles (ULB) 50 51 63 Brussels, Belgium 52 53 64 E-mail: [email protected] 54 65 55 56 66  Co-author: Hans Verstraelen 57 58 67 Department of Obstetrics & Gynaecology 59 60 68 Ghent University Hospital 69 Ghent, BelgiumFor peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 37 BMJ Open

1 2 3 70 E-mail: [email protected] 4 71 5 6 72  Keywords: peripartum hysterectomy, arterial embolisation, interventional radiology, major obstetric 7 8 73 haemorrhage, maternal near miss, severe maternal morbidity 9 10 74 11 75  Word count: The main text consists of 5952 words 12 13 76 14 15 77 Background:For peer 514 words review only 16 78 Methods: 1007 words 17 18 79 Results: 2630 words 19 20 80 Discussion: 1670 words 21 22 81 Conclusion: 131 words 23 24 82 Total: 5952 words 25 83 26 27 84  Tables and Figures: The manuscript includes 4 tables and 4 figures. Figures are uploaded, in TIFF 28 29 85 format with a resolution of 300 DPI, separately from the text. 30 86 31 32 87 33 34 88 35 36 89 37 90 38 39 91 40 41 92 42 93 43 44 94 45 46 95 47 48 96 49 97 50 51 98 52 53 99 54 55 100 56 101 57 58 102 59 60 103 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 37

1 2 3 104 ABSTRACT 4 5 105 Objectives: 6 7 106 To assess the prevalence of major obstetric haemorrhage managed with peripartum hysterectomy and/or 8 9 107 interventional radiology (IR) in Belgium. To describe women characteristics, circumstances in which the 10 108 interventions took place, management of the obstetric haemorrhage, outcome and additional morbidity of 11 12 109 these women. 13 14 110 Design: 15 For peer review only 16 111 Nationwide population-based prospective cohort study. 17 112 Setting: 18 19 113 Emergency obstetric care. Participation of 97% of maternities covering 98.6% of the deliveries in Belgium. 20 21 114 Participants: 22 115 All women who underwent peripartum hysterectomy and/or IR procedures in Belgium between January 2012 23 24 116 and December 2013. 25 26 117 Results: 27 28 118 We obtained data on 166 women who underwent peripartum hysterectomy (n=84) and/or IR procedures 29 119 (n=102) to manage or prevent major obstetric haemorrhage. One in 3030 women who gave birth in Belgium 30 31 120 underwent a peripartum hysterectomy, while another 1 in 3030 women was managed by IR thereby preserving 32 33 121 the uterus. Interventional radiology procedures were able to cease the bleeding in 87.8% of the attempts. 34 35 122 Abnormal placentation and/or uterine atony were the reported causes of haemorrhage in 83.7%. Abnormally 36 123 invasive placenta was not detected antenatally in 34% of cases. The interventions were planned in 15 women. 37 38 124 Three women were transferred antenatally and seventeen women postnatally to a hospital with emergency IR 39 40 125 service. Urgent peripartum hysterectomy could be averted in 72% of women who were transferred, with no 41 126 significant difference in need for transfusion. Disseminated intravascular coagulation secondary to major 42 43 127 haemorrhage was reported in 32 women. One woman died, the case fatality rate was 0.6% (95% CI 0.1-3.3). 44 45 128 Conclusion: 46 47 129 The prevalence in Belgium of major obstetric haemorrhage requiring peripartum hysterectomy and/or IR is 48 130 estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries. Implementing routine screening for abnormal 49 50 131 placentation could further improve management and outcome of major obstetric haemorrhage. A case-by-case 51 52 132 in-depth analysis would be necessary to reveal whether the hysterectomies performed in this series were 53 54 133 appropriate or preventable. 55 134 56 57 58 135 59 60 136 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 37 BMJ Open

1 2 3 137 ARTICLE SUMMARY: Strengths and limitations of this study: 4 5 138 6 139 • This is a nation-wide study of Belgian maternities, with excellent participation and response rates 7 8 9 140 (98.8%). 10 11 141 • The cases were collected on a monthly basis using the standard methodology of the Obstetric 12 13 142 Surveillance Systems. 14 15 For peer review only 16 143 • Data from the background population as control group were limited to those available in the Belgian 17 18 144 perinatal registry. 19 20 145 • This series describes how women with major obstetric haemorrhage were managed with peripartum 21 22 23 146 hysterectomy and/or interventional radiology according to the circumstances of the event, but does 24 25 147 not allow to making firm conclusions on the motivation or the appropriateness of this management. 26 27 148 28 29 149 30 31 150 BACKGROUND 32 33 151 Emergency peripartum hysterectomy has been referred to as the single most dramatic procedure in modern 34 35 152 obstetrics and has been listed by the World Health Organisation (WHO) as an identification criterion for 36 37 38 153 maternal near miss.(1) The near miss concept has been introduced in the face of very low maternal mortality 39 40 154 rates in developed country settings (2) and has been increasingly adopted as an indicator of obstetric care and 41 42 155 as an analytic tool to address health system failures with the overall goal of improving obstetric care.(3) 43 44 45 156 Substandard care in the management of major obstetric haemorrhage continues to be a critical cause of 46 47 157 maternal deaths and severe maternal morbidity in developed countries.(4, 5) Internationally recognized 48 49 158 guidelines (6-8) generally recommend that all obstetric practitioners should be trained in the management of 50 51 52 159 postpartum haemorrhage, to guarantee an immediate response and a multidisciplinary team approach. These 53 54 160 guidelines further indicate that one or more second line measures, including intra-uterine (balloon) 55 56 161 tamponade, haemostatic brace suturing, ligation of the uterine arteries and interventional radiology (IR), 57 58 162 should be available in hospitals with delivery units and that obstetric practitioners should be familiar with 59 60 163 these procedures to control the bleeding while preserving the uterus. Notably, guidelines on postpartum For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 37

1 2 3 164 haemorrhage emphasize that appropriate clinical judgement, and hence, early involvement of an experienced 4 5 165 senior obstetrician, can save lives: the timely decision to turn to a hysterectomy can avoid further blood loss, 6 7 166 risk of disseminated intravascular coagulation (DIC), additional morbidity and finally death.(6) 8 9 10 167 While emergency peripartum hysterectomy serves as the ultimate approach to manage major obstetric 11 12 168 haemorrhage, inevitably causing additional maternal morbidity and, not least, infertility, resorting to IR 13 14 169 procedures on the pathway towards hysterectomy could be both lifesaving and fertility-preserving. IR for major 15 For peer review only 16 170 obstetric haemorrhage basically involves intra-arterial balloon occlusion or arterial embolisation with gel foam, 17 18 19 171 coils or glue of the uterine or the internal iliac arteries accessed via the femoral artery. This approach can be 20 21 172 used in emergency settings as well as for elective procedures, to manage or to prevent obstetric haemorrhage. 22 23 173 Research publications on the effectiveness of IR approaches in obstetric settings are accumulating with 24 25 26 174 reported success rates as high as 85 to 95 percent.(9-12) Success rates should be interpreted with caution 27 28 175 however, as they may be driven by ascertainment bias. Interventional radiology is gently introduced in 29 30 176 guidelines and recommendations,(6-8, 13, 14) that are describing its possible role in obstetric emergency 31 32 33 177 settings, but awaiting further evidence to make firm recommendations. Yet in the United Kingdom a national 34 35 178 recommendation was made to provide a network of emergency IR services available for all trusts with delivery 36 37 179 units in response to a Healthcare Commission Investigation into maternal deaths.(13, 15) 38 39 180 The aims of this study were to investigate the population-based prevalence of peripartum hysterectomy and 40 41 42 181 the indications to proceed to this intervention, likewise the prevalence of major obstetric haemorrhage 43 44 182 managed with IR in Belgium. Furthermore, we aimed to describe the characteristics of the women in Belgium 45 46 183 who required these interventions to treat or prevent major obstetric haemorrhage, the circumstances in which 47 48 49 184 the interventions took place, the management of the obstetric haemorrhage and the outcome and additional 50 51 185 morbidity of these women. 52 53 186 54 55 56 187 METHOD 57 58 188 Belgian Obstetric Surveillance System (B.OSS) . 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 37 BMJ Open

1 2 3 189 Peripartum hysterectomy and interventional radiology in Belgium has been registered as part of a national 4 5 190 reporting system to study rare obstetric complications, as described previously.(16) Participation of 97 percent 6 7 191 of Belgian maternity units (n= 110/113) was obtained, covering 98.6 percent (n = 248 681) of deliveries in the 8 9 10 192 two-year study period. Briefly, an appointed contact person (OB/GYN or senior-midwife) in each participating 11 12 193 maternity unit was invited by monthly mailing to report a selected number of rare obstetric complications 13 14 194 (including peripartum hysterectomy and/or arterial embolisation) that had occurred in the preceding month or 15 For peer review only 16 195 to state that there was ‘nothing to report’. In the event a case was reported in reply, the contact person was 17 18 19 196 asked to complete an extensive data collection form. Confidentiality was guaranteed for mother, provider and 20 21 197 hospital; person-identifiable information was eliminated from data-analysis. In case of incomplete reporting, 22 23 198 the appointed contact person was encouraged repeatedly by email and phone to provide missing data, up to 24 25 26 199 six months following the period of case reporting. 27 28 200 Belgium has a large relative number of tertiary referral centres providing neonatal intensive care (n=19) along 29 30 201 with a high concentration of small-volume maternity units. All tertiary referral centres have an emergency IR 31 32 33 202 service, as well as some of the larger non-tertiary maternity services. 34 35 203 In answer to an inquiry of B.OSS contact persons, 40 maternity units (38 %) declared to have an IR service 36 37 204 available in the hospital and seven more maternity units declared to have an IR service but not 24/7 or to call 38 39 205 upon the vascular surgeons for similar procedures. 40 41 42 206 43 44 207 Study period . 45 46 208 B.OSS aimed to collect all cases of peripartum hysterectomy and IR that have occurred in Belgium between 47 48 49 209 January 2012 and December 2013. 50 51 210 52 53 211 Case definition. 54 55 56 212 Peripartum hysterectomy generally refers to the surgical removal of the uterus at the time of delivery or within 57 58 213 24 hours following delivery. Alternatively, peripartum hysterectomy has been defined as a hysterectomy 59 60 214 performed any time from delivery until discharge from the same hospitalisation, or until six weeks postpartum. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 37

1 2 3 215 Antenatal diagnosis of a gynaecological cancer or a severe abnormally invasive placenta (AIP) is an occasional 4 5 216 indication for an elective hysterectomy following caesarean delivery, called a ‘caesarean hysterectomy’. Most 6 7 217 peripartum hysterectomies are however emergent, life-saving interventions following vaginal or caesarean 8 9 10 218 delivery, most commonly performed to manage major obstetric haemorrhage and sporadically because of a 11 12 219 severe pelvic infection. Hysterectomy is occasionally required in the first or second trimester, termed abortion 13 14 220 hysterectomy, to prevent or manage major obstetric haemorrhage resulting from an abnormally invasive 15 For peer review only 16 221 placenta or from an ectopic pregnancy in the cornua, the cervix or in a caesarean scar. 17 18 19 222 B.OSS aimed to collect all cases of peripartum hysterectomy and IR at any gestational age corresponding to the 20 21 223 definition: any woman giving birth to a fetus or infant and undergoing a hysterectomy and/or IR in the same 22 23 224 clinical episode. Successful IR was defined as the ability to stop or prevent major obstetric haemorrhage or to 24 25 26 225 avert the need for urgent peripartum hysterectomy. 27 28 226 Registered variables . 29 30 227 Data collection forms questioned maternal characteristics, medical, surgical and obstetrical history, details of 31 32 33 228 the index pregnancy, details of the delivery, circumstances of the adverse event, the management and the 34 35 229 outcome for mother and neonate. 36 37 230 The cause of the obstetric haemorrhage was deducted from the clinical signs reported by the clinician and 38 39 231 assorted according to the 4T’s mnemonic: Tonus, Tissue, Trauma, and Thrombin. Tonus: uterine atony. Tissue: 40 41 42 232 abnormally invasive placenta (AIP), placenta praevia, combination of AIP and placenta praevia, placental 43 44 233 remnants or retained placenta without AIP. Trauma: uterine rupture, genital tract lacerations, bleeding caused 45 46 234 by unintended injuries during caesarean delivery or curettage (D&C); uncomplicated caesarean deliveries were 47 48 49 235 not included in the trauma group. Thrombin: abnormal coagulation before onset of bleeding; DIC secondary to 50 51 236 major haemorrhage was not included in the thrombin group. Others: the cause of bleeding remains unclear or 52 53 237 is unknown. 54 55 56 238 57 58 239 Registered variables of the background population. 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 37 BMJ Open

1 2 3 240 The Belgian perinatal registry (Studiecentrum voor Perinatale Epidemiologie (SPE) in Flanders,(17, 18) Centre 4 5 241 d’Epidémiologie Périnatale (CEpiP) in Brussels (19) and Wallonia (20)) registers data on a mandatory basis, 6 7 242 covering nearly 100 percent of births in Belgian maternity units and home births. A selected set of perinatal 8 9 10 243 data are recorded by the obstetrician, midwife and neonatologist immediately after birth. Births are registered 11 12 244 as such, in case of a live birth or in case of a stillbirth at a birth weight of 500 grams or more and/or after 22 13 14 245 completed weeks of gestation, with the exception of Flanders (SPE) where live births or stillbirths with a birth 15 For peer review only 16 246 weight below 500 grams, irrespective of gestational age, are not registered. 17 18 19 247 20 21 248 Statistical Analysis. 22 23 249 The primary study outcome was the prevalence of peripartum hysterectomy and the prevalence of major 24 25 26 250 obstetric haemorrhage managed with IR in Belgium. The prevalence of the obstetric events targeted was 27 28 251 estimated using as denominator the total number of deliveries in Belgium in 2012 and 2013, corrected for 29 30 252 three hospitals that did not participate in B.OSS (n= 248,681 women). Secondary outcomes were the 31 32 33 253 description of the patient characteristics, the circumstances in which the interventions took place, the 34 35 254 management of the obstetric haemorrhage and the outcome and additional morbidity of these women. 36 37 255 Comparison of the distributions of socio-demographic and obstetric characteristics of reported cases relative to 38 39 256 the background population as obtained from the Belgian perinatal registry (2012-2013) was pursued through 40 41 42 257 use of odds ratios. Odds ratios and accompanying 95 percent confidence intervals (95% CI) were calculated in 43 44 258 univariate analysis, using weighted cases where appropriate. Unadjusted and adjusted odds ratios were 45 46 259 calculated to compare patient and organisational characteristics among interventions (hysterectomy versus IR) 47 48 49 260 in univariate analysis and subsequently through use of unconditional binary logistic regression models. 50 51 261 Comparison between cases with successful and those with failed IR was only made in crude analyses, as 52 53 262 numbers of observations were too small to support logistic regression models. Non-parametric tests were used 54 55 2 56 263 to compare distributions of outcomes between groups, specifically χ for categorical variables, and Kruskal- 57 58 264 Wallis and Mann-Whitney U tests for continuous variables. P-values of <0.05 were considered statistically 59 60 265 significant. Data were analysed using the statistical software package IBM SPSS statistics version 22.0.(21) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 37

1 2 3 266 4 5 6 7 267 RESULTS 8 9 268 Prevalence 10 11 269 Data collection forms of 166 confirmed cases of peripartum hysterectomy and/or IR were retrieved and 12 13 14 270 included in the analysis. Details on reported cases and completeness of data collection forms are presented in 15 For peer review only 16 271 figure 1. 17 18 272 Overall, 84 women (50.6%) underwent a peripartum hysterectomy, corresponding to a prevalence of 3.3/10 19 20 21 273 000 deliveries (95% CI 2.7-4.1). Of these, 17 women (10.2%) underwent a hysterectomy following a previous IR 22 23 274 procedure. Eighty-two women (49.4%) were managed with IR only, corresponding to a prevalence of 3.3/10 24 25 275 000 deliveries, 95% CI 2.6-4.0. Three women (1.8%) needed IR because of persistent bleeding despite 26 27 28 276 hysterectomy. 29 30 277 In 160 women the intervention occurred within 24 hours of delivery. In eight women the intervention occurred 31 32 278 late (from 24 hours to 6 weeks postpartum): two of these data collection forms were not returned, one of 33 34 279 them was the only case in this registry of a hysterectomy performed because of severe peritonitis. All other 35 36 37 280 women underwent hysterectomy and/or IR to treat or prevent major obstetric haemorrhage. The 38 39 281 corresponding prevalence in Belgium of major obstetric haemorrhage requiring peripartum hysterectomy 40 41 282 and/or IR is estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries. 42 43 44 283 45 46 284 Patient Characteristics 47 48 285 We compared the women included in this registry with all the women who gave birth in Belgium during the 49 50 51 286 study period. Patient characteristics of the background population were derived from the Belgian Perinatal 52 53 287 registry. In univariate analysis the women who underwent peripartum hysterectomy and/or IR to treat or 54 55 288 prevent a major obstetric haemorrhage were significantly more likely compared to the background population 56 57 58 289 to be older (≥ 35 years), to be multiparous (parity ≥ 3), to have given birth at a preterm gestational age, to have 59 60 290 had a twin pregnancy, to have had a caesarean delivery in a previous pregnancy, to have had a trial of labour For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 37 BMJ Open

1 2 3 291 after previous caesarean delivery, to have had a caesarean delivery in the index pregnancy and to have 4 5 292 conceived with artificial reproductive technology (ART) (Table 1). Ten women were known to have a uterus 6 7 293 myomatosus (6.0%) and twelve experienced haemorrhage in a previous pregnancy (7.2%). 8 9 10 294 11 12 Table 1. Patient characteristics of women with (imminent) obstetric haemorrhage who underwent peripartum 13 hysterectomy and/or IR compared to all women who gave birth in Belgium during the study period. 14 15 Risk factor For peerCases of H and/or review IR Background only population # Unadjusted OR % 16 (n=166) (n=252 272) for H and/or IR 17 n, % n, % (95% CI) 18 19 Maternal age ≥ 35 years 72 (43.4) 43 256 (17.1) 3.7 (2.7-5.0) § 20 2 £ 21 BMI at booking ≥ 30 kg/m 12 (7.2) 29 453 (11.6) 0.7 (0.3-1.2) 22 Singletons ≥22 weeks Singletons ≥22 weeks 23 Gestational age (weeks) 24 (n=144) (n = 241 136) 25 - Extreme preterm (<28) - 8 (5.4) - 1 221 (0.5) 26 - Very preterm (28-32) - 6 (4.0) - 1 474 (0.6) 7.4 (5.3-10.5) § 27 - Late preterm (32-37) - 37 (24.8) - 13 798 (5.7) 28 - Term (>37) - 93 (62.4) - 224 643 (93.1) Reference 29 Parity ≥ 3 30 (18.0) 18 855 (7.4) 2.7 (1.8-4.1) § 30 31 Multiplets (twin) 13 (7.8) 4690 (1.8) 4.4 (2.5-7.9) § 32 33 & £ § Previous CD 61 (36.7) 27 007 (10.7) 4.8 (3.5-6.6) 34 35 TOLAC ( ≥22 weeks) 14 / 57 (24.5) 12 754 / 27 007 (47.2) 0.3 (0.1-0.66) § 36 37 Delivery Mode ( ≥22 weeks) 38 $ 39 - CD £ § 91 / 160 (56.8) 54 369 (21.5) 4.8 (3.5-6.5) 40 compared to VD

41 - CD before onset of labour 66 / 91 (72.5) 28 586 / 54369 (52.5) 2.3 (1.5-3.7) 42 compared to CD after onset of labour 43 - Assisted VD 44 10 / 69 (14.4) 23 100 /198 444 (11.6) 1.3 (0.6-2.5) compared to spontaneous VD 45 46 ART (IVF/ICSI) 19 (11.4) £ 9233 (3.7) 3.4 (2.1-5.4) § 47 48 Birthweight ≥ 4000 g 11 (6.6) 19967 (7.8) 0.8 (0.4-1.5) 49 CD: Caesarean delivery, VD: Vaginal delivery, H: Hysterectomy, IR: Interventional Radiology , TOLAC: Trial o f Labour After Caesarean Delivery, ART: 50 Artificial Reproductive Technologies, IVF: In Vitro Fertilisation, ICSI: Intracytoplasmic Sperm Injection 51 # Data from the background population were retrieved from SPE reports (Flanders) and CEpiP reports (Brussels and Wallonia) from 2012-2013. 52 % unadjusted Odds ratios (with 95% Confidence Interval) for peripartum hysterectomy and/or IR calculated for the different risk factors: the odds 53 (=probability of peripartum hysterectomy and/or IR occurring divided by the probability of H/IR not occurring) in the women having the risk factor, 54 divided by the odds in the women not having the risk factor. 55 £ Missing data in cases of hysterectomy and/or IR : BMI (n=27), previous CD (n=1), mode of delivery (n=1), ART (n=2), haemorrhage in previous 56 pregnancy (n=5), uterus myomatosus (n=5) 57 & Women with previous Caesarean Delivery: 9 with Placenta Praevia, 9 with AIP, 17 with Placenta Praevia and AIP 58 $Women with Caesarean Delivery as delivery mode: 17 withPlacenta Praevia, 8 with AIP, 23 with Placenta Praevia and AIP 59 Reason for Caesarean Delivery: Bleeding (Placenta Praevia, AIP, abruptio placentae): n= 15; Placenta Praevia / AIP not bleeding: n= 29; Maternal shock / 60 cardiac arrest: n=1; Other: n = 46 § p < 0.05 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 37

1 2 3 295 4 5 296 Circumstances of the obstetric haemorrhage: 6 7 297 IR service availability, antenatal and postnatal transfers, urgent and planned interventions. 8 9 10 298 The circumstances in which the interventions occurred in these 166 women are delineated in Figure 2. Three 11 12 299 women were transferred antenatally to another hospital with emergency IR service availability. In result 107 13 14 300 women gave birth in a maternity unit with emergency IR service availability, 20 women gave birth in a 15 For peer review only 16 17 301 maternity unit where IR service is available but not 24/7 or where obstetricians can call upon vascular surgeons 18 19 302 for IR procedures, 39 women gave birth in a maternity unit with no IR service available. 20 21 303 Seventeen women were transferred postnatally to another hospital with emergency IR service availability and 22 23 24 304 underwent an urgent IR procedure, including one woman for persistent bleeding despite hysterectomy and 25 26 305 one woman for persistent bleeding despite planned intra-arterial balloon occlusion and arterial embolisation 27 28 306 by the vascular surgeons in the referring hospital. Three of the postnatally transferred women consecutively 29 30 307 underwent an urgent hysterectomy for persistent bleeding despite arterial embolisation. 31 32 33 308 Considering the three patients who were transferred antenatally, one woman underwent an urgent 34 35 309 embolisation upon arrival because of a bleeding placenta praevia, one woman was planned for caesarean 36 37 310 hysterectomy but underwent an urgent caesarean hysterectomy because of a bleeding AIP before the fixed 38 39 40 311 date, one woman underwent a planned Caesarean delivery combined with a planned IR procedure because of 41 42 312 AIP. Seven more women were transferred postnatally to the intensive care unit of a tertiary referral centre. 43 44 313 Overall, the intervention was planned for imminent obstetric haemorrhage in 15 women, including 1 planned 45 46 47 314 hysterectomy, 7 planned IR procedures, 2 planned hysterectomies supported by a planned IR procedure, 2 48 49 315 planned IR procedures followed by a planned hysterectomy with one month delay and 3 planned IR procedures 50 51 316 that resulted in an urgent hysterectomy. The indication to plan these interventions was abnormally invasive 52 53 54 317 placenta in 13 women and placenta praevia in 2 women. 55 56 318 57 58 319 Reported cause of the obstetric haemorrhage 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 37 BMJ Open

1 2 3 320 Eighteen women (11%) presented with antepartum haemorrhage due to abnormal placentation (placenta 4 5 321 praevia, AIP or both) or abruptio placentae (n=2). Four women (2%) started bleeding during termination of 6 7 322 pregnancy (<24 weeks) due to a cervical pregnancy (n=1) or an AIP (n=3). Hundred thirty-eight women (83%) 8 9 10 323 presented with early postpartum haemorrhage within 24 hours postpartum and six (4%) with late postpartum 11 12 324 haemorrhage from 24 hours to 6 weeks postpartum. 13 14 325 The reported clinical signs put forward as cause of the obstetric haemorrhage were uterine atony (n=91, 15 For peer review only 16 326 54.8%) and abnormally invasive placenta (AIP) (n=38, 22.8%) in the majority of women. In 54 women (32.5%) 17 18 19 327 more than one cause was reported, with atony and abnormal placentation (AIP and/or placenta praevia) in 20 20 21 328 women (12%) as the most frequent combination. The cause of the (imminent) obstetric haemorrhage assorted 22 23 329 according to the 4T’s mnemonic is demonstrated in Figure 3. 24 25 26 330 27 28 331 Management of the obstetric haemorrhage 29 30 332 Hysterectomy was performed in all the cases of uterine rupture (n=12) and in 73.7 percent of the cases of 31 32 33 333 abnormally invasive placenta (n=28/38). Interventional radiology could cease the bleeding and avert the need 34 35 334 for peripartum hysterectomy in the majority of cases of haemorrhage reported to be associated with genital 36 37 335 tract lacerations (n=10/12, 83.3%), placental remnants or retention (n=17/26, 65.4%) and uterine atony 38 39 40 336 (n=28/43, 65.1%). 41 42 337 Other measures that were attempted to manage the obstetric haemorrhage before or while proceeding to the 43 44 338 hysterectomy or IR are listed in Table 2. The first line measures being the administration of uterotonic agents 45 46 47 339 and bimanual uterine compression, other second line measures being intra-uterine (balloon) tamponade, 48 49 340 haemostatic brace suturing (B-lynch or other) and ligation of the uterine or internal iliac arteries, besides blood 50 51 341 transfusion and the administration of fresh frozen plasma (FFP), platelets or clotting factors. An intra-uterine 52 53 54 342 balloon was inserted in 35 women (21%) of whom 7 women before transfer to another hospital with 55 56 343 availability of an IR service. Brace suturing and arterial ligation was attempted in 6 and 4 women respectively, 57 58 344 reflecting the infrequent use of these procedures in Belgium. 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 37

1 2 3 345 In 19 women (11.4%) no other measures were undertaken besides hysterectomy (n=13), IR (n=2) and 4 5 346 hysterectomy following IR (n=4). The reported cause of haemorrhage in these women was abnormally invasive 6 7 347 placenta (n=12), placenta praevia (n=2), uterine rupture (n=2), retained placenta (n=2) and one case of a late 8 9 10 348 postpartum haemorrhage. The intervention was planned in five of these women. Eight of them had minor 11 12 349 haemorrhage and did not require transfusion, while three women undergoing urgent peripartum hysterectomy 13 14 350 needed massive transfusion (defined as receiving eight or more units of red blood cells). 15 For peer review only 16 351 17 18 19 Table 2. First and second line measures in women undergoing hysterectomy (H) and /or IR (IR) according to the 20 circumstances of the obstetric haemorrhage. 21 Total Planned H Urgent H and/or IR 22 and/or IR n=151 23 IR service IR service No IR service Postnatal 24 24/7 not 24/7 Transfer 25 26 n=166 n=15 n=98 n=12 n=25 n=16* 27 n (%) n (%) n (%) n (%) n (%) n (%) 28 Cause of obstetric haemorrhage 29 30 31 Abnormal placentation 56 (34) 15 (100) 29 (30) 5 (42) 5 (20) 2 (13) 32 Atony only 43 (26) - 29 (30) 5 (42) 7 (28) 2 (13) 33 Uterine rupture 12 (7) - 6 (6) 1 (8) 5 (20) - 34 Miscellaneous £ 55 (33) - 34 (34) 1 (8) 8 (32) 10 (62) 35 36 First and second line measure 37 38 Uterotonics 134 (81) 10 (67) 81 (83) 10 (83) 19 (76) 15 (94) 39 Oxytocin 123 (74) 10 (67) 73 (75) 9 (75) 18 (72) 13 (81) 40 PG 109 (66) 5 (33) 64 (65) 17 (68) 8 (67) 15 (94) 41 Cytotec 61 (37) 1 (7) 34 (35) 6 (50) 10 (40) 10 (63) 42 43 Prostin 15M 63 (39) 5 (33) 42 (43) 3 (25) 9 (25) 9 (56) 44 Oversewing 30 (18) 1 (7) 19 (20) 1 (8) 7 (28) 2 (13) 45 (Bimanual) compression 64 (39) 3 (20) 36 (37) 5 (42) 9 (36) 11 (69) 46 Balloon tamponade 35 (21) 2 (13) 21 (22) 2 (17) 3 (12) 7 (44) 47 48 B-lynch or ligation 10 (6) - 5 (5) 1 (8) 4 (16) - 49 Transfusion 139 (84) 7 (47) 85 (87) 12 (100) 21 (84) 15 (88) 50 Massive transfusion $ 51 (31) 2 (13) 24 (24) 7 (58) 10 (40) 8 (50) 51 FFP 108 (65) 4 (27) 64 (65) 11 (92) 18 (72) 11 (69) 52 Platelets 57 (34) 1 (7) 29 (30) 9 (75) 10 (40) 8 (50) 53 54 Clotting factors % 23 (14) - 8 (8) 5 (42) 7 (28) 3 (19) 55 * On e woman was transferred postnatally for persistent bleeding despite planned IR in the referring hospital; she was included in the ‘planned’ group. 56 £ Miscellaneous group: Placental remnants or retention n=25, Surgical trauma n=16, Genital tract laceration n=11, Intra-abdominal arterial bleeding n=4, 57 Coagulation disorders n= 3, Unknown n= 6. The total number is exceeding n = 55 due to overlap in between cases. 58 $ Massive transfusion: transfusion of 8 or more units of red blood cells % 59 Clotting factors: Fibrinogen, recombinant Factor VIIa, PPSB (Prothrombin-Proconvertin-Stuart Factor-Antihemophilic Factor B) or a combination 60 352 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 37 BMJ Open

1 2 3 353 Management of abnormally invasive placenta (AIP) 4 5 354 The management of the (imminent) obstetric haemorrhage in the women with abnormally invasive placenta 6 7 355 (AIP) (n=38) is illustrated in Figure 4. Seventeen women (45%) were susceptible for AIP due to previous 8 9 10 356 caesarean delivery (CD) and placenta praevia in the index pregnancy, another 7 women (18%) had placenta 11 12 357 praevia in the index pregnancy and 9 women (24%) had previous caesarean delivery. In 25 women (65.7%) 13 14 358 there was antenatal suspicion of the AIP, of whom 16 had placenta praevia and previous caesarean delivery, 5 15 For peer review only 16 17 359 had placenta praevia and 4 had previous caesarean delivery only. All women with antenatal suspicion of AIP 18 19 360 were planned to undergo an elective caesarean delivery in a maternity unit with IR service availability (24/7 20 21 361 n=20, not 24/7 n=5), which occurred in 21 women. Yet 3 women had an emergency caesarean delivery because 22 23 24 362 of antepartum haemorrhage and 1 woman delivered vaginally following acute onset of labour. Further 25 26 363 preventive measures were undertaken in 13 women (31.5%): antenatal transfer to a centre with IR service 27 28 364 availability (n=2), placement of intra-arterial catheters pre-caesarean delivery (n=12), planned caesarean 29 30 365 hysterectomy (n=3) or planned hysterectomy at a later stage (n=2). 31 32 33 366 In 13 women (34%) there was no antenatal suspicion of the AIP. In retrospect all of them had risk factors: 1 34 35 367 woman had placenta praevia and previous caesarean delivery, 4 women had placenta praevia, 6 had previous 36 37 368 caesarean delivery, 1 woman had previous myomectomy and adhesiolysis for Asherman syndrome and 38 39 40 369 another woman had a previous manual removal of the placenta. Nine women delivered in a maternity unit 41 42 370 with IR service availability (24/7 n=7, not 24/7 n=2) and 4 delivered in a maternity unit with IR service. Mode of 43 44 371 delivery in those 13 women who had no antenatal suspicion of AIP was elective caesarean delivery in 2 women, 45 46 47 372 emergent caesarean delivery in 5 women and vaginal delivery in 6 women. 48 49 373 The placenta was left in situ in 10 women with AIP: 3 in an emergency setting where 1 woman underwent 50 51 374 immediate hysterectomy and 2 women underwent arterial embolisation followed by urgent hysterectomy. 52 53 54 375 Another 7 women had their placenta left in situ in an elective setting combined with planned IR: 2 women 55 56 376 consecutively underwent planned hysterectomy, 2 underwent an urgent hysterectomy within 24 hours, 2 57 58 377 underwent a planned hysterectomy after one month and 1 woman underwent elective evacuation of retained 59 60 378 products of conception after one month. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 37

