388 Practical Neurology
Pract Neurol 2008; 8: 388–392
A difficult case of postpartum collapse David P Breen, Anna Williams, Peter Keston, Ian R Whittle, Peter A G Sandercock
D P Breen Senior House Officer in Neurology
A Williams Wellcome Clinical Scientist and Honorary Consultant Neurologist
P Keston Consultant Neuroradiologist
I R Whittle Forbes Professor of Surgical Neurology
P A G Sandercock Professor of Medical Neurology Department of Clinical Neurosciences, Western General Hospital, Edinburgh, UK Correspondence to: Dr D P Breen Department of Clinical Neurosciences, Western General Hospital, Crewe Road South, Edinburgh EH4 2XU, UK; [email protected]
29-year-old woman was brought to hospital on 24 December, having gone into labour with her third child at 34 weeks gestation. Her first two pregnancies had been uneventful and she had enjoyed an uncomplicated third pregnancy, continuing to work a farm with her husband. Past medical history included an uncorrected Ebstein’s anomaly (a congenital heart defect where the opening of the tricuspid valve is displaced towards the apex of the right ventricle due to abnormal valve leaflet formation) and spinal stabilisation for childhood scoliosis. At 6.21am, she delivered a healthy baby girl by normal vaginal delivery. At 8.00am, she walked to the shower. While there, she developed a ‘‘thumping headache’’ and ‘‘twitching’’ of her left hand. She had no further recollection of events, but was found collapsed in the shower at 8.40am. When the resuscitation team arrived, her Glasgow Coma Score (GCS) was 9 (E2V2M5). She was breathing spontaneously, haemodynamically stable, but too drowsy to maintain her own airway. She was localising to pain with her right arm, but appeared not to be moving her left side.
Plantar response was extensor on the left. Pupils were equal and reactive. She had a CT brain scan at 9.53am.
Question 1 What are the possible diagnoses?
COMMENT This woman has most likely had a stroke. There is probably an increased risk of stroke during pregnancy and for approximately six weeks after delivery; the risk in developed countries varies between 11 and 26 per 100 000 deliveries.1 Studies have repeatedly identified the immediate postpartum period as the greatest period of risk. Most of these are ischaemic strokes, in part perhaps because pregnancy and the puerperium are pro-thrombotic states. Some ischaemic strokes in pregnancy can be attributed to specific causes such as pre-eclampsia or eclampsia, arterial dissection, antiphospholipid syndrome, postpartum angiopathy, thrombotic thrombocytopenic
Breen, Williams, Keston, et al 389 purpura, amniotic fluid or air embolism, or peripartum cardiomyopathy. Our patient’s Ebstein’s anomaly (which is known to be associated with atrial septal defect) made paradoxical embolism secondary to deep vein thrombosis (DVT) more likely. Intracranial haemorrhage too can present in this way, often related to uncontrolled hypertension. Intracerebral haemorrhage, aneurysmal or perimesencephalic subarachnoid haemorrhage, and bleeding from an underlying arteriovenous malformation are all potential diagnoses in this case, and can only be excluded by brain imaging. If she had presented with ophthalmoplegia, a visual field defect or reduced visual acuity, we would have considered pituitary apoplexy as a cause. Intracranial venous thrombosis is more common in the postpartum period. When the cortical veins are involved, a focal neurological deficit can occur due to venous infarction. However, depressed conscious level is unusual unless there are multiple areas of infarction, thalamic involvement or brainstem compression. Given the history, it is possible that our patient had an unwitnessed secondary-generalised seizure, resulting in post-ictal drowsiness and Todd’s paresis. There are other causes of postpartum collapse, but these are less likely given the patient’s persistent focal neurological deficit. These include hypoglycaemia, postpartum haemorrhage, pulmonary embolism, myocardial infarction, cardiac arrhythmia, Addisonian crisis, thyroid disease or drug overdose. Her admission CT brain scan showed thrombus at the right internal carotid artery (ICA) bifurcation extending into the middle cerebral artery (MCA), but no convincing evidence of ischaemic change (none would be expected so quickly following onset) (fig 1). Acute ischaemic stroke was diagnosed, she was intubated and ventilated, and transferred to the regional neurosciences unit for further management. She arrived there just over three hours after stroke onset.
