Objectives Early neurological deterioration prompting urgent brain imaging occurs in nearly 15% of patients with ischaemic stroke receiving intravenous tissue plasminogen activator (tPA). We aim to determine risk factors associated with symptomatic intracranial haemorrhage (sICH) in patients with ischaemic stroke undergoing emergent brain imaging for early neurological deterioration after receiving tPA.
Methods We abstracted data from our prospective stroke database and included all patients receiving tPA for ischaemic stroke between 1 March 2015 and 1 March 2017. We then identified patients with neurological deterioration who underwent urgent brain imaging prior to their per-protocol surveillance imaging and divided patients into two groups: those with and without sICH. We compared baseline demographics, clinical variables, in-hospital treatments and functional outcomes at 90 days between the two groups.
Results We identified 511 patients who received tPA, of whom 108 (21.1%) had an emergent brain CT. Of these patients, 17.5% (19/108) had sICH; 21.3% (23/108) of emergent scans occurred while tPA was infusing, though only 4.3% of these scans (1/23) revealed sICH. On multivariable analyses, the only predictor of sICH was a change in level of consciousness (OR 6.62, 95% CI 1.64 to 26.70, P=0.008).
Conclusion Change in level of consciousness is associated with sICH among patients undergoing emergent brain imaging after receiving tPA. In this group of patients, preparation of tPA reversal agents while awaiting brain imaging may reduce reversal times. Future studies are needed to study the cost-effectiveness of this approach.
- ischemic stroke
- thrombolytic therapy
- neurological deterioration
Statistics from Altmetric.com
Intravenous tissue plasminogen activator (tPA) improves functional outcome in patients with acute ischaemic stroke when given within 4.5 hours from symptoms onset.1 2 The most feared complication of this treatment, however, is symptomatic intracranial haemorrhage (sICH), which occurs in 3%–7% of patients and is associated with significant mortality and morbidity.3 4
In patients with ischaemic stroke who receive thrombolytic therapy, previous studies have shown that 5%–14% exhibit early neurological deterioration, defined as an increase in the National Institute of Health Stroke Scale (NIHSS) by four points or more within 24 hours of thrombolytic infusion.5 6 Intracranial haemorrhage (ICH) is ultimately identified in nearly 25% of cases with early neurological deterioration.5 6 When early neurological deterioration does occurs during tPA infusion, the infusion is typically held until repeat imaging excludes ICH, at which point the infusion is usually resumed.7 However, interruption of the tPA infusion may potentially be deleterious to patients whose early neurological deterioration is related to ischaemia. Therefore, it is essential to identify clinical predictors of sICH in patients receiving tPA who subsequently develop early neurological deterioration. In this study, we aim to investigate demographic factors, clinical risk factors and signs/symptoms that most strongly correlated with sICH diagnosis after early neurological deterioration.
Institutional review board approval was obtained to perform a single-centre retrospective cohort study using prospective data collected as part of our institutional stroke registry. This registry serves as a quality improvement database for ischaemic stroke and uses a REDCap (Vanderbilt University, Nashville, Tennessee) platform for data entry. Our study included all patients with ischaemic stroke admitted to our facility between 1 March 2015 and 1 March 2017. All patients were evaluated by a neurology resident and emergency department attending, and initial evaluations in the emergency department included clinical assessments, laboratory tests and imaging studies. The latter included both non-contrast CT of the brain along with a CT angiogram of the head and neck. Thrombolytic therapy was given, when clinically indicated, to all patients presenting within 4.5 hours from symptom onset who did not have clear evidence of haemorrhage on their initial CT. Other inclusion/exclusion criteria8 were discussed on a case-by-case basis given the recent changes in the tPA insert package and recent suggestions to revisit tPA exclusion criteria.9 We identified all patients who received tPA, then included only those with neurological deterioration who had an emergent repeat CT scan performed to exclude ICH prior to the usual 24-hour per-protocol surveillance CT.
The symptoms prompting urgent imaging were abstracted from medical records (including physician and nursing documentation) and were divided into three categories: any change in level of consciousness from baseline, any worsening of existing symptoms or new symptoms without any change in level of consciousness, or any new headache or nausea or vomiting without accompanied neurological symptoms.
The following covariates were abstracted:
Clinical and demographic variables: age, sex, history of comorbidities (including hypertension, diabetes, hyperlipidaemia, congestive heart failure, coronary artery disease, stroke and current smoking) and NIHSS scores at baseline and at 24 hours from admission.
Medication use prior to admission: any antiplatelet agents, warfarin or novel oral anticoagulant.
In-hospital treatment: tPA infusion that was transiently or permanently discontinued, mechanical thrombectomy.
Functional outcome: modified Rankin Scale (mRS) obtained at 90 days via a structured telephone call, with good outcome defined as mRS 0–1 at 90 days.
All initial and follow-up brain CT scans were reviewed by the treating vascular neurologist and neuroradiologist for the presence of acute haemorrhage. The outcome was sICH defined as any neurological deterioration with evidence of haemorrhage on the follow-up brain CT scan1 as determined by the treating vascular neurologist.
We divided patients into two groups: early neurological deterioration without ICH (non-ICH) and early neurological deterioration with ICH (ICH). We compared baseline demographics, clinical characteristics, medication use, in-hospital treatments and functional outcomes between the two groups. We used t-tests for continuous variables and Fisher’s test for categorical variables. We then performed multivariate regression analysis including all factors achieving statistical significance on univariate analyses to identify clinically relevant risk factors associated with sICH. Statistical analysis was performed using SPSS V.18.0 and P <0.05 was considered statistically significant.
