Incidence, outcomes and management of spontaneous haemoperitoneum in pregnancy: a UK population-based study

Introduction

Spontaneous haemoperitoneum in pregnancy (SHiP) is the occurrence during pregnancy of sudden intra-abdominal haemorrhage that is unrelated to extrauterine pregnancy, trauma, or uterine rupture. SHiP most commonly occurs in the third trimester but can occur at any time during pregnancy or in the weeks after birth^1,2, and is associated with preterm birth, high perinatal mortality and, more rarely, maternal mortality^1–3.

Known maternal risk factors for SHIP are maternal age over 35 years² and factors associated with endometriosis^1,4,5, including pregnancies conceived via artificial reproductive techniques (ART)^6–8 as well as multiple pregnancy². Women experiencing SHiP typically present with sudden acute abdominal pain, frequently in combination with hypovolaemic shock, decreased haemoglobin levels, and/or signs of fetal distress⁷. Bleeding may originate from abdominal parenchymal organs, abdominal arterial or venous vessels such as splenic artery aneurysms or pelvic varicose veins, from extrauterine endometroid or decidual tissue implants, or other causes^1,9,10. Often, placental abruption is initially suspected^1,7. Diagnosis usually involves detection of free peritoneal fluid by ultrasound¹, sometimes in combination with other modalities such as computed tomography^2,3,5. Management is typically surgical, with laparotomy the most common intervention^1,2, although expectant management was the most frequently used treatment approach in the Netherlands³. The mode of birth is caesarean section in a majority of cases^2,3,7.

Because of the rarity of the condition, it is challenging to estimate the incidence of SHiP. Most reports of SHiP in the literature are individual case studies or small case series¹. The MBRRACE-UK surveillance programme identified six maternal deaths attributed to spontaneous intra-abdominal bleeding in the three years from 2019 to 2021⁹, but information on non-fatal case incidence and morbidity related to SHiP in the UK is lacking. In light of the evidence gap, International Network of Obstetric Survey Systems (INOSS) initiated a series of prospective nationwide surveillance studies in multiple countries¹¹.

A two-year study in the Netherlands found an estimated incidence of SHiP of 4.9 per 100,000 births, with no maternal mortality among 14 women, and three perinatal deaths³. A similar study in Italy estimated the incidence at 4 per 100,000 births, with two maternal deaths, one neonatal death, and five stillbirths among 29 cases². Here, we present the findings of the UK Obstetric Surveillance System (UKOSS) study of SHiP.

The aim of this two-year, prospective, UK population-based surveillance study was to describe the incidence, diagnosis, and management of SHiP, and the characteristics and outcomes of affected women and their babies.

Methods

Results

Discussion

This two-year, prospective, UK population-based surveillance study estimated the incidence of SHiP as 1.3 cases per 100,000 maternities, equating to 1 case per 75,614 maternities. Aneurysms and rupture or haematoma in the liver, spleen or kidney were the most common sources of bleeding. The diagnosis was most commonly made and treated during open abdominal surgery. A minority of affected women were receiving anticoagulant agents for prophylaxis (2/20) or treatment (4/20) of thromboembolism. Severe maternal and perinatal outcomes were common; these included stillbirths, neonatal intensive care unit admission, maternal intensive care unit admission, and one maternal death.

This estimated incidence of SHiP is considerably lower than that estimated using similar national surveillance systems in the Netherlands (4.9 cases per 100,000 births) and Italy (4 cases per 100,000 births). All countries collaborate in the INOSS network and based their study on a joint definition of SHIP. However, local adaptations to this definition need to be considered in interpretation of differences across countries. The Dutch study ultimately deviated from the INOSS definition of SHiP¹¹ and decided to exclude abdominal bleeding from aneurysms or liver, spleen and kidney. After these exclusions the management was expectant in over a third of included women, while fewer than two thirds underwent either surgery (6/14) or embolisation (2/14)³. Our study also deviated from the INOSS definition, but by including only cases requiring surgery (rather than those requiring either surgery or embolisation), which could have led us to underestimate incidence. Management by endovascular intervention depends on availability of expertise, including in interventional radiology, in the hospital. Our inclusion criteria may account for the some of the disparity between our findings and those of the Italian surveillance², which used the INOSS definition, although the authors do not report the number of women who were eligible for inclusion as a result of having undergone embolisation.

