Predicted impact of intravenous thrombolysis on prognosis of general population of stroke patients: simulation modelBMJ 1999; 319 doi: https://doi.org/10.1136/bmj.319.7205.288 (Published 31 July 1999) Cite this as: BMJ 1999;319:288
- Henrik Stig Jørgensen, consultant ()a,
- Hirofumi Nakayama, senior registrara,
- Lars Peter Kammersgaard, senior registrara,
- Hans Otto Raaschou, consultanta,
- Tom Skyhøj Olsen, associate professorb
- a Department of Neurology, Bispebjerg Hospital, Bispebjerg bakke 23, DK-2400 Copenhagen NV, Denmark
- b Department of Neurology, Gentofte Hospital, Copenhagen, Denmark
- Correspondence to: H S Jørgensen
- Accepted 18 February 1999
Alteplase (recombinant tissue plasminogen activator) can be used to dissolve blood clots and achieve reperfusion in some stroke patients. Three randomised controlled trials have studied its clinical effect.1–3 A US trial studied patients who were treated within three hours of onset of stroke and reported a 32% (95% confidence interval 1% to 70%) relative increase in the proportion of patients with full recovery but no effect on overall mortality.1 This led to approval of alteplase for stroke patients in the United States. A European trial of patients treated within six hours of stroke onset was negative,2 and a second trial, published recently, reported no significant positive effect.3 An application for European approval of alteplase treatment within three hours of stroke onset is being considered. Alteplase often leads to bleeding and should be given only by specialised teams.3 Reorganisation of stroke treatment is therefore necessary if alteplase is approved. The trials were performed in highly selected patients.1–3 To elucidate the effect of alteplase on the general stroke population we conducted a retrospective analysis, applying the exclusion criteria and treatment effect reported in the US trial1 to the unselected stroke population of the Copenhagen stroke study (COST).4
Subjects, methods, and results
COST comprised 1197 patients with acute stroke recruited during 1991-3 from an area of Copenhagen.4 These patients constituted 88% of stroke patients in the area: the remaining 12% were not admitted because they had very mild stroke or died before reaching hospital; none would have qualified for alteplase treatment. The figure shows the impact of each exclusion criterion from the US trial1 on the COST population. Only 64 (5%) of the patients fulfilled all criteria. Nineteen of these patients died, and 17 had full recovery (defined as 95 or 100 points on the Barthel index at discharge1) These patients would therefore not have benefited from alteplase treatment. The 28 patients who survived but did not achieve full recovery could have benefited from treatment. With a 32% relative increase in patients with full recovery, five patients (17×0.32=5 (95% confidence interval 1 to 12)) or 0.4% (0.1% to 1.0%) of the stroke population would have benefited if alteplase had been available.
Time was the most critical criterion; only 14% of the patients were admitted in due time. However, this percentage would probably increase if alteplase were approved, so we also estimated treatment effect assuming the ideal—all patients admitted in due time. In this case 539 patients (45%) would have been eligible (figure). Of these patients, 119 died and 150 had full recovery, leaving 270 patients who survived but did not achieve full recovery who could have benefited from treatment. An estimated 48 patients (150×0.32=48 (1 to 105)), or 4% (0.1% to 8%) of the stroke population, would have benefited from treatment.
Combining data from the US alteplase trial1 and the Copenhagen stroke study, we estimated that 0.4% of unselected stroke patients would benefit from alteplase treatment. With no time constraints for treatment, still only 4% would benefit. These estimates may be too generous, as we could not exclude patients with rapidly improving symptoms, a criterion excluding 10% in the US trial. In conclusion, treatment with alteplase may benefit single patients but will have no impact on the general prognosis of stroke. Because time is crucial and because evaluation of patient and paraclinical data requires a specialist setting, treatment with alteplase will need large investments and reorganisation of the care for stroke patients. Before it is decided to offer this expensive, potentially harmful, and possibly only marginally effective treatment we suggest that another, much larger, European trial is needed to test the results of the US trial.
Contributors: HSJ was principal investigator of the Copenhagen stroke study. All the authors participated in designing the study, data collection, and writing this paper. HSJ and TSO analysed and interpreted data for this paper and are guarantors for the paper.
Funding The study was supported by grants from The Danish Health Foundation and The Danish Heart Foundation.
Competing interests None declared.