Theophylline as an add-on to thrombolytic therapy in acute ischaemic stroke (TEA-Stroke): A randomized, double-blinded, placebo-controlled, two-centre phase II study

Abstract

Introduction

Early reperfusion of brain tissue at risk of injury (penumbra salvage) is crucial in treating acute ischaemic stroke. Neuroprotective agents may extend the time window for the reperfusion. The vasoactive agent theophylline redistributes the perfusion to ischaemic brain tissue and thus reduces brain damage, brain tissue oedema and mortality in animal stroke models. Furthermore, treatment with theophylline has been shown to result in considerable and rapid clinical improvement, albeit only temporary, in some stroke patients.

We hypothesize that treatment with theophylline will improve the collateral supply in acute ischaemic brain tissue and thus facilitate reperfusion despite proximal vessel occlusion. The primary study objective is to evaluate whether theophylline is safe and efficient in acute ischaemic stroke patients as an add-on to thrombolytic therapy.

Methods

The TEA-Stroke Trial is a two-centre, proof of concept phase II clinical study with a randomized, double-blinded, placebo-controlled design. One hundred and twenty patients with acute ischaemic stroke and significant perfusion–diffusion mismatch, as determined by magnetic resonance imaging, are randomized 1:1 to either theophylline or placebo as an add-on to standard thrombolytic therapy.

Study outcome

The dual primary outcome measures include penumbra salvage (penumbral tissue not developing into infarcted tissue) and clinical improvement at the 24-h follow-up.

Discussion

Results from studies of theophylline in stroke animal models, clinical case series and randomized clinical trials are controversial. A Cochrane analysis from 2004 concluded that there was not enough evidence to assess whether theophylline is safe and improves outcomes in patients with acute ischaemic stroke. The TEA-Stroke Trial will clarify whether theophylline as an add-on to standard thrombolytic therapy improves penumbra salvage with a reduced risk of reperfusion damage, reduced final infarct size, and improved clinical outcome.

Introduction

Early reperfusion of brain tissue at risk for injury (penumbra salvage) is crucial in treating patients with acute ischaemic stroke as persistent vessel occlusion leads to an enlargement of the ischaemic core over time¹ and is associated with a poor clinical outcome.² The current standard therapy consists of thrombolytic therapy with intravenous recombinant tissue plasminogen activator (rtPA)^3,4 and endovascular treatment in cases of large vessel occlusion.⁵ A typical stroke patient will lose 1.9 million neurons each minute that the stroke is left untreated.⁶ The recanalization effect of rtPA is delayed by approximately 23 min⁷ on average, and endovascular therapy is delayed due to the preparation required for the procedure. This problem might be solved by the immediate improvement of the collateral supply to enable immediate reperfusion while waiting for recanalization.

Theophylline is licensed in Denmark for the treatment of asthma and bronchial spasm due to its relaxing effect on bronchial smooth muscles. Furthermore, theophylline is also a cerebral vasoactive agent that reduces regional cerebral blood flow in areas of healthy brain tissue and increases regional cerebral blood flow in ischaemic tissue.^8–10 Theophylline is known as a phosphodiesterase inhibitor and a competitive adenosine receptor antagonist, but the effects of theophylline on cerebral blood flow are not fully understood. Preconditioning with theophylline in animal stroke models reduced brain damage, brain tissue oedema and mortality.^11–13

The treatment of stroke patients with theophylline has been described in several case reports.^14–18 Briefly, these case series studies showed considerable and rapid clinical improvement, albeit only temporary, in some patients. Thus far, two randomized, double-blinded, placebo-controlled clinical trials have failed to demonstrate any benefit of theophylline treatment in stroke patients. In the study by Geismar et al.¹⁹ with 79 patients, early improvement was observed significantly more frequently in patients treated with theophylline (38% versus 15%), but there was no significant difference at follow-up (three weeks). The study by Britton et al.²⁰ included 46 patients but did not detect significant differences at follow-up. Nonetheless, both studies had important limitations as follows:

1. The diagnosis of stroke was determined purely on clinical grounds and was only supported by the results of a lumbar puncture. Brain imaging was not available to exclude intracerebral haemorrhage or stroke mimics;

