Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Does caffeine treatment for apnoea of prematurity improve neurodevelopmental outcome in later life?

Three Part Question

In [premature neonates under 32 weeks gestation] does [caffeine]improve [subsequent neurodevelopment]?

Clinical Scenario

A 29-week gestation neonate is having frequent apnoeas while on continuous positive airway pressure, so you prescribe caffeine to help prevent apnoeas of prematurity. You notice another baby of the same gestation is not on caffeine but recall reading caffeine treatment may reduce future neurodevelopmental disability. You wonder how strong the evidence is and whether all neonates under 32 weeks gestation should be on caffeine treatment.

Search Strategy

Medline via PubMed was the primary source of articles. The search was via ‘Clinical queries’ using the broad, sensitive filter and search terms: [caffeine AND develop*]. A second search using [methylxanthine AND develop*] was also performed. Limits were: humans and All children (from birth to 12 years). Dates included 1966 to October 2009.

Secondary searches on the Cochrane Database, Clinical Evidence and SUMsearch were performed using the same search terms.

Search Outcome

The search was performed independently by two people. A total of 27 papers were found via PubMed, three of which were relevant. The references of the above papers were scanned, along with the linked articles, but no further articles were found. No additional articles were found via secondary sources, by using the second search terms or by extending the upper age limit to 18 years.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Schmidt et al,
2006 neonates, birth weight 500–1250 g (full data for 1869) Caffeine n=937 Placebo n=932 Deemed to need caffeine and treated until caffeine no longer needed for AOP Assessments at 18–21 months corrected age Further follow-up assessments planned at 5 yearsMulticentre, international double blind RCT (level 1b)Death, cerebral palsy, blindnessDeath or NDI Caffeine=40.2% (377/937) Placebo=46.2% (431/932) OR 0.77 (95% CI 0.64 to 0.93) p=0.008 NNT=16 (95% CI 5 to 56)

Cerebral palsy Caffeine=4.4% (40/909) Placebo=7.3% (66/901) p=0.009, RR=0.6 OR 0.58 (95% CI 0.39 to 0.87) NNT=34 (95% CI 20 to 132)

Cognitive delay, Caffeine=33.8% (293/867) Placebo=38.3% (329/858) p=0.04, RR=0.88 OR 0.81 (95% CI 0.66 to 0.99) NNT=23 (95% CI 11 to 4387)

No signifi cant effect on growth in caffeine group (p=0.66) Improved scores on psychomotor DIS (mean difference 1.66, 95% CI 0.26 to 3.06) and mental DIS (mean difference 2.54, 95% CI 1 to 4.06) in caffeine treated group
Corticosteroids given in first 4 days increase risk of CP (NNH=22, 95% CI 12 to 133) Earlier discontinuation of positive pressure ventilation in caffeine group (p<0.001)ORs quoted, rather than relative risks (slightly less significant findings when calculated) Borderline statistical significance of reduction in cognitive delay with very wide NNT confi dence intervals Large preanalysis drop-out rate (up to 15%). ‘Worst case scenario’ analysis reverses the significant effects, with improved neurodevelopment in the control group:RR of CP in caffeine group 1.4 (NNH 15, 95% CI 10 to 23), and of cognitive delay 1.2 (NNH 10, 95% CI 7 to 17)
Ment et al,
102 neonates, birth weight <1250 g (full data for 73) MTX group=43 No MTX group=30 Assessment at 18 months corrected ageCohort study (level 2b)Neurodevelopmental outcome (Bayley score) Examiner blinded to neonatal historyBayley MDI scores MTX, no IVH (n=17)=95.82 MTX, IVH (n=26)=90.54 No MTX, no IVH (n=18)=86.72 No MTX, IVH (n=12)=82.42

Increased scores in neonates treated with MTX, regardless of haemorrhage status (p=0.033)

No difference in abnormal neurology at 18 months corrected age MTX group=4/43 No MTX group=0/30 p=0.14
Which methylxanthine used not documented so subgroup analysis not possible No intergroup matching IVH/GMH may be a confounding factor, but not possible to analyse due to data presentation and small numbers
Gunn et al,
42 neonates, birth weight <2500 g, with apnoeas Caffeine group=21 Placebo group=21 Assessment every 3 months between 12 and 24 months of ageCohort study (retrospective matching) (level 2b)Growth, neurological and developmental outcome (using Griffi ths tests) Ophthalmology and EEG findingsNo signifi cant difference between groups in development, growth or neurological sequelae Development quotient Caffeine=100±19 Control=101.5±18 p>0.6 (95% CI 93.2 to 109.8) Study 30 years old so neonatal populations and other treatments have changed Caffeine started after other options unsuccessful, with no uniformity in caffeine treatment Confounding factors such as severe CNS damage may have already occurred Follow-up assessors may not have been blinded to group or study hypothesis