1 2 3 379 As anticipated, the women for whom measures were planned antenatally had significant better outcome than 4 5 380 women where AIP was diagnosed peripartum: less estimated blood loss (median 625 mL (range 300-1250) 6 7 381 versus median 4500 mL (range 500-6000), p=0.08), less need for transfusion (46.2% versus 92.3%, p=0.056), 8 9 10 382 less need for massive transfusion (≥8 units) (8.3% versus 53.8%, p=0.045) and less hysterectomies (61.5% 11 12 383 versus 76.9%, p=0.6). The maternal outcome was not significantly different when comparing women with AIP 13 14 384 undergoing IR as first intervention versus hysterectomy as first intervention: estimated blood loss (median 15 For peer review only 16 385 4000 mL (range 500-15000) versus median 1000 mL (range 300-8500), p=0.3) need for transfusion (76.5% 17 18 19 386 versus 70%, p=0.9), need for massive transfusion (≥8 units) (29.4% versus 30%, p=1.0). All women with planned 20 21 387 interventions were managed in tertiary referral centres. 22 23 388 24 25 26 389 Motive to attempt interventional radiology. 27 28 390 The decision of an obstetrician in the acute moment dealing with a major obstetric haemorrhage to attempt IR 29 30 391 rather than turning to a hysterectomy immediately, will be led by many clinical and organisational factors. 31 32 33 392 Some risk factors that may have contributed to this decision in this series are presented in Table 3. The women 34 35 393 with double interventions (n=20) were excluded from this analysis to facilitate interpretation. 36 37 394 High parity (≥ 3), previous caesarean delivery, caesarean delivery in the index pregnancy, abnormally invasive 38 39 40 395 placenta and the lack of emergency IR service availability are significantly positively associated with peripartum 41 42 396 hysterectomy in univariate analysis. In unconditional binary logistic regression, the odds to undergo a 43 44 397 hysterectomy for obstetric haemorrhage is 7.2 times higher in women with high parity and 8.2 times higher in 45 46 47 398 women delivering in a hospital with no emergency IR service available, independent of the other risk factors in 48 49 399 the model (Hosmer and Lemeshow Test p =0.67). 50 51 400 We repeated the analysis in a subgroup, namely the women who delivered in a maternity unit with an 52 53 54 401 emergency IR service available, with exclusion of the women with planned interventions, consisting of 31 55 56 402 women who underwent hysterectomy only and 60 women who underwent embolisation only. In 57 58 403 unconditional binary logistic regression, the odds to undergo a peripartum hysterectomy for obstetric 59 60 404 haemorrhage is significantly higher in women with high parity (odds 4.5 (95% CI 1.2-16.5)) and nearly For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 37 BMJ Open

1 2 3 405 significantly higher in women with previous caesarean delivery (odds 3.7 (95% CI 0.98-14.6)) independent of 4 5 406 the other risk factors in the model (Hosmer and Lemeshow Test p = 0.69). 6 7 407 8 9 10 Table 3. Risk factors possibly contributing to the choice of intervention in women undergoing hysterectomy or 11 interventional radiology. Unadjusted and adjusted OR (95% CI) for hysterectomy (only) per risk factor . 12 Risk factors Number of women Unadjusted OR Adjusted OR 13 14 Hysterectomy Interventional (95%CI) (95%CI) 15 For peeronly reviewradiology only only for for 16 n=64 n=82 Hysterectomy Hysterectomy 17 n, % n, % only only 18 Demographic factors 19 20 Maternal age ≥ 35 years 32 (50.0) 30 (36.5) 1.7 (0.8-3.3) 21 BMI at booking ≥ 30 kg/m2 6 (9.3) 3 (3.6) 2.7 (0.6-11.3) 22 Obstetric factors 23 Preterm (<37w) 28 (43.8) 24 (29.3) 1.8 (0.9-3.7) 24 Parity ≥ 3 18 (28.1) 7 (8.5) 2.7 (1.4-9.4) § 7.2 (2.1-24.6) § 25 § 26 Previous caesarean delivery 31 (48.4) 18 (21.9) 3.3 (1.6-6.8) 2.3 (0.9-6.3) § 27 Caesarean delivery ( ≥22w) * 38 / 59 (64.4) 37 / 82 (45.1) 2.2 (1.1-4.3) 1.6 (0.6-4.0) 28 Cause of bleeding 29 AIP and AIP+ praevia 16 (25.0) 9 (10.9) 2.7 (1.1-6.6) § 3.0 (0.9-10.0) 30 Atony 29 (45.3) 52 (63.4) 0.4 (0.2-0.9) § 0.9 (0.3-2.3) 31 32 Organisational factors 33 Weekend and/or non-office hours # 24/60 (40.0) 34/80 (42.5) 0.9 (0.4-1.7) 34 IR NOT available $ 30 (46.8) 15 (18.2) 3.9 (1.8-8.2) § 8.2 (3.1-21.6) § 35 * Exclusion of 5 cases < 22 weeks 36 £ Missing data: BMI (n=27), previous caesarean delivery (n=1), Delivery Mode (n=1), placenta location (n=13), Date of delivery (n=2), Time of 37 delivery (n=5), week/weekend (n=3) 38 # based on date and time of delivery 39 IR : interventional radiology 40 $ Emergency IR service availability in the maternity unit where the woman gave birth 41 § p<0.05 42 408 43 44 45 409 46 47 410 Peripartum hysterectomy despite interventional radiology 48 49 411 One hundred and two women in this registry underwent IR, of whom three women because of persistent 50 51 52 412 bleeding despite peripartum hysterectomy and 99 women as primary intervention. The IR procedure could 53 54 413 prevent or cease the bleeding in 87 of them (87.8%). Nevertheless, five of these 87 women had a planned 55 56 414 hysterectomy performed consecutively because of abnormally invasive placenta, two of them with a delay of 57 58 59 415 one month after the caesarean delivery and planned IR. In the other twelve women an urgent hysterectomy 60 416 was performed because of persistent bleeding despite the IR procedure. We compared the characteristics of For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 37

1 2 3 417 these 12 women who needed urgent peripartum hysterectomy despite IR with the characteristics of the 87 4 5 418 women in whom IR was successful including those five women with a planned hysterectomy. In univariate 6 7 419 analysis, an increased risk of urgent hysterectomy despite IR was observed for women with caesarean delivery 8 9 10 420 in a previous pregnancy, for women with abnormally invasive placenta and for women who had no uterotonic 11 12 421 agents administered. These findings are presented in Table 4. 13 14 Table 4. Risk factors possibly contributing to failure of IR (IR). Unadjusted odds ratio’s (95% CI) 15 For peer review only 16 for failed IR per risk factor. 17 Riskfactors Number of women Unadjusted OR 18 Failed IR Successful IR (95% CI) 19 20 n=12 n=87 for failed IR 21 n, % n, % 22 Demographic factors 23 Maternal age ≥ 35 years 5 (41.7) 33 (37.9) 1.1 (0.3-3.9) 24 2 25 BMI at booking ≥ 30 kg/m 2 (18.2) 3 (3.9) 5.6 (0.8-37.6) 26 Obstetric factors 27 Twins 1 (8.3) 6 ( 6.9) 1.2 (0.1-11.1) 28 29 Preterm gestational age < 37w 7 (58.3) 28 (32.2) 2.9 (0.8-10.1) 30 Preterm gestational age < 34w 3 (25.0) 6 (6.8) 4.5 (0.9-21.1) 31 Parity ≥3 3 (25.0) 8 (9.2) 3.2 (0.7-14.6) 32 33 Previous caesarean delivery 7 (58.3) 21 (24.1) 4.4 (1.2-15.3) § 34 Caesarean delivery (≥22 weeks) 8 (72.7) 42 (48.3) 2.1 (0.6-7.6) 35 Induction of labour 3 (25.0) 25 (28.7) 0.8 (0.2-3.3) 36 Augmentation of labour 2 (16.7) 26 (29.9) 0. (0.09-2.2) 37 38 Cause of haemorrhage 39 Atony 7 (63.6) 52 (62.7) 0.9 (0.2-3.2) 40 AIP and AIP+ praevia 7 (58.3) 14 (16.1) 7.3 (2.0-26.3) § 41 Surgical trauma 2 (16.7) 6 (6.9) 2.7 (0.4-15.2) 42 43 Genital tract laceration 1 (8.3) 10 (11.5) 0.7 (0.08-6.0) 44 1st and 2 nd line treatment 45 46 Uterotonic agents 8 (66.7) 82 (94.3) 0.1 (0.02-0.5) § 47 No use of uterotonic agents 4 (33.3) 5 (5.7) 8.2 (1.8-36.8) § 48 49 Intra-uterine balloon 3 (25.0) 26 (29.9) 0.7 (0.1-3.1) 50 Clotting factors IV 2 (16.7) 8 (9.2) 1.9 (0.3-10.6) 51 52 FFP 7 (58.3) 54 (62.1) 0.8 (0.2-2.9) 53 Organisational factors 54 Planned intervention 2 (14) 10 (12) 1.5 (0.3-8.0) 55 # 56 Weekend and/or non-office hours 5 / 10 (50.0) 51 / 85 (60.0) 0.6 (0.1-2.4) 57 IR NOT available $ 3 (25.0) 10 (11.4) 2.5 (0.6-11.1) 58 Emergency IR NOT available $ 6 (50.0) 20 (22.9) 3.3 (0.9-11.5) 59 60 Postnatal transfer 3 (25.0) 13 (14.9) 1.9 (0.4-7.9) Combination of riskfactors For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 37 BMJ Open

1 2 3 Previous CD + AIP 6 (50.0) 7 (8.0) 11.4 (2.9-44.9) § 4 Caesarean delivery + AIP 5 (41.6) 11 (12.6) 4.9 (1.3-18.2) § 5 No use of uterotonic agents + AIP 4 (33.3) 10 (13.7) 3.8 (0.9-15.1) 6 7 IR: interventional radiology; AIP: abnormally invasive placenta; FFP: Fresh Frozen Plasma ; CD: Caesarean delivery # based on date and time of delivery 8 $ 9 IR service availability (24/7 and not 24/7) in the maternity unit where the woman gave birth: $ 10 Emergency IR service availability (24/7) in the maternity unit where the woman gave birth 11 § p < 0.05 12 13 422 14 15 423 For peer review only 16 17 424 Maternal outcome and additional morbidity 18 19 20 425 Blood transfusion was given to 139 women (83.7%, missing data for 2 women), with a median of 6.0 units of 21 22 426 red blood cells (RBC) (range 0-31), 4.0 units of fresh frozen plasma (FFP) (range 0-31) and 0.5 units of platelets 23 24 427 (range 0-29). Fifty-one women (30.7%) were transfused ≥ 8 units of RBC and 23 women were administered 25 26 27 428 clotting factors (fibrinogen, recombinant factor VIIa, PPSB (Prothrombin-Proconvertin-Stuart Factor- 28 29 429 Antihemophilic Factor B), or a combination) . Hundred-fourteen women (68.6%, missing data in 4 women) were 30 31 430 admitted to an intensive care unit (ICU) for a median of 2.0 days (range 0-18 days); the median hospital stay 32 33 34 431 was 8.0 days (range 2-33 days). Five women needed relaparotomy and 3 women needed arterial embolisation 35 36 432 following hysterectomy because of persistent bleeding. The cause of haemorrhage in these women was AIP 37 38 433 (n=1), AIP and atony (n=2), atony (n=3) and in 1 case the cause was unclear (n=1). 39 40 41 434 Disseminated intravascular coagulation (DIC) secondary to major haemorrhage was reported in 32 women. 42 43 435 Other reported maternal complications associated with major haemorrhage were: renal failure (n=2), 44 45 436 pulmonary oedema (n=4), acute respiratory distress syndrome (n=2), myocardial ischemia (n=1), pleural and 46 47 437 pericardial effusion (n=1). Complications associated with peripartum hysterectomy were ureteral injuries (n=2) 48 49 50 438 and ileus (n=4). One woman suffered from pain in loins and legs, this was the only reported complication 51 52 439 attributed to the IR procedure. 53 54 440 One mother died (case fatality rate 0.6% (95% CI 0.1-3.3)). She underwent an emergency caesarean delivery 55 56 57 441 because of cardiac arrest caused by massive pulmonary embolism. She survived the hysterectomy performed 58 59 442 because of major bleeding due to uterine atony and was transferred to the intensive care unit of a tertiary 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 37

1 2 3 443 referral centre using Extracorporeal Life Support (ECLS). She underwent a re-laparotomy because of major 4 5 444 hepatic haemorrhage, but died the subsequent day during pulmonary thrombectomy. 6 7 445 Considering the whole group, the outcome in women who underwent hysterectomy did not significantly differ 8 9 10 446 from women who underwent IR, regarding transfusion rate and ICU admission, except for a significant 11 12 447 difference in total duration of hospitalisation (median 9 days, range (2-30) versus median 7 days (range 2-33)). 13 14 448 When comparing the women according to whether the intervention was planned, the number of women who 15 For peer review only 16 449 underwent a hysterectomy was comparable (8 women (53%) in the planned versus 76 women (50%) in the 17 18 19 450 urgent group), but there was a significant difference in estimated blood loss (median 560 mL (range 300-2000) 20 21 451 versus median 2500 mL (range 400-15000), p=0.0) and need for transfusion (median number of units 0.5 22 23 452 (range 0-11) versus 5.5 (range 0-31), p=0.002) in the planned versus the urgent group. 24 25 26 453 When comparing the women according to the emergency IR service availability in the maternity unit where 27 28 454 they gave birth, the number of women who underwent a peripartum hysterectomy (29 women (74%) 29 30 455 compared to 44 women (41%)) and the number of women who received massive transfusion (defined as ≥ 8 31 32 33 456 units of RBC) (17 women (45%) compared to 25 women (24%)) was significantly higher when emergency IR 34 35 457 service was not available compared to when IR service was available 24/7. 36 37 458 When comparing the women who gave birth in a maternity unit without emergency IR service availability 38 39 459 according to whether they were transferred postpartum, there was no significant difference in estimated blood 40 41 42 460 loss (median 2850 mL (range 500-5000) versus median 3500 mL (range 3000-5000), p=0.4) and need for 43 44 461 transfusion (median number of units 6.5 (range 0-19) versus 9.5 (range 0-31), p=0.2), but all women who were 45 46 462 not transferred (n=22/22) underwent a hysterectomy compared to 28 percent (n=4/14) in the transferred 47 48 49 463 group (p=0.0). 50 51 464 52 53 465 DISCUSSION 54 55 56 466 The main findings of this study are, firstly, that the prevalence in Belgium of major obstetric haemorrhage 57 58 467 requiring peripartum hysterectomy and/or IR is estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries. 59 60 468 Secondly, that AIP was not detected antenatally in 34 percent of the cases in this series. Thirdly, that women For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 37 BMJ Open

1 2 3 469 with major obstetric haemorrhage who gave birth in a maternity unit without emergency IR service were more 4 5 470 likely to undergo an urgent peripartum hysterectomy (74%) compared to women who delivered in a maternity 6 7 471 unit with 24/7 IR service availability (41%) and, fourthly, that the need for urgent peripartum hysterectomy was 8 9 10 472 averted in 72 percent of the women who delivered in a maternity unit without emergency IR service and who 11 12 473 were transferred postnatally, with no significant difference in estimated blood loss and transfusion rate, 13 14 474 compared to the women who were not transferred. 15 For peer review only 16 475 This study is a nationwide population-based report of the use of IR to prevent or treat postpartum 17 18 19 476 haemorrhage and avert the need for peripartum hysterectomy. Strengths of the study are its methodology by 20 21 477 collecting cases prospectively on a monthly basis and its national design with excellent participation rate and 22 23 478 response rate (98.9%) of the Belgian maternity units. The small number of cases, intrinsic to rare but severe 24 25 26 479 obstetric complications, warrant caution however in making strong inferences. The numbers were too small to 27 28 480 support multivariate analysis of failure of IR; accordingly the correlations found are uncertain because of low 29 30 481 statistical power and risk of confounding. Underreporting of cases, due to the non-mandatory reporting of 31 32 33 482 severe pregnancy complications may have biased our results to some extent. We believe that the ambiguous 34 35 483 meaning of ‘clinical episode’, which was not clearly defined in our study definition, may have led to 36 37 484 underreporting of cases. A clinical episode can be interpreted either as ‘until discharge from the same 38 39 485 hospitalisation’ or as ‘until 6 weeks postpartum, being the puerperium’. Furthermore, the addition of the study 40 41 42 486 of IR besides peripartum hysterectomy may have restrained clinicians from reporting cases of hysterectomy for 43 44 487 other causes than postpartum haemorrhage. Other limitations of this study are the inability to control for 45 46 488 several variables not available from the background population through the Belgian perinatal registry. 47 48 49 489 Furthermore, we have not systematically requested technical details of the IR procedures (which vessel, which 50 51 490 material), the hysterectomy procedure (total versus subtotal) nor the histopathology results of the uterus. 52 53 491 Lastly and most importantly, one should be aware that obstetricians are forced to make rapid decisions when 54 55 56 492 dealing with a major obstetric haemorrhage. Belgian obstetricians cannot rely on national guidelines for 57 58 493 postpartum haemorrhage or abnormally invasive placenta. Small volume maternity units even lack local 59 60 494 hospital guidelines for these matters. Besides, there is no emergency IR service available in more than half of For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 37

1 2 3 495 the maternity units. Therefore, the management of major obstetric haemorrhage will differ greatly in between 4 5 496 hospitals and in between gynaecologists. This series describes how women with major obstetric haemorrhage 6 7 497 were managed according to the circumstances of the event, but does not allow making firm conclusions on the 8 9 10 498 motivation or the appropriateness of this management. 11 12 499 Abnormally invasive placenta (AIP) accounts for 23 percent of the women and 86 percent of the planned 13 14 500 interventions in this registry. Targeted papers report good sensitivity and positive predictive value (PPV) of 15 For peer review only 16 501 grey-scale ultrasound as the screening tool for AIP in women with a high index of suspicion, enhanced by the 17 18 19 502 use of transvaginal ultrasound and colour Doppler.(22, 23) Nonetheless, the diagnosis of AIP was missed 20 21 503 antenatally in 34 percent of the women in this study, while all of them had risk factors that should have 22 23 504 increased awareness. Several population-based studies report rather low antenatal diagnosis rates of AIP: 50 24 25 26 505 percent of AIP cases were missed in a study of the UK Obstetric Surveillance System (UKOSS), of these 30 27 28 506 percent had a previous caesarean delivery and placenta praevia.(24) Similarly, in the Nordic Obstetric 29 30 507 Surveillance Study (NOSS) 70 percent of AIP cases were missed, while 33 percent had placenta praevia and 39 31 32 33 508 percent had previous caesarean delivery.(25) Knowing this, we believe that antenatal diagnosis of AIP may be 34 35 509 improved by implementing routinely screening in all pregnant women. Firstly, by being aware of risk factors 36 37 510 and actively asking for previous uterine surgery at the first antenatal visit. Secondly, by the evaluation of the 38 39 511 position of the placenta at the moment of the second trimester ultrasound in all pregnant women. Thirdly, by 40 41 42 512 the evaluation of ultrasonographic features suggestive of AIP (26) in all pregnant women, with great attention 43 44 513 in women with risk factors. Importantly, there is little information about the sensitivity or positive predictive 45 46 514 value (PPV) of ultrasound or MRI as the screening tool for AIP in women who had no previous uterine surgery 47 48 49 515 or in women in whom the placenta is not implanted anteriorly in the lower uterine segment,(22) which would 50 51 516 be an interesting topic for future research. Despite routine screening a number of cases of abnormal 52 53 517 placentation will still be missed, therefore all maternity units should ensure a state of preparedness to deal 54 55 56 518 with an unanticipated AIP causing a major haemorrhage at any time. 57 58 519 Although numbers are small, this study aligns with the findings in previous studies that maternal outcome is 59 60 520 better when women with suspicion of AIP give birth in specialised tertiary centres and when preventive For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 37 BMJ Open

1 2 3 521 measures are taken.(24, 27, 28) A planned preterm caesarean hysterectomy with the placenta left in situ is 4 5 522 the indubitable treatment of choice in women with AIP incorporated in recommendations of RCOG,(14) ACOG, 6 7 523 (29) and SMFM.(30) The placenta was left in situ in ten women with AIP (26%) in this registry, of whom seven 8 9 10 524 in an elective setting combined with planned IR. Two of these women underwent a planned hysterectomy with 11 12 525 one month delay after the caesarean delivery. A small number of papers support this conservative 13 14 526 management, leaving the placenta in situ with a planned delayed hysterectomy, in women having a placenta 15 For peer review only 16 527 percreta with infiltration in the bladder and/or other surrounding tissues and therefore high risk of 17 18 19 528 haemorrhage or adjacent tissue injury.(31-33) The rationale being that this delay will allow the decrease of the 20 21 529 peri-uterine vascularisation in order to prevent haemorrhage. Studies with larger numbers are needed to 22 23 530 increase evidence of this methodology claimed to be safer compared to caesarean hysterectomy in women 24 25 26 531 with severe placenta percreta. 27 28 532 The role of IR in the management of AIP is still a matter of debate. Most guidelines agree that IR can be 29 30 533 lifesaving in the treatment of major postpartum haemorrhage.(14, 30) Guidelines currently do not recommend 31 32 33 534 the routine use of pre-operative placement of intra-arterial catheters to enable balloon occlusion or arterial 34 35 535 embolisation before caesarean hysterectomy or conservative treatment of AIP, based on lack of evidence of 36 37 536 benefit.(14, 29, 30, 34) The number of papers reporting benefit of prophylactic IR is growing ever since.(35, 36) 38 39 537 Our population-based study reports a diverse management of women with AIP, inherent to a multi-centre 40 41 42 538 study. Twenty-one women with AIP were managed with IR, in an elective setting in 12 women and in an 43 44 539 emergency setting in 9 women. The need for hysterectomy was averted in 10 women. We cannot make firm 45 46 540 conclusions on the success rate of IR in the elective setting prior to planned caesarean hysterectomy or 47 48 49 541 conservative management, because of small numbers and lack of controls. 50 51 542 Other than AIP, most cases of postpartum haemorrhage are not antenatally predictable. This study confirms 52 53 543 the existence of an informal network of emergency IR services in Belgium: twenty women (12%) were 54 55 56 544 transferred to a centre with IR availability, of whom 3 antenatally and 17 postnatally. This study demonstrates 57 58 545 that the obstetricians were significantly more inclined to perform a hysterectomy in multiparous women 59 60 546 (parity ≥ 3) and when emergency IR service was not available. This result may not surprise, moreover, For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 37

1 2 3 547 performing a hysterectomy sooner rather than later must be considered good practice,(6) because losing time 4 5 548 trying another conservative measure can cause further blood loss and further morbidity. The decision is more 6 7 549 easily made in women having no further fertility desire, likewise, the transfer of a hemodynamically unstable 8 9 10 550 patient to an IR service is hardly ever considered a safe alternative if this IR service is located in a nearby 11 12 551 hospital. A case-by-case in-depth analysis would be necessary to reveal whether the hysterectomies performed 13 14 552 in this study were appropriate or preventable. 15 For peer review only 16 553 This study demonstrated a high success rate (87.8%) of IR in controlling postpartum haemorrhage, consistent 17 18 19 554 with success rates reported by others, ranging between 85 percent and 95 percent.(9-12, 37-39) Interventional 20 21 555 radiology was more likely to fail in women with a caesarean delivery in a previous pregnancy (failure rate 25%), 22 23 556 women with AIP (33%) and women who had no uterotonic agents administered (44%) in univariate analysis, 24 25 26 557 taking into account the small numbers. Several large hospital-based retrospective studies analysed predictive 27 28 558 factors associated with failure of IR in the management of postpartum haemorrhage.(12, 37-40) Most papers 29 30 559 report a lower success rate of IR in women with AIP,(37, 38) which explains to a great extent the lower success 31 32 33 560 rates in women with previous caesarean delivery and in women with caesarean delivery in the current 34 35 561 pregnancy, both associated with AIP. A population-based study in the Netherlands (10) reported a 25 percent 36 37 562 failure rate for IR following caesarean delivery in the current pregnancy, which was just 16 percent in this study 38 39 563 and barely 10 percent in a large study reported by Lee et al. (n=176).(41) Other researchers consistently 40 41 42 564 reported a higher failure rate of IR in women with different clinical determinants related to major 43 44 565 haemorrhage: hemodynamic shock,(40, 42) diffuse intravascular coagulopathy,(12, 42), low fibrinogen and 45 46 566 prothrombin time (38) and need for massive transfusion.(12, 37, 38) Yet these factors may be both cause and 47 48 49 567 consequence of failure of IR, which was not clearly differentiated in some of the aforementioned studies. Key 50 51 568 message of this knowledge is to start transfusion of packed cells, fresh frozen plasma and platelets in time, to 52 53 569 avoid DIC and hemodynamic shock and consequent failure of IR. Interestingly, hospital to hospital transfer was 54 55 56 570 not associated with a higher failure rate of IR in this study, as reported by others.(12, 37, 38) 57 58 571 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 37 BMJ Open

1 2 3 572 CONCLUSION 4 5 573 The prevalence in Belgium of major obstetric haemorrhage requiring peripartum hysterectomy and/or IR is 6 7 574 estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries: roughly 1 in 3030 women who give birth in Belgium 8 9 10 575 undergoes a peripartum hysterectomy, while another 1 in 3030 women is successfully managed by IR thereby 11 12 576 preserving the uterus. Abnormal placentation and/or uterine atony are reported as the cause of haemorrhage 13 14 577 and reason for intervention in the majority of the women (83.7%). Further improvement in the management of 15 For peer review only 16 17 578 obstetric haemorrhage could possibly be realised by implementing routine screening for abnormal 18 19 579 placentation, both placenta praevia and abnormally invasive placenta, and antenatal transfer to tertiary 20 21 580 centres in case of suspicion. A case-by-case in-depth analysis is necessary to reveal whether the hysterectomies 22 23 24 581 performed in this series were appropriate or preventable. 25 26 582 27 28 583 ACKNOWLEDGEMENTS 29 30 31 584 This study would not have been possible without the voluntary participation of the Belgian maternity units: 32 33 585 GZA Sint-Jozef, Mortsel; GZA Sint-Augustinus, Wilrijk; GZA Sint-Vincentius, Antwerpen; ZNA, Sint-Erasmus, 34 35 586 Borgerhout; ZNA, Jan Palfijn , Merksem; ZNA, Middelheim, Antwerpen; ASZ, campus Aalst; ASZ, campus 36 37 38 587 ; AZ Lokeren; Imelda Ziekenhuis, Bonheiden; Sint-Jozef ziekenhuis, Bornem; AZ Klina, 39 40 588 Brasschaat; AZ Sint-Jan, Brugge; AZ Sint-Lucas, Assebroek; Centre de Santé des Fagnes; Centre Hospitalier de 41 42 589 l’Ardenne; Centre Hospitalier EpiCura, site Hornu; Centre Hospitalier de Mouscron; Centre Hospitalier du Bois 43 44 45 590 de l’Abbaye et de Hesbay ; Centre Hospitalier Peltzer, La Tourelle ; CHC Saint-Vincent; CHC Saint-Joseph; CHC 46 47 591 Sainte-Elisabeth; CHIREC, Clinique Sainte-Anne, Saint-Remi; CHIREC Braine-l’Alleud-Waterloo; CHR de la 48 49 592 Citadelle; CHR de Namur; CHR du Val de Sambre; CHR Haute Senne-Le Tilleriau; CHR Mons Clinique Saint- 50 51 593 Joseph; CHU Ambroise Paré; CHU Brugmann; CHU Charleroi, André Vésale; CHU Notre-Dame des Bruyères; 52 53 54 594 CHU Saint-Pierre; CHU Tivoli; CHwapi-Site Notre-Dame; Clinique Edith Cavell; Clinique et Maternité Sainte 55 56 595 Elisabeth; Clinique Notre dame de Grâce; Clinique Reine Astrid; Clinique Sainte-Piere; Cliniques de l’Europe, 57 58 596 Saint-Michel; Cliniques de l’Europe, Sainte-Elisabeth; Cliniques du Sud ; Cliniques Universitaires 59 60 597 Saint-Luc; AZ Sint-Vincentius, Deinze; AZ Sint-Blasius , Dendermonde; AZ Monica, Deurne; AZ Diest; AZ Sint- For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 37

1 2 3 598 Maarten, Duffel; AZ Alma, Eeklo; Dimpna Ziekenhuis, Geel; AZ Jan Palfijn, Gent; AZ Maria Middelares, Gent; AZ 4 5 599 Sint-Lucas, Gent; GHDC - Notre Dame, Charleroi; AZ Maria ziekenhuis, Halle; Jessaziekenhuis, ; AZ Sint 6 7 600 Elisabeth, Herentals; CAZ Midden-, Heusden-Zolder; Hôpital Civil Marie Curie; Hôpital Erasme, 8 9 10 601 Bruxelles; Hôpitaux Iris Sud – Site Ixelles ; Jan Yperman ziekenhuis, Ieper; IFAC Hôpital Princesse Paola; INDC 11 12 602 Entité Jolimontoise; Sint-Jozef ziekenhuis, Izegem; Kliniek Sint-Jan; Klinik Saint-Josef; AZ Zeno, campus Knokke- 13 14 603 Heist; AZ Groeninge, Kortrijk; Heilig Hart ziekenhuis, Leuven; UZ Leuven; Heilig Hart ziekenhuis, Lier; AZ Sint- 15 For peer review only 16 604 Jozef, Malle; AZ Sint-Maarten, Mechelen; AZ Delta, campus Menen; Heilig Hart ziekenhuis, Mol; Onze-Lieve- 17 18 19 605 Vrouw ziekenhuis, campus Aalst; Onze-Lieve-Vrouw ziekenhuis, campus Asse; AZ Damiaan, campus Sint-Jozef, 20 21 606 Oostende; AZ Sint-Jan, campus Oostende; AZ Oudenaarde; Mariaziekenhuis Noord-Limburg, Overpelt; AZ 22 23 607 Heilige Familie, Reet AZ Delta, campus Brugsesteenweg, Roeselare; AZ Delta, campus Wilgenstraat, Roeselare; 24 25 26 608 AZ Glorieux, Ronse; AZ Nikolaas, Sint-Niklaas; Sint-Trudo ziekenhuis, Sint-Truiden; Saint-Nikolaus Hospital; Sint- 27 28 609 Andries ziekenhuis, Tielt; Heilig Hartziekenhuis, Tienen; AZ Vesalius, Tongeren; Sint-Rembertziekenhuis, 29 30 610 Torhout; AZ Turnhout; UZ Antwerpen; UZ Brussel; AZ Sint-Augustinus, Veurne; AZ Jan Portaels, Vilvoorde; 31 32 33 611 Onze-Lieve-Vrouw van Lourdes ziekenhuis, Waregem; Ziekenhuis Oost-Limburg, campus Sint-Jan, Genk; AZ 34 35 612 Sint-Elisabeth, . 36 37 613 We would like to thank Fatima Bercha, Marlies Deblaere and Ann Langedock for their contribution in collecting 38 39 614 the data. 40 41 42 615 43 44 616 AUTHOR CONTRIBUTIONS 45 46 47 617 Conceived the project: MH, YE. Designed the study: GV, VVL, JVK, MH, YE. Collected the data: GV, MG, IR, VVL, 48 49 618 JVK. Analysed the data: GV, MG, IJ, ER. Interpreted the data: GV, MG, IJ, KR, MH, ER, YE, HV. Wrote the first 50 51 619 draft of the manuscript: GV. Contributed to the writing of the manuscript: MG, IJ, KR, MH, YE, HV. Critically 52 53 54 620 revised the manuscript: MG, IJ, VVL, KR, MH, ER, JV, YE, HV. Approved the final version: GV, MG, IJ, VVL, KR, 55 56 621 MH, ER, JV, YE, HV. Agreed to be accountable for all aspects of the work: GV, MG, IJ, VVL, KR, MH, ER, JV, YE, 57 58 622 HV. ICMJE criteria for authorship met: GV, MG, IJ, VVL, KR, MH, ER, JV, YE, HV. 59 60 623 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 37 BMJ Open

