A blinded, randomized, placebo-controlled trial was conducted to evaluate the effectiveness of theophylline and doxapram therapy in 31 infants with significant apnea of prematurity. Of 10 infants, two had a short-term response to placebo, 8 of 10 infants to theophylline, and 7 of 11 infants to doxapram (placebo vs treatment with theophylline or doxapram: p = 0.01). The two infants who initially responded to placebo remained responsive for the duration of the study. Of the eight infants in whom treatment with placebo failed, five were randomly assigned to receive theophylline, for a total of 15 infants treated with theophylline, and two of the eight were randomly assigned to receive doxapram, for a total of 13 infants treated with doxapram; the remaining infant required tracheal intubation. Of the 15 infants randomly assigned to receive theophylline, seven responded for the duration of the study; of the eight infants who did not respond to treatment with theophylline, five responded to doxapram, one responded to a combination of theophylline and doxapram, and two remained resistant to treatment. Of the 13 infants randomly assigned to receive doxapram four responded for the duration of the study; of the nine who did not respond to doxapram, seven responded to theophylline, one responded to a combination of theophylline and doxapram, and one remained resistant to treatment. This study demonstrates that although therapy with theophylline or doxapram is associated with a significant short-term reduction in the incidence of apnea compared with that in placebo-treated infants, the long-term response to treatment is frequently incomplete and is not sustained more than 1 week.