Data from the Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico (GISSI) trial were reviewed to describe the epidemiology of pericardial involvement in patients treated with or without thrombolysis, and to establish its role as a marker of the extent of myocardial infarction and its prognostic value. In both GISSI-1 (n = 11,806) and 2 (n = 12,381), a specific item regarding presence/absence of clinically detected pericardial involvement was included in the study forms. In GISSI-1, patients with ST elevation and depression at the onset of myocardial infarction were admitted, whereas GISSI-2 included only those with ST elevation. Results of univariate analysis are presented as Mantel-Haenszel-Peto odds ratios with 95% confidence intervals. Cox proportional hazards models were used to assess the independent prognostic significance of pericardial involvement for in-hospital and long-term mortality. The main results indicate that: (1) the incidence of pericardial involvement in patients treated with thrombolytic agents is approximately half of that in the control group (6.7 vs 12.0%); (2) the earlier is the treatment, the lower is the incidence of pericardial involvement; (3) pericardial involvement is strongly associated with infarction size, evaluated by electrocardiograms, creatine kinase peak and echo assessments; and (4) pericardial involvement is associated with a higher long-term mortality, but is not an independent prognostic factor (RR 1.02; 95% confidence interval 0.82-1.26). Pericardial involvement is a reliable bedside, cost-free marker of myocardial infarction size and poorer outcome. Because it may elude detection owing to its transitory and often short duration, it should be given greater attention.