Endometriosis-associated spontaneous major haemoperitoneum in pregnancy

Abstract

Spontaneous haemoperitoneum in pregnancy (SHiP) related to endometriosis is a rare and life-threatening complication. We report a case of a patient presenting to our department with major haemoperitoneum at 23+3 weeks of gestation due to a large rectovaginal endometriotic nodule. The patient required a midline laparotomy to evacuate 1 L of haemoperitoneum and achieve haemostasis. A large rectovaginal nodule was seen bleeding and was packed with haemostatic material and a large swab. After 24 hours, the swab was removed and haemostasis was confirmed. The patient was monitored very closely by a multidisciplinary team and the pregnancy was allowed to continue to try and achieve a better outcome for the baby and at 28 weeks of gestation, a girl was delivered in good condition via caesarean section.

Background

Endometriosis-associated spontaneous haemoperitoneum in pregnancy (SHiP) is a challenging and life-threatening condition for both mother and baby. Although in most cases the baby is delivered at the time of presentation, in cases of extreme prematurity, a conservative approach may be considered to improve the baby’s outcome. A multidisciplinary team (MDT) approach is required to achieve optimal outcome.

Case presentation

A patient in her late 20s, para 1, who conceived naturally was transferred at 23+3 weeks of gestation to our unit due to lack of level 3 neonatal intensive care at the referring hospital. The patient presented with generalised abdominal pain and antepartum haemorrhage. Her medical history included a previous full-term elective caesarean section and deep infiltrating rectovaginal endometriosis, for which she never had an operation. At 20 weeks of gestation, the routine anomaly ultrasound revealed major placenta praevia.

On arrival, auscultation confirmed a normal foetal heart. Speculum examination revealed in the posterior fornix mild bleeding arising from a full-thickness vaginal endometriotic nodule. The patient was haemodynamically stable and the cervix was long and closed, with no blood arising from its canal. Blood test revealed normal coagulation profile but significant anaemia, with a haemoglobin of 78 g/L.

The first dose of corticosteroids (dexamethasone 12 mg) was given at presentation at 23+3 weeks with a second and final dose administered 12 hours after.

As the patient was stable and considering the early gestation of 23+3 weeks, it was decided to try conservative management in a high-dependency unit and investigate further with imaging.

An ultrasound scan identified free fluid in the pelvis and abdomen and CT scan (figure 1) revealed no active bleeding but moderate haemoperitoneum of approximately 500 mL, extending into the pouch of Douglas. Two units of blood were requested and transfused.

Figure 1.

(A) CTAP showing foetus and haemoperitoneum. (B) CTAP showing foetus and haemoperitoneum. (C) CTAP showing pelvis with moderate haemoperitoneum. No active bleeding. CTAP, CT abdomen and pelvis.

The differential diagnosis included rupture of heterotopic pregnancy, rupture of aneurism (such as splenic or mesenteric) and intra-abdominal bleeding due to endometriosis. The bleeding observed at the posterior vaginal fornix endometriotic nodule suggested that the same process may have occurred intra-abdominally as well, leading to the haemoperitoneum. Hoping that the internal bleeding may be self-limited, and given the extreme prematurity and patient being haemodynamically stable, the multidisciplinary team agreed to proceed with conservative management.

On day 3 following admission, the patient developed type 1 respiratory failure secondary to abdominal pain and splinting, with evidence of hypovolaemic shock. Chest X-ray revealed right lower lobe collapse with consolidation and effusion, indicating possible hospital-acquired pneumonia. Progressive deterioration, increasing abdominal pain, distension and further decline in haemoglobin levels (79 g/L from 88 g/L) led to the decision for exploratory laparotomy. A midline incision was performed and 1000 mL of blood was evacuated from the abdomen. Few areas of mild bleeding were identified at the uterovesical pouch, directly originating from peritoneal endometriotic lesions. These were controlled with bipolar diathermy. The main source of bleeding, in the form of moderate venous ooze, was seen in the pouch of Douglas, which was obliterated by friable endometriotic tissue. Due to the large size, friability and invasion of the endometriotic nodule in the rectum, it was decided to achieve haemostasis with pressure. Haemostatic material (oxidised regenerated cellulose) was applied directly on the bleeding nodule and large swab was used to pack the pouch.

A decision was taken to not deliver the foetus, and to return, 24 hours later, for a relook procedure. This showed a dry field and the swab was removed. The patient recovered in the intensive care unit (ICU) for 3 days before returning to a high-dependency unit in delivery suite.

Investigations

The investigations have been extensively described in the ‘case presentation’ paragraph.

Differential diagnosis

The differential diagnosis includes more common pathologies that present in the second trimester of pregnancy such as placental abruption, heterotopic pregnancy or uterine rupture.

