Several mechanisms appear to be involved in the decrease in apnea frequency seen after methylxanthine administration. These include the following:

Several studies have shown that the plasma clearance and elimination of theophylline and caffeine are both prolonged in newborn babies compared with adults (31,32,33). The representative kinetic profiles of these two drugs are shown in Table 20.1. The obvious difference between the two drugs is the remarkably slow elimination of caffeine relative to theophylline. The plasma half-life is about 100 hours for caffeine and about 30 hours for theophylline. This difference in drug elimination indicates that caffeine can be given more sparingly (i.e., once daily), and that drug monitoring is probably not as crucial with caffeine as with theophylline. Caffeine half-life may be further prolonged in infants with cholestatic jaundice and breastfed infants (34). The recommended therapeutic plasma concentrations for theophylline and caffeine are about 5 to 15 and 5 to 20 mg per L, respectively. To achieve and maintain these
plasma concentrations, a loading dose of 4 to 8 mg per kg of theophylline (active base) followed by a maintenance dose of 2 to 4 mg per kg per day in two to four divided doses may be required. There exists substantial interindividual variability in the pharmacokinetic properties of theophylline; thus, it is necessary to monitor plasma concentrations and adjust the dose accordingly (saliva has been proposed as an alternative site for therapeutic drug monitoring in the preterm, with good correlation with blood levels) (35,36). Similarly, caffeine is recommended as a loading dose of 10 to 20 mg per kg of active base or 20 to 40 mg per kg of caffeine citrate salt, intravenously or orally. A maintenance dose of 2.5 to 4 mg per kg per day (or 5 to 8 mg per kg per day of caffeine citrate) is usually needed to maintain plasma concentrations of 5 to 20 mg per L of caffeine (Table 20.2). About 25% of theophylline is methylated to caffeine (37,38), with plasma theophylline-to-caffeine ratios sometimes reaching 0.30 to 0.40 at steady state. A small (3% to 8%) proportion of caffeine is converted to theophylline via CYP1A2. Thus, the overall methylxanthine effect has to account for the sum of the two drugs because both agents are pharmacologically active.

The methylxanthines are powerful CNS stimulants and may interact with anticonvulsants such as phenobarbital at a kinetic or a pharmacodynamic level. Babies given theophylline and phenobarbital have been shown to require higher doses of theophylline to control apnea and higher doses of phenobarbital to control seizures (39). The methylxanthines also have the potential to interact with drugs that are substrates for CYP1A2 such as cimetidine and ketoconazole.

Caffeine and theophylline exert similar pharmacodynamic effects but may vary in their potency concerning a specific organ receptor. Moreover, the differences in their kinetic properties alter the dosing schedules and the need for therapeutic drug monitoring. Table 20.3 lists some of the differences between the two drugs. Controlled comparative trials between theophylline and caffeine indicate that although both drugs are effective in the management of apnea, more adverse effects such as tachycardia are observed with theophylline (40,41,42). Because caffeine has a more prolonged plasma half-life, the dosing schedule is less frequent and the need for therapeutic monitoring less crucial. Although frequent monitoring is advisable for theophylline, plasma caffeine measurement only if there is lack of clinical response or suspected toxicity is generally acceptable during the neonatal period (43,44). In cases of overdosing with caffeine, the prolonged drug elimination may result in sustained high plasma concentrations of caffeine for a prolonged period. However, observations suggest that caffeine plasma concentrations of up to 50 mg per L may occur with no adverse effects, whereas plasma concentrations of theophylline greater than 15 mg per L may be associated with tachycardia. Some investigators even suggest much higher doses of caffeine to achieve a therapeutic effect. This suggests that caffeine might have a wider therapeutic index relative to theophylline. In practice, caffeine has emerged as the preferred alternative in infants with apnea of prematurity.

In 1998, Steer and Henderson-Smart reviewed three trials comparing theophylline and caffeine in reducing
recurrent apnea and the need for mechanical ventilation (45). There was no difference in the failure rate (<50% reduction in apnea/bradycardia) of treatment with caffeine or theophylline at 1 to 3 days (two studies) or 5 to 7 days (three studies). There was a higher rate of apnea in the standard caffeine group at 1 to 3 days [three studies, mean difference 0.398 (0.334 to 0.72)/100 minutes] but not at 5 to 7 days (two studies). Side effects, as indicated by tachycardia or feed intolerance leading to changed dosing, were lower with caffeine, relative risk (RR) = 0.17 (0.04 to 0.43). One additional study, which has been reported only in abstract form, found that mean rates of apnea (15 seconds or more) and episodes of oxygen desaturation (<85%) were no different in the 13 infants treated with theophylline than in the 11 infants treated with caffeine. However, mean rates of bradycardia were lower with theophylline at 1, 3, and 7 days after the initiation of therapy (46,47,48). In aggregate, the data favor the use of caffeine for neonatal apnea.