1 2 3 624 FUNDING 4 5 625 The study has been funded by the College for Mother and Newborn, a consultative body of the Federal Public 6 7 626 Service of Health in Belgium. GV has been funded by the Research Foundation Flanders (FWO). 8 9 10 627 11 12 628 COMPETING INTERESTS 13 14 15 629 All authors have completedFor the ICMJEpeer uniform disclo reviewsure form at www.icmje.org/coi_disclosure.pdf only and 16 17 630 declare: all authors had no financial relationships with any organisations that might have an interest in the 18 19 631 submitted work in the previous three years; no other relationships or activities that could appear to have 20 21 22 632 influenced the submitted work. 23 24 633 25 26 634 DATA SHARING 27 28 29 635 No additional data are available. 30 31 636 32 33 34 637 ETHICS APPROVAL 35 36 638 The B.OSS methodology and this study were approved by the Medical Ethics Committee of the Ghent 37 38 639 University Hospital (EC UZG 2012/734; B670201215359 and EC UZG 2015/1470; B670201526875) and by the 39 40 41 640 Medical Ethics Committee of the University Hospital Erasme, Brussels (EC ULB 2012/111; B406201213660). 42 43 641 Data were retrieved in retrospect from the case notes by means of data collection forms. No personally 44 45 642 identifiable information was obtained. Participants were informed and enabled to opt-out. 46 47 48 643 49 50 644 REFERENCES 51 52 645 1. Evaluating the quality of care for severe pregnancy complications: the WHO near-miss approach for 53 54 55 646 maternal health. Geneva, Switzerland: WHO press 2011. Retrieved from 56 57 647 http://apps.who.int/iris/bitstream/10665/44692/1/9789241502221_eng.pdf. Accessed 22 Dec. 16. 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 28 of 37

1 2 3 648 2. Stones W, Lim W, Al-Azzawi F, Kelly M. An investigation of maternal morbidity with identification of 4 5 649 life-threatening 'near miss' episodes. Health trends. 1991;23(1):13-5. 6 7 650 3. Say L, Souza JP, Pattinson RC. Maternal near miss--towards a standard tool for monitoring quality of 8 9 10 651 maternal health care. Best practice & research Clinical obstetrics & gynaecology. 2009;23(3):287-96. 11 12 652 4. Lewis, G (ed) 2007. The Confidential Enquiry into Maternal and Child Health (CEMACH). Saving 13 14 653 Mother's Lives: reviewing maternal deaths to make motherhood safer-2003-2005. The Seventh Report on 15 For peer review only 16 654 Confidential Enquiries into Maternal Deaths in the United Kingdom. London: CEMACH. 17 18 19 655 5. Knight M, Tuffnell D, Kenyon S, , Kenyon S, Shakespear J, Gray R, Kurinczuk JJ (Eds.) on behalf of 20 21 656 MBRRACE-UK. Saving Lives, Improving Mother's Care - Surveillance of maternal deaths in the UK 2011-13 and 22 23 657 lessons learned to inform maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths 24 25 26 658 and Morbidity 2009-13. Oxford: National Perinatal Epidemiology Unit, University of Oxford 2015. 27 28 659 6. The Royal College of Obstetricians and Gynaecologists. Postpartum Haemorrhage, Prevention and 29 30 660 Management. Green-top Guideline No. 52. Published: 11/05/2009. 31 32 33 661 7. Postpartum Hemorrhage. The American College of Obstetricians and Gynaecologistis. ACOG Practice 34 35 662 Bulletin. Number 76, October 2006. Reaffirmed 2015. 36 37 663 8. Dahlke JD, Mendez-Figueroa H, Maggio L, Hauspurg AK, Sperling JD, Chauhan SP, et al. Prevention and 38 39 664 management of postpartum hemorrhage: a comparison of 4 national guidelines. American journal of obstetrics 40 41 42 665 and gynecology. 2015;213(1):76.e1-10. 43 44 666 9. Kayem G, Kurinczuk JJ, Alfirevic Z, Spark P, Brocklehurst P, Knight M. Specific second-line therapies for 45 46 667 postpartum haemorrhage: a national cohort study. BJOG : an international journal of obstetrics and 47 48 49 668 gynaecology. 2011;118(7):856-64. 50 51 669 10. Zwart JJ, Dijk PD, van Roosmalen J. Peripartum hysterectomy and arterial embolization for major 52 53 670 obstetric hemorrhage: a 2-year nationwide cohort study in the Netherlands. American journal of obstetrics and 54 55 56 671 gynecology. 2010;202(2):150.e1-7. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 37 BMJ Open

1 2 3 672 11. Inoue S, Masuyama H, Hiramatsu Y. Efficacy of transarterial embolisation in the management of post- 4 5 673 partum haemorrhage and its impact on subsequent pregnancies. The Australian & New Zealand journal of 6 7 674 obstetrics & gynaecology. 2014;54(6):541-5. 8 9 10 675 12. Lee HY, Shin JH, Kim J, Yoon HK, Ko GY, Won HS, et al. Primary postpartum hemorrhage: outcome of 11 12 676 pelvic arterial embolization in 251 patients at a single institution. Radiology. 2012;264(3):903-9. 13 14 677 13. The Royal College of Obstetricians and Gynaecologists. The role of emergency and elective 15 For peer review only 16 678 interventional radiology in postpartum haemorrhage. (Good Practice No. 6). June 2007. 17 18 19 679 14. The Royal College of Obstetricians and Gynaecologists. Placenta Praevia, Placenta Praevia Accreta and 20 21 680 Vasa Praevia : Diagnosis and Management. Green-top Guideline No. 27. Published January 2011. 22 23 681 15. Investigation into 10 Maternal Deaths at, or following Delivery at, Northwick Park Hospital, North West 24 25 26 682 London Hospitals NHS Trust, between April 2002 and April 2005. Commission for Healthcare Audit and 27 28 683 Inspection. London, August 2006. 29 30 684 16. Vandenberghe G, De Blaere M, Van Leeuw V, Roelens K, Englert Y, Hanssens M, et al. Nationwide 31 32 33 685 population-based cohort study of uterine rupture in Belgium: results from the Belgian Obstetric Surveillance 34 35 686 System. BMJ open. 2016;6(5):e010415. 36 37 687 17. H. Cammu EM, G. Martens, C. Van Mol, Y Jacquemyn. Perinatale activiteiten in Vlaanderen 2013. 2014. 38 39 688 Retrieved from: https://www.zorg-en- 40 41 42 689 gezondheid.be/sites/default/files/atoms/files/SPE_jaarrapport%202013.pdf. Accessed on 22 Dec. 16 43 44 690 18. H. Cammu EM, G. Martens, C. Van Mol, Y Jacquemyn. Perinatale Activiteiten in Vlaanderen 2012. 2013. 45 46 691 Retrieved from: https://www.zorg-en- 47 48 49 692 gezondheid.be/sites/default/files/atoms/files/SPE_jaarrapport%202012.pdf. Accessed on 22 Dec. 16. 50 51 693 19. Van Leeuw V LC, Englert Y. Perinatale gegevens in het Brussels Gewest - Jaren 2013. Centre 52 53 694 d'Epidémiologie périnatale, 2014. Retrieved from: http://www.cepip.be/pdf/rapport_CEPIP_Bxl2013_NL.pdf. 54 55 56 695 Accessed on 22 Dec. 16. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 30 of 37

1 2 3 696 20. Leroy Ch VLV, Englert Y. Données périnatales en Wallonie - Année 2013. Centre d'Epidémiologie 4 5 697 Périnatale, 2014. Retrieved from : http://www.cepip.be/pdf/rapport_CEPIP_wallonie2013_tma.pdf. Accessed 6 7 698 on 22 Dec. 16. 8 9 10 699 21. Armonk NIC. IBM SPSS Statistics for Windows, Version 22.0. In: Corp I, editor. 2013. 11 12 700 22. Comstock CH, Bronsteen RA. The antenatal diagnosis of placenta accreta. BJOG : an international 13 14 701 journal of obstetrics and gynaecology. 2014;121(2):171-81; discussion 81-2. 15 For peer review only 16 702 23. D'Antonio F, Iacovella C, Bhide A. Prenatal identification of invasive placentation using ultrasound: 17 18 19 703 systematic review and meta-analysis. Ultrasound in obstetrics & gynecology: the official journal of the 20 21 704 International Society of Ultrasound in Obstetrics and Gynecology. 2013;42(5):509-17. 22 23 705 24. Fitzpatrick KE, Sellers S, Spark P, Kurinczuk JJ, Brocklehurst P, Knight M. The management and 24 25 26 706 outcomes of placenta accreta, increta, and percreta in the UK: a population-based descriptive study. BJOG : an 27 28 707 international journal of obstetrics and gynaecology. 2014;121(1):62-70; discussion -1. 29 30 708 25. Thurn L, Lindqvist PG, Jakobsson M, Colmorn LB, Klungsoyr K, Bjarnadottir RI, et al. Abnormally invasive 31 32 33 709 placenta-prevalence, risk factors and antenatal suspicion: results from a large population-based pregnancy 34 35 710 cohort study in the Nordic countries. BJOG : an international journal of obstetrics and gynaecology. 36 37 711 2016;123(8):1348-55. 38 39 712 26. Collins SL, Ashcroft A, Braun T, Calda P, Langhoff-Roos J, Morel O, et al. Proposal for standardized 40 41 42 713 ultrasound descriptors of abnormally invasive placenta (AIP). Ultrasound in obstetrics & gynecology : the 43 44 714 official journal of the International Society of Ultrasound in Obstetrics and Gynecology. 2016;47(3):271-5. 45 46 715 27. Chantraine F, Braun T, Gonser M, Henrich W, Tutschek B. Prenatal diagnosis of abnormally invasive 47 48 49 716 placenta reduces maternal peripartum hemorrhage and morbidity. Acta obstetricia et gynecologica 50 51 717 Scandinavica. 2013;92(4):439-44. 52 53 718 28. Eller AG, Bennett MA, Sharshiner M, Masheter C, Soisson AP, Dodson M, et al. Maternal morbidity in 54 55 56 719 cases of placenta accreta managed by a multidisciplinary care team compared with standard obstetric care. 57 58 720 Obstetrics and gynecology. 2011;117(2 Pt 1):331-7. 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 31 of 37 BMJ Open

1 2 3 721 29. Placenta Accreta. The American College of Obstetricians and Gynaecologistis. Committee Opinion. 4 5 722 Number 529. Published July 2012. Reaffirmed 2015. 6 7 723 30. Belfort MA. Placenta accreta. American journal of obstetrics and gynecology. 2010;203(5):430-9. 8 9 10 724 31. Fox KA, Shamshirsaz AA, Carusi D, Secord AA, Lee P, Turan OM, et al. Conservative management of 11 12 725 morbidly adherent placenta: expert review. American journal of obstetrics and gynecology. 2015;213(6):755- 13 14 726 60. 15 For peer review only 16 727 32. Chantraine F, Nisolle M, Petit P, Schaaps JP, Foidart JM. Individual decisions in placenta increta and 17 18 19 728 percreta: a case series. Journal of perinatal medicine. 2012;40(3):265-70. 20 21 729 33. Sentilhes L, Ambroselli C, Kayem G, Provansal M, Fernandez H, Perrotin F, et al. Maternal outcome 22 23 730 after conservative treatment of placenta accreta. Obstetrics and gynecology. 2010;115(3):526-34. 24 25 26 731 34. Dilauro MD, Dason S, Athreya S. Prophylactic balloon occlusion of internal iliac arteries in women with 27 28 732 placenta accreta: literature review and analysis. Clinical radiology. 2012;67(6):515-20. 29 30 733 35. Cali G, Forlani F, Giambanco L, Amico ML, Vallone M, Puccio G, et al. Prophylactic use of intravascular 31 32 33 734 balloon catheters in women with placenta accreta, increta and percreta. European journal of obstetrics, 34 35 735 gynecology, and reproductive biology. 2014;179:36-41. 36 37 736 36. Duan XH, Wang YL, Han XW, Chen ZM, Chu QJ, Wang L, et al. Caesarean section combined with 38 39 737 temporary aortic balloon occlusion followed by uterine artery embolisation for the management of placenta 40 41 42 738 accreta. Clinical radiology. 2015;70(9):932-7. 43 44 739 37. Sentilhes L, Gromez A, Clavier E, Resch B, Verspyck E, Marpeau L. Predictors of failed pelvic arterial 45 46 740 embolization for severe postpartum hemorrhage. Obstetrics and gynecology. 2009;113(5):992-9. 47 48 49 741 38. Poujade O, Zappa M, Letendre I, Ceccaldi PF, Vilgrain V, Luton D. Predictive factors for failure of pelvic 50 51 742 arterial embolization for postpartum hemorrhage. International journal of gynaecology and obstetrics: the 52 53 743 official organ of the International Federation of Gynaecology and Obstetrics. 2012;117(2):119-23. 54 55 56 744 39. Yamasaki Y, Morita H, Miyahara Y, Ebina Y, Okada T, Yamaguchi M, et al. The factors associated with 57 58 745 the failure of transcatheter pelvic arterial embolization for intractable postpartum hemorrhage. Journal of 59 60 746 perinatal medicine. 2014;42(3):359-62. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 37

1 2 3 747 40. Touboul C, Badiou W, Saada J, Pelage JP, Payen D, Vicaut E, et al. Efficacy of selective arterial 4 5 748 embolisation for the treatment of life-threatening post-partum haemorrhage in a large population. PloS one. 6 7 749 2008;3(11):e3819. 8 9 10 750 41. Lee HJ, Jeon GS, Kim MD, Kim SH, Lee JT, Choi MJ. Usefulness of pelvic artery embolization in cesarean 11 12 751 section compared with vaginal delivery in 176 patients. Journal of vascular and interventional radiology : JVIR. 13 14 752 2013;24(1):103-9. 15 For peer review only 16 753 42. Kim YJ, Yoon CJ, Seong NJ, Kang SG, An SW, Kim YS, et al. Failed pelvic arterial embolization for 17 18 19 754 postpartum hemorrhage: clinical outcomes and predictive factors. Journal of vascular and interventional 20 21 755 radiology : JVIR. 2013;24(5):703-9. 22 23 756 24 25 26 757 27 28 758 29 30 31 759 32 33 34 760 35 36 761 37 38 39 762 40 41 763 42 43 44 764 45 46 765 47 48 49 766 FIGURE LEGENDS 50 51 767 Figure 1. Flowchart of case reporting and data collection. 52 53 768 Figure 2. Circumstances in which peripartum hysterectomy and interventional radiology occurred in 166 54 55 56 769 women. 57 58 770 Figure 3. Cause of obstetric haemorrhage according to the 4T’s Mnemonics. 59 60 771 Figure 4. Antenatal suspicion and preventive measures in women with abnormally invasive placenta (AIP). For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 33 of 37 BMJ Open

1 2 3 772 4 5 773 6 7 774 8 9 10 775 11 12 776 13 14 777 15 For peer review only 16 778 17 18 19 779 20 21 780 22 23 781 24 25 26 782 27 28 783 29 30 784 31 32 33 785 34 35 786 36 37 787 38 39 788 40 41 42 789 43 44 790 45 46 47 791 48 49 792 Figure 1. Flowchart of case reporting and data collection. 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 34 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 793 44 45 46 794 47 48 795 49 50 51 796 52 53 797 54 55 56 798 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 35 of 37 BMJ Open

1 2 3 799 Figure 2. Circumstances in which peripartum hysterectomy and interventional radiology occurred in 166 4 5 800 women. 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 801 44 45 802 46 47 48 803 49 50 804 51 52 53 805 54 55 56 806 57 58 807 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 36 of 37

1 2 3 808 Figure 3. Cause of obstetric haemorrhage according to the 4T’s mnemonics: Tonus, Tissue, Trauma, 4 5 809 Thrombin. 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 810 42 811 Tonus: uterine atony (n=91); Tissue: AIP (n=14), placenta praevia (n=18), combination of AIP and placenta 43 44 812 praevia (n=24), placental remnants or retained placenta without AIP (n=26). Trauma: uterine rupture (n=12), 45 46 47 813 genital tract lacerations (n=12), bleeding caused by unintended injuries during caesarean delivery (n=14) or 48 49 814 D&C (n=3). Uncomplicated caesarean deliveries are not included in the trauma group. Thrombin: abnormal 50 51 815 coagulation before onset of bleeding (n=6). DIC secondary to major haemorrhage was not included in the 52 53 54 816 thrombin group. Others: the cause of bleeding remains unclear or is unknown. 55 56 817 57 58 818 59 60 819 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 37 of 37 BMJ Open

1 2 3 820 Figure 4. Antenatal suspicion and preventive measures in women with abnormally invasive placenta (AIP). 4 5 821 § < 0.05 for difference in between groups. 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 822 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open

A nationwide population -based cohort study of peripartum hysterectomy and arterial embolisation in Belgium: results from the Belgian Obstetric Surveillance System.

For peer review only Journal: BMJ Open

Manuscript ID bmjopen-2017-016208.R1

Article Type: Research

Date Submitted by the Author: 10-May-2017

Complete List of Authors: Vandenberghe, Griet; Ghent University Hospital, Department of Obstetrics and Gynaecology Guisset, Marine; Leuven University Hospital, Department of Obstetrics and Gynaecology Janssens, Iris; Ghent University, Faculty of Medicine Van Leeuw, Virginie; Université Libre de Bruxelles (ULB), School of Public Health; Perinatal Epidemiology Centre (Centre d'Épidémiologie Périnatale, CEpiP) Roelens, Kristien; Ghent University Hospital, Department of Obstetrics and Gynaecology Hanssens, Myriam; University Hospital Leuven, Department of Obstetrics & Gynaecology Russo, Erika; Intercommunale de Santé Publique du Pays de Charleroi (ISSPC), Department of Obstetrics and Gynaecology Vankeirsbilck, Joachim; St Jan's Hospital, Department of Obstetrics and Gynaecology Englert, Yvon; Université Libre de Bruxelles, Faculty of Medicine, Research Laboratory on Human Reproduction; Perinatal Epidemiology Center (Centre d'Épidémiologie Périnatale, CEpiP) Verstraelen, Hans; Ghent University Hospital, Department of Obstetrics and Gynaecology

Primary Subject Obstetrics and gynaecology Heading:

Secondary Subject Heading: Epidemiology

peripartum hysterectomy, arterial embolisation, Interventional radiology < Keywords: RADIOLOGY & IMAGING, maternal near miss, severe maternal morbidity

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 41 BMJ Open

1 2 3 1 TITLE PAGE 4 5 2  Title of the article: 6 7 3 A nationwide population-based cohort study of peripartum hysterectomy and arterial embolisation in 8 9 4 Belgium: results from the Belgian Obstetric Surveillance System. 10 5 11 12 6  Author and Co-Author’s name: 13

14 7 Vandenberghe G, Guisset M, Janssens I, Van Leeuw V, Roelens K, Hanssens M, Russo E, Van Keirsbilck J, 15 For peer review only 16 8 Englert Y, Verstraelen H. 17 9 18 19 10  Corresponding author: Griet Vandenberghe 20 21 11 Department of Obstetrics & Gynaecology 22 12 Ghent University Hospital 23 24 13 De Pintelaan 185, 9000 Ghent, Belgium 25 26 14 E-mail: [email protected] 27 28 15 Telephone number: +32 9 33 27 849 29 16 30 31 17  Co-author: Marine Guisset 32 33 18 Department of Obstetrics & Gynaecology 34 35 19 Leuven University Hospital 36 20 Leuven, Belgium 37 38 21 E-mail: [email protected] 39 40 22 41 23  Co-author: Iris Janssens 42 43 24 Faculty of Medicine 44 45 25 Ghent University 46 47 26 Ghent, Belgium 48 27 E-mail: [email protected] 49 50 28 51 52 29  Co-author: Virginie Van Leeuw 53 54 30 Perinatal Epidemiology Center (Centre d’Épidémiologie Périnatale, CEpiP) 55 31 Université Libre de Bruxelles (ULB), School of Public Health. 56 57 32 Brussels, Belgium 58 59 33 E-mail: [email protected] 60 34 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 41

1 2 3 35  Co-author: Kristien Roelens 4 36 Department of Obstetrics & Gynaecology 5 6 37 Ghent University Hospital 7 8 38 Ghent, Belgium 9 10 39 E-mail: [email protected] 11 40 12 13 41  Co-author: Myriam Hanssens 14 15 42 Department Forof Obstetrics peer & Gynaecology review only 16 43 University Hospital Leuven 17 18 44 Leuven, Belgium 19 20 45 E-mail: [email protected] 21 22 46 23 47  Co-author: Erika Russo 24 25 48 Department of Obstetrics & Gynaecology 26 27 49 Intercommunale de Santé Publique du Pays de Charleroi (ISPPC) 28 29 50 Hôpital Civil Marie Curie 30 51 Charleroi, Belgium 31 32 52 E-mail: [email protected] 33 34 53 35 54  Co-author: Joachim Vankeirsbilck 36 37 55 Department of Obstetrics & Gynaecology 38 39 56 St Jan’s Hospital 40 41 57 Bruges, Belgium 42 58 E-mail: [email protected] 43 44 59 45 46 60  Co-author: Yvon Englert 47 48 61 Perinatal Epidemiology Center (Centre d’Épidémiologie Périnatale, CEpiP) 49 62 Université Libre de Bruxelles (ULB), Faculty of Medicine, Research Laboratory on Human Reproduction. 50 51 63 Brussels, Belgium 52 53 64 E-mail: [email protected] 54 65 55 56 66  Co-author: Hans Verstraelen 57 58 67 Department of Obstetrics & Gynaecology 59 60 68 Ghent University Hospital 69 Ghent, BelgiumFor peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 41 BMJ Open

1 2 3 70 E-mail: [email protected] 4 71 5 6 72  Keywords: peripartum hysterectomy, arterial embolisation, interventional radiology, major obstetric 7 8 73 haemorrhage, maternal near miss, severe maternal morbidity 9 10 74 11 75  Word count: The main text consists of 5738 words 12 13 76 14 15 77 Background:For peer 449 words review only 16 78 Methods: 952 words 17 18 79 Results: 2510 words 19 20 80 Discussion: 1695 words 21 22 81 Conclusion: 132 words 23 24 82 Total: 5738 words 25 83 26 27 84  Tables and Figures: The manuscript includes 4 tables and 4 figures and 1 supplementary figure. Figures 28 29 85 are uploaded, in JPEG format with a resolution of 300 DPI, separately from the text. 30 86 31 32 87 33 34 88 35 36 89 37 90 38 39 91 40 41 92 42 93 43 44 94 45 46 95 47 48 96 49 97 50 51 98 52 53 99 54 55 100 56 57 101 58 59 102 60 103 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 41

1 2 3 104 ABSTRACT 4 5 105 Objectives: 6 7 106 To assess the prevalence of major obstetric haemorrhage managed with peripartum hysterectomy and/or 8 9 107 interventional radiology (IR) in Belgium. To describe women characteristics, circumstances in which the 10 108 interventions took place, management of the obstetric haemorrhage, outcome and additional morbidity of 11 12 109 these women. 13 14 110 Design: 15 For peer review only 16 111 Nationwide population-based prospective cohort study. 17 112 Setting: 18 19 113 Emergency obstetric care. Participation of 97% of maternities covering 98.6% of the deliveries in Belgium. 20 21 114 Participants: 22 115 All women who underwent peripartum hysterectomy and/or IR procedures in Belgium between January 2012 23 24 116 and December 2013. 25 26 117 Results: 27 28 118 We obtained data on 166 women who underwent peripartum hysterectomy (n=84) and/or IR procedures 29 30 119 (n=102), corresponding to 1 in 3030 women undergoing a peripartum hysterectomy and another 1 in 3030 31 32 120 women being managed by IR thereby preserving the uterus. Seventeen women underwent hysterectomy 33 34 35 121 following IR and three women needed further IR despite hysterectomy. Abnormal placentation and/or uterine 36 37 122 atony were the reported causes of haemorrhage in 83.7%. Abnormally invasive placenta was not detected 38 39 123 antenatally in 34% of cases. The interventions were planned in 15 women. Three women were transferred 40 41 124 antenatally and seventeen women postnatally to a hospital with emergency IR service. Urgent peripartum 42 43 44 125 hysterectomy could be averted in 72% of women who were transferred, with no significant difference in need 45 46 126 for transfusion. Interventional radiology procedures were able to cease the bleeding in 87.8% of the attempts. 47 48 127 Disseminated intravascular coagulation secondary to major haemorrhage was reported in 32 women (19%). 49 50 51 128 Conclusion: 52 129 The prevalence in Belgium of major obstetric haemorrhage requiring peripartum hysterectomy and/or IR is 53 54 55 130 estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries. Increased awareness of clinicians for risk factors of 56 57 131 abnormal placentation could further improve management and outcome of major obstetric haemorrhage. A 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 41 BMJ Open

1 2 3 132 case-by-case in-depth analysis is necessary to reveal whether the hysterectomies and arterial embolisations 4 5 133 performed in this series were appropriate or preventable. 6 7 134 8 9 10 135 11 12 136 13 14 15 137 ARTICLE SUMMARY:For Strengths peer and limitations review of this study: only 16 17 138 18 19 139 • This is a nation-wide study of Belgian maternities, with excellent participation and response rates 20 21 140 (98.8%). The cases were collected on a monthly basis using the standard methodology of the Obstetric 22 23 141 Surveillance Systems. 24 25 142 • This series describes how women with major obstetric haemorrhage were managed with peripartum 26 27 28 143 hysterectomy and/or interventional radiology according to the circumstances of the event. 29 30 144 • The study lacks a perspective on the overall picture of postpartum haemorrhage in Belgium, namely 31 32 145 the total number of women with severe postpartum haemorrhage, near-misses or maternal deaths 33 34 35 146 due to postpartum haemorrhage. 36 37 147 • The study lacks denominator data of the women with postpartum haemorrhage that were successfully 38 39 148 managed with other second line measures. Therefore we can not make conclusions on the failure rate 40 41 42 149 of intra-uterine balloon tamponade, haemostatic brace suturing and ligation of the uterine arteries. 43 44 45 150 46 151 47 48 152 49 50 153 51 52 154 53 155 54 55 156 56 57 157 58 59 158 60 159 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 41

1 2 3 160 BACKGROUND 4 5 161 Emergency peripartum hysterectomy is the single most dramatic procedure in modern obstetrics. The 6 7 162 procedure is stressful and surgically challenging, inevitably causing additional maternal morbidity and, not 8 9 10 163 least, infertility. The World Health Organisation (WHO) listed emergency peripartum hysterectomy as an 11 12 164 identification criterion for maternal near miss.(1) The maternal near miss concept was introduced in view of 13 14 165 the very low maternal mortality rates in developed country settings(2) as an analytic tool to address health 15 For peer review only 16 17 166 system failures with the overall goal of improving obstetric care.(3) Substandard care in the management of 18 19 167 major obstetric haemorrhage continues to be a critical cause of maternal death and severe maternal morbidity 20 21 168 in developed countries.(4, 5) Every obstetric practitioner must be trained in the management of postpartum 22 23 24 169 haemorrhage to guarantee an immediate response and a multidisciplinary team approach. Internationally 25 26 170 recognized guidelines (6-8) indicate that one or more second line measures, including intra-uterine (balloon) 27 28 171 tamponade, haemostatic brace suturing, ligation of the uterine arteries and interventional radiology (IR), 29 30 172 should be available in hospitals with delivery units and that obstetric practitioners should be familiar with 31 32 33 173 these procedures. Notably, the appropriate clinical judgement and hence early involvement of an experienced 34 35 174 senior obstetrician can save lives: the timely decision to turn to a hysterectomy can avoid further blood loss, 36 37 175 risk of disseminated intravascular coagulation (DIC), additional morbidity and finally death.(6) 38 39 40 176 While emergency peripartum hysterectomy serves as the ultimate approach to manage major obstetric 41 42 177 haemorrhage, resorting to IR procedures on the pathway towards hysterectomy could be both lifesaving and 43 44 178 fertility-preserving. IR for major obstetric haemorrhage basically involves intra-arterial balloon occlusion or 45 46 47 179 arterial embolisation of the uterine or the internal iliac arteries. Research publications on the effectiveness of 48 49 180 IR procedures in obstetric settings are accumulating reporting high success rates of 85 to 95 percent.(9-12) 50 51 181 Success rates should be interpreted with caution however, as they may be driven by ascertainment bias. 52 53 54 182 Interventional radiology is gently introduced in guidelines and recommendations, (6-8, 13, 14) describing its 55 56 183 possible role in obstetric emergency settings, but awaiting further evidence to make firm recommendations. 57 58 184 Yet in the United Kingdom a national recommendation was made to provide a network of emergency IR 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 41 BMJ Open

1 2 3 185 services available for all trusts with delivery units in response to a Healthcare Commission Investigation into 4 5 186 maternal deaths.(13, 15) 6 7 187 The aims of this study were to investigate the population-based prevalence of peripartum hysterectomy and 8 9 10 188 the indications to proceed to this intervention, likewise the prevalence of major obstetric haemorrhage 11 12 189 managed with IR in Belgium. Furthermore, we aimed to describe the characteristics of the women in Belgium 13 14 190 who required these interventions to treat or prevent major obstetric haemorrhage, the circumstances in which 15 For peer review only 16 191 the interventions took place, the management of the obstetric haemorrhage and the outcome and additional 17 18 19 192 morbidity of these women. 20 21 193 22 23 194 METHOD 24 25 26 195 Belgian Obstetric Surveillance System (B.OSS) . 27 28 196 Peripartum hysterectomy and interventional radiology in Belgium were registered as part of a national 29 30 197 reporting system to study rare obstetric complications. The methodology of the Belgian Obstetric Surveillance 31 32 33 198 System has been described previously.(16) Briefly, an appointed contact person (OB/GYN or senior-midwife) in 34 35 199 each participating maternity unit was invited by monthly mailing to report a selected number of rare obstetric 36 37 200 complications that had occurred in the preceding month or to state that there was ‘nothing to report’. In the 38 39 40 201 event a case was reported in reply, the contact person was asked to complete an extensive data collection 41 42 202 form. Confidentiality was guaranteed for mother, provider and hospital. Person-identifiable information was 43 44 203 eliminated from data-analysis. In case of incomplete reporting, the appointed contact person was encouraged 45 46 47 204 repeatedly by email and phone to provide missing data, up to six months following the period of case 48 49 205 reporting. 50 51 206 Participation of 97 percent of Belgian maternity units (n= 110/113) was obtained, covering 98.6 percent of 52 53 54 207 deliveries (n = 248 681) in the two-year study period. Belgium has a large relative number of tertiary referral 55 56 208 centres providing neonatal intensive care (n=19) along with a high concentration of small-volume maternity 57 58 209 units (less than 1000 annual births in more than half of the maternity units). All tertiary referral centres have 59 60 210 an emergency IR service, as well as some of the larger non-tertiary maternity services. In answer to an inquiry For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 41

1 2 3 211 of B.OSS contact persons, 40 maternity units (38 %) declared to have an IR service available in the hospital and 4 5 212 seven more maternity units declared to have an IR service available but not 24/7 or that they could call upon 6 7 213 the vascular surgeons for similar procedures. 8 9 10 214 Study period . 11 12 215 Between January 2012 and December 2013. 13 14 216 Case definition. 15 For peer review only 16 217 B.OSS aimed to collect all cases of peripartum hysterectomy and IR at any gestational age corresponding to the 17 18 19 218 definition: any woman giving birth to a fetus or infant and undergoing a hysterectomy and/or IR procedure in 20 21 219 the same clinical episode. Most peripartum hysterectomies are emergent, life-saving interventions following 22 23 220 vaginal or caesarean delivery, most commonly performed to manage major obstetric haemorrhage and 24 25 26 221 sporadically because of a severe pelvic infection. Hysterectomy is occasionally required in the first or second 27 28 222 trimester, termed abortion hysterectomy, to prevent or manage major obstetric haemorrhage resulting from 29 30 223 an abnormally invasive placenta or from an ectopic pregnancy in the cornua, the cervix or in a caesarean scar. 31 32 33 224 A peripartum hysterectomy can be performed electively during caesarean delivery, termed ‘caesarean 34 35 225 hysterectomy’, following antenatal diagnosis of a severe abnormally invasive placenta (AIP) or a gynaecological 36 37 226 cancer. 38 39 227 Failure of IR was defined as the need for urgent peripartum hysterectomy to control the haemorrhage. 40 41 42 228 Registered variables . 43 44 229 Data collection forms questioned maternal characteristics, medical, surgical and obstetrical history, details of 45 46 230 the index pregnancy, details of the delivery, circumstances of the adverse event, the management and the 47 48 49 231 outcome for mother and neonate. 50 51 232 The cause of the obstetric haemorrhage was deducted from the clinical signs reported by the clinician and 52 53 233 assorted according to the 4T’s mnemonic: Tonus, Tissue, Trauma, and Thrombin. Tonus: uterine atony. Tissue: 54 55 56 234 abnormally invasive placenta (AIP), placenta praevia, combination of AIP and placenta praevia, placental 57 58 235 remnants or retained placenta without AIP. Trauma: uterine rupture, genital tract lacerations, bleeding caused 59 60 236 by unintended injuries during caesarean delivery or curettage (D&C); uncomplicated caesarean deliveries were For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 41 BMJ Open

