CPR has not been well studied in children and little is known about factors predictive of outcome. We conducted a study with three goals: longitudinal determination of demographic and laboratory data characterizing pediatric arrest victims; identification of factor(s) predictive of outcome; and determination of the prevalence of ionized hypocalcemia in pediatric arrest victims. All resuscitation efforts were documented during a one-year period in a 240-bed tertiary care children's hospital. Patients were classified into two groups--respiratory arrest (RA, requiring only assisted ventilation), and cardiac arrest (CA, absence of palpable cardiac activity requiring closed-chest CPR). Collected data and laboratory tests were analyzed using a step-wise discriminant analysis to determine which factors were predictive of outcome. There were 113 arrests in 93 children; 53 were CA victims and 40 were RA victims. CA had a high in-hospital mortality (90.6%) compared to RA (32.5%). The population was young (55% less than 1 year old) and 87% had at least one chronic underlying disease. No laboratory or demographic value was significantly associated with eventual outcome. The number of doses of epinephrine in CA victims, or bicarbonate in RA victims, was associated with eventual outcome. None of 31 CA victims receiving more than two doses of epinephrine survived to discharge. Low ionized calcium concentrations (less than 3.5 mg/dL) were identified in ten patients; septic shock was present in seven, and chronic renal failure in two.(ABSTRACT TRUNCATED AT 250 WORDS)