Apnea, bradycardia, and neuromuscular hyperirritability have been associated with magnesium (Mg) deficiency in young human infants and weanling animals. This is a retrospective review of a clinical experience of Mg therapy among 200 premature neonates who showed physical and clinical chemical changes compatible with Mg deficiency. The 200 infants all had idiopathic apnea neonatorum, and 93% also had the respiratory distress syndrome (RDS). This review was conducted to learn whether Mg therapy was associated with a significant reduction in apnea. The author suggested that the dose of Mg be 0.4 mEq/kg body weight/day, as 50% MgSO4.7H2O intramuscularly (IM) for 5 days; or as 1.0 mEq/kg/day, as 10% MgCl2.6H2O by mouth for 2 or more weeks, with appropriate monitoring of plasma Mg values in all infants. Sixty-one infants received a minimum of 5 days of Mg by either route (mean, 11.4 +/- 0.9, Group A); five received 3-4 doses IM (mean, 3.6 +/- 0.2, Group B); and 134 received 0-2 doses IM (0.5 +/- 0.1, Group C). Group A infants Mg-treated before Day 20 showed earlier cessation of apnea and bradycardia than those treated after Day 20. In Group A patients, 7 dose-days [corrected] of Mg therapy was associated with continuation of apnea; 14 dose-days [corrected], with cessation of apnea. Compared with Group A, Group C continued to develop apnea (P less than .003) and bradycardia (P less than 0.03) over longer periods of time. Group A infants showed no record of death or of hospital readmission for recurrent apnea, while 32 of 134 Group C infants had one or both of those unfavorable outcomes (P less than 0.001), with four of the five deaths in Group C (NS) as the sudden infant death syndrome (SIDS). In conclusion, Mg was associated with a reduction of apnea in this population. Emphasis was placed on the need to closely observe infants receiving supplementary Mg, with monitoring of plasma Mg levels.