Facebook Twitter YouTube


Caffeine for Apnea of Prematurity: Is More Better?

Friday, October 19, 2012
Room R02-R05 (Morial Convention Center)
Fara Davalian, MD, Doug Lake, PhD, J. Randall Moorman, MD, John Kattwinkel, MD and Santina Zanelli, MD, Pediatrics, University of Virginia, Charlottesville, VA


Caffeine is used for preterm infants with apnea of prematurity (AOP) and has a wide therapeutic index. The utility of caffeine levels to monitor therapeutic safety is controversial and whether efficacy is related to caffeine levels is unknown. Despite the lack of evidence that higher levels improve apnea rates, caffeine boluses are used in clinical practice for infants with worsening apnea and a perceived “low” caffeine level.


To evaluate the effects of a caffeine bolus on apnea rates in symptomatic preterm infants on caffeine for AOP.


Bedside monitor waveforms were collected prospectively from spontaneously breathing VLBW infants admitted to a single NICU from January 2009 to June 2011.  Continuous recordings were analyzed for central apnea using a novel validated computer algorithm. Apnea events were defined as apnea (>10 sec (ABD10), >20 sec (ABD20), or >30 sec (ABD30)) accompanied by bradycardia (HR<100 bpm) and desaturation (SpO2<80%).  Events reported by nurses were also recorded.  Charts were retrospectively reviewed for occurrence of a caffeine bolus in infants with perceived low levels. All decisions to obtain a caffeine level or give additional caffeine were made by clinicians unaware of the automated ABD count.  ABD events were reviewed for the 24h preceding and the 48h following the caffeine bolus. Statistical significance was assessed using a large sample (n=54) paired z=test.


Data were analyzed for 54 caffeine boluses in 43 patients (mean gestational age 26.3±2.0wks, mean birthweight 887±25g). Caffeine boluses were given at a mean postmenstrual age of 32.0±2.7wks and mean weight of 1480±512g. The mean caffeine level was 13.2±2.8 (median 13.1) and the mean caffeine bolus was 10.2±4.7 mg/kg (median 10 mg/kg). Respiratory support at the time of caffeine bolus included CPAP (48%), nasal cannula (50%), and no support (2%). .ABD10, 20, 30 and nursing reported apnea rates prior to and after a caffeine bolus are shown in the figure. The mean ABD10 and ABD20 rates per day were significantly reduced following a caffeine bolus: from 5.8±7.3 to 3.7±4.7 for ABD 10 (p=0.03, paired z-test) and from 4.1±6.1 to 2.5±3.6 for ABD20 (p=0.04, paired z-test)  The decrease in ABD events was strongest in the 24-48h period after the bolus.  ABD rates changes are not detected in nursing records. 


The administration of a caffeine bolus results in a significant decrease in ABD events in preterm infants on caffeine for AOP.  The effect is stronger 24 to 48 hours after the bolus when detected with a novel automated ABD detection system, but not nursing records. Additional studies are ongoing to evaluate the duration of effect as well as identify patients' characteristics of responders vs. non-responders including effects of post-menstrual age, caffeine levels, and apnea rates.

Funded by NICHD 5RCZHD064488