Long-term Effects of Caffeine Therapy for Apnea of Prematurity.
N Engl J Med. 2007 Nov 8;357(19):1893-902.
Schmidt B et al. Caffeine for Apnea of Prematurity Trial Group
BACKGROUND and AIMS
Although methylxanthines are widely used in neonatal intensive care, their effects have been evaluated only in a few, small, short-term studies. In this paper the authors report the long-term results of the CAP (Caffeine for Apnea of Prematurity) trial with a specific focus on neurodevelopmental outcome. This outcome is of particular importance because methylxanthines inhibit adenosine receptors. In experimental animals, adenosine protects the brain from energy failure and cell death when there are hypoxic or ischemic insults. Therefore blocking adenosine can potentially be harmful to the preterm brain and this has lead people to wonder whether caffeine is safe, and if it adversely affects the preterm brain.
Multicenter trial, with mostly Canadian units, and just one institution from the US.
Birth weight 500 to 1250 grams, age 10 days or less, whose clinicians considered them to be candidates for methylxanthine therapy (for prevention or treatment of apnea or to facilitate extubation).
Caffeine citrate: Loading dose of 20 mg/kg body weight, with daily maintenance dose of 5 mg/kg (could be increased up to 10 mg/kg/day if apnea persisted). First dose of drug was given at a median of three days. No drug levels were monitored. It was weaned off before a median post-menstrual age of 35 weeks.
Normal saline placebo
2006 infants were randomized, 1006 to caffeine, and 1000 to placebo.
Primary outcome: This was a composite outcome, assessed at 18 months corrected age, of death, or survival with one or more of the following: cerebral palsy, cognitive delay (on Bayley Scales of Infant Development), hearing loss requiring amplification, and bilateral blindness (visual acuity less than 20/200). Follow-up could be done up to 21 months of age.
Outcome Caffeine Group Placebo Group Odds Ratio
Death or disability 40% 46% 0.79
Cerebral palsy 4.4% 7.3% 0.59
Cognitive delay 34% 38% 0.83
Severe ROP 5% 8% 0.61
There was no difference in the rates of death, deafness, and blindness. The mean percentiles for height, weight and head circumference at follow-up did not differ significantly between the two groups.
Number-needed -to-treat: Seventeen infants would have to be treated with caffeine to prevent one adverse outcome at 18 months of age. In other words, treating 17 infants routinely with caffeine would yield one additional "intact survivor."
Can We Believe the Results of The
Random assignment of treatment? Yes.
Randomization concealed? Yes
Groups similar at start of trial? Yes.
Follow-up sufficiently long? Yes.
Follow-up complete? Yes, adequate data was available for 93% of patients in each arm.
Intention-to-treat analysis? Cannot tell from the article
Blinding of clinicians and study personnel? Yes
Groups treated equally, apart from experimental intervention? Not stated, but presumably yes.
This is an excellent study. It is methodologically sound, and we should take its results seriously and start using caffeine routinely. Likely mechanisms of action of caffeine are (1) early discontinuation of ventilation, less ventilator associated lung injury and less BPD, which is associated with neurodevelopmental impairment (2) fewer hypoxic apnea episodes than in the placebo group.
It's a little unclear how the control group babies were managed. If they had significant apnea, were there clear criteria for giving open-label caffeine or some other methylxanthine? Or were they managed solely with CPAP or other respiratory support and continued to receive the study drug? In that case, the control group may not be representative of the current population of NICU babies (who almost always get a methylxanthine for significant apnea). And if the control group did not receive optimal pharmacologic management, did that make the caffeine look better than it actually is, and does that explain the higher rate of BPD in the control group (more ventilation and lung injury because non-pharmacologic support used)? Nevertheless, I believe this study provides a 'pure' estimate of the effect of caffeine versus no caffeine. It's not a comparison of "caffeine for all" versus "selective caffeine."
So, which babies should be treat with caffeine? Personally, I think we have enough evidence from the CAP trial's short-term and 18 month results to routinely place all babies less than 1250 grams who are AGA (appropriate for gestational age) on caffeine by day 3 of life. Late preterm or term babies who are in this weight category because they are small for gestational age need not routinely get caffeine. Most NICU's already treat symptomatic (apneic) babies with methylxanthines. The results of this study suggest that we should (a) use caffeine and not other methlyxanthines in preterm babies (b) use it early and universally in small premies, not just prior to extubation or just for symptomatic apnea.