1 2 3 237 not included in the trauma group. Thrombin: abnormal coagulation before onset of bleeding; DIC secondary to 4 5 238 major haemorrhage was not included in the thrombin group. Others: the cause of bleeding remains unclear or 6 7 239 is unknown. 8 9 10 240 Registered variables of the background population. 11 12 241 The Belgian perinatal registries (Studiecentrum voor Perinatale Epidemiologie (SPE) in Flanders,(17, 18) Centre 13 14 242 d’Epidémiologie Périnatale (CEpiP) in Brussels (19) and Wallonia (20)) register data on a mandatory basis, 15 For peer review only 16 243 covering nearly 100 percent of births in Belgian maternity units and home births. A selected set of perinatal 17 18 19 244 data are recorded by the obstetrician, midwife and neonatologist immediately after birth. Births are registered 20 21 245 as such, in case of a live birth or in case of a stillbirth at a birth weight of 500 grams or more and/or after 22 22 23 246 completed weeks of gestation, with the exception of Flanders (SPE) where live births or stillbirths with a birth 24 25 26 247 weight below 500 grams, irrespective of gestational age, are not registered. 27 28 248 Statistical Analysis. 29 30 249 The aim was to register all cases of peripartum hysterectomy and arterial embolization during a 2-year study 31 32 33 250 period, therefore a fixed sample size was net set at start of the study. The prevalence of the obstetric events 34 35 251 targeted was estimated using as denominator the total number of deliveries in Belgium in 2012 and 2013, 36 37 252 corrected for three hospitals that did not participate in B.OSS (n= 248,681 women). Comparison of the 38 39 253 distributions of socio-demographic and obstetric characteristics of reported cases relative to the background 40 41 42 254 population as obtained from the Belgian perinatal registry (2012-2013) was pursued through use of odds 43 44 255 ratios. Odds ratios and accompanying 95 percent confidence intervals (95% CI) were calculated in univariate 45 46 256 analysis, using weighted cases where appropriate. Unadjusted and adjusted odds ratios were calculated to 47 48 49 257 compare patient and organisational characteristics among interventions (hysterectomy versus IR) in univariate 50 51 258 analysis and subsequently through use of unconditional binary logistic regression models. Comparison between 52 53 259 cases with successful and those with failed IR was only made in crude analyses, as numbers of observations 54 55 56 260 were too small to support logistic regression models. Non-parametric tests were used to compare distributions 57 58 261 of outcomes between groups, specifically χ2 for categorical variables, and Kruskal-Wallis and Mann-Whitney U 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 41

1 2 3 262 tests for continuous variables. P-values of <0.05 were considered statistically significant. Data were analysed 4 5 263 using the statistical software package IBM SPSS statistics version 22.0.(21) 6 7 264 The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting 8 9 10 265 observational studies were followed in this manuscript. 11 12 266 13 14 15 For peer review only 16 267 RESULTS 17 18 268 Prevalence 19 20 21 269 Data collection forms of 166 confirmed cases of peripartum hysterectomy and/or IR were retrieved and 22 23 270 included in the analysis. Details on reported cases and completeness of data collection forms are presented in 24 25 271 figure 1. 26 27 28 272 Overall, 84 women (50.6%) underwent a peripartum hysterectomy, corresponding to a prevalence of 3.3 per 29 30 273 10 000 deliveries (95% CI 2.7-4.1). Of these, 17 women (10.2%) underwent a hysterectomy following a previous 31 32 274 IR procedure. Eighty-two women (49.4%) were managed with IR only, corresponding to a prevalence of 3.3 per 33 34 275 10 000 deliveries (95% CI 2.6-4.0). Three women (1.8%) needed IR because of persistent bleeding despite 35 36 37 276 hysterectomy. 38 39 277 In 160 women the intervention occurred within 24 hours of delivery. In eight women the intervention occurred 40 41 278 late (from 24 hours to 6 weeks postpartum): two of these data collection forms were not returned, one of 42 43 44 279 them was the only case in this registry of a hysterectomy performed because of severe peritonitis. All other 45 46 280 women underwent hysterectomy and/or IR to treat or prevent major obstetric haemorrhage. The 47 48 281 corresponding prevalence in Belgium of major obstetric haemorrhage requiring peripartum hysterectomy 49 50 51 282 and/or IR is estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries. 52 53 283 54 55 284 Patient Characteristics 56 57 58 285 We compared the women included in this registry with all the women who gave birth in Belgium during the 59 60 286 study period. Patient characteristics of the background population were derived from the Belgian perinatal For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 41 BMJ Open

1 2 3 287 registries. In univariate analysis the women who underwent peripartum hysterectomy and/or IR for major 4 5 288 obstetric haemorrhage were significantly more likely compared to the background population to be older (≥ 35 6 7 289 years), to be multiparous (parity ≥ 3), to have given birth at a preterm gestational age, to have had a twin 8 9 10 290 pregnancy, to have had a caesarean delivery in a previous pregnancy, to have had a caesarean delivery in the 11 12 291 index pregnancy and to have conceived with artificial reproductive technology (ART) (Table 1). Ten of these 166 13 14 292 women were known to have a uterus myomatosus (6.0%) and twelve experienced haemorrhage in a previous 15 For peer review only 16 293 pregnancy (7.2%). 17 18 19 294 20 21 Table 1. Patient characteristics of women with (imminent) obstetric haemorrhage who underwent peripartum 22 23 hysterectomy and/or IR compared to all women who gave birth in Belgium during the study period. 24 Risk factor Cases of H and/or IR Background population # Unadjusted OR % 25 (n=166) (n=252 272) for H and/or IR 26 27 n, % n, % (95% CI) 28 Maternal age ≥ 35 years 72 (43.4) 43 256 (17.1) 3.7 (2.7-5.0) § 29 30 BMI at booking ≥ 30 kg/m 2 12 (7.2) £ 29 453 (11.6) 0.7 (0.3-1.2) 31 Singletons ≥22 weeks Singletons ≥22 weeks 32 Gestational age (weeks) 33 (n=144) (n = 241 136) 34 - Extreme preterm (<28) - 8 (5.4) - 1 221 (0.5) 35 - Very preterm (28-32) - 6 (4.0) - 1 474 (0.6) 7.4 (5.3-10.5) § 36 - Late preterm (32-37) - 37 (24.8) - 13 798 (5.7) 37 - Term (>37) - 93 (62.4) - 224 643 (93.1) Reference 38 § 39 Parity ≥ 3 30 (18.0) 18 855 (7.4) 2.7 (1.8-4.1)

40 § 41 Multiplets (twin) 13 (7.8) 4690 (1.8) 4.4 (2.5-7.9)

42 & £ § 43 Previous CD 61 (36.7) 27 007 (10.7) 4.8 (3.5-6.6)

44 § 45 TOLAC ( ≥22 weeks) 14 / 57 (24.5) 12 754 / 27 007 (47.2) 0.3 (0.1-0.66) 46 47 Delivery Mode ( ≥22 weeks) $ 48 - CD £ § 91 / 160 (56.8) 54 369 (21.5) 4.8 (3.5-6.5) 49 compared to VD 50 - CD before onset of labour 51 66 / 91 (72.5) 28 586 / 54369 (52.5) 2.3 (1.5-3.7) compared to CD after onset of labour 52 - Assisted VD 53 10 / 69 (14.4) 23 100 /198 444 (11.6) 1.3 (0.6-2.5) 54 compared to spontaneous VD 55 ART (IVF/ICSI) 19 (11.4) £ 9233 (3.7) 3.4 (2.1-5.4) § 56 57 Birthweight ≥ 4000 g 11 (6.6) 19967 (7.8) 0.8 (0.4-1.5) 58 59 CD: Caesarean delivery, VD: Vaginal delivery, H: Hysterectomy, IR: Interventional Radiology , TOLAC: Trial o f Labour After Caesarean Delivery, ART: Artificial Reproductive Technologies, IVF: In Vitro Fertilisation, ICSI: Intracytoplasmic Sperm Injection 60 # Data from the background population were retrieved from SPE reports (Flanders) and CEpiP reports (Brussels and Wallonia) from 2012-2013. % unadjusted Odds ratiosFor (with peer 95% review Confidence only Inter val)- http://bmjopen.bmj.com/site/about/guidelines.xhtml for peripartum hysterectomy and/or IR calculated for the different risk factors: the odds BMJ Open Page 12 of 41

1 2 (=probability of peripartum hysterectomy and/or IR occurring divided by the probability o f H/IR not occurring ) in the women having the risk factor, 3 divided by the odds in the women not having the risk factor. 4 £ Missing data in cases of hysterectomy and/or IR : BMI (n=27), previous CD (n=1), mode of delivery (n=1), ART (n=2), haemorrhage in previous 5 pregnancy (n=5), uterus myomatosus (n=5) 6 & Women with previous Caesarean Delivery: 9 with Placenta Praevia, 9 with AIP, 17 with Placenta Praevia and AIP 7 $Women with Caesarean Delivery as delivery mode: 17 with Placenta Praevia, 8 with AIP, 23 with Placenta Praevia and AIP 8 Reason for Caesarean Delivery: Bleeding (Placenta Praevia, AIP, abruptio placentae): n= 15; Placenta Praevia / AIP not bleeding: n= 29; Maternal shock / 9 cardiac arrest: n=1; Other: n = 46 10 § p < 0.05 11 295 12 13 14 296 Circumstances of the obstetric haemorrhage: 15 For peer review only 16 297 IR service availability, antenatal and postnatal transfers, urgent and planned interventions. 17 18 19 298 The circumstances in which the interventions occurred in these 166 women are delineated in Figure 2. Three 20 21 299 women were transferred antenatally to another hospital with emergency IR service availability. In result 107 22 23 300 women gave birth in a maternity unit with emergency IR service availability, 20 women gave birth in a 24 25 301 maternity unit where IR service is available but not 24/7 or where obstetricians can call upon vascular surgeons 26 27 28 302 for IR procedures, 39 women gave birth in a maternity unit with no IR service available. Throughout the paper 29 30 303 we will use the terms 24/7-IR unit, not-24/7-IR unit and non-IR unit respectively. 31 32 304 Seventeen women were transferred postnatally to another hospital with emergency IR service availability and 33 34 35 305 underwent an urgent IR procedure, including one woman for persistent bleeding despite hysterectomy and 36 37 306 one woman for persistent bleeding despite planned intra-arterial balloon occlusion and arterial embolisation 38 39 307 by the vascular surgeons in the referring hospital. Three of the postnatally transferred women consecutively 40 41 42 308 underwent an urgent hysterectomy for persistent bleeding despite arterial embolisation. 43 44 309 Considering the three patients who were transferred antenatally, one woman underwent an urgent 45 46 310 embolisation upon arrival because of a bleeding placenta praevia, one woman was planned for caesarean 47 48 311 hysterectomy but underwent an urgent caesarean hysterectomy because of a bleeding AIP before the fixed 49 50 51 312 date, one woman underwent a planned caesarean delivery combined with a planned IR procedure because of 52 53 313 AIP. Seven more women were transferred postnatally to the intensive care unit of a tertiary referral centre. 54 55 314 Overall, the intervention was planned for imminent obstetric haemorrhage in 15 women. The indication was 56 57 58 315 abnormally invasive placenta in 13 women and placenta praevia in 2 women. 59 60 316 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 41 BMJ Open

1 2 3 317 Reported cause of the obstetric haemorrhage 4 5 318 Eighteen women (11%) presented with antepartum haemorrhage due to abnormal placentation (placenta 6 7 319 praevia, AIP or both) or abruptio placentae (n=2). Four women (2%) started bleeding during termination of 8 9 10 320 pregnancy (<24 weeks) due to a cervical pregnancy (n=1) or an AIP (n=3). Hundred thirty-eight women (83%) 11 12 321 presented with early postpartum haemorrhage within 24 hours postpartum and six (4%) with late postpartum 13 14 322 haemorrhage from 24 hours to 6 weeks postpartum. 15 For peer review only 16 17 323 The reported cause of the obstetric haemorrhage was uterine atony (n=91, 54.8%) and abnormally invasive 18 19 324 placenta (AIP) (n=38, 22.8%) in the majority of women. In 54 women (32.5%) more than one cause was 20 21 325 reported, with atony and abnormal placentation (AIP and/or placenta praevia) in 20 women (12%) as the most 22 23 24 326 frequent combination. The cause of the (imminent) obstetric haemorrhage assorted according to the 4T’s 25 26 327 mnemonic is demonstrated in Figure 3. 27 28 328 29 30 31 329 Management of the obstetric haemorrhage 32 33 330 Hysterectomy was performed in all the cases of haemorrhage reported to be associated with uterine rupture 34 35 331 (n=12) and in 73.7 percent of the cases associated with abnormally invasive placenta (n=28/38). Interventional 36 37 38 332 radiology could cease the bleeding and avert the need for peripartum hysterectomy in the majority of cases of 39 40 333 haemorrhage reported to be associated with genital tract lacerations (n=10/12, 83.3%), placental remnants or 41 42 334 retention (n=17/26, 65.4%) and uterine atony (n=28/43, 65.1%). 43 44 45 335 Other measures that were attempted to manage the obstetric haemorrhage before or while proceeding to the 46 47 336 hysterectomy or IR are listed in Table 2. The first line measures being the administration of uterotonic agents 48 49 337 and bimanual uterine compression, other second line measures being intra-uterine (balloon) tamponade, 50 51 338 haemostatic brace suturing (B-lynch or other) and ligation of the uterine or internal iliac arteries, besides blood 52 53 54 339 transfusion and the administration of fresh frozen plasma (FFP), platelets or clotting factors. An intra-uterine 55 56 340 balloon was inserted in 35 women (21%) of whom 7 women before transfer to another hospital with 57 58 341 availability of an IR service. Brace suturing and arterial ligation was attempted in 6 and 4 women respectively. 59 60 342 In 19 women (11.4%) no other measures were undertaken besides hysterectomy (n=13), IR (n=2) and For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 41

1 2 3 343 hysterectomy following IR (n=4). The reported cause of haemorrhage in these women was abnormally invasive 4 5 344 placenta (n=12), placenta praevia (n=2), uterine rupture (n=2), retained placenta (n=2) and one case of a late 6 7 345 postpartum haemorrhage. The intervention was planned in five of these women. Eight of them had minor 8 9 10 346 haemorrhage and did not require transfusion, while three women undergoing urgent peripartum hysterectomy 11 12 347 needed massive transfusion (defined as receiving eight or more units of red blood cells). 13 14 348 15 For peer review only 16 Table 2. First and second line measures in women undergoing hysterectomy (H) and /or IR (IR) according to the 17 18 circumstances of the obstetric haemorrhage. 19 Total Planned H Urgent H and/or IR 20 and/or IR n=151 21 IR service IR service No IR service Postnatal 22 24/7 not 24/7 Transfer 23 24 n=166 n=15 n=98 n=12 n=25 n=16* 25 n (%) n (%) n (%) n (%) n (%) n (%) 26 Cause of obstetric haemorrhage 27 28 Abnormal placentation 56 (34) 15 (100) 29 (30) 5 (42) 5 (20) 2 (13) 29 Atony only 43 (26) - 29 (30) 5 (42) 7 (28) 2 (13) 30 31 Uterine rupture 12 (7) - 6 (6) 1 (8) 5 (20) - 32 Miscellaneous £ 55 (33) - 34 (34) 1 (8) 8 (32) 12 (75) 33 First and second line measure 34 35 Uterotonics 134 (81) 10 (67) 81 (83) 10 (83) 19 (76) 15 (94) 36 37 Oxytocin 123 (74) 10 (67) 73 (75) 9 (75) 18 (72) 13 (81) 38 PG 109 (66) 5 (33) 64 (65) 17 (68) 8 (67) 15 (94) 39 Misoprostol 61 (37) 1 (7) 34 (35) 6 (50) 10 (40) 10 (63) 40 Carboprost 63 (39) 5 (33) 42 (43) 3 (25) 9 (25) 9 (56) 41 # 42 Oversewing 30 (18) 1 (7) 19 (20) 1 (8) 7 (28) 2 (13) 43 (Bimanual) compression 64 (39) 3 (20) 36 (37) 5 (42) 9 (36) 11 (69) 44 Balloon tamponade 35 (21) 2 (13) 21 (22) 2 (17) 3 (12) 7 (44) 45 B-lynch or ligation 10 (6) - 5 (5) 1 (8) 4 (16) - 46 Transfusion 139 (84) 7 (47) 85 (87) 12 (100) 21 (84) 15 (88) 47 $ 48 Massive transfusion 51 (31) 2 (13) 24 (24) 7 (58) 10 (40) 8 (50) 49 FFP 108 (65) 4 (27) 64 (65) 11 (92) 18 (72) 11 (69) 50 Platelets 57 (34) 1 (7) 29 (30) 9 (75) 10 (40) 8 (50) 51 Clotting factors % 23 (14) - 8 (8) 5 (42) 7 (28) 3 (19) 52 53 * On e woman was transferred postnatally for persistent bleeding due to AIP despite planned IR in the referring hospital; she was included in the ‘planned’ 54 group. 55 £ Miscellaneous group: Placental remnants or retention n=25, Surgical trauma n=16, Genital tract laceration n=11, Intra-abdominal arterial bleeding n=4, Coagulation disorders n= 3, Unknown n= 6. The total number is exceeding n = 55 due to overlap in between cases. 56 PG: Prostaglandin. This can be any or a combination of the following: misoprostol, sulprostone, dinoproston, carboprost. 57 # Oversewing of the placental implantation site, bleeding points, tears or ruptures. 58 $ Massive transfusion: transfusion of 8 or more units of red blood cells 59 % Clotting factors: Fibrinogen, recombinant Factor VIIa, PPSB (Prothrombin-Proconvertin-Stuart Factor-Antihemophilic Factor B) or a combination 60 349 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 41 BMJ Open

1 2 3 350 Management of abnormally invasive placenta (AIP) 4 5 351 The management of the (imminent) obstetric haemorrhage in the women with abnormally invasive placenta 6 7 352 (AIP) (n=38) is illustrated in Figure 4 (for details see supplementary Figure 1). Seventeen women (45%) were 8 9 10 353 susceptible for AIP due to previous caesarean delivery (CD) and placenta praevia in the index pregnancy, 11 12 354 another 7 women (18%) had placenta praevia in the index pregnancy and 9 women (24%) had previous 13 14 355 caesarean delivery. In 25 women (65.7%) there was antenatal suspicion of the AIP, of whom 16 had placenta 15 For peer review only 16 17 356 praevia and previous caesarean delivery, 5 had placenta praevia and 4 had previous caesarean delivery only. All 18 19 357 women with antenatal suspicion of AIP were planned to undergo an elective caesarean delivery in a maternity 20 21 358 unit with IR service availability (24/7 n=20, not 24/7 n=5), which occurred in 21 women. Yet 3 women had an 22 23 24 359 emergency caesarean delivery because of antepartum haemorrhage and 1 woman delivered vaginally following 25 26 360 acute onset of labour. Further preventive measures were undertaken in 13 women (31.5%): antenatal transfer 27 28 361 to a centre with IR service availability (n=2), placement of intra-arterial catheters pre-caesarean delivery 29 30 362 (n=12), planned caesarean hysterectomy (n=3) or planned hysterectomy at a later stage (n=2). 31 32 33 363 In 13 women (34%) there was no antenatal suspicion of the AIP. In retrospect all of them had risk factors: 1 34 35 364 woman had placenta praevia and previous caesarean delivery, 4 women had placenta praevia, 6 had previous 36 37 365 caesarean delivery, 1 woman had previous myomectomy and adhesiolysis for Asherman syndrome and 38 39 40 366 another woman had a previous manual removal of the placenta. Nine women delivered in a maternity unit 41 42 367 with IR service availability (24/7 n=7, not-24/7 n=2) and 4 delivered in a non-IR unit.. Mode of delivery in those 43 44 368 13 women who had no antenatal suspicion of AIP was elective caesarean delivery in 2 women, emergent 45 46 47 369 caesarean delivery in 5 women and vaginal delivery in 6 women. 48 49 370 The placenta was left in situ in 11 women with AIP: 4 in an emergency setting where 3 women underwent 50 51 371 immediate hysterectomy and 1 woman underwent urgent arterial embolisation followed by urgent 52 53 54 372 hysterectomy because of bleeding after 3 days. Another 7 women had their placenta left in situ in an elective 55 56 373 setting combined with planned IR: 2 women consecutively underwent planned hysterectomy, 2 underwent an 57 58 374 urgent hysterectomy within 24 hours, 2 underwent a planned hysterectomy after one month and 1 woman 59 60 375 underwent elective evacuation of retained products of conception after one month. In conclusion, For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 41

1 2 3 376 conservative management defined as aiming to preserve the placenta and the uterus in a multidisciplinary 4 5 377 approach, was attempted in 6 women and was successful in 3 women. 6 7 378 As anticipated, the women for whom measures were planned antenatally had significant better outcome than 8 9 10 379 women where AIP was diagnosed peripartum: less estimated blood loss (median 625 mL (range 300-1250) 11 12 380 versus median 4500 mL (range 500-6000), p=0.08), less need for transfusion (46.2% versus 92.3%, p=0.056), 13 14 381 less need for massive transfusion (≥8 units) (8.3% versus 53.8%, p=0.045) and less hysterectomies (61.5% 15 For peer review only 16 382 versus 76.9%, p=0.6). The maternal outcome was not significantly different when comparing women with AIP 17 18 19 383 undergoing IR as first intervention versus hysterectomy as first intervention: estimated blood loss (median 20 21 384 4000 mL (range 500-15000) versus median 1000 mL (range 300-8500), p=0.3) need for transfusion (76.5% 22 23 385 versus 70%, p=0.9), need for massive transfusion (≥8 units) (29.4% versus 30%, p=1.0). All women with planned 24 25 26 386 interventions were managed in tertiary referral centres. 27 28 387 29 30 388 Risk factors possibly contributing to the choice of intervention 31 32 33 389 The decision of an obstetrician dealing with a major obstetric haemorrhage to attempt IR rather than turning 34 35 390 to a hysterectomy immediately, will be led by many clinical and organisational factors. Some risk factors that 36 37 391 may have contributed to this decision are presented in Table 3. The women with double interventions (n=20) 38 39 40 392 were excluded from this analysis to facilitate interpretation. 41 42 393 High parity (≥ 3), previous caesarean delivery, caesarean delivery in the index pregnancy, abnormally invasive 43 44 394 placenta and the lack of emergency IR service availability are significantly positively associated with peripartum 45 46 47 395 hysterectomy in univariate analysis. In unconditional binary logistic regression, the odds to undergo a 48 49 396 hysterectomy for obstetric haemorrhage is 7.2 times higher in women with high parity and 8.2 times higher in 50 51 397 women delivering in a non-IR unit, independent of the other risk factors in the model (Hosmer and Lemeshow 52 53 54 398 Test p =0.67). 55 56 399 We repeated the analysis in a subgroup, namely the women who delivered in a 24/7-IR unit, with exclusion of 57 58 400 the women with planned interventions, consisting of 31 women who underwent hysterectomy only and 60 59 60 401 women who underwent embolisation only. In unconditional binary logistic regression, the odds to undergo a For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 41 BMJ Open

1 2 3 402 peripartum hysterectomy for obstetric haemorrhage is significantly higher in women with high parity (odds 4.5 4 5 403 (95% CI 1.2-16.5)) and nearly significantly higher in women with previous caesarean delivery (odds 3.7 (95% CI 6 7 404 0.98-14.6)) independent of the other risk factors in the model (Hosmer and Lemeshow Test p = 0.69). 8 9 405 10 11 12 Table 3. Risk factors possibly contributing to the choice of intervention in women undergoing hysterectomy or 13 interventional radiology for obstetric haemorrhage. Unadjusted and adjusted OR (95% CI) for hysterectomy (only) 14 per risk factor . 15 Risk factors For peer reviewNumber of women onlyUnadjusted OR Adjusted OR 16 Hysterectomy Interventional (95%CI) (95%CI) 17 18 only radiology only for for 19 n=64 n=82 Hysterectomy Hysterectomy 20 n, % n, % only only 21 Demographic factors 22 Maternal age ≥ 35 years 32 (50.0) 30 (36.5) 1.7 (0.8-3.3) 23 24 BMI at booking ≥ 30 kg/m2 6 (9.3) 3 (3.6) 2.7 (0.6-11.3) 25 Obstetric factors 26 Preterm (<37w) 28 (43.8) 24 (29.3) 1.8 (0.9-3.7) 27 Parity ≥ 3 18 (28.1) 7 (8.5) 2.7 (1.4-9.4) § 7.2 (2.1-24.6) § 28 Previous caesarean delivery 31 (48.4) 18 (21.9) 3.3 (1.6-6.8) § 2.3 (0.9-6.3) 29 § 30 Caesarean delivery ( ≥22w) * 38 / 59 (64.4) 37 / 82 (45.1) 2.2 (1.1-4.3) 1.6 (0.6-4.0) 31 Cause of bleeding 32 AIP and AIP+ praevia 16 (25.0) 9 (10.9) 2.7 (1.1-6.6) § 3.0 (0.9-10.0) 33 Atony 29 (45.3) 52 (63.4) 0.4 (0.2-0.9) § 0.9 (0.3-2.3) 34 35 Organisational factors # 36 Weekend and/or non-office hours 24/60 (40.0) 34/80 (42.5) 0.9 (0.4-1.7) 37 IR NOT available $ 30 (46.8) 15 (18.2) 3.9 (1.8-8.2) § 8.2 (3.1-21.6) § 38 * Exclusion of 5 cases < 22 weeks 39 £ Missing data: BMI (n=27), previous caesarean delivery (n=1), Delivery Mode (n=1), placenta location (n=13), Date of delivery (n=2), Time of 40 delivery (n=5), week/weekend (n=3) 41 # based on date and time of delivery 42 IR : interventional radiology 43 $ Emergency IR service availability in the maternity unit where the woman gave birth 44 § p<0.05 45 406 46 47 407 48 49 50 408 Peripartum hysterectomy despite interventional radiology 51 52 409 Seventeen women in this registry had a peripartum hysterectomy performed following a previous IR 53 54 55 410 procedure. In five of these women the hysterectomy was performed in a programmed setting following 56 57 411 planned IR because of abnormally invasive placenta, two of them with a delay of one month. These five IR 58 59 412 procedures were considered successful in preventing or ceasing the bleeding. In the other twelve women an 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 41

1 2 3 413 urgent hysterectomy was performed because of persistent bleeding despite the IR procedure. We compared 4 5 414 the characteristics of these 12 women in whom IR failed with 87 women in whom IR was successful, including 6 7 415 the five women who underwent an elective hysterectomy. In univariate analysis, an increased risk of urgent 8 9 10 416 hysterectomy despite IR was observed for women with caesarean delivery in a previous pregnancy, for women 11 12 417 with abnormally invasive placenta and for women who had no uterotonic agents administered. These findings 13 14 418 are presented in Table 4. 15 For peer review only 16 17 Table 4. Risk factors possibly contributing to failure of IR (IR). Unadjusted odds ratio’s (95% CI) 18 for failed IR per risk factor. 19 20 Riskfactors Number of women Unadjusted OR 21 Failed IR Successful IR (95% CI) 22 n=12 n=87 for failed IR 23 n, % n, % 24 Demographic factors 25 26 Maternal age ≥ 35 years 5 (41.7) 33 (37.9) 1.1 (0.3-3.9) 27 BMI at booking ≥ 30 kg/m 2 2 (18.2) 3 (3.9) 5.6 (0.8-37.6) 28 Obstetric factors 29 30 Twins 1 (8.3) 6 ( 6.9) 1.2 (0.1-11.1) 31 Preterm gestational age < 37w 7 (58.3) 28 (32.2) 2.9 (0.8-10.1) 32 Preterm gestational age < 34w 3 (25.0) 6 (6.8) 4.5 (0.9-21.1) 33 34 Parity ≥3 3 (25.0) 8 (9.2) 3.2 (0.7-14.6) 35 Previous caesarean delivery 7 (58.3) 21 (24.1) 4.4 (1.2-15.3) § 36 Caesarean delivery (≥22 weeks) 8 (72.7) 42 (48.3) 2.1 (0.6-7.6) 37 38 Induction of labour 3 (25.0) 25 (28.7) 0.8 (0.2-3.3) 39 Augmentation of labour 2 (16.7) 26 (29.9) 0. (0.09-2.2) 40 Cause of haemorrhage 41 Atony 7 (63.6) 52 (62.7) 0.9 (0.2-3.2) 42 43 AIP and AIP+ praevia 7 (58.3) 14 (16.1) 7.3 (2.0-26.3) § 44 Surgical trauma 2 (16.7) 6 (6.9) 2.7 (0.4-15.2) 45 Genital tract laceration 1 (8.3) 10 (11.5) 0.7 (0.08-6.0) 46 st nd 1 and 2 line treatment 47 48 Uterotonic agents 8 (66.7) 82 (94.3) 0.1 (0.02-0.5) § 49 No use of uterotonic agents 4 (33.3) 5 (5.7) 8.2 (1.8-36.8) § 50 51 Intra-uterine balloon 3 (25.0) 26 (29.9) 0.7 (0.1-3.1) 52 Clotting factors IV 2 (16.7) 8 (9.2) 1.9 (0.3-10.6) 53 54 FFP 7 (58.3) 54 (62.1) 0.8 (0.2-2.9) 55 Organisational factors 56 57 Planned intervention 2 (14) 10 (12) 1.5 (0.3-8.0) 58 Weekend and/or non-office hours # 5 / 10 (50.0) 51 / 85 (60.0) 0.6 (0.1-2.4) 59 IR NOT available $ 3 (25.0) 10 (11.4) 2.5 (0.6-11.1) 60 Emergency IR NOT available $ 6 (50.0) 20 (22.9) 3.3 (0.9-11.5) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 41 BMJ Open

1 2 3 Postnatal transfer 3 (25.0) 13 (14.9) 1.9 (0.4-7.9) 4 IR: interventional radiology; AIP: abnormally invasive placenta; FFP: Fresh Frozen Plasma ; CD: Caesarean delivery 5 # based on date and time of delivery $ 6 IR service availability (24/7 and not 24/7) in the maternity unit where the woman gave birth: $ 7 Emergency IR service availability (24/7) in the maternity unit where the woman gave birth 8 § p < 0.05 9 10 419 11 12 420 13 14 15 421 Maternal outcomeFor and additional peer morbidity review only 16 17 422 Blood transfusion was given to 139 women (83.7%, missing data for 2 women), with a median of 6.0 units of 18 19 423 red blood cells (RBC) (range 0-31), 4.0 units of fresh frozen plasma (FFP) (range 0-31) and 0.5 units of platelets 20 21 22 424 (range 0-29). Fifty-one women (30.7%) were transfused ≥ 8 units of RBC and 23 women were administered 23 24 425 clotting factors (fibrinogen, recombinant factor VIIa, PPSB (Prothrombin-Proconvertin-Stuart Factor- 25 26 426 Antihemophilic Factor B), or a combination) . Hundred-fourteen women (68.6%, missing data in 4 women) were 27 28 29 427 admitted to an intensive care unit (ICU) for a median of 2.0 days (range 0-18 days); the median hospital stay 30 31 428 was 8.0 days (range 2-33 days). Five women needed relaparotomy and 3 women needed arterial embolisation 32 33 429 following hysterectomy because of persistent bleeding. The cause of haemorrhage in these women was AIP 34 35 36 430 (n=1), AIP and atony (n=2), atony (n=3) and in 1 case the cause was unclear (n=1). 37 38 431 Disseminated intravascular coagulation (DIC) secondary to major haemorrhage was reported in 32 women 39 40 432 (19%). Other reported maternal complications associated with major haemorrhage were: renal failure (n=2), 41 42 433 pulmonary oedema (n=4), acute respiratory distress syndrome (n=2), myocardial ischemia (n=1), pleural and 43 44 45 434 pericardial effusion (n=1). Complications associated with peripartum hysterectomy were ureteral injuries (n=2) 46 47 435 and ileus (n=4). One woman suffered from pain in loins and legs, which was the only reported complication 48 49 436 attributed to the IR procedure. 50 51 52 437 One mother died (case fatality rate 0.6% (95% CI 0.1-3.3)). She underwent an emergency caesarean delivery 53 54 438 because of cardiac arrest caused by massive pulmonary embolism. She survived the hysterectomy performed 55 56 439 because of major bleeding due to uterine atony and was transferred to the intensive care unit of a tertiary 57 58 59 440 referral centre using Extracorporeal Life Support (ECLS). She underwent a re-laparotomy because of major 60 441 hepatic haemorrhage, but died the subsequent day during pulmonary thrombectomy. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 41