Question 2 Imagine you are the admitting neurologist. What action would you take?
Figure 1 Non-enhanced admission CT brain showing hyperdense internal carotid artery bifurcation and proximal middle cerebral artery.
COMMENT There is no ‘‘standard practice’’ for this scenario; it is an evidence-free zone. There is no licence in the UK to give intravenous thrombolysis more than three hours after stroke onset. Pregnancy and the first four weeks postpartum are also generally regarded as exclusion criteria for thrombolysis following acute ischaemic stroke. This is mainly due to the increased risk of intrauterine haemorrhage—the same reason that anticoagulation is used with caution. There are, however, case studies of successful intra-arterial thrombolysis in the immediate postpartum period in patients with acute MCA occlusion.2 Our patient had evidence of intra-arterial thrombus on brain imaging. Since she arrived more than three hours after stroke onset, we felt she was a better candidate for endovascular clot retrieval than for intravenous thrombolysis, which would also avoid the systemic bleeding risk. After obtaining arterial access, a large bore (8 Fr) balloon guide catheter was passed to the right ICA. Angiography confirmed complete occlusion of the distal ICA (fig 2), with a dominant posterior communicating artery providing total supply to the posterior cerebral artery (PCA). Initial attempts to aspirate clot from the distal cervical ICA were unsuccessful. The radiologist administered 6 mg of intra-arterial abciximab (the standard systemic bolus dose was calculated based on the patient’s weight) via a microcatheter directly into the thrombus. He then used a MERCI clot retrieval device (Concentric Medical Inc, Mountain View, USA) www.practical-neurology.com
390 Practical Neurology
Figure 2 Catheter angiogram showing occluded right internal carotid artery.
to make a total of three passes beyond the clot to the M2 segment of the MCA, but retrieved only very small fragments of red clot. He repositioned the guide catheter in the common carotid artery (CCA), and angiography at this level revealed further clot at the CCA bifurcation. Although it was possible to aspirate a large volume of this through the guide catheter, the ICA remained occluded. As it was now six hours after onset of symptoms, we felt that there was little point in further attempts to retrieve the thrombus and the procedure was terminated.
Question 3 What would you do next?
This woman was at high risk of developing neurological complications (due to raised intracranial pressure (ICP)) and general medical complications (such as aspiration pneumonia) and so needed close monitoring. She was transferred to the intensive care unit, where the neurosurgeons inserted an ICP bolt. She remained ventilated (aiming for a PaCO2 of 4 kPa) and sedated (to prevent coughing or agitation which might increase the ICP). Central and arterial lines were inserted, and inotropic support was required to maintain blood pressure. Her GCS was now 7 (E1V1(T)M5). Although she was able to localise with her right hand, she had a dense left hemiparesis. The ICP remained satisfactory overnight and a repeat CT brain scan on 25 December showed an established infarct affecting the right MCA and PCA territories. Unfortunately, her right pupil became fixed and dilated later that day, and she stopped moving her right side. The ICP was only mildly raised (17 mm Hg), perhaps because the midbrain compression was relatively focal, and movement of the right arm and leg returned after treatment with mannitol. The next day, her pupil was again dilated and ICP rose to 39 mmHg. A further CT brain scan showed increased swelling around the infarct, greater midline displacement, contralateral hydrocephalus, and midbrain distortion secondary to uncal herniation (fig 3).
Question 4 Figure 3 Repeat CT showing middle cerebral artery/posterior cerebral artery infarction with mass effect and brain shift to the left.
Is there any further therapy to consider?
