Recurrent apnea is common in preterm infants, particularly at very early gestational ages. These episodes of loss of effective breathing can lead to hypoxemia and bradycardia which may be severe enough to require resuscitation including use of positive pressure ventilation. In infants with apnea, methylxanthines have been successful as treatment to prevent further episodes. It is possible that prophylactic therapy, given to all very preterm infants from soon after birth, might prevent apnea, its associated hypoxemia and bradycardia and need for added ventilatory support.

To determine the effect of prophylactic treatment with methylxanthine on apnea, bradycardia, episodes of hypoxemia, use of mechanical ventilation, and morbidity in preterm infants at risk for apnea of prematurity

The standard search strategy of the Neonatal Review Group was used. This included searches of the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 2, 2008), Oxford Database of Perinatal Trials, MEDLINE (1966 - June 2008), CINAHL (1982 - June 2008) and EMBASE (1988 - June 2008) using MeSH term infant-newborn, and text terms methylxanthines, caffeine, theophylline, apnea, and premature as well as a search of previous reviews including cross references, and conference proceedings including abstracts from the Society for Pediatric Research meeting 2001 - 2008.

All trials utilising random or quasi-random patient allocation, in which prophylactic methylxanthine (caffeine or theophylline) was compared with placebo or no treatment were eligible. Outcomes sought included the rate of apnea, bradycardia, hypoxemic episodes, use of IPPV, morbidity and side effects such as tachycardia or feed intolerance, as well as longer term abnormal growth and development.

The standard methods of the Cochrane Collaboration and its Neonatal Review Group were used. The methodological quality of each trial was reviewed independently by each review author. Each review author extracted data separately, then results were compared and differences resolved. The standard method of the Cochrane Neonatal Review Group was used to analyze the data, utilizing relative risk (RR) and risk difference (RD) .

Two studies examining a total of 104 infants were found. Both studied the effects of prophylactic caffeine. There were no meaningful differences between the caffeine and placebo groups in the number of infants with apnea, bradycardia, hypoxemic episodes, use of IPPV or side effects in either of the studies. Only two outcomes (use of IPPV and tachycardia) were common to the two studies and meta-analysis showed no substantive differences between the groups. One large trial (CAP Trial 2006) of caffeine therapy in a heterogeneous group of infants at risk for and having apnea of prematurity has demonstrated an improved rate of survival without developmental disability at 18 - 21 months of age and is awaiting assessment for inclusion in this review.

The results of this review do not support the use of prophylactic caffeine for preterm infants at risk of apnea, bradycardia or hypoxemic episodes.

Any future studies need to examine the effects of prophylactic methylxanthines in preterm infants at higher risk of apnea, bradycardia or hypoxemic episodes. This should include examination of important clinical outcomes such as need for IPPV, neonatal morbidity, length of hospital stay and long-term development. Much of this data will be available on further subgroup analysis of the Caffeine for Apnea of Prematurity Trial (CAP Trial 2006).

Recurrent episodes of apnea are common in preterm infants and the incidence as well as the severity increase at lower gestational ages (Henderson-Smart 1995). Although apnea can occur spontaneously and be attributed to prematurity alone, it can also be provoked or made more severe if there is some additional insult such as infection, hypoxemia or intracranial pathology. The American Academy of Pediatrics defines infant apnea as a pause in breathing of greater than 20 seconds, or one of less than 20 seconds and associated with bradycardia and/or cyanosis (AAP 2003). The definition, diagnosis, and treatment (drugs and assisted ventilation therapy) of apnea of prematurity has been reviewed (Finer 2006). Finer et al (Finer 2006) review the clinical care and therapeutic research needs regarding apnea of prematurity.

If prolonged, apnea can lead to hypoxemia and reflex bradycardia which may require active resuscitative efforts to reverse. There are clinical concerns that these episodes might be harmful to the developing brain or cause dysfunction of the gut or other organs (reviewed by Henderson-Smart 1995). Frequent episodes may be accompanied by respiratory failure of sufficient severity as to lead to intubation and the use of intermittent positive pressure ventilation (IPPV).