1 2 3 442 Considering the whole group, the outcome in women who underwent hysterectomy did not significantly differ 4 5 443 from women who underwent IR, regarding transfusion rate and ICU admission, except for a significant 6 7 444 difference in total duration of hospitalisation (median 9 days, range (2-30) versus median 7 days (range 2-33)). 8 9 10 445 When comparing the women according to the IR service availability in the maternity unit where they gave 11 12 446 birth, the rate of peripartum hysterectomy was significantly higher in the group of women who delivered in a 13 14 447 non-IR unit (29 women (74%)) compared to the group of women who delivered in a 24/7-IR unit (44 women 15 For peer review only 16 448 (41%)). And so was the rate of massive transfusion (defined as ≥ 8 units of RBC): 17 women (45%) who 17 18 19 449 delivered in a non-IR unit compared to 25 women (24%) who delivered in a 24/7-IR unit. 20 21 450 When comparing the women who gave birth in a non-IR unit according to whether they were transferred 22 23 451 postpartum, there was no significant difference in estimated blood loss (median 2850 mL (range 500-5000) 24 25 26 452 versus median 3500 mL (range 3000-5000), p=0.4) and need for transfusion (median number of units 6.5 27 28 453 (range 0-19) versus 9.5 (range 0-31), p=0.2), but all women who were not transferred (n=25/25) underwent a 29 30 454 hysterectomy compared to 28 percent (n=4/14) in the transferred group (p=0.0). 31 32 33 455 34 35 456 DISCUSSION 36 37 457 The main findings of this study are, firstly, that the prevalence in Belgium of major obstetric haemorrhage 38 39 40 458 requiring peripartum hysterectomy and/or IR is estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries. 41 42 459 Secondly, that AIP was not detected antenatally in 34 percent of the cases in this series, despite the presence 43 44 460 of risk factors. Thirdly, that women with major obstetric haemorrhage who gave birth in a non-IR unit were 45 46 47 461 more likely to undergo an urgent peripartum hysterectomy (74%) compared to women who delivered in a 48 49 462 24/7-IR unit (41%) and, fourthly, that the need for urgent peripartum hysterectomy was averted in 72 percent 50 51 463 of the women who delivered in a non-IR unit who were transferred postnatally, with no significant difference 52 53 54 464 in estimated blood loss and transfusion rate, compared to the women who were not transferred. 55 56 465 Strengths of the study are its methodology by collecting cases prospectively on a monthly basis and its national 57 58 466 design with excellent participation rate and response rate (98.9%) of the Belgian maternity units. 59 60 467 Underreporting of cases, due to the non-mandatory reporting of severe pregnancy complications may have For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 41 BMJ Open

1 2 3 468 biased our results to some extent. The small number of cases, intrinsic to rare but severe obstetric 4 5 469 complications, warrant caution in making strong inferences. Another limitation is the lack of a perspective on 6 7 470 the overall picture of postpartum haemorrhage in Belgium. We have no accurate data of the total number of 8 9 10 471 women with severe postpartum haemorrhage, the near-misses and maternal deaths due to severe 11 12 472 haemorrhage who did not undergo embolisation or hysterectomy. We have no denominator data of the 13 14 473 women that were successfully managed with other second line measures (intra-uterine (balloon) tamponade, 15 For peer review only 16 474 haemostatic brace suturing and ligation of the uterine arteries) during the same study period in Belgium, 17 18 19 475 therefore we can not draw conclusions on the failure rate of these second line measures. These are interesting 20 21 476 topics that can be investigated by B.OSS in future studies. 22 23 477 This nationwide population based series describes how women with major obstetric haemorrhage were 24 25 26 478 managed with either hysterectomy or IR according to the circumstances of the event. However, this study does 27 28 479 not allow making firm conclusions on the motivation or the appropriateness of this management. One should 29 30 480 be aware that obstetricians are forced to make rapid decisions when dealing with a major obstetric 31 32 33 481 haemorrhage. Belgian obstetricians cannot rely on national guidelines for postpartum haemorrhage or 34 35 482 abnormally invasive placenta. Small volume maternity units even lack local hospital guidelines for these 36 37 483 matters. Therefore, the management of major obstetric haemorrhage will differ greatly in between hospitals 38 39 484 and in between gynaecologists. An emergency IR service is available in about 38 percent of Belgian maternity 40 41 42 485 units, complemented with an informal network between hospitals as shown in this registry: twenty women 43 44 486 (12%) were transferred to a centre with IR availability. We can not rule out that this study in fact demonstrates 45 46 487 an over-use of IR procedures and under-use of other second-line measures in the management of postpartum 47 48 49 488 haemorrhage in Belgium. This reasonable hypothesis has been previously suggested by Kayem et al. based on 50 51 489 the results of a population-based study of women with postpartum haemorrhage managed with invasive 52 53 490 therapies in France.(22) 54 55 56 491 The use of peripartum hysterectomy was significantly higher in multiparous women (parity ≥ 3) and in women 57 58 492 who delivered in a non-IR unit. This result may not surprise, moreover, performing a hysterectomy sooner 59 60 493 rather than later must be considered good practice,(6) because losing time trying another second-line measure For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 41

1 2 3 494 can cause further blood loss and further morbidity. The decision is more easily made in women having no 4 5 495 further fertility desire, likewise, the transfer of a hemodynamically unstable patient to an IR service is hardly 6 7 496 ever considered a safe alternative if this IR service is located in a nearby hospital. However, the higher rate of 8 9 10 497 women with massive blood transfusion in the group of women who delivered in a non-IR unit strengthens the 11 12 498 hypothesis that the rate of hysterectomy is not only associated to the IR service availability, but also to the 13 14 499 expert knowledge and experience at larger units in the management of massive obstetric haemorrhage. A 15 For peer review only 16 500 case-by-case in-depth analysis would be necessary to reveal whether the hysterectomies and arterial 17 18 19 501 embolisations performed in this study were appropriate or preventable. A structured audit based on medical 20 21 502 records of 50 peripartum hysterectomies was recently performed in Denmark. This population based audit 22 23 503 revealed that more than 50% of peripartum hysterectomies were avoidable by simple measures, such as the 24 25 26 504 use of intrauterine balloons.(23) 27 28 505 Abnormally invasive placenta (AIP) accounts for 23 percent of the interventions and 86 percent of the planned 29 30 506 interventions in this registry. Targeted papers report good sensitivity and positive predictive value (PPV) of 31 32 33 507 grey-scale ultrasound as the screening tool for AIP in women with a high index of suspicion, enhanced by the 34 35 508 use of transvaginal ultrasound and colour Doppler.(24, 25) Nonetheless, the diagnosis of AIP was missed 36 37 509 antenatally in 34 percent of the women in this study, while all of them had risk factors that should have 38 39 510 increased awareness. Several population-based studies report rather low antenatal diagnosis rates of AIP: 50 40 41 42 511 percent of AIP cases were missed in a study of the UK Obstetric Surveillance System (UKOSS), while 30 percent 43 44 512 of them had a previous caesarean delivery and placenta praevia.(26) Similarly, in the Nordic Obstetric 45 46 513 Surveillance Study (NOSS) 70 percent of AIP cases were missed, while 33 percent had placenta praevia and 39 47 48 49 514 percent had previous caesarean delivery.(27) Knowing this, we believe that antenatal diagnosis of AIP may be 50 51 515 improved by raising the awareness of clinicians of risk factors for AIP. This involves actively asking for previous 52 53 516 uterine surgery at the first antenatal visit, checking the position of the placenta and focussed imaging in the 54 55 56 517 pregnancies at risk.(28) Importantly, there is little information about the sensitivity or positive predictive value 57 58 518 (PPV) of ultrasound or MRI as the screening tool for AIP in women who had no previous uterine surgery or in 59 60 519 women in whom the placenta is not implanted anteriorly in the lower uterine segment,(24) which would be an For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 41 BMJ Open

1 2 3 520 interesting topic for future research. Despite increased awareness a number of cases of abnormal placentation 4 5 521 will still be missed, therefore all maternity units should ensure a state of preparedness to deal with an 6 7 522 unanticipated AIP causing a major haemorrhage at any time. 8 9 10 523 Although numbers are small, this study aligns with the findings in previous studies that maternal outcome is 11 12 524 better when women with suspicion of AIP give birth in specialised tertiary centres and when preventive 13 14 525 measures are taken.(26, 29, 30) A planned preterm caesarean hysterectomy with the placenta left in situ is the 15 For peer review only 16 526 indubitable treatment of choice in women with AIP incorporated in recommendations of RCOG,(14) ACOG, (31) 17 18 19 527 and SMFM.(32) Alternatively, a conservative approach is advocated for women who want to preserve their 20 21 528 fertility and are hemodynamically stable. This means a multidisciplinary approach aiming to leave the placenta 22 23 529 in situ and to preserve the uterus.(33, 34) This study demonstrates that a conservative management of AIP is 24 25 26 530 not well incorporated in the obstetric care in Belgium. The placenta was left in situ in eleven women with AIP 27 28 531 (29%) in this registry. A conservative approach was attempted in 6 women and was successful in 3 women. 29 30 532 Two of these women underwent a planned hysterectomy with one month delay after the caesarean delivery. A 31 32 33 533 small number of papers support this management, leaving the placenta in situ with a planned delayed 34 35 534 hysterectomy, in women having a placenta percreta with infiltration in the bladder and/or other surrounding 36 37 535 tissues and therefore high risk of haemorrhage or adjacent tissue injury.(34-36) The rationale being that this 38 39 536 delay will allow the decrease of the peri-uterine vascularisation in order to prevent haemorrhage. Studies with 40 41 42 537 larger numbers are needed to increase evidence of this methodology claimed to be safer compared to 43 44 538 caesarean hysterectomy in women with severe placenta percreta. 45 46 539 The role of IR in the management of AIP is still a matter of debate. Most guidelines agree that IR can be 47 48 49 540 lifesaving in the treatment of major postpartum haemorrhage.(14, 32) Guidelines currently do not recommend 50 51 541 the routine use of pre-operative placement of intra-arterial catheters to enable balloon occlusion or arterial 52 53 542 embolisation before caesarean hysterectomy or conservative treatment of AIP, based on lack of evidence of 54 55 56 543 benefit.(14, 31, 32, 37) The number of papers reporting benefit of prophylactic IR is growing ever since.(38, 39) 57 58 544 Our population-based study reports a diverse management of women with AIP, inherent to a multi-centre 59 60 545 study. Twelve women with AIP were managed with programmed IR, the need for hysterectomy was averted in For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 41

1 2 3 546 six of them. We cannot make firm conclusions on the success rate of IR in the elective setting because of small 4 5 547 numbers and lack of controls. 6 7 548 This study demonstrated a high success rate (87.8%) of IR in controlling postpartum haemorrhage, consistent 8 9 10 549 with success rates reported by others, ranging between 85 percent and 95 percent.(9-12, 40-42) Interventional 11 12 550 radiology was more likely to fail in women with a caesarean delivery in a previous pregnancy (failure rate 25%), 13 14 551 women with AIP (33%) and women who had no uterotonic agents administered (44%) in univariate analysis, 15 For peer review only 16 552 taking into account the small numbers. Several large hospital-based retrospective studies analysed predictive 17 18 19 553 factors associated with failure of IR in the management of postpartum haemorrhage.(12, 40-43) Most papers 20 21 554 report a lower success rate of IR in women with AIP,(40, 41) which explains to a great extent the lower success 22 23 555 rates in women with previous caesarean delivery and in women with caesarean delivery in the current 24 25 26 556 pregnancy. A population-based study in the Netherlands (10) reported a 25 percent failure rate for IR following 27 28 557 caesarean delivery in the current pregnancy, which was just 16 percent in this study and barely 10 percent in a 29 30 558 large study reported by Lee et al. (n=176).(44) Other researchers consistently reported a higher failure rate of 31 32 33 559 IR in women with different clinical determinants related to major haemorrhage: hemodynamic shock,(43, 45) 34 35 560 diffuse intravascular coagulopathy,(12, 45), low fibrinogen and prothrombin time (41) and need for massive 36 37 561 transfusion.(12, 40, 41) Yet these factors may be both cause and consequence of failure of IR, which was not 38 39 562 clearly differentiated in some of the aforementioned studies. Interestingly, hospital to hospital transfer was not 40 41 42 563 associated with a higher failure rate of IR in this study, consistent with what has been reported by others.(12, 43 44 564 40, 41) 45 46 47 565 48 49 50 566 CONCLUSION 51 52 53 567 The prevalence in Belgium of major obstetric haemorrhage requiring peripartum hysterectomy and/or IR is 54 55 568 estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries: roughly 1 in 3030 women who give birth in Belgium 56 57 569 undergoes a peripartum hysterectomy, while another 1 in 3030 women is successfully managed by IR thereby 58 59 60 570 preserving the uterus. Abnormal placentation and/or uterine atony are reported as the cause of haemorrhage For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 41 BMJ Open

1 2 3 571 and reason for intervention in the majority of these women (83.7%). Further improvement in the management 4 5 572 of obstetric haemorrhage could possibly be realised by increased awareness of clinicians of risk factors for 6 7 573 abnormal placentation and antenatal transfer to tertiary centres in case of suspicion. A case-by-case in-depth 8 9 10 574 analysis is necessary to reveal whether the hysterectomies and arterial embolisations performed in this series 11 12 575 were appropriate or preventable. 13 14 576 15 For peer review only 16 17 577 ACKNOWLEDGEMENTS 18 19 578 This study would not have been possible without the voluntary participation of the Belgian maternity units: 20 21 579 GZA Sint-Jozef, Mortsel; GZA Sint-Augustinus, Wilrijk; GZA Sint-Vincentius, Antwerpen; ZNA, Sint-Erasmus, 22 23 24 580 Borgerhout; ZNA, Jan Palfijn , Merksem; ZNA, Middelheim, Antwerpen; ASZ, campus Aalst; ASZ, campus 25 26 581 Geraardsbergen; AZ Lokeren; Imelda Ziekenhuis, Bonheiden; Sint-Jozef ziekenhuis, Bornem; AZ Klina, 27 28 582 Brasschaat; AZ Sint-Jan, Brugge; AZ Sint-Lucas, Assebroek; Centre de Santé des Fagnes; Centre Hospitalier de 29 30 583 l’Ardenne; Centre Hospitalier EpiCura, site Hornu; Centre Hospitalier de Mouscron; Centre Hospitalier du Bois 31 32 33 584 de l’Abbaye et de Hesbay ; Centre Hospitalier Peltzer, La Tourelle ; CHC Saint-Vincent; CHC Saint-Joseph; CHC 34 35 585 Sainte-Elisabeth; CHIREC, Clinique Sainte-Anne, Saint-Remi; CHIREC Braine-l’Alleud-Waterloo; CHR de la 36 37 586 Citadelle; CHR de Namur; CHR du Val de Sambre; CHR Haute Senne-Le Tilleriau; CHR Mons Clinique Saint- 38 39 40 587 Joseph; CHU Ambroise Paré; CHU Brugmann; CHU Charleroi, André Vésale; CHU Notre-Dame des Bruyères; 41 42 588 CHU Saint-Pierre; CHU Tivoli; CHwapi-Site Notre-Dame; Clinique Edith Cavell; Clinique et Maternité Sainte 43 44 589 Elisabeth; Clinique Notre dame de Grâce; Clinique Reine Astrid; Clinique Sainte-Piere; Cliniques de l’Europe, 45 46 47 590 Saint-Michel; Cliniques de l’Europe, Sainte-Elisabeth; Cliniques du Sud Luxembourg; Cliniques Universitaires 48 49 591 Saint-Luc; AZ Sint-Vincentius, Deinze; AZ Sint-Blasius , Dendermonde; AZ Monica, Deurne; AZ Diest; AZ Sint- 50 51 592 Maarten, Duffel; AZ Alma, Eeklo; Dimpna Ziekenhuis, Geel; AZ Jan Palfijn, Gent; AZ Maria Middelares, Gent; AZ 52 53 54 593 Sint-Lucas, Gent; GHDC - Notre Dame, Charleroi; AZ Maria ziekenhuis, Halle; Jessaziekenhuis, Hasselt; AZ Sint 55 56 594 Elisabeth, Herentals; CAZ Midden-Limburg, Heusden-Zolder; Hôpital Civil Marie Curie; Hôpital Erasme, 57 58 595 Bruxelles; Hôpitaux Iris Sud – Site Ixelles ; Jan Yperman ziekenhuis, Ieper; IFAC Hôpital Princesse Paola; INDC 59 60 596 Entité Jolimontoise; Sint-Jozef ziekenhuis, Izegem; Kliniek Sint-Jan; Klinik Saint-Josef; AZ Zeno, campus Knokke- For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 41

1 2 3 597 Heist; AZ Groeninge, Kortrijk; Heilig Hart ziekenhuis, Leuven; UZ Leuven; Heilig Hart ziekenhuis, Lier; AZ Sint- 4 5 598 Jozef, Malle; AZ Sint-Maarten, Mechelen; AZ Delta, campus Menen; Heilig Hart ziekenhuis, Mol; Onze-Lieve- 6 7 599 Vrouw ziekenhuis, campus Aalst; Onze-Lieve-Vrouw ziekenhuis, campus Asse; AZ Damiaan, campus Sint-Jozef, 8 9 10 600 Oostende; AZ Sint-Jan, campus Oostende; AZ Oudenaarde; Mariaziekenhuis Noord-Limburg, Overpelt; AZ 11 12 601 Heilige Familie, Reet AZ Delta, campus Brugsesteenweg, Roeselare; AZ Delta, campus Wilgenstraat, Roeselare; 13 14 602 AZ Glorieux, Ronse; AZ Nikolaas, Sint-Niklaas; Sint-Trudo ziekenhuis, Sint-Truiden; Saint-Nikolaus Hospital; Sint- 15 For peer review only 16 603 Andries ziekenhuis, Tielt; Heilig Hartziekenhuis, Tienen; AZ Vesalius, Tongeren; Sint-Rembertziekenhuis, 17 18 19 604 Torhout; AZ Turnhout; UZ Antwerpen; UZ Brussel; AZ Sint-Augustinus, Veurne; AZ Jan Portaels, Vilvoorde; 20 21 605 Onze-Lieve-Vrouw van Lourdes ziekenhuis, Waregem; Ziekenhuis Oost-Limburg, campus Sint-Jan, Genk; AZ 22 23 606 Sint-Elisabeth, Zottegem. 24 25 26 607 We would like to thank Fatima Bercha, Marlies Deblaere and Ann Langedock for their contribution in collecting 27 28 608 the data. 29 30 609 31 32 33 610 AUTHOR CONTRIBUTIONS 34 35 611 Conceived the project: MH, YE. Designed the study: GV, VVL, JVK, MH, YE. Collected the data: GV, MG, IR, VVL, 36 37 38 612 JVK. Analysed the data: GV, MG, IJ, ER. Interpreted the data: GV, MG, IJ, KR, MH, ER, YE, HV. Wrote the first 39 40 613 draft of the manuscript: GV. Contributed to the writing of the manuscript: MG, IJ, KR, MH, YE, HV. Critically 41 42 614 revised the manuscript: MG, IJ, VVL, KR, MH, ER, JV, YE, HV. Approved the final version: GV, MG, IJ, VVL, KR, 43 44 45 615 MH, ER, JV, YE, HV. Agreed to be accountable for all aspects of the work: GV, MG, IJ, VVL, KR, MH, ER, JV, YE, 46 47 616 HV. ICMJE criteria for authorship met: GV, MG, IJ, VVL, KR, MH, ER, JV, YE, HV. 48 49 617 50 51 52 618 FUNDING 53 54 619 The study has been funded by the College for Mother and Newborn, a consultative body of the Federal Public 55 56 620 Service of Health in Belgium. GV has been funded by the Research Foundation Flanders (FWO). 57 58 59 621 60 622 COMPETING INTERESTS For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 41 BMJ Open

1 2 3 623 All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and 4 5 624 declare: all authors had no financial relationships with any organisations that might have an interest in the 6 7 625 submitted work in the previous three years; no other relationships or activities that could appear to have 8 9 10 626 influenced the submitted work. 11 12 627 13 14 628 DATA SHARING 15 For peer review only 16 17 629 No additional data are available. 18 19 630 20 21 631 ETHICS APPROVAL 22 23 24 632 The B.OSS methodology and this study were approved by the Medical Ethics Committee of the Ghent 25 26 633 University Hospital (EC UZG 2012/734; B670201215359 and EC UZG 2015/1470; B670201526875) and by the 27 28 634 Medical Ethics Committee of the University Hospital Erasme, Brussels (EC ULB 2012/111; B406201213660). 29 30 31 635 Data were retrieved in retrospect from the case notes by means of data collection forms. No personally 32 33 636 identifiable information was obtained. Participants were informed and enabled to opt-out. 34 35 637 36 37 38 638 REFERENCES 39 40 639 1. Evaluating the quality of care for severe pregnancy complications: the WHO near-miss approach for 41 42 43 640 maternal health. Geneva, Switzerland: WHO press 2011. Retrieved from 44 45 641 http://apps.who.int/iris/bitstream/10665/44692/1/9789241502221_eng.pdf. Accessed 22 Dec. 16. 46 47 642 2. Stones W, Lim W, Al-Azzawi F, Kelly M. An investigation of maternal morbidity with identification of 48 49 50 643 life-threatening 'near miss' episodes. Health trends. 1991;23(1):13-5. 51 52 644 3. Say L, Souza JP, Pattinson RC. Maternal near miss--towards a standard tool for monitoring quality of 53 54 645 maternal health care. Best practice & research Clinical obstetrics & gynaecology. 2009;23(3):287-96. 55 56 646 4. Lewis, G (ed) 2007. The Confidential Enquiry into Maternal and Child Health (CEMACH). Saving 57 58 59 647 Mother's Lives: reviewing maternal deaths to make motherhood safer-2003-2005. The Seventh Report on 60 648 Confidential EnquiriesFor peer into review Maternal only Deaths - http://bmjopen.bmj.com/site/about/guidelines.xhtml in the United Kingdom. London: CEMACH. BMJ Open Page 28 of 41

1 2 3 649 5. Knight M, Tuffnell D, Kenyon S, , Kenyon S, Shakespear J, Gray R, Kurinczuk JJ (Eds.) on behalf of 4 5 650 MBRRACE-UK. Saving Lives, Improving Mother's Care - Surveillance of maternal deaths in the UK 2011-13 and 6 7 651 lessons learned to inform maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths 8 9 10 652 and Morbidity 2009-13. Oxford: National Perinatal Epidemiology Unit, University of Oxford 2015. 11 12 653 6. The Royal College of Obstetricians and Gynaecologists. Postpartum Haemorrhage, Prevention and 13 14 654 Management. Green-top Guideline No. 52. Published: 11/05/2009. 15 For peer review only 16 655 7. Postpartum Hemorrhage. The American College of Obstetricians and Gynaecologistis. ACOG Practice 17 18 19 656 Bulletin. Number 76, October 2006. Reaffirmed 2015. 20 21 657 8. Dahlke JD, Mendez-Figueroa H, Maggio L, Hauspurg AK, Sperling JD, Chauhan SP, et al. Prevention and 22 23 658 management of postpartum hemorrhage: a comparison of 4 national guidelines. American journal of obstetrics 24 25 26 659 and gynecology. 2015;213(1):76.e1-10. 27 28 660 9. Kayem G, Kurinczuk JJ, Alfirevic Z, Spark P, Brocklehurst P, Knight M. Specific second-line therapies for 29 30 661 postpartum haemorrhage: a national cohort study. BJOG : an international journal of obstetrics and 31 32 33 662 gynaecology. 2011;118(7):856-64. 34 35 663 10. Zwart JJ, Dijk PD, van Roosmalen J. Peripartum hysterectomy and arterial embolization for major 36 37 664 obstetric hemorrhage: a 2-year nationwide cohort study in the Netherlands. American journal of obstetrics and 38 39 665 gynecology. 2010;202(2):150.e1-7. 40 41 42 666 11. Inoue S, Masuyama H, Hiramatsu Y. Efficacy of transarterial embolisation in the management of post- 43 44 667 partum haemorrhage and its impact on subsequent pregnancies. The Australian & New Zealand journal of 45 46 668 obstetrics & gynaecology. 2014;54(6):541-5. 47 48 49 669 12. Lee HY, Shin JH, Kim J, Yoon HK, Ko GY, Won HS, et al. Primary postpartum hemorrhage: outcome of 50 51 670 pelvic arterial embolization in 251 patients at a single institution. Radiology. 2012;264(3):903-9. 52 53 671 13. The Royal College of Obstetricians and Gynaecologists. The role of emergency and elective 54 55 56 672 interventional radiology in postpartum haemorrhage. (Good Practice No. 6). June 2007. 57 58 673 14. The Royal College of Obstetricians and Gynaecologists. Placenta Praevia, Placenta Praevia Accreta and 59 60 674 Vasa Praevia : Diagnosis and Management. Green-top Guideline No. 27. Published January 2011. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 41 BMJ Open

1 2 3 675 15. Investigation into 10 Maternal Deaths at, or following Delivery at, Northwick Park Hospital, North West 4 5 676 London Hospitals NHS Trust, between April 2002 and April 2005. Commission for Healthcare Audit and 6 7 677 Inspection. London, August 2006. 8 9 10 678 16. Vandenberghe G, De Blaere M, Van Leeuw V, Roelens K, Englert Y, Hanssens M, et al. Nationwide 11 12 679 population-based cohort study of uterine rupture in Belgium: results from the Belgian Obstetric Surveillance 13 14 680 System. BMJ open. 2016;6(5):e010415. 15 For peer review only 16 681 17. H. Cammu EM, G. Martens, C. Van Mol, Y Jacquemyn. Perinatale activiteiten in Vlaanderen 2013. 2014. 17 18 19 682 Retrieved from: https://www.zorg-en- 20 21 683 gezondheid.be/sites/default/files/atoms/files/SPE_jaarrapport%202013.pdf. Accessed on 22 Dec. 16 22 23 684 18. H. Cammu EM, G. Martens, C. Van Mol, Y Jacquemyn. Perinatale Activiteiten in Vlaanderen 2012. 2013. 24 25 26 685 Retrieved from: https://www.zorg-en- 27 28 686 gezondheid.be/sites/default/files/atoms/files/SPE_jaarrapport%202012.pdf. Accessed on 22 Dec. 16. 29 30 687 19. Van Leeuw V LC, Englert Y. Perinatale gegevens in het Brussels Gewest - Jaren 2013. Centre 31 32 33 688 d'Epidémiologie périnatale, 2014. Retrieved from: http://www.cepip.be/pdf/rapport_CEPIP_Bxl2013_NL.pdf. 34 35 689 Accessed on 22 Dec. 16. 36 37 690 20. Leroy Ch VLV, Englert Y. Données périnatales en Wallonie - Année 2013. Centre d'Epidémiologie 38 39 691 Périnatale, 2014. Retrieved from : http://www.cepip.be/pdf/rapport_CEPIP_wallonie2013_tma.pdf. Accessed 40 41 42 692 on 22 Dec. 16. 43 44 693 21. Armonk NIC. IBM SPSS Statistics for Windows, Version 22.0. In: Corp I, editor. 2013. 45 46 694 22. Kayem G, Dupont C, Bouvier-Colle MH, Rudigoz RC, Deneux-Tharaux C. Invasive therapies for primary 47 48 49 695 postpartum haemorrhage: a population-based study in France. BJOG : an international journal of obstetrics and 50 51 696 gynaecology. 2016;123(4):598-605. 52 53 697 23. Colmorn LB, Krebs L, Langhoff-Roos J. Potentially Avoidable Peripartum Hysterectomies in Denmark: A 54 55 56 698 Population Based Clinical Audit. PloS one. 2016;11(8):e0161302. 57 58 699 24. Comstock CH, Bronsteen RA. The antenatal diagnosis of placenta accreta. BJOG : an international 59 60 700 journal of obstetrics and gynaecology. 2014;121(2):171-81; discussion 81-2. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 30 of 41

1 2 3 701 25. D'Antonio F, Iacovella C, Bhide A. Prenatal identification of invasive placentation using ultrasound: 4 5 702 systematic review and meta-analysis. Ultrasound in obstetrics & gynecology : the official journal of the 6 7 703 International Society of Ultrasound in Obstetrics and Gynecology. 2013;42(5):509-17. 8 9 10 704 26. Fitzpatrick KE, Sellers S, Spark P, Kurinczuk JJ, Brocklehurst P, Knight M. The management and 11 12 705 outcomes of placenta accreta, increta, and percreta in the UK: a population-based descriptive study. BJOG : an 13 14 706 international journal of obstetrics and gynaecology. 2014;121(1):62-70; discussion -1. 15 For peer review only 16 707 27. Thurn L, Lindqvist PG, Jakobsson M, Colmorn LB, Klungsoyr K, Bjarnadottir RI, et al. Abnormally invasive 17 18 19 708 placenta-prevalence, risk factors and antenatal suspicion: results from a large population-based pregnancy 20 21 709 cohort study in the Nordic countries. BJOG : an international journal of obstetrics and gynaecology. 22 23 710 2016;123(8):1348-55. 24 25 26 711 28. Collins SL, Ashcroft A, Braun T, Calda P, Langhoff-Roos J, Morel O, et al. Proposal for standardized 27 28 712 ultrasound descriptors of abnormally invasive placenta (AIP). Ultrasound in obstetrics & gynecology : the 29 30 713 official journal of the International Society of Ultrasound in Obstetrics and Gynecology. 2016;47(3):271-5. 31 32 33 714 29. Chantraine F, Braun T, Gonser M, Henrich W, Tutschek B. Prenatal diagnosis of abnormally invasive 34 35 715 placenta reduces maternal peripartum hemorrhage and morbidity. Acta obstetricia et gynecologica 36 37 716 Scandinavica. 2013;92(4):439-44. 38 39 717 30. Eller AG, Bennett MA, Sharshiner M, Masheter C, Soisson AP, Dodson M, et al. Maternal morbidity in 40 41 42 718 cases of placenta accreta managed by a multidisciplinary care team compared with standard obstetric care. 43 44 719 Obstetrics and gynecology. 2011;117(2 Pt 1):331-7. 45 46 720 31. Placenta Accreta. The American College of Obstetricians and Gynaecologistis. Committee Opinion. 47 48 49 721 Number 529. Published July 2012. Reaffirmed 2015. 50 51 722 32. Belfort MA. Placenta accreta. American journal of obstetrics and gynecology. 2010;203(5):430-9. 52 53 723 33. Kayem G, Davy C, Goffinet F, Thomas C, Clement D, Cabrol D. Conservative versus extirpative 54 55 56 724 management in cases of placenta accreta. Obstetrics and gynecology. 2004;104(3):531-6. 57 58 725 34. Sentilhes L, Ambroselli C, Kayem G, Provansal M, Fernandez H, Perrotin F, et al. Maternal outcome 59 60 726 after conservative treatment of placenta accreta. Obstetrics and gynecology. 2010;115(3):526-34. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 31 of 41 BMJ Open

1 2 3 727 35. Fox KA, Shamshirsaz AA, Carusi D, Secord AA, Lee P, Turan OM, et al. Conservative management of 4 5 728 morbidly adherent placenta: expert review. American journal of obstetrics and gynecology. 2015;213(6):755- 6 7 729 60. 8 9 10 730 36. Chantraine F, Nisolle M, Petit P, Schaaps JP, Foidart JM. Individual decisions in placenta increta and 11 12 731 percreta: a case series. Journal of perinatal medicine. 2012;40(3):265-70. 13 14 732 37. Dilauro MD, Dason S, Athreya S. Prophylactic balloon occlusion of internal iliac arteries in women with 15 For peer review only 16 733 placenta accreta: literature review and analysis. Clinical radiology. 2012;67(6):515-20. 17 18 19 734 38. Cali G, Forlani F, Giambanco L, Amico ML, Vallone M, Puccio G, et al. Prophylactic use of intravascular 20 21 735 balloon catheters in women with placenta accreta, increta and percreta. European journal of obstetrics, 22 23 736 gynecology, and reproductive biology. 2014;179:36-41. 24 25 26 737 39. Duan XH, Wang YL, Han XW, Chen ZM, Chu QJ, Wang L, et al. Caesarean section combined with 27 28 738 temporary aortic balloon occlusion followed by uterine artery embolisation for the management of placenta 29 30 739 accreta. Clinical radiology. 2015;70(9):932-7. 31 32 33 740 40. Sentilhes L, Gromez A, Clavier E, Resch B, Verspyck E, Marpeau L. Predictors of failed pelvic arterial 34 35 741 embolization for severe postpartum hemorrhage. Obstetrics and gynecology. 2009;113(5):992-9. 36 37 742 41. Poujade O, Zappa M, Letendre I, Ceccaldi PF, Vilgrain V, Luton D. Predictive factors for failure of pelvic 38 39 743 arterial embolization for postpartum hemorrhage. International journal of gynaecology and obstetrics: the 40 41 42 744 official organ of the International Federation of Gynaecology and Obstetrics. 2012;117(2):119-23. 43 44 745 42. Yamasaki Y, Morita H, Miyahara Y, Ebina Y, Okada T, Yamaguchi M, et al. The factors associated with 45 46 746 the failure of transcatheter pelvic arterial embolization for intractable postpartum hemorrhage. Journal of 47 48 49 747 perinatal medicine. 2014;42(3):359-62. 50 51 748 43. Touboul C, Badiou W, Saada J, Pelage JP, Payen D, Vicaut E, et al. Efficacy of selective arterial 52 53 749 embolisation for the treatment of life-threatening post-partum haemorrhage in a large population. PloS one. 54 55 56 750 2008;3(11):e3819. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 41

1 2 3 751 44. Lee HJ, Jeon GS, Kim MD, Kim SH, Lee JT, Choi MJ. Usefulness of pelvic artery embolization in cesarean 4 5 752 section compared with vaginal delivery in 176 patients. Journal of vascular and interventional radiology : JVIR. 6 7 753 2013;24(1):103-9. 8 9 10 754 45. Kim YJ, Yoon CJ, Seong NJ, Kang SG, An SW, Kim YS, et al. Failed pelvic arterial embolization for 11 12 755 postpartum hemorrhage: clinical outcomes and predictive factors. Journal of vascular and interventional 13 14 756 radiology : JVIR. 2013;24(5):703-9. 15 For peer review only 16 17 757 18 19 758 20 21 22 759 23 24 760 25 26 27 761 28 29 762 30 31 32 763 33 34 764 35 36 37 765 38 39 766 40 41 42 767 43 44 768 45 46 47 769 48 49 770 50 51 52 771 53 54 772 55 56 57 773 58 59 774 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 33 of 41 BMJ Open

1 2 3 775 FIGURE LEGENDS 4 5 776 Figure 1. Flowchart of case reporting and data collection. 6 7 777 Figure 2. Circumstances in which peripartum hysterectomy and interventional radiology occurred in 166 8 9 10 778 women. 11 12 779 Figure 3. Cause of obstetric haemorrhage according to the 4T’s Mnemonics. 13 14 780 Figure 4. Antenatal suspicion and preventive measures in women with abnormally invasive placenta (AIP). 15 For peer review only 16 17 781 Supplementary Figure 1. Antenatal suspicion and preventive measures in women with abnormally invasive 18 19 782 placenta (AIP). 20 21 783 22 23 24 784 25 26 785 27 28 786 29 30 787 31 32 33 788 34 35 789 36 37 790 38 39 40 791 41 42 792 43 44 793 45 46 47 794 48 49 795 50 51 796 52 53 54 797 55 56 798 57 58 799 59 60 800 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 34 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 293x157mm (300 x 300 DPI) 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 35 of 41 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Figure 1. Flowchart of case reporting and data collection.

