‘Hearing hooves, finding zebras’: the differential diagnosis of cardiac arrest precipitated by chest pain in the postpartum woman

Abstract

We describe the case of a 3-week postpartum 27-year-old woman who presented with chest pain to the emergency department shortly before developing cardiac arrest with refractory ventricular fibrillation. She was initially misdiagnosed and treated for presumed pulmonary embolism (PE) with thrombolysis. A total of 14 direct current cardioversion shocks were given and return of spontaneous circulation (ROSC) was achieved post thrombolysis. Subsequent CT pulmonary angiography excluded PE. A post-ROSC ECG demonstrated anterolateral ST elevation and she was transferred to the local cardiac unit for angiography. This revealed extensive dissection of the left anterior descending artery (LAD) with proximal occlusion. The diagnosis therefore was pregnancy-associated spontaneous artery dissection, a type of acute coronary syndrome. She received percutaneous intervention to her LAD with five drug-eluting stents. The patient survived and was discharged 5 days later. Her ventricular function is now grossly impaired, and had the correct diagnosis been arrived at sooner, this loss of function would have been less severe.

Background

Acute coronary syndrome (ACS) and pulmonary embolism (PE) are life-threatening differential diagnoses in pregnancy. Within the context of pregnancy and cardiac arrest, they demonstrate few discriminating diagnostic features. In this case, an awareness of the relevance of refractory ventricular fibrillation (VF) alongside a presenting complaint of chest pain radiating down the left arm may have allowed for earlier diagnosis. Keeping an open mind when encountering diagnostic uncertainty is vital, as this reduces cognitive bias and the subsequent likelihood of misinterpreting investigations. A delay in recognition of cardiac ischaemia carries a significantly higher mortality and morbidity profile both in the short and long term. A greater awareness of pregnancy-associated spontaneous artery dissection (PASCAD) as a differential diagnosis in this set of circumstances may allow for earlier recognition in the future.

Case presentation

A 27-year-old woman developed central crushing chest pain radiating to the left upper arm. She was 3 weeks post partum, gravity 4 parity 3+1. She had delivered as a normal vaginal delivery via water birth with no complications and an estimated blood loss of 500 mL. The baby’s birth weight was within the normal range at 7.2 lb. She had no medical, drug or family history of note. She had experienced no health problems across the pregnancy, was a non-smoker and displayed a normal Body Mass Index (BMI).

An hour after symptom onset, she collapsed while standing in the emergency department reception prior to triage. She was found to be in cardiac arrest with VF as the initial rhythm. Advanced life support resuscitation was commenced immediately, but despite this, she remained in extended VF arrest. Fifteen minutes into the arrest, she was discussed with the local cardiac centre and was accepted for transfer for angiography despite diagnostic uncertainty. Transfer was not taken forward at this time.

A bedside echo was performed which was felt to show dilatation of the right ventricle; no documented assessment was made of the left ventricle. A provisional diagnosis of PE was made and the patient was thrombolysed with Alteplase 50 mg intravenously 24 min into the arrest. Across the resuscitation, she received 14 direct current cardioversion (DCCV) shocks for refractory VF before sustaining an extended run of pulseless electrical activity (PEA) prior to the achievement of return of spontaneous circulation (ROSC) which occurred at 59 min. ECG immediately after ROSC was felt to represent right bundle branch block (figure 1). This in fact shows widespread ST elevation with reciprocal change. Subsequent CT pulmonary angiography was negative for PE. A repeat ECG was recognised to demonstrate anterolateral ST elevation (figure 2). The patient was discussed again with the local cardiac centre and accepted for immediate transfer for angiographic assessment. She arrived to the cardiac centre around 2 hours after initial arrest.

Figure 1.

ECG after return of spontaneous circulation, widespread ST elevation with reciprocal changes, interpreted initially as right bundle branch block.

Figure 2.

ECG after return of spontaneous circulation, widespread ST elevation.

Treatment

During angiography, the patient sustained three additional VF arrests requiring three further DCCV shocks. Angiogram revealed ostial dissection of the left anterior descending artery (LAD) with proximal occlusion (figure 3). Intravascular optical coherence tomography was used to further assess the false lumen and plan subsequent angioplasty (figure 4). All other coronaries were intact. There was widespread intramural haematoma throughout the LAD. An initial attempt was made to balloon the vessel; however, flow could not be maintained after balloon removal. The vessel was therefore stented with a total of five drug-eluting stents. This achieved restoration of flow through the LAD with no further compromise (figure 5).

Figure 3.

Left: initial angiogram, right anterior oblique caudal view. Right: initial angiogram with projected anatomy. LAD, left anterior descending artery; LCx, left circumflex artery; LMS, left main stem.

Figure 4.

Optical coherence tomography. White arrows demonstrate the false lumen created from the artery’s dissection.

Figure 5.

