In a prospective study of 283 myocardial infarction (MI) patients, state-dependent feelings of anxiety, depression, and irritability were assessed twice during hospital stay, and four times during a 3-5 yr follow-up. A K-mean cluster analysis identified six subgroups of MI patients with different pattern of emotional reactions. Two groups, containing nearly half of the sample, had low average levels of emotional upset at all assessments. Two groups showed an intermediate long-term outcome: one of these had a high level of initial emotional upset that subsided during the first six months after discharge, whereas the other group showed increasing levels of emotional distress long-term follow-up. Finally, two groups failed to achieve long-term emotional readjustment. For one of these, a high level of emotional upset was evident from the first in-hospital assessment, whereas the other one had a sharp increase in emotional distress after discharge. In a series of psychological, social and medical variables, the former pattern was associated with more pre-MI medical and psychosocial problems, whereas the delayed emotional reaction was related to lower levels of cardiac health knowledge. Furthermore, high levels of emotional upset preceded both failure in resuming work and increased long-term rehospitalization. The findings indicate that emotional reactions after a MI should be monitored during convalescence to identify patients at risk for a failure in emotional readjustment. Furthermore, effective treatment of initial emotional reactions could promote resumption of work and reduce long-term morbidity.