Baseline characteristics of the study sample
We identified 511 patients who received intravenous tPA during the study period; the mean age in years was 70.5±16.1% and 50.1% were men. Among patients receiving tPA, 108 (21.1%) had an emergent brain CT performed to exclude haemorrhage, of whom 17.5% (19/108) were found to have sICH. The deterioration occurred within 1 hour after tPA bolus and during infusion in 21.2% (23/108) of patients, leading to interruption of the infusion in all cases. However, of these patients, only one (4.3%) was found to have sICH whereas the infusion was resumed in all others. Figure 1 shows a flow chart of the study sample. Furthermore, the mean time from tPA infusion to symptoms prompting brain imaging in sICH was 7.3±5.3 hours.
All patients received tPA at our institution that is the same institution where mechanical thrombectomy was performed; 43 patients with neurological deterioration underwent mechanical thrombectomy; 6/43 had neurological deterioration prior to thrombectomy while the rest had it during or after the thrombectomy. None of the patients who deteriorated prior to thrombectomy had sICH.
On univariate analyses, patients with sICH on emergent brain CT scan were older (80.2±15.8 years vs 70.8±15.4 years, P=0.026), had higher admission median NIHSS score (14 vs 7, P=0.0016) and higher 24-hour median NIHSS score (24 vs 5, P<0.001). They were more likely to have a change in level of consciousness (78.9% vs 21.3%, P<0.001) and less likely to have worsening or new symptoms without a change level of consciousness (5.3% vs 42.7%, P=0.002). sICH was independently associated with mortality (87.5% vs 22.4%, P<0.001) (table 1). The percentage of patients with good functional outcome (mRS 0–1 at 90 days) was similar between patients with early neurological deterioration related to ischaemia whose tPA infusion was interrupted versus not (35.3% vs 34.0%, P=1.00). The median time to neurological deterioration was longer in patients receiving mechanical thrombectomy versus those who did not (5 hours IQR 2–13 vs 1.5 hours IQR 1–4.5, P<0.001).
On multivariate analysis including age, admission NIHSS, change in level of consciousness and worsening or new symptoms (without a change in level of consciousness), the only predictor of sICH was a change in level of consciousness (OR 6.62, 95% CI 1.64 to 26.70, P=0.008). Other variables were not associated with sICH (table 2).
Our findings suggest that a change in level of consciousness is an independent risk factor associated with sICH in patients with early neurological deterioration after receiving tPA for acute ischaemic stroke, while other symptoms did not appear to have a significant association. Of the patients who developed sICH, 78% were noted to have decreased responsiveness that prompted repeat emergent imaging. This is in line with prior studies that showed an association between decreased responsiveness following tPA administration and death.10 Meanwhile, taking all patients presenting with acute neurological symptoms into consideration, a change in level of consciousness has been shown to be independently associated with ICH as opposed to ischaemic stroke.11
Based on the results of our study, it may therefore be worthwhile to consider empirically preparing tPA reversal agents in patients with a change in level of consciousness, in parallel with obtaining repeat emergent imaging, so that any potential delays may be reduced. On the other hand, patients in our cohort with worsening or new symptoms were more likely to have worsening ischaemia as opposed to sICH, while our results suggest that patients with headache or nausea/vomiting likely represent a group at lower risk of sICH.
Importantly, our study also demonstrated that among patients who have their tPA infusion stopped as a result of early neurological deterioration, only a small minority had evidence of sICH. In the rest of cases, their symptoms may have been related to worsening or ongoing ischaemia, and interruption of the tPA infusion may have caused harm. In our study, the percentage of patients with good functional outcome was similar between patients with early neurological deterioration related to ischaemia whose tPA infusion was held versus not. Our study, however, is underpowered to compare these outcomes, and larger studies are needed to demonstrate a difference. Until there is a more expedient way of ruling out sICH, however, the safest plan of action in patients with early neurological deterioration remains interrupting the tPA infusion until imaging is obtained to exclude ICH.
Our study has several limitations, including its retrospective nature and that it is a single-centre study, thus limiting its generalisability. In addition, sICH was determined by different providers using the NINDS definition, which has been shown to have only fair inter-rater reliability.12 This definition, however, is the one most commonly used among stroke centres in the USA. Another limitation is that, unfortunately, a full NIHSS score was not performed at the time of deterioration. It would be intuitive though that those patients with a change in level of consciousness will have a more dramatic increase in the NIHSS score as compared with those without level of consciousness. Our study does have several strengths, foremost being its relatively large sample size. It is also one of only a few studies investigating risk factors associated with sICH in this specific patient population.
A change in level of consciousness is associated with sICH among tPA-treated patients who develop early neurological deterioration. In this group of patients, treating physicians may consider preparation of tPA reversal agents while awaiting brain imaging, which may possibly reduce tPA reversal times and improve outcomes. Future studies are needed to confirm this association and to investigate the cost-effectiveness of this approach. Meanwhile, a small minority of patients who develop early neurology deterioration during tPA infusion are found to have sICH, though future studies are needed prior to any potential changes in clinical practice.
Contributors BJ: data collection, manuscript preparation. ADC: data analysis, manuscript revision. RAMT: manuscript revision. MH: manuscript revision. BMG: manuscript revision. SMC: manuscript revision. TMB: manuscript revision. MER: manuscript revision. BT: manuscript revision. LW: manuscript revision. AM: manuscript revision. MS: manuscript revision. TEM: manuscript revision. KNS: manuscript revision. AN: manuscript revision. KLF: manuscript revision, study design. MVJ: manuscript revision. PK: manuscript revision, study design. SY: data collection, drafting of manuscript, study design.
Competing interests None declared.
Patient consent Detail has been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.
Ethics approval The study obtained ethics approval by the Rhode Island Hospital Institution Board review.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.