2. It is well known that a substantial expansion of the infarction occurs within the first hours after stroke onset, and favourable clinical outcomes correlate with adequate early treatment. Accordingly, beneficial clinical effects of theophylline have been observed when theopylline was administered soon after the stroke onset.¹⁹ However, in both of the aforementioned trials, the intervention was administered at later time points. Geismar et al. administered treatment 20 h after stroke onset on average with ⅓ of the patients treated more than 24 h after the stroke onset. Britton et al. treated patients 40 h after the onset of the stroke on average with a range of 18–114 h. Thus, the majority of patients in these two trials were treated far beyond the crucial ‘golden hours’ of acute ischaemic stroke. Little or no penumbra remained, and the brain damage was presumably irreversible at the time of the intervention;

3. Geismar et al. observed a temporary benefit from theophylline treatment in patients with preferably minor or moderate stroke symptoms. However, the mortality rate of 22–23% within three weeks was high in both studies, suggesting the presence of patient selection bias towards severe stroke with predominantly irreversible brain damage at the time of intervention; and

4. No thrombolytic agent or other revascularization therapy was available at the time. The possible penumbra salvage and clinical improvement due to theophylline would presumably only be temporary as theophylline has no known thrombolytic effects and vessel occlusion would continue.

In summary, both randomized, double-blinded, placebo-controlled clinical trials were unlikely to demonstrate any benefit of theophylline treatment due to inadequate patient selection, delayed intervention, and lack of revascularization therapy.

Accordingly, a Cochrane analysis concluded that there was not enough evidence to assess whether theophylline is safe and improves outcomes in patients with acute ischaemic stroke.²¹ None of the studies reported severe side effects, such as uncontrolled blood pressure, arrhythmia or seizure, a relevant safety concern considering the positive inotropic, chronotropic and central stimulating effects of theophylline. There was no indication for side effects in the acute stage of stroke in these clinical trials.

The bolus dose of theophylline that was used in most of the case series studies of stroke patients was less than the bolus dose of 5 mg/kg body weight that is usually used in patients with acute bronchial spasm. A fixed dose of 230 mg (Geismar et al.) and 200 mg (Britton et al.) was used in the two randomized clinical trials mentioned above. To our knowledge, no interaction between theophylline and rtPa has been described.

Based on this information, we designed the TEA-Stroke Trial, which is a two-centre, double-blinded, placebo-controlled phase II proof of concept study. A dose of 220 mg Teofylamin (200 mg theophyllamine) as a short infusion over 15 min was chosen, similar to the dose used in previous stroke studies. We propose that theophylline administration will improve collateral supply in the distal territory of ischaemia despite proximal vessel occlusion. The immediate effect of theophylline after infusion may perfuse the penumbra while patients are waiting for recanalization. Optimized capillary flow with improved oxygen extraction in the area of ischaemia may explain the effects of theophylline. Reperfusion of the penumbra despite the occlusion of the culprit vessel may protect the neuronal and vascular integrity with less infarct expansion and a reduced risk of reperfusion damage.

Methods

Study objective

The aim of this proof of concept study is to evaluate whether theophylline is safe and efficient as an add-on to thrombolytic therapy in patients with acute ischaemic stroke and penumbral mismatch.

Study design

The TEA-Stroke Trial is a two-centre, proof of concept, phase II clinical study with a randomized, double-blinded, placebo-controlled design. The study design is outlined in Figure 1.

Figure 1.

Patient flow and assessment.

ECG: electrocardiography; MRI: magnetic resonance imaging; mRs: modified Rankin Scale; NIHSS: National Institute of Health Stroke Scale.

The Good Clinical Practice-Unit at Aalborg University and Aarhus University monitor the trial carried out in accordance with the Declaration of Helsinki. The first patient was recruited in October 2014 and the study is currently actively recruiting.