Other, less common, differential diagnoses include haemolysis, elevated liver enzymes and low platelets (HELLP) syndrome, severe pre-eclampsia and acute fatty liver disease. The blood tests were not pointing in the direction of these diagnoses.

Ruptured liver capsule, spleen or its vasculature were also taken into consideration due to the level of haemoperitoneum and the significant drop in haemoglobin level.^{1 2}

Treatment

This was extensively described in the case report.

Outcome and follow-up

In the following weeks, her recovery was complicated by urosepsis, acute kidney injury, return to ICU and subacute bowel obstruction, likely to be small bowel. Both her urine and blood cultures showed Klebsiella pneumoniae urosepsis. Conservative management continued with appropriate intravenous antibiotics and under close guidance of our MDT, including a microbiologist. An ultrasound performed at 27+1 weeks of gestation confirmed that the placenta was still low, not covering but, less than 2 cm from the cervical os. The ultrasound showed consistent foetal growth with estimated weight of 846 g. The patient remained stable until 28+3 weeks of gestation when increasing abdominal pain and tenderness indicated potential placental abruption. A baby girl (weight 846 g) was delivered by classical caesarean section. Insufflation and ventilation breaths were given. At 5 min, she was intubated for only 1 min, then extubated because of good respiratory effort. The baby was transferred to the neonatal intensive care unit in good condition, with Apgar 8 at 10 min. Surgically, no further intraperitoneal bleeding was identified and a large retroplacental clot seen at delivery confirmed the diagnosis of placental abruption. Both mother and baby were discharged after a total of 43 days in the hospital. The baby was discharged to the referring hospital with no cardiovascular concerns and established enteral feeding. Neurology and cranial ultrasound were normal at discharge.

The mother was followed up with an MRI (figure 2) performed at 9 months following the delivery. This showed a 4–5 cm rectovaginal nodule with rectal stricture.

Figure 2.

MRI of the patient’s pelvis (9 months postnatally) which shows the endometriotic rectovaginal nodule, behind the cervix. The vaginal cavity is depicted white as it was filled with gel.

The patient declined laparoscopic treatment of endometriosis but, in 2022, she presented in clinic with significant chronic pelvic pain not responding to conservative measurements. She underwent a diagnostic laparoscopy in May 2023 which revealed severe intra-abdominal adhesions and obliterated pouch of Douglas. A repeat MRI confirmed full-thickness vaginal and rectal disease with a large nodule causing rectal stricture.

Radical surgery has been offered to the patient and she is currently considering her options.

Discussion

We present a case of endometriosis causing life-threatening SHiP in a singleton pregnancy at 23 weeks of gestation. Careful decision-making by a multidisciplinary team of specialists including obstetricians, endometriosis specialists, intensivists, general surgeons and neonatologists resulted in a favourable outcome for both mother and baby. The case prompted us to perform an initial literature review, which provided us with useful information while managing the case.

We challenge the traditional view that pregnancy always benefits endometriosis sufferers due to suppression of the disease by the effect of internally produced progesterone. Endometriosis is associated with higher rates of pregnancy complications such as ectopic pregnancy, miscarriage, preterm delivery and caesarean section.² Furthermore, as shown here, several case reports now suggest rare but potentially catastrophic complications including SHiP and spontaneous bowel perforation.^{1 3}

Women with severe endometriosis have an increased risk of SHiP. Previous systematic reviews, not included in our search period, identified deep infiltrating endometriosis in the posterior pelvic compartment and in vitro fertilisation as the main risk factors for SHiP, with an overall frequency of 0.4% in this patient population. Although rare, SHiP is an extremely serious event that can endanger the life of both mother and foetus, with mortality rates of nearly 2% and 27%, respectively.⁴

The pathophysiology of endometriosis-associated SHiP is probably complex and may not be explained by one mechanism alone.^{1 5} Intraperitoneal bleeding in pregnancy should be associated with endometriosis when it originates either from ruptured vessels in close proximity to disease or from endometriotic lesions themselves. The anatomical changes of pregnancy, in combination with the adhesions caused by endometriosis, most likely have the potential to tear vessels involved in these adhesions. This may explain bleeding seen in cases of SHiP associated with severe endometriosis and adhesions. It does not explain though bleeding seen from peritoneal endometriotic lesions. Recent evidence suggests that endometriosis is associated with progesterone resistance, characterised by suboptimal expression of target genes. Brosens et al speculated that ‘functional’’ progesterone withdrawal triggers involution of the decidual phenotype of the ectopic endometrium, leading to peritoneal bleeding of unpredictable severity.⁶ Such a mechanism of apoptotic cell death may explain the mild bleeding that we observed from superficial peritoneal implants on the uterovesical peritoneum in our case. Brosens et al also note that spontaneous haemoperitoneum associated with mild endometriosis has also been described in the postpartum period and at the time of menstruation.⁶