COMMENT This type of massive cerebral infarction with surrounding oedema has been termed ‘‘malignant MCA syndrome’’. Case fatality is extremely high despite maximal intensive care treatment, with neurological complications related to raised ICP generally appearing 2–5 days following stroke onset. Decompressive hemicraniectomy is a surgical procedure in which a generous bone flap is removed on the same side as the infarct, and the dura is opened to relieve pressure and 10.1136/jnnp.2008.164137
Breen, Williams, Keston, et al 391 prevent transtentorial and uncal herniation. The bone flap is usually implanted in a subcutaneous pocket within the anterior abdominal wall until it can be replaced. There is now good evidence from three European randomised-controlled trials (DESTINY, DECIMAL, HAMLET) that decompressive hemicraniectomy reduces mortality and increases the number of patients with a favourable outcome in malignant MCA syndrome. In a pooled analysis of the three trials, more patients in the decompressive surgery group compared with the control group survived (78% vs 29%; absolute risk reduction 49%).3 Importantly, hemicraniectomy did not increase the risk of patients being left severely disabled or dependant. The outcome measure used was the modified Rankin score (mRS) at year 1. The absolute increase in the proportion of patients rated as mRS 2 was 12%, mRS 3 was 10%, and mRS 4 was 29%. In other words, for every 10 hemicraniectomies performed for malignant MCA syndrome, approximately five will escape death, one will have mild disability at year 1, one will have moderate disability, and three will have moderate-severe disability. The studies included patients aged between 18 and 60, and the timing of surgery was within 48 hours of stroke onset. The neurosurgeons performed a right hemispheric decompressive craniotomy on the evening of 26 December. Particular care was taken to remove bone right down to the floor of the right temporal fossa, to permit full decompression. She had no more problems with raised ICP, and repeat CT brain on 31 December showed good decompression and less midline shift (fig 4). She gradually improved and was transferred to the general neurology ward.
Question 5 What further investigations are required?
COMMENT Her thrombophilia screen was normal on admission. We noted that she had a loud pansystolic murmur and an abnormal electrocardiogram (right bundle branch block and
Figure 4 Post-surgery CT showing gliotic infarction with persistent brain herniation through the craniectomy.
Q waves in the anterior chest leads). A possible explanation would be a right to left shunt secondary to her underlying congenital heart defect, with the potential for paradoxical embolism. While in the ICU, she had an echocardiogram that confirmed atrialisation of the right ventricle consistent with Ebstein’s anomaly, moderate tricuspid regurgitation, and a dilated right atrium. There was also flow across the inter-atrial septum secondary to an atrial septal defect. Doppler ultrasound, however, showed no evidence of thrombus in the leg veins, pelvic veins or inferior vena cava. It is unknown whether a thrombus here had passed, or whether a thrombus had formed within the abnormal heart itself. The cardiologists planned an outpatient review to decide whether to close the defect, and whether to recommend long-term anticoagulation or antiplatelet treatment. After returning to the neurology ward, she gradually improved and started to regain some power in her left arm and leg. She developed unpleasant neuropathic pain down her left side, but this improved on gabapentin. Her husband and newborn baby visited regularly and she remained positive and motivated. Gradually her swallow returned and she was able to speak again once we removed the tracheostomy. Repeat CT brain scan on 7 January showed significantly less swelling around the craniotomy site and less parenchymal herniation. After 31 days in hospital, she was transferred to a rehabilitation ward where she continued to make an impressive recovery. Despite her www.practical-neurology.com
392 Practical Neurology
Figure 5 Our patient with her husband and children on the farm following discharge.
Learning points l
Right to left cardiac shunts can appear for the first time during labour (probably due to repeated Valsalva manoeuvres) and put patients at risk of paradoxical embolism. Speed is of the essence after stroke, but even clot retrieval and intraarterial thrombolysis can be ineffective. Decompressive hemicraniectomy in selected patients can be extremely helpful in limiting pressure complications. A good outcome is possible in young patients with stroke.
good quality of life. Bone flap replacement was delayed because of scalp colonisation with methicillin-resistant staphylococcus aureus, but this was also successfully completed in May 2008. She is back living at home with her family (fig 5).
REFERENCES 1. 2.
residual spastic hemiparesis (predominantly facio-brachial), she is now able to walk independently; she is cognitively intact; and her 3rd nerve palsy has resolved. She manages many activities of daily living despite a persistent left-sided hemianopia and has a
Davie CA, O’Brien P. Stroke and pregnancy. J Neurol Neurosurg Psychiatry 2008;79:240–5. Mendez JC, Masjuan J, Garcia N, et al. Successful intra-arterial thrombolysis for acute ischaemic stroke in the immediate postpartum period. Cardiovasc Intervent Radiol 2008;31:193–5. Vahedi K, Hofmeijer J, Juettler E, et al. Early decompressive surgery in malignant infarction of the middle cerebral artery: a pooled analysis of three randomised controlled trials. Lancet Neurol 2007;6:215–22.