48 143x212mm (300 x 300 DPI) 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 36 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Figure 2. Circumstances in which peripartum hysterectomy and interventional radiology took place in 166 women. 48 49 219x267mm (300 x 300 DPI) 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 37 of 41 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Figure 3. Cause of obstetric haemorrhage according to the 4T's mnemonics. 36 Tonus: uterine atony (n=91); Tissue: AIP (n=14), placenta praevia (n=18), combination of AIP and placenta 37 praevia (n=24), placental remnants or retained placenta without AIP (n=26). Trauma: uterine rupture 38 (n=12), genital tract lacerations (n=12), bleeding caused by unintended injuries during caesarean delivery 39 (n=14) or D&C (n=3). Uncomplicated caesarean deliveries are not included in the trauma group. Thr ombin: abnormal coagulation before onset of bleeding (n=6). DIC secondary to major haemorrhage was not 40 included in the thrombin group. Others: the cause of bleeding remains unclear or is unknown. 41 42 171x143mm (300 x 300 DPI) 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 38 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Figure 4. Antenatal suspicion and preventive measures in women with abnormally invasive placenta (AIP). 30 § < 0.05 for difference in between groups. 31 32 373x243mm (300 x 300 DPI) 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 39 of 41 BMJ Open

1 2 3 4 STROBE 2007 (v4) Statement—Checklist of items that should be included in reports of cohort studies 5 6 Item 7 Section/Topic Recommendation Reported on page # 8 # 9 Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the title or the abstract Title Page 10 11 (b) ForProvide in the abstract peer an informative and balancreviewed summary of what was doneonly and what was found Abstract 12 Introduction 13 Background/rationale 2 Explain the scientific background and rationale for the investigation being reported Background p6-7 14 15 Objectives 3 State specific objectives, including any prespecified hypotheses Background p7 16 Methods 17 18 Study design 4 Present key elements of study design early in the paper Method p7 19 Setting 5 Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data Methods p7-8 20 collection 21 Participants 6 (a) Give the eligibility criteria, and the sources and methods of selection of participants. Describe methods of follow-up Not applicable 22 23 Methods p7 24 (b) For matched studies, give matching criteria and number of exposed and unexposed / 25 Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if Methods p8-10 26 applicable 27 28 Data sources/ 8* For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe Methods p9-10 29 measurement comparability of assessment methods if there is more than one group 30 Bias 9 Describe any efforts to address potential sources of bias / 31 Study size 10 Explain how the study size was arrived at Methods P9 and 32 33 Figure 1 34 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and / 35 why 36 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding Methods p9-10 37 38 (b) Describe any methods used to examine subgroups and interactions Methods p9-10 39 (c) Explain how missing data were addressed / 40 (d) If applicable, explain how loss to follow-up was addressed / 41 42 (e) Describe any sensitivity analyses / 43

44 45 46 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 40 of 41

1 2 3 4 Results 5 6 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers potentially eligible, examined for eligibility, confirmed Figure 1 7 eligible, included in the study, completing follow-up, and analysed 8 (b) Give reasons for non-participation at each stage / 9 (c) Consider use of a flow diagram Figure 1 10 11 Descriptive data 14* (a) ForGive characteristics peer of study participants (egreview demographic, clinical, social) and only information on exposures and potential Table 1 12 confounders 13 (b) Indicate number of participants with missing data for each variable of interest Results, Table 1, p11 14 (c) Summarise follow-up time (eg, average and total amount) / 15 16 Outcome data 15* Report numbers of outcome events or summary measures over time Results , prevalence 17 p10 18 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (eg, 95% confidence Results , prevalence 19 interval). Make clear which confounders were adjusted for and why they were included p10 20 (b) Report category boundaries when continuous variables were categorized Table 1 21 22 (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period / 23 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and sensitivity analyses Results, 24 - management of 25 AIP, p15-16, figure 4 26 27 - IR versus H, table 3 28 - Failed E versus 29 succesfull E, table 4 30 31 Discussion 32 Key results 18 Summarise key results with reference to study objectives Discussion p20 33 Limitations 34 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from Discussion p21-24 35 similar studies, and other relevant evidence 36 37 Generalisability 21 Discuss the generalisability (external validity) of the study results / 38 Other information 39 Funding 22 Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on Funding p26 40 41 which the present article is based 42 43

44 45 46 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 41 of 41 BMJ Open

1 2 3 4 5 *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. 6

7 8 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and published examples of transparent reporting. The STROBE 9 checklist is best used in conjunction with this article (freely available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 10 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is available at www.strobe-statement.org. 11 For peer review only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

44 45 46 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open

A nationwide population -based cohort study of peripartum hysterectomy and arterial embolisation in Belgium: results from the Belgian Obstetric Surveillance System.

For peer review only Journal: BMJ Open

Manuscript ID bmjopen-2017-016208.R2

Article Type: Research

Date Submitted by the Author: 06-Jun-2017

Complete List of Authors: Vandenberghe, Griet; Ghent University Hospital, Department of Obstetrics and Gynaecology Guisset, Marine; Leuven University Hospital, Department of Obstetrics and Gynaecology Janssens, Iris; Ghent University, Faculty of Medicine Van Leeuw, Virginie; Université Libre de Bruxelles (ULB), School of Public Health; Perinatal Epidemiology Centre (Centre d'Épidémiologie Périnatale, CEpiP) Roelens, Kristien; Ghent University Hospital, Department of Obstetrics and Gynaecology Hanssens, Myriam; University Hospital Leuven, Department of Obstetrics & Gynaecology Russo, Erika; Intercommunale de Santé Publique du Pays de Charleroi (ISSPC), Department of Obstetrics and Gynaecology Vankeirsbilck, Joachim; St Jan's Hospital, Department of Obstetrics and Gynaecology Englert, Yvon; Université Libre de Bruxelles, Faculty of Medicine, Research Laboratory on Human Reproduction; Perinatal Epidemiology Center (Centre d'Épidémiologie Périnatale, CEpiP) Verstraelen, Hans; Ghent University Hospital, Department of Obstetrics and Gynaecology

Primary Subject Obstetrics and gynaecology Heading:

Secondary Subject Heading: Epidemiology

peripartum hysterectomy, arterial embolisation, Interventional radiology < Keywords: RADIOLOGY & IMAGING, maternal near miss, severe maternal morbidity

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 41 BMJ Open

1 2 3 1 TITLE PAGE 4 5 2  Title of the article: 6 7 3 A nationwide population-based cohort study of peripartum hysterectomy and arterial embolisation in 8 9 4 Belgium: results from the Belgian Obstetric Surveillance System. 10 5 11 12 6  Author and Co-Author’s name: 13

14 7 Vandenberghe G, Guisset M, Janssens I, Van Leeuw V, Roelens K, Hanssens M, Russo E, Van Keirsbilck J, 15 For peer review only 16 8 Englert Y, Verstraelen H. 17 9 18 19 10  Corresponding author: Griet Vandenberghe 20 21 11 Department of Obstetrics & Gynaecology 22 12 Ghent University Hospital 23 24 13 De Pintelaan 185, 9000 Ghent, Belgium 25 26 14 E-mail: [email protected] 27 28 15 Telephone number: +32 9 33 27 849 29 16 30 31 17  Co-author: Marine Guisset 32 33 18 Department of Obstetrics & Gynaecology 34 35 19 Leuven University Hospital 36 20 Leuven, Belgium 37 38 21 E-mail: [email protected] 39 40 22 41 23  Co-author: Iris Janssens 42 43 24 Faculty of Medicine 44 45 25 Ghent University 46 47 26 Ghent, Belgium 48 27 E-mail: [email protected] 49 50 28 51 52 29  Co-author: Virginie Van Leeuw 53 54 30 Perinatal Epidemiology Center (Centre d’Épidémiologie Périnatale, CEpiP) 55 31 Université Libre de Bruxelles (ULB), School of Public Health. 56 57 32 Brussels, Belgium 58 59 33 E-mail: [email protected] 60 34 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 41

1 2 3 35  Co-author: Kristien Roelens 4 36 Department of Obstetrics & Gynaecology 5 6 37 Ghent University Hospital 7 8 38 Ghent, Belgium 9 10 39 E-mail: [email protected] 11 40 12 13 41  Co-author: Myriam Hanssens 14 15 42 Department Forof Obstetrics peer & Gynaecology review only 16 43 University Hospital Leuven 17 18 44 Leuven, Belgium 19 20 45 E-mail: [email protected] 21 22 46 23 47  Co-author: Erika Russo 24 25 48 Department of Obstetrics & Gynaecology 26 27 49 Intercommunale de Santé Publique du Pays de Charleroi (ISPPC) 28 29 50 Hôpital Civil Marie Curie 30 51 Charleroi, Belgium 31 32 52 E-mail: [email protected] 33 34 53 35 54  Co-author: Joachim Vankeirsbilck 36 37 55 Department of Obstetrics & Gynaecology 38 39 56 St Jan’s Hospital 40 41 57 Bruges, Belgium 42 58 E-mail: [email protected] 43 44 59 45 46 60  Co-author: Yvon Englert 47 48 61 Perinatal Epidemiology Center (Centre d’Épidémiologie Périnatale, CEpiP) 49 62 Université Libre de Bruxelles (ULB), Faculty of Medicine, Research Laboratory on Human Reproduction. 50 51 63 Brussels, Belgium 52 53 64 E-mail: [email protected] 54 65 55 56 66  Co-author: Hans Verstraelen 57 58 67 Department of Obstetrics & Gynaecology 59 60 68 Ghent University Hospital 69 Ghent, BelgiumFor peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 41 BMJ Open

1 2 3 70 E-mail: [email protected] 4 71 5 6 72  Keywords: peripartum hysterectomy, arterial embolisation, interventional radiology, major obstetric 7 8 73 haemorrhage, maternal near miss, severe maternal morbidity 9 10 74 11 75  Word count: The main text consists of 5772 words 12 13 76 14 15 77 Background:For peer 450 words review only 16 78 Methods: 964 words 17 18 79 Results: 2509 words 19 20 80 Discussion: 1714 words 21 22 81 Conclusion: 135words 23 24 82 Total: 5772 words 25 83 26 27 84  Tables and Figures: The manuscript includes 4 tables and 4 figures and 1 supplementary figure. Figures 28 29 85 are uploaded, in JPEG format with a resolution of 300 DPI, separately from the text. 30 86 31 32 87 33 34 88 35 36 89 37 90 38 39 91 40 41 92 42 93 43 44 94 45 46 95 47 48 96 49 97 50 51 98 52 53 99 54 55 100 56 57 101 58 59 102 60 103 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 41

1 2 3 104 ABSTRACT 4 5 105 Objectives: 6 7 106 To assess the prevalence of major obstetric haemorrhage managed with peripartum hysterectomy and/or 8 9 107 interventional radiology (IR) in Belgium. To describe women characteristics, the circumstances in which the 10 108 interventions took place, the management of the obstetric haemorrhage, the outcome and additional 11 12 109 morbidity of these women. 13 14 110 Design: 15 For peer review only 16 111 Nationwide population-based prospective cohort study. 17 112 Setting: 18 19 113 Emergency obstetric care. Participation of 97% of maternities covering 98.6% of deliveries in Belgium. 20 21 114 Participants: 22 115 All women who underwent peripartum hysterectomy and/or IR procedures in Belgium between January 2012 23 24 116 and December 2013. 25 26 117 Results: 27 28 118 We obtained data on 166 women who underwent peripartum hysterectomy (n=84) and/or IR procedures 29 30 119 (n=102), corresponding to 1 in 3030 women undergoing a peripartum hysterectomy and another 1 in 3030 31 32 120 women being managed by IR, thereby preserving the uterus. Seventeen women underwent hysterectomy 33 34 35 121 following IR and three women needed further IR despite hysterectomy. Abnormal placentation and/or uterine 36 37 122 atony were the reported causes of haemorrhage in 83.7%. Abnormally invasive placenta was not detected 38 39 123 antenatally in 34% of cases. The interventions were planned in 15 women. Three women were transferred 40 41 124 antenatally and seventeen women postnatally to a hospital with emergency IR service. Urgent peripartum 42 43 44 125 hysterectomy was averted in 72% of the women who were transferred, with no significant difference in need 45 46 126 for transfusion. IR procedures were able to stop the bleeding in 87.8% of the attempts. Disseminated 47 48 127 intravascular coagulation secondary to major haemorrhage was reported in 32 women (19%). 49 50 51 128 Conclusion: 52 129 The prevalence in Belgium of major obstetric haemorrhage requiring peripartum hysterectomy and/or IR is 53 54 55 130 estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries. Increased clinician awareness of the risk factors of 56 57 131 abnormal placentation could further improve the management and outcome of major obstetric haemorrhage. 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 41 BMJ Open

1 2 3 132 A case-by-case in-depth analysis is necessary to reveal whether the hysterectomies and arterial embolisations 4 5 133 performed in this study were appropriate or preventable. 6 7 134 8 9 10 135 11 12 136 13 14 15 137 ARTICLE SUMMARY:For Strengths peer and limitations review of this study: only 16 17 138 18 19 139 • This is a nationwide study of Belgian maternities, with excellent participation and response rates 20 21 140 (98.8%). The cases were collected on a monthly basis using the standard methodology of obstetric 22 23 141 surveillance systems. 24 25 142 • This study describes how women with major obstetric haemorrhage were managed with peripartum 26 27 28 143 hysterectomy and/or IR according to circumstances. 29 30 144 • The study lacks a perspective on the overall picture of postpartum haemorrhage in Belgium, namely 31 32 145 the total number of women suffering severe postpartum haemorrhage, near-misses or death due to 33 34 35 146 postpartum haemorrhage. 36 37 147 • The study lacks denominator data of the women with postpartum haemorrhage that were successfully 38 39 148 managed with other second-line measures. Therefore we cannot draw conclusions on the failure rate 40 41 42 149 of intra-uterine balloon tamponade, haemostatic brace suturing and ligation of the uterine arteries. 43 44 45 150 46 151 47 48 152 49 50 153 51 52 154 53 155 54 55 156 56 57 157 58 59 158 60 159 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 41

1 2 3 160 BACKGROUND 4 5 161 Emergency peripartum hysterectomy is the single most dramatic procedure in modern obstetrics. The 6 7 162 procedure is stressful and surgically challenging, inevitably causing additional maternal morbidity and, not 8 9 10 163 least, infertility. It is listed by the World Health Organisation (WHO) as an identification criterion for maternal 11 12 164 near miss,(1) a concept which, in view of the very low maternal mortality rates in developed country 13 14 165 settings(2), was introduced as an analytical tool to address health system failures with the overall goal of 15 For peer review only 16 17 166 improving obstetric care.(3) Substandard care in the management of major obstetric haemorrhage continues 18 19 167 to be a critical cause of maternal death and severe maternal morbidity in developed countries.(4, 5) Every 20 21 168 obstetric practitioner needs to be trained in the management of postpartum haemorrhage in order to 22 23 24 169 guarantee an immediate response and a multidisciplinary team approach. Internationally recognized guidelines 25 26 170 (6-8) indicate that one or more second-line measures, including intra-uterine (balloon) tamponade, 27 28 171 haemostatic brace suturing, ligation of the uterine arteries and interventional radiology (IR), should be 29 30 172 available in hospitals with delivery units and that obstetric practitioners should be familiar with these 31 32 33 173 procedures. Notably, the early involvement of an experienced senior obstetrician hence appropriate clinical 34 35 174 judgement can save lives: the timely decision to turn to a hysterectomy can avoid further blood loss, risk of 36 37 175 disseminated intravascular coagulation (DIC), additional morbidity and finally death.(6) 38 39 40 176 While emergency peripartum hysterectomy serves as the ultimate response to major obstetric haemorrhage, 41 42 177 resorting to IR procedures on the pathway towards hysterectomy can be both lifesaving and fertility- 43 44 178 preserving. IR in this context basically involves intra-arterial balloon occlusion or arterial embolisation of the 45 46 47 179 uterine or the internal iliac arteries. Research publications on the effectiveness of IR procedures in obstetric 48 49 180 settings reporting high success rates of 85 to 95 percent are accumulating.(9-12) Success rates should be 50 51 181 interpreted with caution however, as they may be driven by ascertainment bias. IR is cautiously suggested in 52 53 54 182 guidelines and recommendations (6-8, 13, 14) describing its possible role in obstetric emergency settings, but 55 56 183 further evidence is awaited before firm recommendations can be made. Yet in the United Kingdom a national 57 58 184 recommendation was made to provide a network of emergency IR services available to all trusts with delivery 59 60 185 units in response to a Healthcare Commission Investigation into maternal deaths.(13, 15) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 41 BMJ Open

1 2 3 186 The aims of this study are to investigate the population-based prevalence of peripartum hysterectomy and the 4 5 187 indications to proceed to this intervention, likewise the prevalence of major obstetric haemorrhage managed 6 7 188 with IR in Belgium. Furthermore, we aim to describe the characteristics of the women in Belgium who required 8 9 10 189 these interventions in order to treat or prevent major obstetric haemorrhage, the circumstances in which the 11 12 190 interventions took place, the management of the obstetric haemorrhage and the outcome and additional 13 14 191 morbidity of these women. 15 For peer review only 16 192 17 18 19 193 METHOD 20 21 194 Belgian Obstetric Surveillance System (B.OSS) . 22 23 24 195 Peripartum hysterectomy and IR in Belgium were registered as part of a national reporting system for the study 25 26 196 of rare obstetric complications. The methodology of the Belgian Obstetric Surveillance System has been 27 28 197 described previously.(16) Briefly, an appointed contact person (OB/GYN or senior-midwife) in each 29 30 198 participating maternity unit was invited by monthly mailing to report a selected number of rare obstetric 31 32 33 199 complications that had occurred in the preceding month or, alternatively, to state that there was ‘nothing to 34 35 200 report’. In the event of such a case being reported, the contact person was asked to complete an extensive 36 37 201 data collection form. Confidentiality was guaranteed for mother, provider and hospital and person-identifiable 38 39 40 202 information was eliminated from data-analysis. In cases of incomplete reporting, the appointed contact person 41 42 203 was encouraged repeatedly by email and phone to provide the missing data, up to six months following the 43 44 204 date of the original report. 45 46 47 205 The participation of 97 percent of Belgian maternity units (n= 110/113) was obtained, covering 98.6 percent of 48 49 206 deliveries (n = 248 681) in the two-year study period. Belgium has a relatively large number of tertiary referral 50 51 207 centres providing neonatal intensive care (n=19) along with a high concentration of small-volume maternity 52 53 54 208 units (more than half of all units) in which there are fewer than 1000 annual births. All tertiary referral centres, 55 56 209 as well as some of the larger non-tertiary maternity services, have an emergency IR service. In answer to an 57 58 210 inquiry of B.OSS contact persons, 40 maternity units (38 %) declared they had an IR service available in the 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 41

1 2 3 211 hospital and a further seven units declared that they had an IR service available but not 24/7 or that they could 4 5 212 call upon vascular surgeons for similar procedures. 6 7 213 Study period . 8 9 10 214 Between January 2012 and December 2013. 11 12 215 Case definition. 13 14 216 B.OSS aimed to collect all cases of peripartum hysterectomy and IR at any gestational age corresponding to the 15 For peer review only 16 217 definition: any woman giving birth to a fetus or infant and undergoing a hysterectomy and/or IR procedure in 17 18 19 218 the same clinical episode. Most peripartum hysterectomies are emergency, life-saving interventions following 20 21 219 vaginal or caesarean delivery, most commonly performed to manage major obstetric haemorrhage but 22 23 220 occasionally because of a severe pelvic infection. Hysterectomy in the first or second trimester, termed 24 25 26 221 abortion hysterectomy, is occasionally required to prevent or manage major obstetric haemorrhage resulting 27 28 222 from an abnormally invasive placenta or an ectopic pregnancy in the cornua, the cervix or a caesarean scar. A 29 30 223 peripartum hysterectomy during caesarean delivery, termed ‘caesarean hysterectomy’, can be performed 31 32 33 224 electively following antenatal diagnosis of a severe abnormally invasive placenta (AIP) or a gynaecological 34 35 225 cancer. 36 37 226 Failure of IR was defined as the need for urgent peripartum hysterectomy to control the haemorrhage. 38 39 227 Registered variables . 40 41 42 228 Data collection forms sought information on maternal characteristics, medical, surgical and obstetric history, 43 44 229 details of the index pregnancy, details of the delivery, the circumstances of the adverse event, its management 45 46 230 and the outcome for mother and neonate. 47 48 49 231 The cause of the obstetric haemorrhage was deduced from the clinical signs reported by the clinician and 50 51 232 classified according to the 4T’s mnemonic: Tonus, Tissue, Trauma, and Thrombin. These labels were taken to 52 53 233 stand for the following characteristics. Tonus: uterine atony. Tissue: abnormally invasive placenta (AIP), 54 55 56 234 placenta praevia, combination of AIP and placenta praevia, placental remnants or retained placenta without 57 58 235 AIP. Trauma: uterine rupture, genital tract lacerations, bleeding caused by unintended injuries during 59 60 236 caesarean delivery or curettage (D&C). (Uncomplicated caesarean deliveries were not included in this For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 41 BMJ Open

1 2 3 237 category.) Thrombin: abnormal coagulation before onset of bleeding. (DIC secondary to major haemorrhage 4 5 238 was not included in this category.) Cases where the cause of bleeding remained unclear or unknown were 6 7 239 classified as ‘Other’. 8 9 10 240 Registered variables of the background population. 11 12 241 The Belgian perinatal registries - Studiecentrum voor Perinatale Epidemiologie (SPE) in Flanders,(17, 18) 13 14 242 Centre d’Epidémiologie Périnatale (CEpiP) in Brussels (19) and Wallonia (20) - record data on a mandatory 15 For peer review only 16 243 basis, covering nearly 100 percent of births in Belgium. A selected set of perinatal data are recorded by the 17 18 19 244 obstetrician, midwife and neonatologist immediately after birth. Births are registered as such in cases of a live 20 21 245 birth or in cases of a stillbirth at a birth weight of 500 grams or more and/or after 22 completed weeks of 22 23 246 gestation, although in Flanders (SPE) live births or stillbirths with a birth weight below 500 grams, irrespective 24 25 26 247 of gestational age, are not registered. 27 28 248 Statistical Analysis. 29 30 249 The aim was to record all cases of peripartum hysterectomy and arterial embolisation during a 2-year study 31 32 33 250 period, meaning that no fixed sample size was set at start of the study. The prevalence of the obstetric events 34 35 251 targeted was estimated using as denominator the total number of deliveries in Belgium in 2012 and 2013, 36 37 252 corrected for the three hospitals that did not participate in B.OSS (n= 248,681 women). Comparison of the 38 39 253 distributions of socio-demographic and obstetric characteristics of reported cases relative to the background 40 41 42 254 population as obtained from the Belgian perinatal registry (2012-2013) was pursued through use of odds 43 44 255 ratios. Odds ratios and accompanying 95 percent confidence intervals (95% CI) were calculated in univariate 45 46 256 analysis, using weighted cases where appropriate. Unadjusted and adjusted odds ratios were calculated in 47 48 49 257 order to compare patient and organisational characteristics between interventions (hysterectomy versus IR) in 50 51 258 univariate analysis and subsequently through the use of unconditional binary logistic regression models. 52 53 259 Comparison between successful and failed IR cases was made only in crude analyses, as numbers of 54 55 56 260 observations were too small to support logistic regression models. Non-parametric tests were used to compare 57 58 261 distributions of outcomes between groups, specifically χ2 for categorical variables, and Kruskal-Wallis and 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 41

1 2 3 262 Mann-Whitney U tests for continuous variables. P-values of <0.05 were considered statistically significant. Data 4 5 263 were analysed using the statistical software package IBM SPSS statistics version 22.0.(21) 6 7 264 The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting 8 9 10 265 observational studies were followed. 11 12 266 13 14 15 For peer review only 16 267 RESULTS 17 18 268 Prevalence 19 20 21 269 Data collection forms of 166 confirmed cases of peripartum hysterectomy and/or IR were retrieved and 22 23 270 included in the analysis. Details of reported cases and completeness of data collection forms are presented in 24 25 271 figure 1. 26 27 28 272 Overall, 84 women (50.6%) underwent a peripartum hysterectomy, corresponding to a prevalence of 3.3 per 29 30 273 10 000 deliveries (95% CI 2.7-4.1). Of these, 17 women (10.2%) underwent a hysterectomy following a previous 31 32 274 IR procedure. Eighty-two women (49.4%) were managed with IR only, corresponding to a prevalence of 3.3 per 33 34 275 10 000 deliveries (95% CI 2.6-4.0). Three women (1.8%) needed IR because of persistent bleeding despite 35 36 37 276 hysterectomy. 38 39 277 In 160 women the intervention occurred within 24 hours of delivery. In eight women it occurred later (from 24 40 41 278 hours to 6 weeks postpartum). The discrepancy between the resultant sum of 168 and the 166 figure is above 42 43 44 279 is because two of the data collection forms were not returned, one of these being the only case in this dataset 45 46 280 of a hysterectomy performed because of severe peritonitis. All other women underwent hysterectomy and/or 47 48 281 IR to treat or prevent major obstetric haemorrhage. The prevalence in Belgium of major obstetric haemorrhage 49 50 51 282 requiring peripartum hysterectomy and/or IR is estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries. 52 53 283 54 55 284 Patient Characteristics 56 57 58 285 We compared the women included in this registry with all the women who gave birth in Belgium during the 59 60 286 study period. Patient characteristics of the background population were derived from the Belgian perinatal For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 41 BMJ Open

1 2 3 287 registries. In univariate analysis the women who underwent peripartum hysterectomy and/or IR for major 4 5 288 obstetric haemorrhage were significantly more likely compared to the background population to be older (≥ 35 6 7 289 years), to be multiparous (parity ≥ 3), to have given birth at a preterm gestational age, to have had a twin 8 9 10 290 pregnancy, to have had a caesarean delivery in a previous pregnancy, to have had a caesarean delivery in the 11 12 291 index pregnancy and to have conceived with artificial reproductive technology (ART) (Table 1). Ten of these 166 13 14 292 women were known to have a uterus myomatosus (6.0%) and twelve experienced haemorrhage in a previous 15 For peer review only 16 293 pregnancy (7.2%). 17 18 19 294 20 21 Table 1. Patient characteristics of women with (imminent) obstetric haemorrhage who underwent peripartum 22 23 hysterectomy and/or IR compared to all women who gave birth in Belgium during the study period. 24 Risk factor Cases of H and/or IR Background population # Unadjusted OR % 25 (n=166) (n=252 272) for H and/or IR 26 27 n, % n, % (95% CI) 28 Maternal age ≥ 35 years 72 (43.4) 43 256 (17.1) 3.7 (2.7-5.0) § 29 30 BMI at booking ≥ 30 kg/m 2 12 (7.2) £ 29 453 (11.6) 0.7 (0.3-1.2) 31 Singletons ≥22 weeks Singletons ≥22 weeks 32 Gestational age (weeks) 33 (n=144) (n = 241 136) 34 - Extreme preterm (<28) - 8 (5.4) - 1 221 (0.5) 35 - Very preterm (28-32) - 6 (4.0) - 1 474 (0.6) 7.4 (5.3-10.5) § 36 - Late preterm (32-37) - 37 (24.8) - 13 798 (5.7) 37 - Term (>37) - 93 (62.4) - 224 643 (93.1) Reference 38 § 39 Parity ≥ 3 30 (18.0) 18 855 (7.4) 2.7 (1.8-4.1)

40 § 41 Multiplets (twin) 13 (7.8) 4690 (1.8) 4.4 (2.5-7.9)

42 & £ § 43 Previous CD 61 (36.7) 27 007 (10.7) 4.8 (3.5-6.6)

44 § 45 TOLAC ( ≥22 weeks) 14 / 57 (24.5) 12 754 / 27 007 (47.2) 0.3 (0.1-0.66) 46 47 Delivery Mode ( ≥22 weeks) $ 48 - CD £ § 91 / 160 (56.8) 54 369 (21.5) 4.8 (3.5-6.5) 49 compared to VD 50 - CD before onset of labour 51 66 / 91 (72.5) 28 586 / 54369 (52.5) 2.3 (1.5-3.7) compared to CD after onset of labour 52 - Assisted VD 53 10 / 69 (14.4) 23 100 /198 444 (11.6) 1.3 (0.6-2.5) 54 compared to spontaneous VD 55 ART (IVF/ICSI) 19 (11.4) £ 9233 (3.7) 3.4 (2.1-5.4) § 56 57 Birthweight ≥ 4000 g 11 (6.6) 19967 (7.8) 0.8 (0.4-1.5) 58 59 CD: Caesarean delivery, VD: Vaginal delivery, H: Hysterectomy, IR: Interventional Radiology , TOLAC: Trial o f Labour After Caesarean Delivery, ART: Artificial Reproductive Technologies, IVF: In Vitro Fertilisation, ICSI: Intracytoplasmic Sperm Injection 60 # Data from the background population were retrieved from SPE reports (Flanders) and CEpiP reports (Brussels and Wallonia) from 2012-2013. % unadjusted Odds ratiosFor (with peer 95% review Confidence only Inter val)- http://bmjopen.bmj.com/site/about/guidelines.xhtml for peripartum hysterectomy and/or IR calculated for the different risk factors: the odds BMJ Open Page 12 of 41

