The effectiveness of calcium in electromechanical dissociation (EMD) has been challenged. Retrospective studies have been contradictory. To determine its effectiveness a prospective, randomized, blinded study comparing calcium chloride and saline in refractory EMD was carried out in the pre-hospital setting from October 1982 to October 1983. Only patients who had received epinephrine and bicarbonate and were refractory were entered in the study. All trauma and pediatric arrests were excluded. Ninety patients presented in refractory EMD. Overall, eight of 48 who received calcium were resuscitated successfully in the field; two of 42 who received saline were resuscitated successfully (P less than .07). A successful resuscitation was defined as the conveyance of a patient with a pulse and a rhythm to an emergency department. Patients were analyzed for age, sex, and witnessing of arrest. There was no statistical difference in demographic data. When the group of EMD patients was broken down into subgroups based on the width of QRS, it was noted that patients with a QRS width less than 0.12 did not respond to calcium, whereas the successfully resuscitated in the group with widened QRS or ischemic changes (N = 70) was eight of 39, compared with one of 31 not receiving calcium (P less than .028). Only one patient who was resuscitated successfully was discharged from the hospital alive. Calcium has been shown to be effective in the cardiac resuscitation of patients in refractory EMD. There may be a subset of patients with widened QRS complexes or ischemic changes who will benefit to a greater extent from the use of calcium chloride.