Thrombolysis with alteplase is the most effective medical therapy for acute ischaemic stroke. Its application has been increasingly permeating medical practice around the globe ever since the publication of the NINDS trial data in 1995 establishing safety and efficacy if the medication is administered within 3 hours of symptom onset adhering to strict in- and exclusion criteria.1

Subsequent trials and meta-analyses have confirmed the utility of this medication in the tertiary setting under expert guidance.2,3 In 2008, the window was extended from 3 to 4.5 hours after publication of ECASS III using slightly more stringent criteria4.

Most international data come from tertiary centres equipped with stroke units and 24 hour/7days a week (24/7) on-call stroke neurologists casting some doubts on the transferability of this data to community hospitals and countries such as New Zealand where even tertiary centers struggle to provide an on-call roster of stroke neurologists.

Fink et al published their Christchurch audit data in 2009 indicating that this therapy can be administered safely and efficaciously within New Zealand with general neurology oversight.5

A similar paper comes from Australia.6 However, these studies once more come from tertiary centres that enjoy a sufficient number of neurologists/stroke physicians on staff to provide such a service.

A recent population-based study from South Australia once confirmed that the vast majority of stroke thrombolysis occurs in the tertiary setting and that distance to a tertiary hospital inversely correlates with access to thrombolysis.7 A few smaller studies report thrombolysis use at community hospitals8–10 and associated ethical dilemmas are discussed.9,11,12

Other papers addressing thrombolysis at non-tertiary or rural/provincial centers focus primarily on telemedicine utilising two way videoconferencing systems to provide tertiary centre back-up9,13,14 rather than the provision of on-site expert support. In New Zealand, stroke thrombolysis is offered at a number of non-tertiary, provincial, and/or rural centres (data unpublished, Ranta, 2011) although data to support this practice in New Zealand is lacking.

Since March 2008 stroke thrombolysis has been offered at Palmerston North Hospital, a 350-bed secondary teaching hospital located on New Zealand’s North Island serving a population of 167,000. The acute stroke service, established in July 2007, operates a 5-bed stroke unit and comprises a multi-disciplinary team lead by two full-time neurologists providing routine Monday through Friday 8am to 5pm coverage and an informal after hour (24 hours, 7 days per week) thrombolysis roster on a goodwill basis (i.e. without remuneration) with only rare occasions of no coverage being provided. All thrombolysis cases are attended in person and directly supervised by either one of the two staff neurologists both during regular and after hours.

This paper presents a complete retrospective audit spanning three years from mid-March 2008 through mid-March 2011 capturing all stroke patients thrombolysed at Palmerston North Hospital starting from the local introduction of this therapy through completion of this audit. This audit was intended to assess safety and efficacy of stroke thrombolysis at a secondary centre in New Zealand.

Between March 2008 and 13 April 2011, Alteplase was considered in 58 patients, 27 of whom received the treatment (45.6%). The total number of ischaemic strokes admitted to Palmerston North Hospital over the same period was 552 indicating a low thrombolysis rate of only 4.9%. However, case frequency has progressively increased over time (Figure 1) and the thrombolysis rate rose to 7.3% during the final 12 months of this audit (16 thrombolysed; 220 total ischaemic strokes). Eleven patients (40.7%) were treated outside of regular hours and one of these was treated after midnight.

Average ± standard deviation (SD) onset to needle time was 170 ± 40 minutes, with a significant proportion spent after completion of CT (average of 52 ± 30.1 minutes) in comparison to a relatively short time span from door to CT (48 ± 32.2 minutes).

The average time of the ambulance crew spent on the scene was 21 ± 6.4 minutes. Priority scores during the ambulance and ED phase ranged from 1–3 with 65% (in ambulance) and 57% (in ED) receiving a priority of ≤2 (i.e. high acuity transported with sirens activated).

Patients with a lower acuity score experienced on average a non-significantly longer time on location (1.8 minutes), but an average 12-minute longer door to needle time in ED. Ambulance transport times could not adequately be compared as patient distance from location to hospital varies widely given the size of the district.

Patient demographics, time delays, and a comparison to tertiary hospital data are summarised in Tables 1 and 2.

Overall, 17 (62.3%) patients improved after receiving alteplase, 10 (37%) of whom had a very favourable outcome (mRS = 0–1), six (22.2%) patients remained unchanged, and four (14.8%) patients worsened. The symptomatic intracranial haemorrhage (ICH) rate was 3.7% (Table 3).

The average degree of improvement observed in responders was 2.41 ± 1.33 points on the mRS. Overall change in mRS including the patients who worsened, died (including non-alteplase related deaths), or did not respond still showed an average net improvement of 1.26 ± 1.95 points (Table 4 and Figure 2).

Six patients were thrombolysed within the extended 3–4.5 hour window. One of these died of a subsequent myocardial infarction, one patient did not improve, and four improved by an average mRS of 3.25 ± 2.06.

All four patients who worsened died. Three of these deaths were unrelated to Alteplase use. Two patients died from ischaemic stroke complications (i.e. aspiration pneumonia) without change in neurologic function after receiving Alteplase and follow-up head CTs did not show any evidence of bleeding. The third patient’s infusion was prematurely stopped after just a third of the dose had been administered because the nurse thought she had made an error in administering the drug. Shortly after stopping the infusion the patient suffered a cardiac arrest. Autopsy demonstrated no evidence of an alteplase-related complication (including no evidence of ICH) and root cause analysis revealed that no medication administration error had taken place. The cause of death was determined to be a myocardial infarction.

