Ventricular fibrillation during termination of pregnancy

Abstract

In June, 2006, a 29-year-old woman underwent a termination of pregnancy at 35 weeks of gestation, because the fetus had been found to have porencephaly. She had had three previous pregnancies, of which two had proceeded to term; her medical history was otherwise unremarkable. The fetus was killed with sufentanil and lidocaine. Misoprostol was administered intravaginally to induce labour, and the membranes were artifi cially ruptured. 15 min later, the patient suddenly lost consciousness and began to gasp for breath. She entered ventricular fi brillation. She was immediately intubated and ventilated; cardiac massage was done; and she was given three DC shocks. The period between onset of cardiac arrest and the start of life support was estimated at around 3 min; life support was given for about 10 min. Immediate transthoracic echocardiography, using an obstetric ultrasound device, showed a massively dilated hypokinetic right ventricle, with fl oating echo-dense masses (fi gure). Blood test results indicated disseminated intravascular coagulation (platelet and fi brinogen concentrations very low at 20×109/L and 0·59 μmol/L respectively; activated partial thromboplastin time >120 s). Within 1 h of the onset of cardiac arrest, the fetus was delivered—with subsequent profuse vaginal bleeding. Manual exploration of the uterus, uterine massage, intravenous oxytocin, and uterine packing all failed to stem the bleeding, so a sulprostone infusion was started, and the patient was given several units of blood, which stabilised her condition. The combination of cardio vascular collapse, coagulopathy, and the echocardiographic fi ndings indicated that an amnioticfl uid embolism (AFE) was likely; a blood sample was immediately analysed by the pathologist, who observed amniotic and fetal cells after staining the sample with Wright’s stain and Nile blue stain. The patient was transferred to another hospital, where more intensive management was possible. She was given bilateral uterine embolisation, which caused the bleeding to cease. However, her haemodynamic stability deteriorated, despite increasing doses of epinephrine and dobutamine. Transoesophageal echo cardiography showed left ventricular failure (ejection fraction <15%). 8 h after cardiac arrest, extracorporeal life support (ECLS) was initiated, by use of a femoro-femoral bypass. The patient’s left ventricular function did not improve. She developed a fever and a raised white cell count that appeared to be caused not by an infection, but a systemic infl ammatory reaction. 12 days after cardiac arrest, the decision was made to give intravenous hydrocortisone (50 mg every 6 h). Left ventricular function began to recover within hours, and we were able to discontinue ECLS 20 days after cardiac arrest. The patient was extubated on day 26, and discharged on day 40. When last seen, in March, 2007, the patient was well, with no evidence of neurological damage; she had resumed work, and was preparing to run the Paris marathon. AFE is a rare complication of pregnancy, occurring in one in 20 000–50 000 deliveries. However, it is one of the leading causes of maternal mortality in developed countries; the proportion of patients surviving without neurological damage is only 15%. The presence of amniotic fl uid in the maternal bloodstream causes severe vasoconstriction and pulmonary hypertension. There follows a cascade of infl ammatory activation, similar to the systemic infl ammatory response caused by sepsis or burns. This process can cause multiple organ dysfunction, and typically depresses the myocardium. This scenario needs to be anticipated, so that alternative treatments—which can include ECLS—can be initiated or prepared. In this case, steroids appeared instrumental in reversing left ventricular dysfunction, perhaps because of their anti-infl ammatory activity. Had ventricular function not recovered, a heart transplant might have been necessary.