Apnoeas occur in up to 90% of preterm neonates and are commonly treated with respiratory stimulants such as methylxanthines. Caffeine reduces the number of apnoeic episodes, rates of bronchopulmonary dysplasia and the need for mechanical ventilation,1 and although apnoeas may be related to subsequent neurodevelopmental impairment (Janvier), until recently there has been little robust evidence of any long-term neuroprotective effects of caffeine.

Some animal studies have reported negative effects related to caffeine, including disruption of sleep, spatial learning and growth (Zimmerberg), transient inhibition of astrocytogenesis (Desfrere) and increased neuronal dendritic length (Juárez-Méndez). One of the most worrying traits as regards neonatal medicine is the potential for vasoconstriction, especially in the brain and gut (Hoecker). However, recent studies have not shown adverse effects on the volume (Dani) or velocity (Soraisham) of blood flow to these areas and caffeine has even been proposed to be a suitable treatment for hypotensive neonates, increasing cardiac output and blood pressure (Soloveychik).

Schmidt et al showed caffeine treatment reduces the rates of cognitive delay and cerebral palsy on first assessment around 2 years of age. However, although the study was large with a robust design, a substantial number of patients were lost to follow-up and if the unlikely ‘extreme worst case scenario’ analysis was undertaken, then higher rates of cerebral palsy and cognitive delay in the caffeine-treated group were found. Two earlier studies were smaller and had less rigorous designs. Ment et al found increased Bayley scores at 18 months of age in methylxanthine-treated neonates, but subgroup analysis of caffeine-only treated subjects was not possible due to their data presentation. Gunn et al did not have the effect of caffeine on neurodevelopment as a primary outcome, and with such small numbers the study was unlikely to be powered to detect benefits. Unfortunately, due to the differing methods, outcome measures and presentation of data, it was not possible to combine the results from all three studies to obtain an overall measure of significance for any neurodevelopmental benefits of caffeine or a number needed to treat.

A meta-analysis of 21 studies (Stein) found no significant deleterious effects on cognition or behaviour in methylxanthine-treated children and the studies above (Schmidt, Ment, Gunn) indicate caffeine is unlikely to increase the risk of neurodevelopmental disability in the preterm population and is likely to be of benefit. The extent to which caffeine is directly neuroprotective or acts via secondary mechanisms (such as earlier discontinuation of positive pressure ventilation or prevention of apnoeas) remains unknown, although an ongoing study (Inder) may provide some answers. Post hoc analysis by Schmidt et al concluded that nearly half of the ‘caffeine effect’ could not be explained by the reduced respiratory support needed, even though their subgroup of neonates receiving positive pressure ventilation gained the most neurodevelopmental benefit from caffeine (Davis). Caffeine-treated hypoxia-exposed rat pups exhibit reduced ventriculomegaly and increased myelination (Back) and neonates on higher dose caffeine (20 vs 5 mg/kg/day) have lower incidences of major disabilities (18% vs 7%, RR 0.42, p=0.05) (Steer). Subclinical apnoeic episodes may, over time, cause damage to the developing brain and by antagonising inappropriate release of adenosine, caffeine may reduce any white matter damage sustained during such hypoxic episodes. Alternatively, caffeine may alter brain haemodynamics, allowing remodelling after an insult.

In summary, caffeine can be neuroprotective in premature infants, despite the described study flaws potentially exaggerating the effect. However, all the neonates investigated were either having apnoeas or deemed to ‘require’ caffeine treatment by the involved clinicians. Prophylactic caffeine use in non-apnoeic neonates has not been shown to significantly reduce the use of positive pressure ventilation (Henderson-Smart) and neuroprotective effects have yet to be investigated or determined. Therefore, in this group, other side-effects of caffeine such as worsening gastro-oesophageal reflux must also be considered.