1 2 (=probability of peripartum hysterectomy and/or IR occurring d ivided by the probability of H/IR not occurring ) in the women having the risk factor, 3 divided by the odds in the women not having the risk factor. 4 £ Missing data in cases of hysterectomy and/or IR : BMI (n=27), previous CD (n=1), mode of delivery (n=1), ART (n=2), haemorrhage in previous 5 pregnancy (n=5), uterus myomatosus (n=5) 6 & Women with previous Caesarean Delivery: 9 with Placenta Praevia, 9 with AIP, 17 with Placenta Praevia and AIP 7 $Women with Caesarean Delivery as delivery mode: 17 with Placenta Praevia, 8 with AIP, 23 with Placenta Praevia and AIP 8 Reason for Caesarean Delivery: Bleeding (Placenta Praevia, AIP, abruptio placentae): n= 15; Placenta Praevia / AIP not bleeding: n= 29; Maternal shock / 9 cardiac arrest: n=1; Other: n = 46 10 § p < 0.05 11 295 12 13 14 296 Circumstances of the obstetric haemorrhage: 15 For peer review only 16 297 IR service availability, antenatal and postnatal transfers, urgent and planned interventions. 17 18 19 298 The circumstances in which the interventions occurred in these 166 women are shown in Figure 2. Three 20 21 299 women having been transferred antenatally to another hospital with emergency IR service availability, a total 22 23 300 of 107 women gave birth in a maternity unit with this availability, while 20 gave birth in a unit where IR service 24 25 301 is available but not 24/7 or where obstetricians can call upon vascular surgeons for IR procedures and 39 in a 26 27 28 302 unit with no IR service. To refer to these different circumstances below, we use the terms 24/7-IR unit, not- 29 30 303 24/7-IR unit and non-IR unit respectively. 31 32 304 Seventeen women were transferred postnatally to a 24/7-IR unit and underwent an urgent IR procedure, 33 34 35 305 including one woman for persistent bleeding despite hysterectomy and one woman for persistent bleeding 36 37 306 despite planned intra-arterial balloon occlusion and arterial embolisation by the vascular surgeons in the 38 39 307 referring hospital. Three of the postnatally transferred women consecutively underwent an urgent 40 41 42 308 hysterectomy for persistent bleeding despite arterial embolisation. 43 44 309 Of the three patients who were transferred antenatally, one underwent an urgent embolisation upon arrival 45 46 310 because of a bleeding placenta praevia, one was first planned for caesarean hysterectomy but then underwent 47 48 311 this procedure urgently because of a bleeding AIP before the fixed date, and one underwent a planned 49 50 51 312 caesarean delivery combined with a planned IR procedure because of AIP. Seven more women were 52 53 313 transferred postnatally to the intensive care unit of a tertiary referral centre. 54 55 314 Overall, intervention for imminent obstetric haemorrhage was planned in 15 women. The indication was 56 57 58 315 abnormally invasive placenta in 13 women and placenta praevia in two women. 59 60 316 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 41 BMJ Open

1 2 3 317 Reported cause of the obstetric haemorrhage 4 5 318 Eighteen women (11%) presented with antepartum haemorrhage due to abnormal placentation (placenta 6 7 319 praevia, AIP or both) or abruptio placentae (n=2). Four women (2%) started bleeding during termination of 8 9 10 320 pregnancy (<24 weeks) due to a cervical pregnancy (n=1) or an AIP (n=3). A hundred-and-thirty-eight women 11 12 321 (83%) presented with early postpartum haemorrhage (within 24 hours) and six (4%) with late postpartum 13 14 322 haemorrhage (from 24 hours to 6 weeks). 15 For peer review only 16 17 323 The reported cause of the obstetric haemorrhage was uterine atony (n=91, 54.8%) or abnormally invasive 18 19 324 placenta (AIP) (n=38, 22.8%) in the majority of women. In 54 women (32.5%) more than one cause was 20 21 325 reported, with atony and abnormal placentation (AIP and/or placenta praevia) in 20 women (12%) as the most 22 23 24 326 frequent combination. The cause of the (imminent) obstetric haemorrhage classified according to the 4T’s 25 26 327 mnemonic is shown in Figure 3. 27 28 328 29 30 31 329 Management of the obstetric haemorrhage 32 33 330 Hysterectomy was performed in all the cases of haemorrhage reported to be associated with uterine rupture 34 35 331 (n=12) and in 73.7 percent of the cases associated with abnormally invasive placenta (n=28/38). IR was able to 36 37 38 332 stop the bleeding and avert the need for peripartum hysterectomy in the majority of cases of haemorrhage 39 40 333 reported to be associated with genital tract lacerations (n=10/12, 83.3%), placental remnants or retention 41 42 334 (n=17/26, 65.4%) and uterine atony (n=28/43, 65.1%). 43 44 45 335 Other measures that were attempted to manage the obstetric haemorrhage before or while proceeding to the 46 47 336 hysterectomy or IR are listed in Table 2. The first-line measures were the administration of uterotonic agents 48 49 337 and bimanual uterine compression, second-line measures being intra-uterine (balloon) tamponade, 50 51 338 haemostatic brace suturing (B-lynch or other) and ligation of the uterine or internal iliac arteries, besides blood 52 53 54 339 transfusion and the administration of fresh frozen plasma (FFP), platelets or clotting factors. An intra-uterine 55 56 340 balloon was inserted in 35 women (21%), in 7 cases before transfer to another hospital with availability of an IR 57 58 341 service. Brace suturing and arterial ligation was attempted in 6 and 4 women respectively. In 19 women 59 60 342 (11.4%) no other measures were undertaken besides hysterectomy (n=13), IR (n=2) or hysterectomy following For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 41

1 2 3 343 IR (n=4). The reported cause of haemorrhage in these women was abnormally invasive placenta (n=12), 4 5 344 placenta praevia (n=2), uterine rupture (n=2), retained placenta (n=2) and in one case a late postpartum 6 7 345 haemorrhage. The intervention was planned in five of these women. Eight of the 19 had minor haemorrhage 8 9 10 346 and did not require transfusion, while three women undergoing urgent peripartum hysterectomy needed 11 12 347 massive transfusion (defined as receiving eight or more units of red blood cells). 13 14 348 15 For peer review only 16 Table 2. First and second line measures in women undergoing hysterectomy (H) and /or IR (IR) according to the 17 18 circumstances of the obstetric haemorrhage. 19 Total Planned H Urgent H and/or IR 20 and/or IR n=151 21 IR service IR service No IR service Postnatal 22 24/7 not 24/7 Transfer 23 24 n=166 n=15 n=98 n=12 n=25 n=16* 25 n (%) n (%) n (%) n (%) n (%) n (%) 26 Cause of obstetric haemorrhage 27 28 Abnormal placentation 56 (34) 15 (100) 29 (30) 5 (42) 5 (20) 2 (13) 29 Atony only 43 (26) - 29 (30) 5 (42) 7 (28) 2 (13) 30 31 Uterine rupture 12 (7) - 6 (6) 1 (8) 5 (20) - 32 Miscellaneous £ 55 (33) - 34 (34) 1 (8) 8 (32) 12 (75) 33 First and second line measure 34 35 Uterotonics 134 (81) 10 (67) 81 (83) 10 (83) 19 (76) 15 (94) 36 37 Oxytocin 123 (74) 10 (67) 73 (75) 9 (75) 18 (72) 13 (81) 38 PG 109 (66) 5 (33) 64 (65) 17 (68) 8 (67) 15 (94) 39 Misoprostol 61 (37) 1 (7) 34 (35) 6 (50) 10 (40) 10 (63) 40 Carboprost 63 (39) 5 (33) 42 (43) 3 (25) 9 (25) 9 (56) 41 # 42 Oversewing 30 (18) 1 (7) 19 (20) 1 (8) 7 (28) 2 (13) 43 (Bimanual) compression 64 (39) 3 (20) 36 (37) 5 (42) 9 (36) 11 (69) 44 Balloon tamponade 35 (21) 2 (13) 21 (22) 2 (17) 3 (12) 7 (44) 45 B-lynch or ligation 10 (6) - 5 (5) 1 (8) 4 (16) - 46 Transfusion 139 (84) 7 (47) 85 (87) 12 (100) 21 (84) 15 (88) 47 $ 48 Massive transfusion 51 (31) 2 (13) 24 (24) 7 (58) 10 (40) 8 (50) 49 FFP 108 (65) 4 (27) 64 (65) 11 (92) 18 (72) 11 (69) 50 Platelets 57 (34) 1 (7) 29 (30) 9 (75) 10 (40) 8 (50) 51 Clotting factors % 23 (14) - 8 (8) 5 (42) 7 (28) 3 (19) 52 53 * On e woman was transferred postnatally for persistent bleeding due to AIP despite planned IR in the referring hospital; she was included in the ‘planned’ 54 group. 55 £ Miscellaneous group: Placental remnants or retention n=25, Surgical trauma n=16, Genital tract laceration n=11, Intra-abdominal arterial bleeding n=4, Coagulation disorders n= 3, Unknown n= 6. The total number exceeds n = 55 due to overlap in between cases. 56 PG: Prostaglandin. This can be any or a combination of the following: misoprostol, sulprostone, dinoproston, carboprost. 57 # Oversewing of the placental implantation site, bleeding points, tears or ruptures. 58 $ Massive transfusion: transfusion of 8 or more units of red blood cells 59 % Clotting factors: Fibrinogen, recombinant Factor VIIa, PPSB (Prothrombin-Proconvertin-Stuart Factor-Antihemophilic Factor B) or a combination 60 349 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 41 BMJ Open

1 2 3 350 Management of abnormally invasive placenta (AIP) 4 5 351 The management of the (imminent) obstetric haemorrhage in the women with abnormally invasive placenta 6 7 352 (AIP) (n=38) is illustrated in Figure 4 (for details see supplementary Figure 1). Seventeen women (45%) were 8 9 10 353 susceptible to AIP due to previous caesarean delivery (CD) and placenta praevia in the index pregnancy, 11 12 354 another 7 women (18%) had placenta praevia in the index pregnancy and 9 women (24%) had previous 13 14 355 caesarean delivery. In 25 women (65.7%) there was antenatal suspicion of the AIP, of whom 16 had placenta 15 For peer review only 16 17 356 praevia and previous caesarean delivery, 5 had placenta praevia and 4 had previous caesarean delivery only. All 18 19 357 women with antenatal suspicion of AIP were planned to undergo an elective caesarean delivery in a maternity 20 21 358 unit with IR service availability (24/7 n=20, not 24/7 n=5). This occurred in 21 women, with 3 women having an 22 23 24 359 emergency caesarean delivery because of antepartum haemorrhage and 1 woman delivering vaginally 25 26 360 following acute onset of labour. Further preventive measures were undertaken in 13 women (31.5%): 27 28 361 antenatal transfer to a centre with IR service availability (n=2), placement of intra-arterial catheters pre- 29 30 362 caesarean delivery (n=12), planned caesarean hysterectomy (n=3) or planned hysterectomy at a later stage 31 32 33 363 (n=2). 34 35 364 In 13 women (34%) there was no antenatal suspicion of the AIP. In retrospect, though, all of them had risk 36 37 365 factors: 1 woman had placenta praevia and previous caesarean delivery, 4 had placenta praevia, 6 had previous 38 39 40 366 caesarean delivery, 1 had previous myomectomy and adhesiolysis for Asherman syndrome and another had a 41 42 367 previous manual removal of the placenta. Nine women delivered in a maternity unit with IR service availability 43 44 368 (24/7 n=7, not-24/7 n=2) and 4 delivered in a non-IR unit. Mode of delivery in those 13 women who had no 45 46 47 369 antenatal suspicion of AIP was elective caesarean delivery in 2 women, emergency caesarean delivery in 5 48 49 370 women and vaginal delivery in 6 women. 50 51 371 The placenta was left in situ in 11 women with AIP: 4 in an emergency setting where 3 women underwent 52 53 54 372 immediate hysterectomy and 1 woman underwent urgent arterial embolisation followed by urgent 55 56 373 hysterectomy because of bleeding after 3 days. Another 7 women had their placenta left in situ electively 57 58 374 combined with planned IR: 2 women consecutively underwent planned hysterectomy, 2 an urgent 59 60 375 hysterectomy within 24 hours, 2 a planned hysterectomy after one month and 1 elective evacuation of For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 41

1 2 3 376 retained products of conception after one month. In conclusion, conservative management, defined as aiming 4 5 377 to preserve the placenta and the uterus in a multidisciplinary approach, was attempted in 6 women and was 6 7 378 successful in 3 women. 8 9 10 379 As anticipated, the women for whom measures were planned antenatally had significantly better outcomes 11 12 380 than those where AIP was diagnosed peripartum: less estimated blood loss (median 625 mL (range 300-1250) 13 14 381 versus median 4500 mL (range 500-6000), p=0.08), less need for transfusion (46.2% versus 92.3%, p=0.056), 15 For peer review only 16 382 less need for massive transfusion (≥8 units) (8.3% versus 53.8%, p=0.045) and fewer hysterectomies (61.5% 17 18 19 383 versus 76.9%, p=0.6). The maternal outcome was not significantly different between women with AIP 20 21 384 undergoing IR as first intervention and those with hysterectomy as first intervention: estimated blood loss 22 23 385 (median 4000 mL (range 500-15000) versus median 1000 mL (range 300-8500), p=0.3) need for transfusion 24 25 26 386 (76.5% versus 70%, p=0.9), need for massive transfusion (≥8 units) (29.4% versus 30%, p=1.0). All women with 27 28 387 planned interventions were managed in tertiary referral centres. 29 30 388 31 32 33 389 Risk factors possibly contributing to the choice of intervention 34 35 390 The decision of an obstetrician dealing with a major obstetric haemorrhage to attempt IR rather than turning 36 37 391 to a hysterectomy immediately will be guided by many clinical and organisational factors. Some of the risk 38 39 40 392 factors that may have contributed to this decision are presented in Table 3. The women with double 41 42 393 interventions (n=20) were excluded from this analysis to facilitate interpretation. 43 44 394 High parity (≥ 3), previous caesarean delivery, caesarean delivery in the index pregnancy, abnormally invasive 45 46 47 395 placenta and lack of emergency IR service availability are significantly positively associated with peripartum 48 49 396 hysterectomy in univariate analysis. In unconditional binary logistic regression, the odds of undergoing a 50 51 397 hysterectomy for obstetric haemorrhage are 7.2 times higher in women with high parity and 8.2 times higher 52 53 54 398 in women delivering in a non-IR unit, independent of the other risk factors in the model (Hosmer and 55 56 399 Lemeshow Test p =0.67). 57 58 400 We repeated the analysis in a subgroup, namely the women who delivered in a 24/7-IR unit, with exclusion of 59 60 401 the women with planned interventions, consisting of 31 women who underwent hysterectomy only and 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 41 BMJ Open

1 2 3 402 women who underwent embolisation only. In unconditional binary logistic regression, the odds of undergoing 4 5 403 a peripartum hysterectomy for obstetric haemorrhage are significantly higher in women with high parity (odds 6 7 404 4.5 (95% CI 1.2-16.5)) and nearly significantly higher in women with previous caesarean delivery (odds 3.7 (95% 8 9 10 405 CI 0.98-14.6)), independent of the other risk factors in the model (Hosmer and Lemeshow Test p = 0.69). 11 12 406 13 14 Table 3. Risk factors possibly contributing to the choice of intervention in women undergoing hysterectomy or 15 interventional radiologyFor for obstetric peer haemorrhage. review Unadjusted and adjusted only OR (95% CI) for hysterectomy (only) 16 per risk factor . 17 18 Risk factors Number of women Unadjusted OR Adjusted OR 19 Hysterectomy Interventional (95%CI) (95%CI) 20 only radiology only for for 21 n=64 n=82 Hysterectomy Hysterectomy 22 n, % n, % only only 23 24 Demographic factors 25 Maternal age ≥ 35 years 32 (50.0) 30 (36.5) 1.7 (0.8-3.3) 26 BMI at booking ≥ 30 kg/m2 6 (9.3) 3 (3.6) 2.7 (0.6-11.3) 27 Obstetric factors 28 Preterm (<37w) 28 (43.8) 24 (29.3) 1.8 (0.9-3.7) 29 § § 30 Parity ≥ 3 18 (28.1) 7 (8.5) 2.7 (1.4-9.4) 7.2 (2.1-24.6) § 31 Previous caesarean delivery 31 (48.4) 18 (21.9) 3.3 (1.6-6.8) 2.3 (0.9-6.3) § 32 Caesarean delivery ( ≥22w) * 38 / 59 (64.4) 37 / 82 (45.1) 2.2 (1.1-4.3) 1.6 (0.6-4.0) 33 Cause of bleeding 34 § AIP and AIP+ praevia 16 (25.0) 9 (10.9) 2.7 (1.1-6.6) 3.0 (0.9-10.0) 35 § 36 Atony 29 (45.3) 52 (63.4) 0.4 (0.2-0.9) 0.9 (0.3-2.3) 37 Organisational factors 38 Weekend and/or non-office hours # 24/60 (40.0) 34/80 (42.5) 0.9 (0.4-1.7) 39 IR NOT available $ 30 (46.8) 15 (18.2) 3.9 (1.8-8.2) § 8.2 (3.1-21.6) § 40 * Exclusion of 5 cases < 22 weeks 41 £ Missing data: BMI (n=27), previous caesarean delivery (n=1), Delivery Mode (n=1), placenta location (n=13), Date of delivery (n=2), Time of 42 delivery (n=5), week/weekend (n=3) 43 # based on date and time of delivery 44 IR : interventional radiology 45 $ Emergency IR service availability in the maternity unit where the woman gave birth 46 § p<0.05 47 407 48 49 50 408 51 52 409 Peripartum hysterectomy despite interventional radiology 53 54 55 410 Seventeen women in this dataset had a peripartum hysterectomy performed following a previous IR 56 57 411 procedure. In five of these women the hysterectomy was performed in a programmed setting following 58 59 412 planned IR because of abnormally invasive placenta, two of them with a delay of one month. These five IR 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 41

1 2 3 413 procedures were considered successful in preventing or stopping the bleeding. In the other twelve women an 4 5 414 urgent hysterectomy was performed because of persistent bleeding despite the IR procedure. We compared 6 7 415 the characteristics of these 12 women for whom IR failed with 87 women (including the five who underwent an 8 9 10 416 elective hysterectomy) for whom IR was successful. In univariate analysis, an increased risk of urgent 11 12 417 hysterectomy despite IR was observed for women with caesarean delivery in a previous pregnancy, for women 13 14 418 with abnormally invasive placenta and for women who had no uterotonic agents administered. These findings 15 For peer review only 16 419 are presented in Table 4. 17 18 19 Table 4. Risk factors possibly contributing to failure of IR (IR). Unadjusted odds ratio’s (95% CI) 20 for failed IR per risk factor. 21 22 Riskfactors Number of women Unadjusted OR 23 Failed IR Successful IR (95% CI) 24 n=12 n=87 for failed IR 25 n, % n, % 26 27 Demographic factors 28 Maternal age ≥ 35 years 5 (41.7) 33 (37.9) 1.1 (0.3-3.9) 29 BMI at booking ≥ 30 kg/m 2 2 (18.2) 3 (3.9) 5.6 (0.8-37.6) 30 Obstetric factors 31 32 Twins 1 (8.3) 6 ( 6.9) 1.2 (0.1-11.1) 33 Preterm gestational age < 37w 7 (58.3) 28 (32.2) 2.9 (0.8-10.1) 34 Preterm gestational age < 34w 3 (25.0) 6 (6.8) 4.5 (0.9-21.1) 35 36 Parity ≥3 3 (25.0) 8 (9.2) 3.2 (0.7-14.6) 37 Previous caesarean delivery 7 (58.3) 21 (24.1) 4.4 (1.2-15.3) § 38 39 Caesarean delivery (≥22 weeks) 8 (72.7) 42 (48.3) 2.1 (0.6-7.6) 40 Induction of labour 3 (25.0) 25 (28.7) 0.8 (0.2-3.3) 41 Augmentation of labour 2 (16.7) 26 (29.9) 0. (0.09-2.2) 42 Cause of haemorrhage 43 44 Atony 7 (63.6) 52 (62.7) 0.9 (0.2-3.2) 45 AIP and AIP+ praevia 7 (58.3) 14 (16.1) 7.3 (2.0-26.3) § 46 Surgical trauma 2 (16.7) 6 (6.9) 2.7 (0.4-15.2) 47 Genital tract laceration 1 (8.3) 10 (11.5) 0.7 (0.08-6.0) 48 st nd 49 1 and 2 line treatment 50 Uterotonic agents 8 (66.7) 82 (94.3) 0.1 (0.02-0.5) § 51 52 No use of uterotonic agents 4 (33.3) 5 (5.7) 8.2 (1.8-36.8) § 53 Intra-uterine balloon 3 (25.0) 26 (29.9) 0.7 (0.1-3.1) 54 55 Clotting factors IV 2 (16.7) 8 (9.2) 1.9 (0.3-10.6) 56 FFP 7 (58.3) 54 (62.1) 0.8 (0.2-2.9) 57 58 Organisational factors 59 Planned intervention 2 (14) 10 (12) 1.5 (0.3-8.0) 60 Weekend and/or non-office hours # 5 / 10 (50.0) 51 / 85 (60.0) 0.6 (0.1-2.4) For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 41 BMJ Open

1

2 $ 3 IR NOT available 3 (25.0) 10 (11.4) 2.5 (0.6-11.1) 4 Emergency IR NOT available $ 6 (50.0) 20 (22.9) 3.3 (0.9-11.5) 5 Postnatal transfer 3 (25.0) 13 (14.9) 1.9 (0.4-7.9) 6 7 IR: interventional radiology; AIP: abnormally invasive placenta; FFP: Fresh Frozen Plasma ; CD: Caesarean delivery # based on date and time of delivery 8 $ 9 IR service availability (24/7 and not 24/7) in the maternity unit where the woman gave birth: $ 10 Emergency IR service availability (24/7) in the maternity unit where the woman gave birth 11 § p < 0.05 12 13 420 14 15 421 Maternal outcomeFor and additional peer morbidity review only 16 17 18 422 Blood transfusion was given to 139 women (83.7%, missing data for 2 women), with a median of 6.0 units of 19 20 423 red blood cells (RBC) (range 0-31), 4.0 units of fresh frozen plasma (FFP) (range 0-31) and 0.5 units of platelets 21 22 424 (range 0-29). Fifty-one women (30.7%) were transfused ≥ 8 units of RBC and 23 women were administered 23 24 25 425 clotting factors (fibrinogen, recombinant factor VIIa, PPSB (Prothrombin-Proconvertin-Stuart Factor- 26 27 426 Antihemophilic Factor B), or a combination) . A hundred-and-fourteen women (68.6%, missing data in 4 28 29 427 women) were admitted to an intensive care unit (ICU) for a median of 2.0 days (range 0-18 days); the median 30 31 32 428 hospital stay was 8.0 days (range 2-33 days). Five women needed relaparotomy and 3 women needed arterial 33 34 429 embolisation following hysterectomy because of persistent bleeding. The cause of haemorrhage in these 35 36 430 women was AIP (n=1), AIP and atony (n=2), atony (n=3) and in 1 case unclear (n=1). 37 38 431 Disseminated intravascular coagulation (DIC) secondary to major haemorrhage was reported in 32 women 39 40 41 432 (19%). Other reported maternal complications associated with major haemorrhage were: renal failure (n=2), 42 43 433 pulmonary oedema (n=4), acute respiratory distress syndrome (n=2), myocardial ischemia (n=1) and pleural 44 45 434 and pericardial effusion (n=1). Complications associated with peripartum hysterectomy were ureteral injuries 46 47 48 435 (n=2) and ileus (n=4). One woman suffered from pain in loins and legs, which was the only reported 49 50 436 complication attributed to the IR procedure. 51 52 437 One mother died (case fatality rate 0.6% (95% CI 0.1-3.3)). She underwent an emergency caesarean delivery 53 54 55 438 because of cardiac arrest caused by massive pulmonary embolism. She survived the hysterectomy performed 56 57 439 because of major bleeding due to uterine atony and was transferred to the intensive care unit of a tertiary 58 59 440 referral centre using Extracorporeal Life Support (ECLS). She underwent a re-laparotomy because of major 60 441 hepatic haemorrhage,For peer but review died the only subsequent - http://bmjopen.bmj.com/site/about/guidelines.xhtml day during pulmonary thrombectomy. BMJ Open Page 20 of 41

1 2 3 442 Considering the whole group, the outcome in terms of transfusion rate and ICU admission in women who 4 5 443 underwent hysterectomy did not significantly differ from those who underwent IR, except for a significant 6 7 444 difference in total duration of hospitalisation (median 9 days, range (2-30) versus median 7 days (range 2-33)). 8 9 10 445 When comparing the women according to the IR service availability in the maternity unit where they gave 11 12 446 birth, the rate of peripartum hysterectomy was significantly higher in the group of women who delivered in a 13 14 447 non-IR unit (29 women (74%)) than in the group who delivered in a 24/7-IR unit (44 women (41%)). And so was 15 For peer review only 16 448 the rate of massive transfusion (defined as ≥ 8 units of RBC): 17 women (45%) who delivered in a non-IR unit 17 18 19 449 compared to 25 women (24%) who delivered in a 24/7-IR unit. 20 21 450 When comparing the women who gave birth in a non-IR unit according to whether they were transferred 22 23 451 postpartum, there was no significant difference in estimated blood loss (median 2850 mL (range 500-5000) 24 25 26 452 versus median 3500 mL (range 3000-5000), p=0.4) and need for transfusion (median number of units 6.5 27 28 453 (range 0-19) versus 9.5 (range 0-31), p=0.2), but all women who were not transferred (n=25/25) underwent a 29 30 454 hysterectomy compared to 28 percent (n=4/14) in the transferred group (p=0.0). 31 32 33 455 34 35 456 DISCUSSION 36 37 457 The main findings of this study are: firstly, that the prevalence in Belgium of major obstetric haemorrhage 38 39 40 458 requiring peripartum hysterectomy and/or IR is estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries; 41 42 459 secondly, that AIP was not detected antenatally in 34 percent of the cases in this study, despite the presence of 43 44 460 risk factors; thirdly, that women with major obstetric haemorrhage who gave birth in a non-IR unit were more 45 46 47 461 likely to undergo an urgent peripartum hysterectomy (74%) than those who delivered in a 24/7-IR unit (41%); 48 49 462 fourthly, that the need for urgent peripartum hysterectomy was averted in 72 percent of the women who 50 51 463 delivered in a non-IR unit but transferred postnatally, with no significant difference in estimated blood loss and 52 53 54 464 transfusion rate compared to the women who were not transferred. 55 56 465 The strengths of the study are its methodology of collecting cases prospectively on a monthly basis and its 57 58 466 national design, with excellent participation and response rates (98.9%) of Belgian maternity units. 59 60 467 Underreporting of cases, due to the non-mandatory reporting of severe pregnancy complications, may have For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 41 BMJ Open

1 2 3 468 biased our results to some extent. The small number of cases, inevitable with rare but severe obstetric 4 5 469 complications, warrant caution in making strong inferences. Another limitation is the lack of a perspective on 6 7 470 the overall picture of postpartum haemorrhage in Belgium. We have no accurate data of the total number of 8 9 10 471 women suffering severe postpartum haemorrhage, the near-misses and maternal deaths due to severe 11 12 472 haemorrhage who did not undergo embolisation or hysterectomy. We have no denominator data of the 13 14 473 women who were successfully managed with other second-line measures (intra-uterine (balloon) tamponade, 15 For peer review only 16 474 haemostatic brace suturing and ligation of the uterine arteries) during the same study period in Belgium, so 17 18 19 475 that we cannot draw conclusions on the failure rate of these second-line measures. These are interesting topics 20 21 476 that can be investigated by B.OSS in future studies. 22 23 477 This nationwide population based study describes how women with major obstetric haemorrhage were 24 25 26 478 managed with either hysterectomy or IR according to circumstances. However, this study does not allow firm 27 28 479 conclusions to be drawn regarding the motivation or the appropriateness of this management. One should be 29 30 480 aware that obstetricians are forced to make rapid decisions when dealing with a major obstetric haemorrhage. 31 32 33 481 Belgian obstetricians cannot rely on national guidelines for postpartum haemorrhage or abnormally invasive 34 35 482 placenta. Small volume maternity units even lack local hospital guidelines for these matters. Therefore, the 36 37 483 management of major obstetric haemorrhage will differ greatly between hospitals and individual 38 39 484 gynaecologists. An emergency IR service is available in about 38 percent of Belgian maternity units, 40 41 42 485 complemented with an informal network between hospitals as shown in dataset: twenty women (12%) were 43 44 486 transferred to a centre with IR availability. We cannot rule out the possibility that this study in fact 45 46 487 demonstrates an over-use of IR procedures and under-use of other second-line measures in the management 47 48 49 488 of postpartum haemorrhage in Belgium. This reasonable hypothesis has been previously suggested by Kayem 50 51 489 et al. based on the results of a population-based study of women with postpartum haemorrhage managed with 52 53 490 invasive therapies in France.(22) 54 55 56 491 The use of peripartum hysterectomy was significantly higher in multiparous women (parity ≥ 3) and in women 57 58 492 who delivered in a non-IR unit. This result may not surprise, and performing a hysterectomy sooner rather than 59 60 493 later must be considered good practice,(6) as losing time trying another second-line measure can cause further For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 41

1 2 3 494 blood loss and further morbidity. The decision is more easily made in cases of women who have no desire for 4 5 495 continued fertility. Likewise, the transfer of a hemodynamically unstable patient to an IR service is hardly ever 6 7 496 considered a safe alternative if this IR service is located in a nearby hospital. However, the higher rate of 8 9 10 497 massive blood transfusion in the group of women who delivered in a non-IR unit strengthens the hypothesis 11 12 498 that the rate of hysterectomy is not only associated with the IR service availability, but also with the expert 13 14 499 knowledge and experience at larger units in the management of massive obstetric haemorrhage. A case-by- 15 For peer review only 16 500 case in-depth analysis would be necessary to reveal whether the hysterectomies and arterial embolisations 17 18 19 501 recorded in this study were appropriate or preventable. A structured audit based on medical records of 50 20 21 502 peripartum hysterectomies was recently performed in Denmark. This population based audit revealed that 22 23 503 more than 50% of peripartum hysterectomies were avoidable by simple measures such as the use of 24 25 26 504 intrauterine balloons.(23) 27 28 505 Abnormally invasive placenta (AIP) accounts for 23 percent of the interventions and 86 percent of the planned 29 30 506 interventions in this dataset. Targeted papers report good sensitivity and positive predictive value (PPV) of 31 32 33 507 grey-scale ultrasound as the screening tool for AIP in women with a high index of suspicion, enhanced by the 34 35 508 use of transvaginal ultrasound and colour Doppler.(24, 25) Nonetheless, the diagnosis of AIP was missed 36 37 509 antenatally in 34 percent of the women in this study, although all of them had risk factors that should have 38 39 510 increased awareness. Several population-based studies report rather low antenatal diagnosis rates of AIP: 50 40 41 42 511 percent of AIP cases were missed in a study of the UK Obstetric Surveillance System (UKOSS), even though 30 43 44 512 percent of these missed cases had a previous caesarean delivery and placenta praevia.(26) Similarly, in the 45 46 513 Nordic Obstetric Surveillance Study (NOSS) 70 percent of AIP cases were missed, despite the fact that 33 47 48 49 514 percent had placenta praevia and 39 percent had previous caesarean delivery.(27) In view of these findings, we 50 51 515 believe that antenatal diagnosis of AIP may be improved by raising the awareness of clinicians of the risk 52 53 516 factors for AIP. This involves actively asking for previous uterine surgery at the first antenatal visit, checking the 54 55 56 517 position of the placenta and focussed imaging in the pregnancies at risk.(28) Importantly, there is little 57 58 518 information about the sensitivity or positive predictive value (PPV) of ultrasound or MRI as the screening tool 59 60 519 for AIP in women who have had no previous uterine surgery or in women in whom the placenta is not For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 41 BMJ Open

