Background

Cocaine is a powerful and addictive stimulant which acts on synaptic transporters by inhibiting the reuptake of monoamines (norepinephrine, serotonin and dopamine), thus prolonging and potentiating their actions. These neurotransmitters are involved in the reward pathway of the brain and cocaine uses this to elicit a euphoric state in users.

The first documented account of cocaine-related subarachnoid haemorrhage (SAH) and intracerebral haemorrhage was published in 1999 as a case series.1 However, previous literature has illustrated that the vasculitic changes related to the consumption of cocaine can be attributed to the use of levamisole. This substance is a contaminant found in 70% of street cocaine samples from the USA and 50% of samples tested in the Netherlands and the UK.2 3 Levamisole-induced vasculitis has two classical pathological findings. The most important being leukocytoclastic vasculitis of the small vessels involving fibrinoid necrosis of the vessel wall and angiocentric inflammation.4 Although the effects of cocaine and levamisole on vasculature have been well documented, only one case report related to cocaine-induced SAH has been published with the evidence presented from a postmortem examination.5 The lack of data on these phenomena may be explained partly to the rarity of such event. Having said this, recent evidence has shown that cocaine-induced vasculitis is more common than previously known. Figure 1 demonstrates the main mechanisms in which cocaine abuse can influence the development of stroke.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

A graphical representation of the main mechanisms by which cocaine abuse induces haemorrhagic and ischaemic stroke.

Due to the illegal nature of cocaine, accurate prevalence rates are difficult to determine. A retrospective study conducted by Martin-Schild et al from the years 2004–2007 analysed 3241 patients admitted to a stroke service. Of those, 4.1% tested positive for cocaine metabolites on a urine drug screen (UDS).6 The prevalence of intracranial haemorrhage (ICH) in cocaine-positive patients was 34% compared with 3.24% in cocaine-negative patients.6 A modified Rankin Scale Score was used to measure the functional outcomes of patients at the time of discharge. The results demonstrated a worse functional outcome for cocaine-positive patients who suffered from an ICH (p<0.0001).6 Additionally, these patients had a threefold increase in mortality during their acute hospitalisation (p=0.017).6 Importantly, cocaine use remained a significant independent predictor of poor outcome after accounting for variables which differed significantly on admission.6

Case presentation

A 45-year-old man with a medical history of drug misuse collapsed in his home and was found 12 hours later by his father around 10:00. On admission, his Glasgow Coma Score (GCS) was 9 out of 15, illustrating a reduced level of consciousness. Despite this, he was still responsive to verbal ques. On examination, he had a respiratory rate of 56 breaths/min, a temperature of 39.2°C and blood glucose of 9.8 mmol/L. He was hypertensive and had equally reactive pupils, scoring 5+ bilaterally. At his current level of consciousness on admission, the patient was able to tolerate an oropharyngeal airway. During the patients stay, he presented with a wet sounding cough and possible crackles on his left lung, given this respiratory finding a chest X-ray (CXR) was conducted which presented as normal. When assessing power, he presented with bilateral weakness more prominent on the right hand side scoring 3 out of 5 on the medical research council muscle power scale and slightly higher with 4 out of 5 on the left hand side.

Prior to admission, the patient was known to the drug and alcohol liaison team (DALT) with his history showing long-term heroin use and polysubstance misuse. He denied cannabis, alcohol and cocaine use once responsive, however, his UDS was positive for a cocktail of drugs including cocaine, benzoylecgonine (cocaine metabolite), methadone, heroin, norbuprenorphine and benzodiazepine. He lives with his father and has no childcare responsibilities.

This patient has a history of depression, deep vein thrombosis and was previously tested positive for hepatitis B and C.

Investigations

An initial CT of the head was carried out on 13 August 2020. This illustrated extensive bleeding in the patients’ brain (figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

A CT head of the patient showing extensive intracranial, intraventricular and intraparenchymal bleeds with basal ganglia subarachnoid haemorrhages.