Patient population

The main entry criterion is eligibility for thrombolytic therapy within 4.5 h of symptom onset (according to the current Danish inclusion/exclusion criteria for standard intravenous rtPA thrombolytic therapy) and evidence of perfusion/diffusion-weighted imaging (PWI/DWI) mismatch on magnetic resonance imaging (MRI) in patients with moderate to severe acute ischaemic stroke. Endovascular treatment (revascularization therapy with mechanical devices) in addition to standard intravenous rtPA thrombolytic therapy (bridging therapy) is not an exclusion criterion. The detailed inclusion and exclusion criteria are shown in Table 1.

Table 1.

TEA-Stroke Trial inclusion/exclusion criteria.

Inclusion criteria

• Subjects eligible for thrombolytic therapy within 4.5 h of symptom onset in accordance to the current Danish inclusion/exclusion criteria for standard IV rtPA thrombolytic therapy

• Acute hemispheric cerebral infarction on MRI

• NIHSS ≥ 4 (amended from NIHSS ≥ 6 to NIHSS ≥ 4 after six patients due to low recruitment rate)

• PWI/DWI mismatch defined as PWI volume ≥ 120% and ≥ 10 ml versus the DWI volume

• Informed consent obtained

Exclusion criteria

• Age < 18 years

• Pre-stroke disability defined by modified Rankin Scale > 1

• Contraindication to MRI or MRI contrast agent

• Acute infarction of the brainstem on MRI

• DWI volume > 80 ml

• Occlusion of the internal carotid artery or carotid-T on MRI

• Epilepsy and/or acute seizure

• Treatment with or allergic reaction to xanthines (theophylline or its derivatives)

• Pregnancy, breastfeeding, or positive pregnancy test (negative pregnancy test required in fertile women up to 55 years of age)

• History of liver disease and/or alanine transaminase two times normal value

• History of thyroidal dysfunction

• Renal failure (glomerular filtration rate < 30 ml/min or dialysis treatment)

• Fever > 38.5℃

• Potassium below normal value

• Electrocardiogram with signs of acute ischemic heart disease

• Congestive heart failure, or acute myocardial infarction within the last six months

• Reduced cooperation or other conditions that make it unlikely to proceed the study or to follow the patient for three months

• Participation in any interventional study within the last 30 days

DWI: diffusion-weighted imaging; MRI: magnetic resonance imaging; NIHSS: National Institute of Health Stroke Scale; PWI: perfusion-weighted imaging; rtPa: recombinant tissue plasminogen activator.

Clinical assessment

Health care professionals trained as investigators blinded to the allocation of the study medication perform the clinical assessments throughout the clinical trial. The National Institute of Health Stroke Scale (NIHSS) is used to assess patients at baseline (Visit 0), 2–3 h (Visit 2), 22–32 h (Visit 3), and together with the modified Rankin score (mRs), at day 90 (Visit 4). The absolute change in NIHSS score at 22–32 h is one of the dual primary outcome measures. Clinical improvement of NIHSS score by ≥4 points from baseline to 2–3 h and major clinical improvement from baseline to 22–32 h defined by an improvement in NIHSS score of ≥50% are both secondary clinical outcome measures.

Imaging assessment

MRI scanning is performed at baseline, at 2–3 h, and at 22–32 h follow up with the same field strength (1.5 or 3.0 T) and with standardized MRI sequences for all imaging assessments and across the recruiting sites. MR sequences include DWI, PWI with intravenous Gadolinium (0.1 mmol per kg body weight, 5 ml/s bolus injection), Circle of Willis time-of-flight MR angiography (MRA), gradient echo weighted imaging (T2*) and T2 fluid-attenuated inversion recovery (T2-FLAIR). DWI and PWI volume are assessed semi-automatically by in-house software or via visual inspection with the MRI scanner software to evaluate DWI/PWI-mismatch. T-max maps with 4-s contrast delay calculate PWI lesion volumes. The DWI/PWI-mismatch is otherwise defined as PWI volume ≥120% and ≥10 ml versus the DWI volume. Patients with an infarct DWI volume >80 ml are excluded.