Subjectivity of diagnosis may also be a contributing factor in the differing estimates of incidence – in the Dutch study, auditors disagreed initially as to whether a SHiP diagnosis was appropriate in almost half (9/20) of potential cases submitted by reporters. There may also be underlying differences between the three countries in the risk profile of the pregnant population: for example, ART, an established risk factor for SHiP^6,7, contributes to 4.2% of births in Italy¹⁷ and 3.0% in the Netherlands¹⁸ versus 1.2% in the UK¹⁹. Unfortunately, we did not explicitly request data on ART in our study, although IVF was mentioned in the supplementary notes of one woman; five of the 14 women in the Dutch study and five of the 29 in the Italian study had conceived by ART.

We found that aneurysms, rupture or hematoma in the spleen, liver or kidney were the most common causes of SHiP. Pregnancy is associated with vascular wall and hemodynamic changes, and also increases the risk of haemorrhage/rupture of intracerebral and aortic aneurysms^20–22. Vascular changes and hypertension in preeclampsia may aggravate this, but only two women in our study had pre-existing hypertensive disorders during the index pregnancy. Only one woman had endometriosis as the source of bleeding; this low frequency is striking even when exempting from the denominator cases arising from liver hematoma/rupture known to be associated with hypertensive disorders of pregnancy and HELLP-syndrome. By contrast, endometriosis was present in around a fifth of women with SHiP identified by the Italian surveillance² and a third of women identified in the Netherlands³.

SHiP was diagnosed by imaging prior to surgery in fewer than half of the women in our study: five of 20 were diagnosed using CT scan, four using ultrasound and none by MRI. This is in contrast to diagnosis in the Netherlands, where all affected women in a 2-year surveillance study underwent ultrasound, around a third (4/13) underwent CT, and a third (4/13) underwent MRI. This likely reflects the higher proportion of hemodynamically stable women included under the Dutch definition of SHiP, whereas CT or MRI is unlikely to be appropriate in the more emergent cases captured by the case definition in our study. The diagnostic modalities in Italy more closely resembled that in the UK: 14 of 29 women underwent ultrasound, six underwent CT and none MRI. Focused assessment sonography for trauma (FAST), used effectively in non-pregnant trauma patients with suspected haemoperitoneum²³, offers a more rapid alternative to MRI or CT that may be more appropriate for diagnosing SHiP in haemodynamically unstable women.

Almost a third of women in our study (6/20) were receiving anticoagulant agents, and only four were receiving therapeutic high-dose LMWH for thrombosis or thromboembolism. Thrombosis and thromboembolism are now the leading cause of maternal deaths in the UK during pregnancy or in the six weeks following birth²⁴. Royal College of Obstetricians and Gynaecologists guidelines recommend prophylactic LMWH during pregnancy for women with previous venous thromboembolism or certain combinations of risk factors²⁵, and LMWH treatment in clinically suspected pulmonary embolism or deep vein thrombosis²⁶, unless strongly contraindicated. Awareness of the potential for SHiP, albeit rare, in these women, is important.

Conclusion

SHiP is rare in the UK but when it does occur, it can be associated with major maternal morbidity and mortality, and perinatal outcomes are poor. Healthcare providers should be aware of the possible range of symptoms and consider it as a possible alternative diagnosis to placental abruption, intra-abdominal pain, uterine rupture, or abdominal infection, particularly in pregnant women presenting with abdominal pain and/or circulatory compromise. Future research would benefit from a strengthened consensus on the definition of SHiP.