Case report: Life-threatening bronchospasm after intramuscular carboprost for postpartum haemorrhage


Dr CR Harber, Department of Anaesthesia, University Hospitals NHS Trust, Queen’s Medical Centre, Nottingham, NG7 2UH, UK. Email

Case report

We present a case of severe, life-threatening bronchospasm following intramuscular (i.m.) carboprost (Hemabate®; Pharmacia Upjohn, Kalamazoo, MI, USA). This synthetic analogue of prostaglandin F2-alpha is licensed in the UK for the treatment of postpartum haemorrhage (PPH) unresponsive to uterine massage, intravenous oxytocin and i.m. ergot alkaloids.1 Although bronchospasm is noted in the product literature and caution advised if the drug is considered for women with asthma,1,2 adverse effects reported hitherto have been described as mild, transient and reversible.3–6 This report describes the development of severe, protracted, life-threatening bronchospasm 15 minutes after i.m. carboprost.

A 39-year-old multiparous woman was admitted by ambulance to the delivery suite at 27 weeks of gestation with an antepartum haemorrhage. She had a history of two previous normal vaginal deliveries and one emergency caesarean section for breech presentation. She had mild asthma, for which she infrequently used inhaled salbutamol.

Her initial symptoms were painless bleeding per vaginam (p.v.) and lower abdominal pain. She received 500 ml intravenous crystalloid solution en route to hospital. Her initial vital signs on admission were stable, and cardiotocograph recorded a fetal heart rate of 100 beats per minute with accelerations. While being assessed, she had another significant p.v. bleed, a diagnosis of placental abruption was made followed by decision for a category one (immediate threat to life)7 caesarean section. It was estimated that 1000 ml blood had been lost, and a further 500 ml intravenous crystalloid was given. Within 10 minutes, general anaesthesia had been induced in the operating theatre. Induction consisted of the administration of intravenous thiopental and succinylcholine (standard rapid sequence induction), and anaesthesia was maintained with the inhalational agent sevoflurane. Three minutes later, a live male fetus was delivered, with umbilical arterial pH 6.65 and base deficit 27 mmol. The neonate’s trachea was intubated, and the baby was transferred to the neonatal unit for mechanical ventilation.

Haemorrhage was continuing in theatre; haemoglobin concentration was 8.5 g/dl. At delivery, she received 5 units of intravenous synthetic oxytocin (Syntocinon®; Alliance Pharmaceuticals, Chippenham, UK) followed by an infusion at 10 units per hour. Uterine atony persisted, and she was given a further 5 units of intravenous Syntocinon followed 5 minutes later by 500 microgram i.m. ergometrine. She continued to bleed steadily but remained cardiovascularly stable under anaesthesia. A diagnosis of uterine atony was made, and after discussion between anaesthetist and surgeon regarding the potential adverse effect of bronchospasm in a known asthmatic, she was given 250 microgram i.m. carboprost (Hemabate®). The surgeon applied a compressive suture (B-Lynch) encircling the uterus. Fifteen minutes later, she received a further dose of carboprost (250 microgram i.m.) and bleeding ceased. Estimated total blood loss was 2500 ml, and intravenous fluid administration intraoperatively consisted of 2000 ml crystalloid, 500 ml colloid and 2 units of packed red blood cells.

After a further 15 minutes, her peak airway pressures steadily increased, and it was noted that her lungs were becoming increasingly difficult to ventilate, despite increasing the inhalational anaesthetic agent (a known bronchodilator). A diagnosis of bronchospasm was made, and senior anaesthetic assistance was summoned. Her lungs were ventilated with 100% oxygen, and intravenous salbutamol was administered by bolus (two 250-microgram doses) and infusion (20 microgram/minute). There was no response, and she was subsequently treated with the following intravenous bronchodilators: aminophylline (5 mg/kg bolus then 0.5 mg/kg/hour infusion), magnesium sulphate (5 g), ketamine (25 mg) and finally adrenaline (20 microgram bolus then infusion at a rate rapidly increased to 13 microgram/minute). Intravenous hydrocortisone (200 mg) was also given. Despite volatile anaesthetic agent and salbutamol (both uterine relaxants), there was no further p.v. bleeding.

