Benefit from earlier thrombolytic therapy is certain, but what is the magnitude of benefit?

BMJ 1996; 312 doi: (Published 27 January 1996) Cite this as: BMJ 1996;312:215
  1. Alain Leizorovicz, director of research, INSERMa
  1. a Service de Pharmacologie Clinique, BP 3041, 69394 Lyon Cedex 03, France

    Abundant evidence exists to confirm that fibrinolytic treatment saves lives in patients with acute myocardial infarction and that the earlier the treatment the higher the benefit obtained.1 There is also evidence to suggest that administration of fibrinolytic treatment, under certain conditions, before hospital admission may lead to further improvement in patients' prognosis with no significant additional risk.2 3 4

    John Rawles attempts to quantify the benefit of earlier fibrinolytic treatment using the data from the 311 patients included in the GREAT trial, which evaluated the feasibility, safety, and efficacy of domiciliary fibrinolysis by general practitioners.5 Such quantification is essential if providers of health care are to make an informed decision on whether to allocate resources to domiciliary fibrinolysis. The magnitude of the benefit is controversial, and the trial that could fully resolve this will never be performed for obvious ethical reasons. This trial would require randomising patients to, say, four or five groups, each group having a predetermined delay from diagnosis of acute myocardial infarction to fibrinolytic treatment. Thus, the only available way to assess the magnitude of benefit of earlier fibrinolytic treatment compared with later treatment is to retrospectively analyse data from fibrinolytic studies, performing indirect comparisons of the randomised groups or using an epidemiological approach.

    In the systematic overview by the Fibrinolytic Therapy Trialists' Collaborative Group an unadjusted, indirect comparison showed that a one hour delay in the time to treatment would lead to an increased mortality of 1.6 (SD 0.6) lives per 1000 within 35 days.1 The underlying assumption in this analysis was that the patients in the different subgroups defined by the time to treatment were comparable, which is, of course, not the case. Dr Rawles presents a classic epidemiological approach with a model using multivariate analysis of the effect of the time gained in the delay to treatment on mortality at 30 days and 30 months. The outcome of the analysis of data combined from both treatment groups is, not surprisingly, in favour of earlier fibrinolytic treatment, and the results are impressive: “In patients presenting two hours after start of symptoms each hour's delay in receiving thrombolysis led to the loss of 21 lives per 1000 within 30 days (95% confidence interval 1 to 94 lives per 1000) (P=0.03) and 69 lives per 1000 within 30 months (16 to 141 lives per 1000) (P=0.0004).”

    However, these results, though significant, should be moderated by the considerable width of the 95% confidence intervals (1 to 94 for 30 day mortality and 16 to 141 for 30 month mortality), which make the results equally compatible with more favourable and less favourable results. Although the point estimator at 30 days is much greater than that reported by the Fibrinolytic Therapy Trialists' Collaborative Group (21 v 1.6), the lower limit of the 95% confidence interval (equivalent to the loss of 1 life per 1000) means that the compatibility of the results cannot be excluded. However, in their meta-analysis of all randomised clinical trials of prehospital treatment versus hospital treatment, the European Myocardial Infarction Project Group's point estimate of the benefit of treatment given one hour earlier is similar to that reported by Dr Rawles (17% (95% confidence interval 2 to 29), P=0.03).2 Furthermore, Dr Rawles' report is the first to show that time to treatment favourably influences the benefit of fibrinolytic treatment on long term mortality, and it would be worth confirming in other studies or in a meta-analysis.

    The clinical implication from Dr Rawles' paper remains that patients with suspected acute myocardial infarction should be given fibrinolytic treatment as early as possible, but the precise magnitude of the benefit is still debatable.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    View Abstract