Early thrombolytic treatment in acute myocardial infarction: reappraisal of the golden hourIntroductionExperimental studiesPre-hospital versus in-hospital thrombolysisFTT analysisAlternative presentation of trial dataMethodsResultsDiscussionReferencesEarly thrombolytic treatment in acute myocardial infarction:reappraisal of the golden hourEric Boersma, Arthur C P Maas, Jaap W Deckers, Maarten L SimoonsTHE LANCETSummaryBackground There is conclusive evidence from clinical trialsthat reduction of mortality by fibrinolytic therapy in acutemyocardial infarction is related to the time elapsingbetween onset of symptoms and commencement oftreatment. However, the exact pattern of this relationcontinues to be debated. This paper discusses whether ornot appreciable additional gain can be achieved with veryearly treatment.Methods The relation between treatment delay and short-term mortality (up to 35 days) was evaluated usingtabulated data from all randomised trials of at least 100patients (n=22; 50 246 patients) that compared fibrinolytictherapy with placebo or control, reported between 1983and 1993.Findings Benefit of fibrinolytic therapy was 65 (SD 14), 37(9), 26 (6) and 29 (5) lives saved per 1000 treated patientsin the 0–1, 1–2, 2–3, and 3–6 h intervals, respectively.Proportional mortality reduction was significantly higher inpatients treated within 2 h compared to those treated later(44% [95% CI 32, 53] vs20% [15, 25]; p=0·001). Therelation between treatment delay and mortality reductionper 1000 treated patients was expressed significantlybetter by a non-linear (19·4–0·6x+29·3x⫺1) than a linear(34·7–1·6x) regression equation (p=0·03).Interpretation The beneficial effect of fibrinolytic therapy issubstantially higher in patients presenting within 2 h aftersymptom onset compared to those presenting later.Lancet1996; 348: 771–75IntroductionThe reduction in mortality that can be achieved withreperfusion therapy in patients with evolving myocardialinfarction depends on the time elapsing between onset ofsymptoms and initiation of treatment or, more specifically,on the duration of coronary occlusion before reperfusion.1Although earlier reperfusion yields a better clinicaloutcome the relation between treatment delay andmortality reduction is controversial. A key question iswhether or not a substantial additional reduction of themortality risk can be achieved with very early treatment,ie, within 2–3 h after onset of symptoms. The concept of a‘first golden hour’2is supported both by experimentalstudies and randomised trials comparing pre-hospital within-hospital therapy.2,3By contrast, the FibrinolyticTherapy Trialists’ (FTT) Collaborative Group, in apooled dataset of randomised trials of more than 1000patients,4reported only a gradual decrease of benefit withlonger delay. We present an alternative analysis of datafrom previous trials. Experimental studiesThe duration of coronary occlusion and the extent ofcollateral circulation are the main determinants of infarctsize in pigs, dogs, cats, and other animals.5–7In animalswith a coronary collateral circulation similar to that ofhumans an occlusion persisting for 15–30 min generallydoes not lead to significant myocardial damage.6,7Thus,necrosis can be prevented provided reperfusion is achievedwithin this period.8A small area of necrosis usually occurswith reperfusion after 45 min occlusion, while the mid-endocardial and subendocardial zones are still viable.6Longer durations of coronary occlusion result inprogressive growth of the infarction and reduction of theamount of salvageable myocardium. At 90 min the extentof cell death involves 40–50% of the area at risk; less thanhalf of the jeopardised myocardium remains viable at thattime.3,66 h after the onset of continuous ischaemia the areaat risk is fully infarcted such that myocardial salvage willbe minimal.In humans, the thrombotic event frequently consists ofmultiple cycles of temporary occlusion and reperfusion.The degree of chest pain (if present) varies among patientsso that it is often difficult to determine the exact durationof the coronary occlusion. Nevertheless, data indicate thatevolution of (enzymatically detectable) infarct size overtime in humans shows a pattern similar to that in animals.9Pre-hospitalversusin-hospital thrombolysisVarious clinical trials have shown that early restoration ofcoronary patency improves survival.10–13Laterrecanalisation may also be beneficial, particularly inpatients with sufficient collateral flow and in those withstuttering infarction. Randomised trials comparing pre-Vol 348 • September 21, 1996 771ArticlesErasmus University, Rotterdam, Netherlands (E Boersma MSc,A C P Maas MD, Prof J W Deckers MD, Prof M L Simoons PhD)Correspondence to: Prof M L Simoons, Thoraxcenter Bd 434,Erasmus University and University Hospital Rotterdam-Dijkzigt,Dr Molewaterplein 40, 3015 GD Rotterdam, NetherlandsTHE LANCETto this analysis little would be gained by extra efforts toachieve very early, prehospital, therapy.Alternative presentation of trial dataSince the FTT analysis is at variance with theexperimental data, and with the larger benefit found indirect, randomised comparison of prehospital with in-hospital therapy, an alternative presentation of the trialdata seems appropriate. Furthermore, estimations ofbenefit for early treatment in the FTT analysis may havebeen influenced by the chosen framework for analysis.The inclusion-threshold of 1000 patients for each of thenine studies included in the FTT analysis is ratherarbitrary. There is no good reason to exclude smallertrials. Although the number of patients in the smallerstudies may not be sufficient to prove the effect ofthrombolytic therapy, these studies may nevertheless assista more powerful estimation of treatment effects insubgroups of patients in a pooled dataset; more so,because several smaller studies excluded from the FTTanalysis included a significant proportion of patients whowere treated very early.The statistical analysis was based on subdivision ofpatients according to delay from symptom onset with onlytwo benefit estimations in the first 3 h (at average delaytimes of 0·98 and 2·50 h, respectively). In this way, part ofa potential large effect in the first 2 h may be obscured bya relatively small effect in the third hour. A moredifferentiated subdivision seems more appropriate to studythe effect of very early therapy.23Furthermore, non-linearmodels might be
View Full Document
Unlocking...