Her severe bronchospasm persisted, and several hours of stabilisation in theatre preceded transfer to the intensive care unit (ICU). Oxygenation was maintained despite increasing difficulty in ventilating her lungs, although extremely high inflation pressures were required to achieve acceptable tidal volumes. Inspiratory pressures of 45–50 cmH2O gave tidal volumes between 180 and 220 ml. This resulted in a severe respiratory acidosis (arterial pH 6.75, PaCo2 26 kPa) that corrected over the following 9 hours. Cardiovascular variables remained stable, albeit with inotropic support, despite the high airway pressures. A chest radiograph showed extensive bilateral pulmonary infiltrates with a pneumopericardium. A diagnosis of severe acute respiratory distress syndrome (ARDS) was made, compatible with acute lung injury secondary to barotrauma. A subsequent echocardiogram showed normal cardiac function.

Recurrent episodes of bronchospasm and airway hyper-reactivity led to persistent ventilatory difficulty. Weaning from mechanical ventilation proved difficult, and an elective tracheostomy was performed on day 5. She was transferred to the medical high-dependency unit after 15 days in ICU, then to a respiratory ward. She was finally discharged from hospital 28 days postpartum.

Following discharge, she was immobile on account of dyspnoea and was re-admitted to hospital with a significant pulmonary thromboembolism proven by computed tomography. Throughout her admission, she had received daily thromboprophylaxis (subcutaneous enoxaparin 40 mg daily). The neonate required prolonged mechanical ventilation and was finally discharged home some 3 months after birth.


This case illustrates a potentially fatal adverse effect of carboprost. The combination of placental abruption, caesarean section and coagulopathy are strong risk factors for major PPH (blood loss > 1000 ml, with or without clinical signs of shock).8 However, our patient showed cardiovascular stability despite continuing haemorrhage, and resuscitation was prompt, as anaesthetic and surgical teams were prepared for major blood loss. It has been highlighted previously that physiological responses rather than estimates of blood loss may best gauge the severity of haemorrhage.9 In the absence of life-threatening haemorrhage, the use of carboprost might have been unwarranted.

The standard drug regimen for induction of general anaesthesia has the potential to cause bronchospasm. The onset of ventilatory difficulty was not, in this case, temporally related to induction of anaesthesia and occurred at a point when most of these drugs would have been redistributed or metabolised. Fluid volume overload was not considered probably in view of her age, lack of cardiovascular co-morbidity, degree of haemorrhage and fluid volume given. Her preoperative arterial blood pressure was normal, and there were no features of pre-eclampsia. Transfusion-related lung injury (TRALI) also seemed unlikely as only 4 units of packed red cells were given. Two units were given intraoperatively after the onset of ventilatory difficulty and 2 units on ICU. Neither fresh-frozen plasma nor platelets (the most common precipitants of TRALI) were given.

Amniotic fluid embolism was considered, but there was no evidence of hypoxaemia or coagulopathy. Moreover, echocardiographic evidence of normal pulmonary artery pressures and right ventricular function made this diagnosis unlikely. Although the possibility of anaphylaxis exists, the temporal relationship to the administration of carboprost was compelling, and the lack of haemodynamic instability again makes this improbable. Serum tryptase was not measured. ARDS probably resulted from ventilator-induced lung injury, possibly compounded by haemorrhage and blood transfusion secondary to placental abruption.

The most frequent reported adverse effects of carboprost are nausea, vomiting, flushing, pyrexia and myalgia.1,2 Increases in pulmonary and systemic vascular resistance together with intrapulmonary shunting can cause arterial haemoglobin oxygen desaturation.8,10,11 Carboprost can cause significant bronchospasm, even in those without asthma.12 This has been documented in the confidential enquiries into maternal deaths9 although is not emphasised in the drug product literature or British National Formulary.1,2 The product licence does, however, stipulate that ‘conventional’ treatment (uterine massage, intravenous. oxytocics and i.m. ergot preparations) must precede its use. The implication is that carboprost should be a last-resort therapy used in an attempt to avoid more invasive surgical intervention. This stance is at odds with published investigations of the drug’s use for prophylaxis of PPH.10,13 Rectal misoprostol (a prostaglandin E1 analogue) might have exerted a uterotonic effect14 with simultaneous bronchodilation.

In this case, carboprost was given knowingly to an asthmatic mother, a joint clinical decision between anaesthetist and surgeon. We hope this heightens awareness of a serious and potentially fatal complication and that uninformed use of the drug will be minimised.