Editor Comment

AOP, apnoea of prematurity; CP, cerebral palsy; DIS, Developmental Index Score; IVH, germinal layer haemorrhage or intraventricular haemorrhage; MDI, Mental Development Index; MTX, methylxanthine; NDI, neurodevelopmental impairment; NNH, number needed to harm; NNT, number needed to treat; RCT, randomised controlled trial.

Clinical Bottom Line

Caffeine given to premature infants does not increase neurodisability rates. (Grade A) Caffeine treatment for apnoea of prematurity in preterm infants reduces the risk of adverse events such as death, cerebral palsy and cognitive delay. (Grade A) Even preterm infants with small self-limiting desaturations or apnoeas may benefit from caffeine, if these apnoeas cause subclinical damage to brain. (Grade B) In asymptomatic (non-apnoeic) neonates the immediate side-effects of caffeine may outweigh potential long-term benefits. (Grade B)


  1. Henderson-Smart DJ, Steer P. Methylxanthine treatment for apnea in preterm infants. Cochrane Database Syst Rev 2001;3:CD000140.
  2. Janvier A, Khairy M, Kokkotis A, et al. Apnea is associated with neurodevelopmental impairment in very low birth weight infants. J Perinatol 2004;24:763–8.
  3. Zimmerberg B, Carr KL, Scott A, et al. The effects of postnatal caffeine exposure on growth, activity and learning in rats. Pharmacol Biochem Behav 1991;39:883–8.
  4. Desfrere L, Olivier P, Schwendimann L, et al. Transient inhibition of astrocytogenesis in developing mouse brain following postnatal caffeine exposure. Pediatr Res 2007;62:604–9.
  5. Juárez-Méndez S, Carretero R, Martínez-Tellez R, et al. Neonatal caffeine administration causes a permanent increase in the dendritic length of prefrontal cortical neurons of rats. Synapse 2006;60:450–5.
  6. Hoecker C, Nelle M, Poeschl J, et al. Caffeine impairs cerebral and intestinal blood flow velocity in preterm infants. Pediatrics 2002;109:784–7.
  7. Dani C, Bertini G, Reali MF, et al. Brain hemodynamic changes in preterm infants after maintenance dose caffeine and aminophylline treatment. Biol Neonate 2000;78:27–32.
  8. Soraisham AS, Elliott D, Amin H. Effect of single loading dose of intravenous caffeine infusion on superior mesenteric artery blood flow velocities in preterm infants. J Paediatr Child Health 2008;44:119–21.
  9. Soloveychik V, Bin-Nun A, Ionchev A, et al. Acute hemodynamic effects of caffeine administration in premature infants. J Perinatol 2009;29:205–8.
  10. Schmidt B, Roberts RS, Davis P, et al. Long-term effects of caffeine therapy for apnea of prematurity. N Engl J Med 2007;357:1893–902.
  11. Ment LR, Scott DT, Ehrenkranz RA, et al. Early childhood developmental follow-up of infants with GMH/IVH: effect of methylxanthine therapy. Am J Perinatol 1985;2:223–7.
  12. Gunn TR, Metrakos K, Riley P, et al. Sequelae of caffeine treatment in preterm infants with apnea. J Pediatr 1979;94:106–9.
  13. Stein MA, Krasowski M, Leventhal BL, et al. Behavioral and cognitive effects of methylxanthines. A meta-analysis of theophylline and caffeine. Arch Pediatr Adolesc Med 1996;150:284–8.
  14. Inder TE, Tjoeng TH, Vavasseur C, et al. Magnetic Resonance Imaging (MRI) and Neurodevelopmental Outcomes in Preterm Infants Following Administration of High-Dose Caffeine. Washington University School of Medicine ( identifier NCT 00809055). Due for completion in 2015.
  15. Davis PG, Schmidt B, Roberts RS, et al. Caffeine for Apnea of Prematurity Trial Group. Caffeine for Apnea of Prematurity trial: benefits may vary in subgroups. J Pediatr 2010;156:382–7.
  16. Back SA, Craig A, Luo NL, et al. Protective effects of caffeine on chronic hypoxia-induced perinatal white matter injury. Ann Neurol 2006;60:696–705.
  17. Steer P, Flenady V, Shearman A, et al. High dose caffeine citrate for extubation of preterm infants: a randomised controlled trial. Arch Dis Child Fetal Neonatal Ed 2004;89:F499–503.
  18. Henderson-Smart DJ, Steer P. Prophylactic methylxanthine for preventing of apnea in preterm infants. Cochrane Database Syst Rev 2000;2:CD000432.