Final angiogram (right anterior oblique caudal view) after revascularisation.

Outcome and follow-up

The patient was monitored on the intensive treatment unit and was extubated after 24 hours. She was discharged after a further 4 days. During this time, an echocardiogram was performed which demonstrated severe left ventricular impairment with an ejection fraction of 30%. The apex was akinetic with hypokinesia anterolaterally. This may have prognostic implications for this patient in the future. She did not experience any further episodes of chest pain and was neurologically intact. Medications on discharge included aspirin 75 mg, clopidogrel 75 mg, bisoprolol 2.5 mg and eplenerone 25 mg.

Repeat angiogram 3 weeks later revealed propagation of the dissection into the left main stem which was sealed with a further stent. The patient was asymptomatic over this time and has been well since.

Discussion

Spontaneous coronary artery dissection (SCAD) is a non-atheromatous, non-traumatic cause of ACS. A rupture within the layers of the vessel wall creates a potential space between them, known as a false lumen. Into this space, blood can flow and form intramural haematoma. This causes obstruction of the true vessel lumen. PASCAD is the most common cause of myocardial infarction in pregnancy. It is defined as SCAD which occurs in the peripartum, antepartum and postpartum periods.¹ It carries a significant risk of mortality and morbidity occurring in previously healthy arteries where no collateral circulation exists; occlusion therefore can be catastrophic.²

A systematic review of 120 cases demonstrated a mean age of 34, with 72.5% of cases occurring post partum.³ Chest pain is the presenting complaint in 94% of cases. Sixty per cent of dissections affect one coronary artery, with the LAD most affected in 72% of cases. Moreover, 75.5% of cases present with ST elevation.⁴ Pathophysiology is not well understood, but associations are made between the hormonal changes of pregnancy and alterations in collagen synthesis. It is theorised that this leads to a thinning of the vessel wall predisposing to dissection.⁵ Traditional cardiac risk factors increase the risk of PASCAD, as does pre-eclampsia, the incidence of which is significantly higher in the PASCAD population. Multiparity is similarly associated with increased risk.³

Conservative management approaches should be taken initially if possible. This is based on data that dissection propagates with percutaneous intervention (PCI) while usually healing on its own with time.⁵ However, in the presence of haemodynamic instability, more aggressive management should be undertaken. Thrombolytic therapy has been reported to cause dissection extension and arterial rupture leading to cardiac tamponade; PCI is therefore favoured.⁵ In this case, PCI should have served as the primary intervention and thrombolysis avoided. Further pregnancy is not recommended with an estimated 20%–40% chance of recurrence.⁵

The differential diagnosis of chest pain in the pregnant woman is broad, but clinical presentation is similar to that of the general population. Life-threatening causes include peripartum cardiomyopathy, aortic root dissection, acute pericarditis, acute pancreatitis, pneumonia, pneumothorax, ACS and PE.⁶ The 2016 MBRRACE-UK report describes a reluctance to diagnose cardiac disease in pregnant women. This is despite cardiac disease being both the leading indirect and overall cause of maternal mortality, carrying a maternal mortality ratio of 2.18 per 100 000. ACS specifically carries an incidence of 1.7 per 100 000. In comparison, PE carries a maternal mortality ratio of 1.08 per 100 000 and an incidence of 1 per 1000.⁷ Pregnancy is a hypercoagulable state which persists for 8–12 weeks after birth. The postpartum period presents the highest risk for PE. Factors that contribute include a complex delivery, reduced mobility and increased BMI.

A review of 1246 general population cardiac arrests between 1991 and 1999 showed PE to be the leading cause of cardiac arrest in 60% (n=747) of cases. The presenting rhythm was PEA in 63%, asystole in 32% and VF in 5% of cases.⁸ No reported cases of PE presenting with refractory VF were found on review of the literature. Maternal cardiac arrest carries an incidence of 1 in 36 000. A review of maternal cardiac arrests between 2011 and 2014 described 66 women; of these, data regarding suspected cause were available in 59 cases. PE was suspected as the cause of arrest in eight cases, whereas a cardiac cause was suspected in six cases.⁹

A study looking into the sensitivity and specificity of right ventricle (RV) dilatation in cases of cardiac arrest found expected dilatation with PE but also with other underlying causes such as primary arrhythmia. The usefulness of this finding with regards to diagnostic accuracy is dependent on the skill and experience of the scanner and clinical context.¹⁰ Presence of RV dilatation in a context not in keeping with PE should not be taken as a firm diagnostic indicator.

Learning points.

• There are no reported cases associating pulmonary embolism (PE) with refractory ventricular fibrillation (VF).

• Refractory VF is strongly associated with cardiac ischaemia.

• Central crushing chest pain radiating down the left arm is atypical for PE.

• Alone, the assessment of the right ventricle by bedside echo in cardiac arrest lacks specificity and sensitivity for concrete diagnosis.