Several other factors can also be at play. Decidualised sites tend to perforate the utero-ovarian vessels, which are already friable due to the chronic inflammation associated with the underlying endometriosis. During pregnancy, there is an increase in the venous pressure in the utero-ovarian circulation. All the above elements, either single or combined, may be responsible for vascular instability and bleeding.^{1 2}

Spontaneous haemoperitoneum in pregnancy occurs usually in the second or early third trimester. It can be diagnosed using clinical, laboratory and imaging investigations. A high level of suspicion in known cases of endometriosis is required. The differential diagnosis includes placental abruption, heterotopic pregnancy, ruptured uterus or appendix, HELLP syndrome, severe pre-eclampsia, acute fatty liver disease, chorioamnionitis and ruptured liver, spleen or its vasculature.^{1 2}

The patient usually presents with abdominal, pelvic or flank pain. Hypotension, tachycardia or any other sign of hypovolaemic shock, fall of haemoglobin levels or any sign of foetal distress are alarming indications of significant bleeding. Ultrasound sonography will confirm the presence of peritoneal haemorrhagic fluid, its amount and possibly its origin. MRI or CT may conclude the diagnosis.^{2 7} Interestingly, our case presented with vaginal bleeding from a rectovaginal endometriotic nodule fully infiltrating the wall of the posterior vaginal fornix in addition to the signs and symptoms of internal haemorrhage. The simultaneous presence of placenta praevia presented an additional diagnostic challenge. Careful clinical examination per speculum revealed that the bleeding originated from the nodule behind the cervix and not through the cervical external os. We therefore advocate a detailed clinical examination, including speculum, and high level of suspicion in pregnant patients with known or suspected severe endometriosis who present with vaginal bleeding with or without signs of intra-abdominal haemorrhage.

The management of SHiP will depend on the clinical presentation, but in most cases, will require surgical intervention. A conservative approach may be reasonable in a patient who is haemodynamically stable and provided high-dependency monitoring is available. Signs of haemodynamic instability should prompt surgical exploration. Laparoscopy may be considered at earlier stages of pregnancy depending on the availability of relevant expertise. Advanced laparoscopic skills will be required in such cases. In most cases, an exploratory laparotomy will be appropriate.⁷ Delivering the foetus may be necessary in some cases to gain access to the pelvis. The decision is easier to take at advanced gestations but will be a difficult one earlier on. Recurrence may occur if the foetus remains undelivered. Delivery, or even hysterectomy in the first trimester with the foetus in utero, may be necessary occasionally to stop the bleeding and save the mother’s life.^{7 8} In our case, the combination of extreme prematurity and our patient being haemodynamically stable prompted us to avoid delivery of the foetus at first presentation. However, we acknowledge the significant clinical challenge of these cases. In other scenarios, where the gestation is more advanced, for example after 26–28 weeks of gestation and/or the mother is haemodynamically unstable, immediate delivery will be appropriate.

In conclusion, our case highlights a rare but potentially catastrophic clinical presentation in pregnancy.^{1 7–9} SHiP has been associated with a number of causes including endometriosis.⁵ In our view, involving an MDT of specialists is paramount in achieving the best possible outcome for both mother and baby. Stopping the bleeding to save the mother’s life is the priority.⁶ The decision to perform a laparotomy or a laparoscopy will depend on a number of factors, two main ones—skills available and gestational age. The surgical manoeuvres that may be needed will include ligating bleeding vessels, using diathermy (for example for superficial peritoneal implants not at close proximity to visceral structures), performing an oophorectomy or hysterectomy in unsalvageable cases, or using haemostatic agents and packing for large nodules overlying important structures, as we had to do with our patient’s bleeding and friable rectovaginal nodule. The decision to deliver the foetus at the time of the first presentation of SHiP is exceptionally challenging in cases of extreme prematurity. We show here that it is possible, at very early gestations, to control the bleeding surgically and prolong the pregnancy. When this decision is taken, the patient will have to be monitored very carefully by the MDT, as recurrence of SHiP is possible and increased morbidity should be anticipated. In our case, we were able to prolong our patient’s pregnancy by approximately 4 weeks following the initial SHiP episode and achieved a good outcome for both mother and baby.

Learning points.

• Spontaneous haemoperitoneum in pregnancy related to endometriosis is a rare and life-threatening complication.

• The baby may often have to be delivered at the time of first presentation but, in cases of extreme prematurity, a conservative approach may be considered to improve foetal outcome.

• A multidisciplinary approach is required to achieve optimal outcome in these cases.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).