Stopping caffeine in premature neonates: how long does it take for the level of caffeine to fall below the therapeutic range?

Abstract

Background

Caffeine is routinely used in preterm infants for apnea of prematurity. Preterm infants are usually monitored for 5 days after discontinuation of caffeine to assess for possible recurrence of apnea. Our objective was to determine if the serum concentration of caffeine decreases to a subtherapeutic level 5 days after its discontinuation.

Methods

This is a retrospective analysis of caffeine levels after the drug was discontinued in preterm neonates (birth weight ≤1500 g) born between January 2010 and June 2017. The primary outcome was the proportion of infants with therapeutic levels of caffeine 5 days after the drug was stopped.

Results

Caffeine levels were measured in 353 samples from 280 infants (birth weight 1246 ± 390 g and gestational age 29.2 ± 2.4 weeks) after discontinuation of the drug. Five and more days after discontinuation of caffeine, 29.3% (82/280) of the infants had caffeine levels ≥5 mg/L. Approximately 41% (75/181) of the caffeine levels measured between 5 and 7 days and 18% (17/95) between 8 and 10 days were ≥5 mg/L. A caffeine dose of >5 mg/kg/day when discontinued was associated with the caffeine level of ≥5 mg/L (OR 2.3, 95% CI 1.28–4.13, p = .005).

Conclusions

Preterm infants treated with caffeine frequently had therapeutic levels of caffeine 5–10 days after discontinuation of the drug. The infants receiving higher doses were more likely to have a therapeutic level of caffeine 5 days after stopping the medication. Preterm infants should be monitored for recurrence of apnea for more than 5 days after stopping caffeine or levels should be monitored prior to discharge.

Keywords:

• Apnea
• caffeine
• therapeutic level
• VLBW

Acknowledgments

This study was presented in part at the Eastern Society for Pediatric Research Annual Meeting, Philadelphia, PA, USA; March 2018, and the Annual Meeting of the Pediatric Academic Societies, Toronto, Canada; May 2018.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Institutional Development Award (IDeA) (Aghai) from the National Institute of General Medical Sciences of the National Institutes of Health under grant numbers [U54-GM104941] (PI: Binder-Macleod) and [NIH COBRE P30GM114736] (PI: Thomas H. Shaffer).