The post-processing of the outcome measurements by the central imaging reading board includes visual inspection of all MRI images to assess the imaging quality and make corrections of patient motion if necessary. Post-processing of diffusion and perfusion images is performed semi-automatically using in-house software. Cerebral blood flow and mean transit time will be calculated by deconvolution of the tissue concentration curves with an automatically determined arterial input function using block-circulant singular value decomposition and a parametric technique. DWI, PWI (baseline and 22–32 h) and T2-FLAIR (22–32 h) are co-registered within subjects. Co-registration is performed in MATLAB 2010b (MathWorks Inc., Natick, MA, USA) using SPM8 (Welcome Trust Centre for Neuroimaging, University College London, London, UK). Penumbral salvage is defined as a voxel by voxel quantification of tissue in the acute PWI/DWI-mismatch zone not progressing to infarction on T2-FLAIR at 22–32 h.

MRA at baseline is compared with MRA at 2–3 and 22–32 h. The thrombolysis in myocardial infarction (TIMI) grading of arterial obstruction (score zero = complete occlusion; one = severe stenosis; two = mild to moderate stenosis; three = normal arterial calibre) is applied. Occlusion of the intracerebral vessel is defined as a TIMI score of 0–1. Recanalization is defined as TIMI scores of 2–3. Final infarct lesions are outlined on T2-FLAIR images. Baseline DWI, T2-FLAIR, and apparent diffusion coefficient images are available to the Central Imaging Reading Board to ensure that only baseline infarcts are outlined.

Randomization

Randomization is centralized and web-based via the online TEA-Stroke Trial electronic case form. The randomization system for the study medication is based on computer-generated randomization code lists with stratification for age ≤ 60/ > 60 and severity of symptoms defined by the NIHSS scores of ≤ 15/ > 15. Patients are randomized 1:1 to the study medication and placebo, respectively.

Intervention

Enrolled patients are randomized to receive 10 ml of theophylline (Teofylamin 22 mg/ml containing 20 mg of theophylline monohydrate and 5.5 mg of solubilized ethylenediamine hydrate) versus placebo (normal saline solution). The study drug is administered intravenously over 15 min no later than 30 min after the start of thrombolytic therapy. The vital signs and neurological symptoms of patients are closely monitored according to the standard of care throughout the hospitalization stay. No standard of care is withheld from the participants.

Primary outcomes

The dual primary outcomes are as follows:

• Proportion of penumbra salvage defined as PWI/DWI-mismatch volume not progressing to infarction evident on the 24-h follow-up T2-FLAIR MRI and

• Clinical improvement as defined by the absolute change in the NIHSS score from baseline to the 24-h follow-up.

In this proof of concept study, the primary dual endpoints of penumbra salvage as a surrogate marker and NIHSS score as an early clinical assessment were chosen to avoid a high sample volume. Penumbra salvage correlates strongly with favourable clinical outcome.^22,23 Furthermore, the same dual endpoint was chosen in a well-designed clinical trial with a similar patient population as the TEA-Stroke Trial.²⁴

Secondary outcomes

• Reduction in the risk of brain tissue infarction at the 24-h follow-up. This measure is calculated for each subject using the fitted model by considering a voxel by voxel difference in the probability of infarct with and without add-on theophylline treatment. A formal test for a significant subject-level risk reduction of brain infarction will be performed on the basis of the subject summaries as described in the remote ischaemic preconditioning study by Hougaard et al.²⁵;

• Proportion of penumbral reperfusion at the 2–3 h follow-up time defined as [(PWI lesion volume at baseline – PWI lesion volume at 2–3 h)/PWI lesion volume at baseline] × 100;

• Rapid clinical improvement (NIHSS score reduced by ≥ 4 points) at the 2–3 h follow-up assessment;

• Major clinical improvement (NIHSS score reduced by ≥ 50%) at the 24-h follow-up assessment;

• Proportion of infarct growth defined by the lesion volume on DWI at the 24-h follow-up assessment;

• Recanalization rate defined as a change in the TIMI score of 0–1 at baseline to a score of 2–3 at the 2–3 h and 24-h follow-up assessments; and

• Dichotomized functional outcome at day 90 mRS 0–1 indicates a favourable functional outcome and mRs 2–6 indicates an unfavourable functional outcome;

• Categorical shift in mRS at day 90.