Methylxanthines are thought to stimulate breathing efforts and have been used in clinical practice to reduce apnea since the 1970's. Theophylline and caffeine are two forms of methylxanthine that have been used and they are effective for the treatment of infants with recurrent apnea (Henderson-Smart 2008b). Their mechanism of action is not certain. Possibilities include increased chemoreceptor responsiveness (based on increased breathing responses to CO2), enhanced respiratory muscle performance and generalised central nervous system excitation.

In the treatment of apnea, caffeine has similar effects to theophylline (Steer 2005) but has potential therapeutic advantages due to the larger gap between therapeutic blood levels and those associated with toxic effects, more reliable enteral absorption and the longer half life which allows once daily administration (Blanchard 1992).

To determine the effect of prophylactic treatment with methylxanthine on apnea, bradycardia, episodes of hypoxemia, use of mechanical ventilation, and morbidity in preterm infants at risk for apnea of prematurity.

To evaluate a difference in response depending on quality of the study, type (caffeine or theophylline) or dose of methylxanthine, gestational age at birth, postnatal age at study entry, or duration of treatment?

Criteria for considering studies for this review

Search methods for identification of studies

Data collection and analysis

Description of studies

Risk of bias in included studies

Effects of interventions

The total number of infants (104) studied in these two trials is small. As a result, only large differences (e.g. 50% relative risk reduction) could be detected, even for outcomes which are common, such as bradycardia (> 24 /day) and hypoxemic episodes (> 12 /day) where, in the study of Bucher 1988, there is more than 60% incidence in the control group.

Although no effect of caffeine used as prophylaxis was found here, methylxanthines, including caffeine, are effective in reducing apnea and the use of IPPV when used to treat infants with apnea (Henderson-Smart 2008b). Another review has suggested that methylxanthines prior to extubation might be of benefit in reducing the rate of respiratory failure, which is due in part to hypoventilation and apnea (Henderson-Smart 2008a).

It is possible that failure to observe an effect of prophylactic caffeine in the Bucher 1988 trial was due to measurement of hypoxemia and mild bradycardia as primary outcomes - events associated with apnea, rather than apnea itself. It could be argued that such events might be increased in the treatment group if caffeine increased arousal, movements and metabolic rate. This would minimise any differences due to a reduction in apnea, if that occurred. In the Levitt 1988 trial, however, the clinical events of apnea and bradycardia were recorded and no differences were found between infants who received prophylactic caffeine and those on placebo.

The trials in this review did not allow subgroup analyses to determine whether the results varied by the type (caffeine or theophylline) or dose of methylxanthine, gestational or postnatal age at study entry, or duration of treatment. A major concern is the small numbers in each study which, while adequate to show the large effect on apnea, would not be able to detect less common adverse effects.

Of particularly concern is the lack of trial data on long-term growth and development. The CAP Trial (CAP Trial 2006) has published outcomes at discharge and growth and development at 18 to 21 months. These results include a large number of very low birthweight infants (Caffeine group 1006, placebo group 1000) with any one of the three indications for trial entry (prophylaxic prevention of apnea in 22%, treatment of apnea in 40% or prophylaxis for extubation in 38%). At present, the results cannot be specifically applied to this review on prophylactic methylxanthine for apnea, although they do provide a generalized effect of caffeine indicating that there is improved outcome at discharge and in neurodevelopment at follow-up. There is improved rate of survival without developmental disability at 18-21 months of age (adjusted odds ratio 0.77, 95% CI, 0.64, 0.93). The CAP trial authors have been requested to evaluate outcomes for each indication which will make the trial eligible for inclusion in this review and also the other two Cochrane reviews dealing with the other two indications (Henderson-Smart 2006a; Henderson-Smart 2006) and allow for a more precise understanding of the effects in these related but different populations.

Implications for practice

Implications for research

Bucher 1988

Levitt 1988

Larsen 1995

CAP Trial 2006

Bucher 1988

Published and unpublished data

Levitt 1988

Published and unpublished data

Larsen 1995

CAP Trial 2006