The patient sustained intracranial, intraventricular and intraparenchymal bleeds, with bilateral basal ganglia SAHs. These findings prompted further investigations into the origin of the patients intracranial bleeding. As a result, an MRI of the head and MR venogram was conducted on 18 August 2020 to further characterise origin of intracranial bleeding. This did not demonstrate any acute findings since the CT head therefore it was advised to conduct an CT angiogram, in order to rule out an underlying vascular malformation or arterial aneurysm, associated with recreational drug use such as cocaine. A CT angiogram of the brain was performed on 22 August 2020. A multifocal arterial narrowing was described; this was reported to represent a drug-induced Central Nervous System (CNS) vasculitis in the context of this patients history and radiological findings of frontal haemorrhage and SAH (figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

A CT angiogram of the patient illustrating multifocal narrowing.

The deeper the stroke team delved into the cause of this patients signs and symptoms, the more apparent it became that this patients’ stroke presentation was more complex and unique than initially thought. Following the CT angiogram report indicating a CNS vasculitis, a partial vasculitis screen was conducted. Blood tests included a lupus anticoagulant and anticardiolipin antibodies, in order to rule out a coagulopathy. Both tests returned a negative result. In hindsight, a further vasculitis screen including C3/C4, Anti-Neutrophilic Cytoplasmic Autoantibody (ANCA), Anti-Nuclear Antibody (ANA) would have potential been beneficial in further characterising a vasculitis, as described by Berlit.7 However, the medical team did not test for further antibodies as it was felt the pathology was secondary to cocaine, in the context of the patients’ signs and symptoms and the result of the CT angiography. These circumstances, the clinical picture and the method of investigation conducted this far were deemed to be sufficient in diagnosing CNS vasculitis.

It must be acknowledged that CNS vasculitis is a rare phenomenon that is not fully understood and currently digital subtraction angiography is the radiological gold standard for diagnosis of CNS vasculitis.8 In reality, this is a highly invasive procedure and advances in the quality of non-invasive neuroimaging may obviate the need for invasive neuroimaging in the diagnosis of CNS vasculitis in the near future.9 As a consequence, the use of CT angiography and MR venography was demonstrated to have correlated with the clinical picture and blood results to constitute a diagnosis of CNS vasculitis. Additionally, the invasive nature of digital subtraction angiography carries a small, but significant risk of stroke and neurological deficit at 1% increased risk when compared with 0% risk from CT or MR angiography.9 In hindsight, there was an indication to use digital subtraction angiography, given it is the gold standard investigation, however, it was deemed little was to be gained from its use compared with what was already available through CT and MR. Moreover, the frontal haemorrhage and SAH, in addition to maintaining the patients’ physical status, were seen as more pertinent issues to contend with and the potential risk of digital subtraction angiography was deemed to outweigh the need for this investigation.

An important bedside investigation that may have aided the diagnosis of vasculitis is a lumbar puncture in order to obtain a cerebrospinal fluid (CSF sample). Usually in CNS vasculitis, a mild lymphomonocytic pleocytosis or protein elevation is apparent, as a reflection of the inflammatory changes occurring in the central nervous system.7 10 In this case, a lumbar puncture was not conducted as the patients CT head illustrated that the patients’ intracranial bleeding had caused a mass effect, with compression of the frontal horns of both lateral ventricles. Given the patients CT head findings and focal neurological deficit, it was decided that a lumbar puncture would not be suitable initially.10 Furthermore, the patients’ diagnosis was felt to be made from the culmination of the clinical history, context of cocaine use and the CT angiography. Therefore, it was felt that the risk of conducting a lumbar puncture was greater than the benefit of obtaining a CSF sample.

A special consideration was given to whether septic emboli may have been the source. This consideration was mainly as a result of the patients’ background of intravenous heroine injection. Intravenous administration of drugs is known to frequently cause serious complications, one of which is septic embolus.11

The possibility of septic emboli and infective endocarditis (IE) was assessed using a transthoracic echocardiocram (TTE). This was performed on 14 August 2020. This investigation demonstrated there to be tricuspid valve thickening/calcification with sclerotic appearance and at least mild aortic regurgitation. There were no obvious large or new vegetations identified on the echo, however, it was not possible to fully exlude prior endocarditis, as IE cannot be fully excluded using this imaging modality. Therefore, further investigation was required to rule out a diagnosis of IE. Seven different blood cultures were taken from the patient, from these only one was reported as contaminated. Therefore, prior to 19 August 2020, there was no clinical/biochemical/haematological evidence to suggest sepsis or IE.