1 2 3 520 implanted anteriorly in the lower uterine segment,(24) which would be an interesting topic for future research. 4 5 521 Despite increased awareness a number of cases of abnormal placentation will still be missed. For this reason, 6 7 522 all maternity units should ensure a state of constant preparedness to deal with an unanticipated AIP which 8 9 10 523 causes a major haemorrhage at any time. 11 12 524 Although numbers are small, this study aligns with the findings in previous studies that maternal outcome is 13 14 525 better when women with suspicion of AIP give birth in specialised tertiary centres and when preventive 15 For peer review only 16 526 measures are taken.(26, 29, 30) A planned preterm caesarean hysterectomy with the placenta left in situ is the 17 18 19 527 indubitable treatment of choice in women with AIP and is incorporated in recommendations of RCOG,(14) 20 21 528 ACOG, (31) and SMFM.(32) Alternatively, a conservative approach is advocated for women who want to 22 23 529 preserve their fertility and are hemodynamically stable, which means a multidisciplinary approach aiming to 24 25 26 530 leave the placenta in situ and to preserve the uterus.(33, 34) This study demonstrates that conservative 27 28 531 management of AIP is not well incorporated in obstetric care in Belgium. The placenta was left in situ in eleven 29 30 532 women with AIP (29%) in this dataset. A conservative approach was attempted in 6 women and was successful 31 32 33 533 in 3 women. Two of these women underwent a planned hysterectomy delayed until one month after the 34 35 534 caesarean delivery. A small number of papers support this management (leaving the placenta in situ with a 36 37 535 planned delayed hysterectomy) in women who have a placenta percreta with infiltration in the bladder and/or 38 39 536 other surrounding tissues and therefore a high risk of haemorrhage or adjacent tissue injury,(34-36) the 40 41 42 537 rationale being that this delay will allow the decrease of the peri-uterine vascularisation and thus prevent 43 44 538 haemorrhage. Studies with larger numbers are needed to strengthen evidence for the claim that this 45 46 539 methodology is safer than caesarean hysterectomy in women with severe placenta percreta. 47 48 49 540 The role of IR in the management of AIP is still a matter of debate. Most guidelines agree that IR can be 50 51 541 lifesaving in the treatment of major postpartum haemorrhage.(14, 32) They currently do not recommend the 52 53 542 routine use of pre-operative placement of intra-arterial catheters to enable balloon occlusion or arterial 54 55 56 543 embolisation before caesarean hysterectomy or conservative treatment of AIP, based on lack of evidence of 57 58 544 benefit.(14, 31, 32, 37) Meanwhile, the number of papers reporting benefit of prophylactic IR is growing.(38, 59 60 545 39) Our population-based study reports diverse management of women with AIP, inherent to a multi-centre For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 41

1 2 3 546 study. Twelve women with AIP were managed with programmed IR, the need for hysterectomy being averted 4 5 547 in six of them. We cannot draw firm conclusions on the success rate of IR in the elective setting because of 6 7 548 small numbers and lack of controls. 8 9 10 549 This study has demonstrated a high success rate (87.8%) of IR in controlling postpartum haemorrhage, 11 12 550 consistent with the success rates ranging between 85 percent and 95 percent reported by others.(9-12, 40-42) 13 14 551 Univariate analysis (taking into account the small numbers) showed IR was more likely to fail in women with a 15 For peer review only 16 552 caesarean delivery in a previous pregnancy (failure rate 25%), women with AIP (33%) and women who had no 17 18 19 553 uterotonic agents administered (44%). Several large hospital-based retrospective studies have analysed the 20 21 554 predictive factors associated with failure of IR in the management of postpartum haemorrhage.(12, 40-43) 22 23 555 Most papers report a lower success rate of IR in women with AIP,(40, 41) which explains to a great extent the 24 25 26 556 lower success rates in women with previous caesarean delivery and in women with caesarean delivery in the 27 28 557 current pregnancy. A population-based study in the Netherlands (10) reported a 25 percent failure rate for IR 29 30 558 following caesarean delivery in the current pregnancy, compared with just 16 percent in this study and barely 31 32 33 559 10 percent in a large study reported by Lee et al. (n=176).(44) Other researchers have consistently reported a 34 35 560 higher failure rate of IR in women with various clinical determinants related to major haemorrhage: 36 37 561 hemodynamic shock,(43, 45) diffuse intravascular coagulopathy,(12, 45), low fibrinogen and prothrombin time 38 39 562 (41) and the need for massive transfusion.(12, 40, 41) However, these factors could be either cause or 40 41 42 563 consequence of the failure of IR, which was not clearly differentiated in some of the aforementioned studies. 43 44 564 Interestingly, hospital to hospital transfer was not associated with a higher failure rate of IR in this study, 45 46 565 consistent with what has been reported by others.(12, 40, 41) 47 48 49 566 50 51 52 53 567 CONCLUSION 54 55 568 The prevalence in Belgium of major obstetric haemorrhage requiring peripartum hysterectomy and/or IR is 56 57 569 estimated at 6.6 (95% CI 5.7-7.7) per 10 000 deliveries: roughly 1 in 3030 women who give birth in Belgium 58 59 60 570 undergoes a peripartum hysterectomy, while another 1 in 3030 women is successfully managed by IR, thereby For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 41 BMJ Open

1 2 3 571 preserving the uterus. Abnormal placentation and/or uterine atony are reported as the cause of haemorrhage 4 5 572 and reason for intervention in the majority of these women (83.7%). Further improvement in the management 6 7 573 of obstetric haemorrhage could possibly be achieved by increased awareness of risk factors for abnormal 8 9 10 574 placentation on the part of clinicians and by antenatal transfer to tertiary centres in case of suspicion. A case- 11 12 575 by-case in-depth analysis is necessary to reveal whether the hysterectomies and arterial embolisations 13 14 576 performed in this study were appropriate or preventable. 15 For peer review only 16 577 17 18 19 578 ACKNOWLEDGEMENTS 20 21 579 This study would not have been possible without the voluntary participation of the following Belgian maternity 22 23 24 580 units: GZA Sint-Jozef, Mortsel; GZA Sint-Augustinus, Wilrijk; GZA Sint-Vincentius, Antwerpen; ZNA, Sint- 25 26 581 Erasmus, Borgerhout; ZNA, Jan Palfijn , Merksem; ZNA, Middelheim, Antwerpen; ASZ, campus Aalst; ASZ, 27 28 582 campus Geraardsbergen; AZ Lokeren; Imelda Ziekenhuis, Bonheiden; Sint-Jozef ziekenhuis, Bornem; AZ Klina, 29 30 583 Brasschaat; AZ Sint-Jan, Brugge; AZ Sint-Lucas, Assebroek; Centre de Santé des Fagnes; Centre Hospitalier de 31 32 33 584 l’Ardenne; Centre Hospitalier EpiCura, site Hornu; Centre Hospitalier de Mouscron; Centre Hospitalier du Bois 34 35 585 de l’Abbaye et de Hesbay ; Centre Hospitalier Peltzer, La Tourelle ; CHC Saint-Vincent; CHC Saint-Joseph; CHC 36 37 586 Sainte-Elisabeth; CHIREC, Clinique Sainte-Anne, Saint-Remi; CHIREC Braine-l’Alleud-Waterloo; CHR de la 38 39 40 587 Citadelle; CHR de Namur; CHR du Val de Sambre; CHR Haute Senne-Le Tilleriau; CHR Mons Clinique Saint- 41 42 588 Joseph; CHU Ambroise Paré; CHU Brugmann; CHU Charleroi, André Vésale; CHU Notre-Dame des Bruyères; 43 44 589 CHU Saint-Pierre; CHU Tivoli; CHwapi-Site Notre-Dame; Clinique Edith Cavell; Clinique et Maternité Sainte 45 46 47 590 Elisabeth; Clinique Notre dame de Grâce; Clinique Reine Astrid; Clinique Sainte-Piere; Cliniques de l’Europe, 48 49 591 Saint-Michel; Cliniques de l’Europe, Sainte-Elisabeth; Cliniques du Sud Luxembourg; Cliniques Universitaires 50 51 592 Saint-Luc; AZ Sint-Vincentius, Deinze; AZ Sint-Blasius , Dendermonde; AZ Monica, Deurne; AZ Diest; AZ Sint- 52 53 54 593 Maarten, Duffel; AZ Alma, Eeklo; Dimpna Ziekenhuis, Geel; AZ Jan Palfijn, Gent; AZ Maria Middelares, Gent; AZ 55 56 594 Sint-Lucas, Gent; GHDC - Notre Dame, Charleroi; AZ Maria ziekenhuis, Halle; Jessaziekenhuis, Hasselt; AZ Sint 57 58 595 Elisabeth, Herentals; CAZ Midden-Limburg, Heusden-Zolder; Hôpital Civil Marie Curie; Hôpital Erasme, 59 60 596 Bruxelles; Hôpitaux Iris Sud – Site Ixelles ; Jan Yperman ziekenhuis, Ieper; IFAC Hôpital Princesse Paola; INDC For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 41

1 2 3 597 Entité Jolimontoise; Sint-Jozef ziekenhuis, Izegem; Kliniek Sint-Jan; Klinik Saint-Josef; AZ Zeno, campus Knokke- 4 5 598 Heist; AZ Groeninge, Kortrijk; Heilig Hart ziekenhuis, Leuven; UZ Leuven; Heilig Hart ziekenhuis, Lier; AZ Sint- 6 7 599 Jozef, Malle; AZ Sint-Maarten, Mechelen; AZ Delta, campus Menen; Heilig Hart ziekenhuis, Mol; Onze-Lieve- 8 9 10 600 Vrouw ziekenhuis, campus Aalst; Onze-Lieve-Vrouw ziekenhuis, campus Asse; AZ Damiaan, campus Sint-Jozef, 11 12 601 Oostende; AZ Sint-Jan, campus Oostende; AZ Oudenaarde; Mariaziekenhuis Noord-Limburg, Overpelt; AZ 13 14 602 Heilige Familie, Reet AZ Delta, campus Brugsesteenweg, Roeselare; AZ Delta, campus Wilgenstraat, Roeselare; 15 For peer review only 16 603 AZ Glorieux, Ronse; AZ Nikolaas, Sint-Niklaas; Sint-Trudo ziekenhuis, Sint-Truiden; Saint-Nikolaus Hospital; Sint- 17 18 19 604 Andries ziekenhuis, Tielt; Heilig Hartziekenhuis, Tienen; AZ Vesalius, Tongeren; Sint-Rembertziekenhuis, 20 21 605 Torhout; AZ Turnhout; UZ Antwerpen; UZ Brussel; AZ Sint-Augustinus, Veurne; AZ Jan Portaels, Vilvoorde; 22 23 606 Onze-Lieve-Vrouw van Lourdes ziekenhuis, Waregem; Ziekenhuis Oost-Limburg, campus Sint-Jan, Genk; AZ 24 25 26 607 Sint-Elisabeth, Zottegem. 27 28 608 We would like to thank Fatima Bercha, Marlies Deblaere and Ann Langedock for their contribution in collecting 29 30 609 the data. 31 32 33 610 34 35 611 AUTHOR CONTRIBUTIONS 36 37 38 612 Project conception: MH, YE. Study design: GV, VVL, JVK, MH, YE. Data collection: GV, MG, IR, VVL, JVK. Data 39 40 613 analysis: GV, MG, IJ, ER. Data interpretation: GV, MG, IJ, KR, MH, ER, YE, HV. Manuscript first draft: GV. 41 42 614 Contributed to the writing of the manuscript: MG, IJ, KR, MH, YE, HV. Manuscript revision: MG, IJ, VVL, KR, MH, 43 44 45 615 ER, JV, YE, HV. All authors approved the final version, agreed to be accountable for all aspects of the work and 46 47 616 met the ICMJE criteria for authorship. 48 49 617 50 51 52 618 FUNDING 53 54 619 The study was funded by the College for Mother and Newborn, a consultative body of the Belgian Public Health 55 56 620 Service. GV was funded by the Flemish Research Foundation (FWO). 57 58 59 621 60 622 COMPETING INTERESTS For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 41 BMJ Open

1 2 3 623 All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and 4 5 624 hereby declare that they have no financial relationships with any organisations that might have an interest in 6 7 625 the submitted work in the past three years, nor that they have partaken of any other relationships or activities 8 9 10 626 that could appear to have influenced the submitted work. 11 12 627 13 14 628 DATA SHARING 15 For peer review only 16 17 629 No additional data are available. 18 19 630 20 21 631 ETHICS APPROVAL 22 23 24 632 The B.OSS methodology and this study were approved by the Medical Ethics Committee of Ghent University 25 26 633 Hospital (EC UZG 2012/734; B670201215359 and EC UZG 2015/1470; B670201526875) and by the Medical 27 28 634 Ethics Committee of the University Hospital Erasme, Brussels (EC ULB 2012/111; B406201213660). Data were 29 30 31 635 retrieved in retrospect from the case notes by means of data collection forms. No personally identifiable 32 33 636 information was obtained. Participants were informed and enabled to opt-out. 34 35 637 36 37 38 638 REFERENCES 39 40 639 1. Evaluating the quality of care for severe pregnancy complications: the WHO near-miss approach for 41 42 43 640 maternal health. Geneva, Switzerland: WHO press 2011. Retrieved from 44 45 641 http://apps.who.int/iris/bitstream/10665/44692/1/9789241502221_eng.pdf. Accessed 22 Dec. 16. 46 47 642 2. Stones W, Lim W, Al-Azzawi F, Kelly M. An investigation of maternal morbidity with identification of 48 49 50 643 life-threatening 'near miss' episodes. Health trends. 1991;23(1):13-5. 51 52 644 3. Say L, Souza JP, Pattinson RC. Maternal near miss--towards a standard tool for monitoring quality of 53 54 645 maternal health care. Best practice & research Clinical obstetrics & gynaecology. 2009;23(3):287-96. 55 56 646 4. Lewis, G (ed) 2007. The Confidential Enquiry into Maternal and Child Health (CEMACH). Saving 57 58 59 647 Mother's Lives: reviewing maternal deaths to make motherhood safer-2003-2005. The Seventh Report on 60 648 Confidential EnquiriesFor peer into review Maternal only Deaths - http://bmjopen.bmj.com/site/about/guidelines.xhtml in the United Kingdom. London: CEMACH. BMJ Open Page 28 of 41

1 2 3 649 5. Knight M, Tuffnell D, Kenyon S, , Kenyon S, Shakespear J, Gray R, Kurinczuk JJ (Eds.) on behalf of 4 5 650 MBRRACE-UK. Saving Lives, Improving Mother's Care - Surveillance of maternal deaths in the UK 2011-13 and 6 7 651 lessons learned to inform maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths 8 9 10 652 and Morbidity 2009-13. Oxford: National Perinatal Epidemiology Unit, University of Oxford 2015. 11 12 653 6. The Royal College of Obstetricians and Gynaecologists. Postpartum Haemorrhage, Prevention and 13 14 654 Management. Green-top Guideline No. 52. Published: 11/05/2009. 15 For peer review only 16 655 7. Postpartum Hemorrhage. The American College of Obstetricians and Gynaecologistis. ACOG Practice 17 18 19 656 Bulletin. Number 76, October 2006. Reaffirmed 2015. 20 21 657 8. Dahlke JD, Mendez-Figueroa H, Maggio L, Hauspurg AK, Sperling JD, Chauhan SP, et al. Prevention and 22 23 658 management of postpartum hemorrhage: a comparison of 4 national guidelines. American journal of obstetrics 24 25 26 659 and gynecology. 2015;213(1):76.e1-10. 27 28 660 9. Kayem G, Kurinczuk JJ, Alfirevic Z, Spark P, Brocklehurst P, Knight M. Specific second-line therapies for 29 30 661 postpartum haemorrhage: a national cohort study. BJOG : an international journal of obstetrics and 31 32 33 662 gynaecology. 2011;118(7):856-64. 34 35 663 10. Zwart JJ, Dijk PD, van Roosmalen J. Peripartum hysterectomy and arterial embolization for major 36 37 664 obstetric hemorrhage: a 2-year nationwide cohort study in the Netherlands. American journal of obstetrics and 38 39 665 gynecology. 2010;202(2):150.e1-7. 40 41 42 666 11. Inoue S, Masuyama H, Hiramatsu Y. Efficacy of transarterial embolisation in the management of post- 43 44 667 partum haemorrhage and its impact on subsequent pregnancies. The Australian & New Zealand journal of 45 46 668 obstetrics & gynaecology. 2014;54(6):541-5. 47 48 49 669 12. Lee HY, Shin JH, Kim J, Yoon HK, Ko GY, Won HS, et al. Primary postpartum hemorrhage: outcome of 50 51 670 pelvic arterial embolization in 251 patients at a single institution. Radiology. 2012;264(3):903-9. 52 53 671 13. The Royal College of Obstetricians and Gynaecologists. The role of emergency and elective 54 55 56 672 interventional radiology in postpartum haemorrhage. (Good Practice No. 6). June 2007. 57 58 673 14. The Royal College of Obstetricians and Gynaecologists. Placenta Praevia, Placenta Praevia Accreta and 59 60 674 Vasa Praevia : Diagnosis and Management. Green-top Guideline No. 27. Published January 2011. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 41 BMJ Open

1 2 3 675 15. Investigation into 10 Maternal Deaths at, or following Delivery at, Northwick Park Hospital, North West 4 5 676 London Hospitals NHS Trust, between April 2002 and April 2005. Commission for Healthcare Audit and 6 7 677 Inspection. London, August 2006. 8 9 10 678 16. Vandenberghe G, De Blaere M, Van Leeuw V, Roelens K, Englert Y, Hanssens M, et al. Nationwide 11 12 679 population-based cohort study of uterine rupture in Belgium: results from the Belgian Obstetric Surveillance 13 14 680 System. BMJ open. 2016;6(5):e010415. 15 For peer review only 16 681 17. H. Cammu EM, G. Martens, C. Van Mol, Y Jacquemyn. Perinatale activiteiten in Vlaanderen 2013. 2014. 17 18 19 682 Retrieved from: https://www.zorg-en- 20 21 683 gezondheid.be/sites/default/files/atoms/files/SPE_jaarrapport%202013.pdf. Accessed on 22 Dec. 16 22 23 684 18. H. Cammu EM, G. Martens, C. Van Mol, Y Jacquemyn. Perinatale Activiteiten in Vlaanderen 2012. 2013. 24 25 26 685 Retrieved from: https://www.zorg-en- 27 28 686 gezondheid.be/sites/default/files/atoms/files/SPE_jaarrapport%202012.pdf. Accessed on 22 Dec. 16. 29 30 687 19. Van Leeuw V LC, Englert Y. Perinatale gegevens in het Brussels Gewest - Jaren 2013. Centre 31 32 33 688 d'Epidémiologie périnatale, 2014. Retrieved from: http://www.cepip.be/pdf/rapport_CEPIP_Bxl2013_NL.pdf. 34 35 689 Accessed on 22 Dec. 16. 36 37 690 20. Leroy Ch VLV, Englert Y. Données périnatales en Wallonie - Année 2013. Centre d'Epidémiologie 38 39 691 Périnatale, 2014. Retrieved from : http://www.cepip.be/pdf/rapport_CEPIP_wallonie2013_tma.pdf. Accessed 40 41 42 692 on 22 Dec. 16. 43 44 693 21. Armonk NIC. IBM SPSS Statistics for Windows, Version 22.0. In: Corp I, editor. 2013. 45 46 694 22. Kayem G, Dupont C, Bouvier-Colle MH, Rudigoz RC, Deneux-Tharaux C. Invasive therapies for primary 47 48 49 695 postpartum haemorrhage: a population-based study in France. BJOG : an international journal of obstetrics and 50 51 696 gynaecology. 2016;123(4):598-605. 52 53 697 23. Colmorn LB, Krebs L, Langhoff-Roos J. Potentially Avoidable Peripartum Hysterectomies in Denmark: A 54 55 56 698 Population Based Clinical Audit. PloS one. 2016;11(8):e0161302. 57 58 699 24. Comstock CH, Bronsteen RA. The antenatal diagnosis of placenta accreta. BJOG : an international 59 60 700 journal of obstetrics and gynaecology. 2014;121(2):171-81; discussion 81-2. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 30 of 41

1 2 3 701 25. D'Antonio F, Iacovella C, Bhide A. Prenatal identification of invasive placentation using ultrasound: 4 5 702 systematic review and meta-analysis. Ultrasound in obstetrics & gynecology : the official journal of the 6 7 703 International Society of Ultrasound in Obstetrics and Gynecology. 2013;42(5):509-17. 8 9 10 704 26. Fitzpatrick KE, Sellers S, Spark P, Kurinczuk JJ, Brocklehurst P, Knight M. The management and 11 12 705 outcomes of placenta accreta, increta, and percreta in the UK: a population-based descriptive study. BJOG : an 13 14 706 international journal of obstetrics and gynaecology. 2014;121(1):62-70; discussion -1. 15 For peer review only 16 707 27. Thurn L, Lindqvist PG, Jakobsson M, Colmorn LB, Klungsoyr K, Bjarnadottir RI, et al. Abnormally invasive 17 18 19 708 placenta-prevalence, risk factors and antenatal suspicion: results from a large population-based pregnancy 20 21 709 cohort study in the Nordic countries. BJOG : an international journal of obstetrics and gynaecology. 22 23 710 2016;123(8):1348-55. 24 25 26 711 28. Collins SL, Ashcroft A, Braun T, Calda P, Langhoff-Roos J, Morel O, et al. Proposal for standardized 27 28 712 ultrasound descriptors of abnormally invasive placenta (AIP). Ultrasound in obstetrics & gynecology : the 29 30 713 official journal of the International Society of Ultrasound in Obstetrics and Gynecology. 2016;47(3):271-5. 31 32 33 714 29. Chantraine F, Braun T, Gonser M, Henrich W, Tutschek B. Prenatal diagnosis of abnormally invasive 34 35 715 placenta reduces maternal peripartum hemorrhage and morbidity. Acta obstetricia et gynecologica 36 37 716 Scandinavica. 2013;92(4):439-44. 38 39 717 30. Eller AG, Bennett MA, Sharshiner M, Masheter C, Soisson AP, Dodson M, et al. Maternal morbidity in 40 41 42 718 cases of placenta accreta managed by a multidisciplinary care team compared with standard obstetric care. 43 44 719 Obstetrics and gynecology. 2011;117(2 Pt 1):331-7. 45 46 720 31. Placenta Accreta. The American College of Obstetricians and Gynaecologistis. Committee Opinion. 47 48 49 721 Number 529. Published July 2012. Reaffirmed 2015. 50 51 722 32. Belfort MA. Placenta accreta. American journal of obstetrics and gynecology. 2010;203(5):430-9. 52 53 723 33. Kayem G, Davy C, Goffinet F, Thomas C, Clement D, Cabrol D. Conservative versus extirpative 54 55 56 724 management in cases of placenta accreta. Obstetrics and gynecology. 2004;104(3):531-6. 57 58 725 34. Sentilhes L, Ambroselli C, Kayem G, Provansal M, Fernandez H, Perrotin F, et al. Maternal outcome 59 60 726 after conservative treatment of placenta accreta. Obstetrics and gynecology. 2010;115(3):526-34. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 31 of 41 BMJ Open

1 2 3 727 35. Fox KA, Shamshirsaz AA, Carusi D, Secord AA, Lee P, Turan OM, et al. Conservative management of 4 5 728 morbidly adherent placenta: expert review. American journal of obstetrics and gynecology. 2015;213(6):755- 6 7 729 60. 8 9 10 730 36. Chantraine F, Nisolle M, Petit P, Schaaps JP, Foidart JM. Individual decisions in placenta increta and 11 12 731 percreta: a case series. Journal of perinatal medicine. 2012;40(3):265-70. 13 14 732 37. Dilauro MD, Dason S, Athreya S. Prophylactic balloon occlusion of internal iliac arteries in women with 15 For peer review only 16 733 placenta accreta: literature review and analysis. Clinical radiology. 2012;67(6):515-20. 17 18 19 734 38. Cali G, Forlani F, Giambanco L, Amico ML, Vallone M, Puccio G, et al. Prophylactic use of intravascular 20 21 735 balloon catheters in women with placenta accreta, increta and percreta. European journal of obstetrics, 22 23 736 gynecology, and reproductive biology. 2014;179:36-41. 24 25 26 737 39. Duan XH, Wang YL, Han XW, Chen ZM, Chu QJ, Wang L, et al. Caesarean section combined with 27 28 738 temporary aortic balloon occlusion followed by uterine artery embolisation for the management of placenta 29 30 739 accreta. Clinical radiology. 2015;70(9):932-7. 31 32 33 740 40. Sentilhes L, Gromez A, Clavier E, Resch B, Verspyck E, Marpeau L. Predictors of failed pelvic arterial 34 35 741 embolization for severe postpartum hemorrhage. Obstetrics and gynecology. 2009;113(5):992-9. 36 37 742 41. Poujade O, Zappa M, Letendre I, Ceccaldi PF, Vilgrain V, Luton D. Predictive factors for failure of pelvic 38 39 743 arterial embolization for postpartum hemorrhage. International journal of gynaecology and obstetrics: the 40 41 42 744 official organ of the International Federation of Gynaecology and Obstetrics. 2012;117(2):119-23. 43 44 745 42. Yamasaki Y, Morita H, Miyahara Y, Ebina Y, Okada T, Yamaguchi M, et al. The factors associated with 45 46 746 the failure of transcatheter pelvic arterial embolization for intractable postpartum hemorrhage. Journal of 47 48 49 747 perinatal medicine. 2014;42(3):359-62. 50 51 748 43. Touboul C, Badiou W, Saada J, Pelage JP, Payen D, Vicaut E, et al. Efficacy of selective arterial 52 53 749 embolisation for the treatment of life-threatening post-partum haemorrhage in a large population. PloS one. 54 55 56 750 2008;3(11):e3819. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 41

1 2 3 751 44. Lee HJ, Jeon GS, Kim MD, Kim SH, Lee JT, Choi MJ. Usefulness of pelvic artery embolization in cesarean 4 5 752 section compared with vaginal delivery in 176 patients. Journal of vascular and interventional radiology : JVIR. 6 7 753 2013;24(1):103-9. 8 9 10 754 45. Kim YJ, Yoon CJ, Seong NJ, Kang SG, An SW, Kim YS, et al. Failed pelvic arterial embolization for 11 12 755 postpartum hemorrhage: clinical outcomes and predictive factors. Journal of vascular and interventional 13 14 756 radiology : JVIR. 2013;24(5):703-9. 15 For peer review only 16 17 757 18 19 758 20 21 22 759 23 24 760 25 26 27 761 28 29 762 30 31 32 763 33 34 764 35 36 37 765 38 39 766 40 41 42 767 43 44 768 45 46 47 769 48 49 770 50 51 52 771 53 54 772 55 56 57 773 58 59 774 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 33 of 41 BMJ Open

1 2 3 775 FIGURE LEGENDS 4 5 776 Figure 1. Flowchart of case reporting and data collection. 6 7 777 Figure 2. Circumstances in which peripartum hysterectomy and interventional radiology occurred in 166 8 9 10 778 women. 11 12 779 Figure 3. Cause of obstetric haemorrhage according to the 4T’s Mnemonic. 13 14 780 Figure 4. Antenatal suspicion and preventive measures in women with abnormally invasive placenta (AIP). 15 For peer review only 16 17 781 Supplementary Figure 1. Antenatal suspicion and preventive measures in women with abnormally invasive 18 19 782 placenta (AIP). 20 21 783 22 23 24 784 25 26 785 27 28 786 29 30 787 31 32 33 788 34 35 789 36 37 790 38 39 40 791 41 42 792 43 44 793 45 46 47 794 48 49 795 50 51 796 52 53 54 797 55 56 798 57 58 799 59 60 800 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 34 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Figure 1. Flowchart of case reporting and data collection.

48 143x212mm (300 x 300 DPI) 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 35 of 41 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Figure 2. Circumstances in which peripartum hysterectomy and interventional radiology took place in 166 women. 48 49 219x267mm (300 x 300 DPI) 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 36 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Figure 3. Cause of obstetric haemorrhage according to the 4T's mnemonic. Tonus: uterine atony (n=91); 36 Tissue: AIP (n=14), placenta praevia (n=18), combination of AIP and placenta praevia (n=24), placental 37 remnants or retained pl acenta without AIP (n=26). Trauma: uterine rupture (n=12), genital tract lacerations 38 (n=12), bleeding caused by unintended injuries during caesarean delivery (n=14) or D&C (n=3). 39 Uncomplicated caesarean deliveries are not included in the trauma group. Thrombin: abnormal coagulation before onset of bleeding (n=6). DIC secondary to major haemorrhage was not included in the thrombin 40 group. Others: the cause of bleeding remains unclear or is unknown. 41 42 171x143mm (300 x 300 DPI) 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 37 of 41 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Figure 4. Antenatal suspicion and preventive measures in women with abnormally invasive placenta (AIP). 30 § < 0.05 for difference in between groups. 31 32 373x243mm (300 x 300 DPI) 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 38 of 41

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 293x157mm (300 x 300 DPI) 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 39 of 41 BMJ Open

1 2 3 4 STROBE 2007 (v4) Statement—Checklist of items that should be included in reports of cohort studies 5 6 Item 7 Section/Topic Recommendation Reported on page # 8 # 9 Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the title or the abstract Title Page 10 11 (b) ForProvide in the abstract peer an informative and balancreviewed summary of what was doneonly and what was found Abstract 12 Introduction 13 Background/rationale 2 Explain the scientific background and rationale for the investigation being reported Background p6-7 14 15 Objectives 3 State specific objectives, including any prespecified hypotheses Background p7 16 Methods 17 18 Study design 4 Present key elements of study design early in the paper Method p7 19 Setting 5 Describe the setting, locations, and relevant dates, including periods of recruitment, exposure, follow-up, and data Methods p7-8 20 collection 21 Participants 6 (a) Give the eligibility criteria, and the sources and methods of selection of participants. Describe methods of follow-up Not applicable 22 23 Methods p7 24 (b) For matched studies, give matching criteria and number of exposed and unexposed / 25 Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, and effect modifiers. Give diagnostic criteria, if Methods p8-10 26 applicable 27 28 Data sources/ 8* For each variable of interest, give sources of data and details of methods of assessment (measurement). Describe Methods p9-10 29 measurement comparability of assessment methods if there is more than one group 30 Bias 9 Describe any efforts to address potential sources of bias / 31 Study size 10 Explain how the study size was arrived at Methods P9 and 32 33 Figure 1 34 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If applicable, describe which groupings were chosen and / 35 why 36 Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding Methods p9-10 37 38 (b) Describe any methods used to examine subgroups and interactions Methods p9-10 39 (c) Explain how missing data were addressed / 40 (d) If applicable, explain how loss to follow-up was addressed / 41 42 (e) Describe any sensitivity analyses / 43

44 45 46 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open Page 40 of 41

1 2 3 4 Results 5 6 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers potentially eligible, examined for eligibility, confirmed Figure 1 7 eligible, included in the study, completing follow-up, and analysed 8 (b) Give reasons for non-participation at each stage / 9 (c) Consider use of a flow diagram Figure 1 10 11 Descriptive data 14* (a) ForGive characteristics peer of study participants (egreview demographic, clinical, social) and only information on exposures and potential Table 1 12 confounders 13 (b) Indicate number of participants with missing data for each variable of interest Results, Table 1, p11 14 (c) Summarise follow-up time (eg, average and total amount) / 15 16 Outcome data 15* Report numbers of outcome events or summary measures over time Results , prevalence 17 p10 18 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates and their precision (eg, 95% confidence Results , prevalence 19 interval). Make clear which confounders were adjusted for and why they were included p10 20 (b) Report category boundaries when continuous variables were categorized Table 1 21 22 (c) If relevant, consider translating estimates of relative risk into absolute risk for a meaningful time period / 23 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and sensitivity analyses Results, 24 - management of 25 AIP, p15-16, figure 4 26 27 - IR versus H, table 3 28 - Failed E versus 29 succesfull E, table 4 30 31 Discussion 32 Key results 18 Summarise key results with reference to study objectives Discussion p20 33 Limitations 34 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, multiplicity of analyses, results from Discussion p21-24 35 similar studies, and other relevant evidence 36 37 Generalisability 21 Discuss the generalisability (external validity) of the study results / 38 Other information 39 Funding 22 Give the source of funding and the role of the funders for the present study and, if applicable, for the original study on Funding p26 40 41 which the present article is based 42 43

44 45 46 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 41 of 41 BMJ Open

1 2 3 4 5 *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and unexposed groups in cohort and cross-sectional studies. 6

7 8 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and published examples of transparent reporting. The STROBE 9 checklist is best used in conjunction with this article (freely available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 10 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is available at www.strobe-statement.org. 11 For peer review only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

44 45 46 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 47 48 49 50 51 52 53 54 55 56 57 58 59 60