The one patient who suffered a symptomatic ICH had a severe ischaemic stroke on presentation with National Institute of Health Stroke Scale (NIHSS) of 27; given the severity of the stroke bleeding risk was considered relatively high, but after consultation with patient’s family it was deemed in the patient’s best interest to provide the treatment, the patient was within the 3 hour window and all NINDS inclusion criteria were met.

One other patient showed evidence of a significant haemorrhage on post-thrombolysis CT, but this did not lead to clinical worsening and is thus not classed as a “symptomatic” ICH.1,4 This patient also met all inclusion/exclusion criteria although the patient had a borderline platelet count of 99,000/ml3, which rose to 102,000/ml3 on rapid re-check prior to giving alteplase (cut off is 100,000/ml3).

Two further patients demonstrated mild “streaking” on CT suggesting a very minor haemorrhage. Both patients were asymptomatic from these minor haemorrhages and experienced significant and rapid recovery from their presenting stroke symptoms. Aside from the debatable deviation from inclusion criteria in the above patient with a borderline platelet count there were no identified violations of the local thrombolysis protocol, which is largely based on the NINDS criteria.

Reasons for deferring thrombolysis in the 31 patients who were referred for thrombolysis but not treated included ICH on CT, rapidly improving or resolved symptoms, current warfarin therapy/elevated INR, persistently elevated blood pressure, increased systemic bleeding risk (e.g. recent bleeding ulcer), and presentation inconsistent with stroke.

The majority of the remaining 492 ischaemic stroke patients admitted to Palmerston North Hospital during the study period were not ‘considered’ for tPA because they arrived outside the treatment window. However, some patients also missed out because of inefficiencies and unawareness of thrombolysis availability by emergency room staff. This was predominantly a problem shortly after introduction of stroke thrombolysis at our centre and since these rates started to be monitored 12 months ago we are aware of only a single patient who missed out due to a poor triage process. This patient was a cancer patient and was initially referred to the oncologists instead of the thrombolysis team and by the time neurology was notified too much time had elapsed.

This is a retrospective audit and is thus limited by its observational, non-randomised, and open label design. Observer bias was limited by recruiting independent chart auditors and adhering to international definitions for determining mRS and classifications such as “symptomatic haemorrhage.”

Keeping the limitations of a retrospective audit and small sample size in mind this study is reassuring as regards safety of stroke thrombolysis with intravenous Alteplase in the secondary setting in New Zealand given that our results closely mimic both tertiary data from a New Zealand centre as well as data from a large international meta-analysis.

Efficacy is more difficult to evaluate from an unblinded series with no control group. However, it is encouraging that our efficacy data not only closely mimics the Christchurch and meta-analysis results, but also the results of the pivotal NINDS placebo-controlled trial. This trial demonstrated at 39% rate of very favourable outcome (compared with our 37%), which compared with placebo indicated a 30% greater likelihood of recovery.1 It is furthermore reassuring that the extension to the 4.5 hours window, based on the ECASS III trial,4 has not adversely affected the overall outcome in our series.

Despite significant benefit in the responder group, as anticipated there are associated risks. While three of the four patients who died suffered unrelated deaths one died as a direct result of treatment. It is noteworthy that under the ECASS III protocol4 this patient would have been excluded due to his high NIHSS score. While some patients and families may prefer to accept the higher bleeding risk and still choose thrombolysis as it may offer the only hope of a meaningful recovery in this situation we no longer routinely offer thrombolysis to patients with an NIHSS of >25.

The patient with the borderline platelet count did not come to harm and whether the ICH was related to the borderline platelet count remains unclear. Several consultant haematolgists have indicated that the link seems unlikely. Nonetheless, despite a lack of similar cases in the literature one might conclude that treating in the setting of a borderline platelet count should perhaps be avoided.

Delays in accomplishing a head CT in a centre with a single available CT scanner and no after-hours on-site CT technicians is understandable and results are in fact quite encouraging. However, the observed delays by ambulance staff at the scene and in ED after the CT is obtained are more difficult to understand and justify. These delays are most likely related to limited awareness of the urgency with which these patients need to be assessed and lack of familiarity with local protocols. This may, in part, be due to a combination of frequent staff turn-over and relative infrequency of these cases.

Some of the delays may also be attributable to poorly written protocols and much effort has gone into collaborating with ED staff to maximise user-friendliness. Similar efforts have gone into ongoing education of ED staff and should go into education of ambulance staff. The latter can be challenging as ambulance crews are typically trained off site and are thus not under the direct influence of the local stroke team. Similarly difficult to reach is the public who often continues to take the attitude of attempting to “sleep off” the symptoms and thereby missing out on the opportunity of treatment.

While this audit is very reassuring, it has to be emphasized that supervision occurred by on site experienced neurologists , in the setting of an organised stroke service with a formalised local thrombolysis protocol and essentially no protocol violations. Thus it is not possible to draw conclusions about the safety and efficacy of thrombolysis provided by less experienced clinicians at even smaller centres around the country.

Even with neurologists on site a one-in-two roster is difficult to maintain and alternative options, such as tele-medicine through regional networks, ought to be explored by district health boards to further increase patient access to this very effective therapy without exposing patients to excessive risks.

Author information: Annemarei Ranta, Neurologist, Department of Neurology, MidCentral Health, Palmerston North; Calvin Chan, House Surgeon, Department of Neurology, MidCentral Health, Palmerston North, ; Dorothea Rump, Medical Elective Student, Department of Neurology, MidCentral Health, Palmerston North; Pietro Cariga, Consultant Neurologist, Department of Neurology, MidCentral Health, Palmerston North.

Correspondence: Dr Annemarei Ranta, Department of Neurology, MidCentral Health, Private Bag 11036, Palmerston North 4442, New Zealand. Fax: +64 (0)6 3508391; email: anna.ranta@midcentraldhb.govt.nz