Given the patients prior history of illicit substance misuse, a urine toxicology screening test was performed on 14 August 2020. This provided a positive result for a cocktail of drugs, including: heroin, methadone, EDDP (2-ethylidene-1,5-dimethyl-3,3-diphenylpyrrolidine).

The patient’s ICHs culminated in the patient having residual motor dysfunction. This was observed in the Initial Postural Assessment Scale (PASS) carried out on 18 August 2020, which came back as a score of 0 out of 36. Throughout the patients’ inpatient stay, this significantly improved and when the patient repeated the PASS on 25 August 2020, they managed to score 14 out of 36. The majority of improvement was illustrated by the patient being able to stand in a supine position and moving from a seated to standing position, which he could not on admission.

Following the patients’ respiratory findings, a CXR was conducted on 25 August 2020. This came back clear, with no obvious consolidation.

Differential diagnosis

Elevated creatine kinase levels were indicative of rhabdomyolysis; a likely consequence of the 12-hour long lie the patient experienced, secondary to his ICH, but was also a differential as a result of his drug use.12

Aneurysms are known to increase the risk of haemorrhage; a study suggests that aneurysms are the cause of haemorrhage in 85% of patients.13 Hence, an intracranial angiogram was requested. Interestingly, the angiogram reported multifocal arterial narrowing’s, and this was attributed to drug use with no evidence of an aneurysm. Simultaneously, investigations were carried out to rule out IE. This involved a TTE which showed no obvious vegetation and so seven blood culture samples were immediately sent following admission over a 12-hour period. Once cultured, a lack of vegetation was demonstrated with six negative results and one contaminated sample. This in turn ruled out a diagnosis of SAH as a result of septic emboli.

During critical care admission, a differential diagnosis of cerebral haemorrhage as a result of hypertension was made. Throughout his stay, this remained a working diagnosis, a CT brain was requested as a result. The imaging showed extensive acute intracerebral haemorrhage and SAH, and urgent neurosurgical opinion was recommended. Further MRI of the brain was requested which did not demonstrate new acute findings since the primary CT brain. On 02 September 2020, a final CT angiogram was taken to ensure no further progression. There was no evidence as such, and the appearance was stable, however, arterial beading was noted on this imaging.

Treatment

Due to the suspicion of septic emboli, the patient was initiated on intravenous piperacillin and tazobactam. This was prescribed on admission and was discontinued after 2 days, ‘as per the advice provided by the microbiology department’. Other than the intravenous antibiotics, the patient was also given the topical antibiotic flamazine, to be used on his burns. Flamazine is a sulphonamide that has been proven to be effective against a wide range of organisms including both gram positive and gram negative bacteria and a number of viruses and fungi.

During his stay in hospital, the patients’ early warning score reached 5 on 24 August 2020. Following a physical assessment of the patient, it was established that there was left lower lung crepitations on auscultation and an audibly moist sounding cough. The patient was commenced on levofloxacin and metronidazole as a result.

Given the patients long lie prior to admission, multiple intravenous medications were prescribed in order to manage dehydration. These involved administering saline, sodium chloride intravenous midline and plasmalyte. This treatment for dehydration could be seen throughout the patients care, as they were unable to swallow for majority of their hospital admission. On 14 August 2020, the primary medications prescribed for this dehydration were: 5% dextrose and plasmalyte intravenous. On 25 August 2020, this was changed to Hartmann’s intravenously, which is more physiological than normal saline.

Outcome and follow-up

At almost 4 months postadmission, a collateral history taken from the patient’s father revealed that his condition had improved significantly since being discharged. However, he has not returned to his baseline prior to admission and as a result has not made a full recovery. He is now independently mobile and can complete his personal hygiene requirements himself. When he was discharged from the hospital, the patient was continent for the majority, but at times he would be incontinent especially when having to rush to the toilet. This aspect of his recovery has improved with the patient currently being fully continent. Despite having mild residual dysarthria and minimal dysphasia, the patient is able to hold a conversation with others. He still requires aid with aspect of his life which requires fine motor movement such as buttoning up shirts and cutting his food into pieces.