Data safety monitoring board

The independent data safety and monitoring board reviews the following safety data:

• Death within three months;

• Symptomatic intracerebral haemorrhage indicating a combined clinical/imaging endpoint as defined by parenchymal haematoma type II and an NIHSS score decline of ≥4 or death within 24 h of follow-up; and

• Parenchymal haematoma type II or I at the 24-h follow-up assessment.

The data are analysed using Fisher’s exact test for every block size of 20 randomized patients. In the case of an overall significance level of 0.1 or other safety concerns, the data safety and monitoring board will make a recommendation to the Steering Committee as to continue, stop, or modify the study. An interim analysis of the dual primary end-points is scheduled after the recruitment of 60 patients.

Sample size estimates

Treatment with theophylline as an add-on to thrombolytic therapy has never been evaluated before. Thus, the sample size in this proof of concept study was estimated based on the MRI surrogate absolute reduction in infarct expansion. To detect a 22% difference between the treatment groups (80% power, alpha = 0.05), the sample size of 120 subjects with 60 subjects in each treatment group was estimated according to the calculation provided by Phan et al.²⁶ The absolute reduction in infarct expansion correlates well with the clinical outcome, and we expect a similar correlation for our dual primary endpoints, the penumbral salvage and clinical improvement.

Statistical analyses

The primary hypothesis is that the theophylline group will be superior to the placebo group with respect to one or both primary outcome measures. The primary analysis of the proportion of penumbra salvage will be tested with Student’s t-test for Gaussian distributed data, and Wilcoxon’s Rank Sum test for non-Gaussian distributed data. The primary analysis for the clinical outcome will be performed with a logistic regression model or a chi-squared test.

Furthermore, the secondary and safety endpoints will be analysed. A subgroup analysis of the primary and secondary endpoints will be performed for subjects with and without persistent symptomatic vessel occlusion and for subjects treated with endovascular therapy in addition to intravenous thrombolysis. A multiple logistic regression will be performed and adjusted for additional potential confounding variables.

Study organization

The study is investigator-driven and supported by the Aalborg University Hospital in cooperation with the Danish Stroke Centre, Aarhus University Hospital and the Centre for Integrative Neuroscience, Aarhus University.

Summary and conclusion

The TEA-Stroke Trial is the first study to evaluate whether the addition of theophylline to thrombolytic therapy is safe and efficient in patients with acute ischaemic stroke. In contrast to previous studies, the TEA-Stroke Trial adds theophylline to thrombolytic therapy in patients with acute ischaemic stroke. Furthermore, patients are selected by multimodal MRI to depict significant penumbral tissue defined by DWI/PWI-mismatch, as several phase II clinical trials have shown that the use of the mismatch concept is suitable to select patients most likely to benefit from acute treatment.

The hypothesis underlying the TEA-Stroke Trial is that theophylline improves collateral supply in the distal territory of the ischaemic area. As a result, this will facilitate reperfusion and neuroprotection in this area until thrombolytic therapy has recanalized the occluded vessel. Reperfusion of the penumbra may protect neuronal and vascular integrity despite the occlusion of the culprit vessel, which will reduce infarct progression, reduce the risk of reperfusion damage, and, most important, improve clinical outcomes. Furthermore, penumbra salvage due to theophylline administration may extend the time window for thrombolytic therapy to benefit stroke patients untreatable today.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study is funded by the Danish Regions (ref. no. 14/217) and the Danish Heart Foundation (ref.no. 13-04-R94-A4619-22792 and 14-R97-A5066-22829).

Ethical approval

The Danish Health and Medicines Authorities (ref. no. 2013050908) and the Regional Scientific Ethic Committee (ref.-no. N-20130034) approved the study.

Informed consent

Written informed consent was obtained from all subjects before enrolment.

Trial Registration

EudraCT number 2013-001989-42.

Guarantor

BM.

Contributorship

BM researched literature and conceived the study. BM, NH, KM, LØ, FWB, and GA were involved in protocol development, gaining ethical approval, and obtaining funding. BM and NH were involved in patient recruitment and data management. BM wrote the first draft of the manuscript. All authors reviewed and edited the manuscript and approved the final version of the manuscript.