Setting A single, outpatient, general pediatric practice inHershey, Pennsylvania.

Participants 105 children, ages 2 to 18 years, with upperrespiratory tract infections, nocturnal symptoms, and illnessduration of fewer than 8 days.

Intervention A single dose of buckwheat honey, honey-fla-vored DM, or no treatment administered 30 minutes beforebedtime.

Outcomes Cough frequency, cough severity, bothersomenature of cough, and child and parent sleep quality.

Main results Significant differences in symptom improve-ment were detected between treatment groups, with honeyconsistently scoring the best and no treatment scoring theworst. In paired comparisons, honey was significantly superiorto no treatment for cough frequency (P � .01) and thecombined symptom score (P � .04), but DM was not betterthan no treatment for any outcome. Comparison of honeywith DM revealed no significant differences.

Conclusions Parents rated honey most favorably for symp-tomatic relief of their child’s nocturnal cough and sleep dif-ficulty caused by upper respiratory tract infection.

Comment In this well-designed and valid study, Paul etal were able to show that honey was significantly superiorto no treatment for improvement in cough severity (47.3%reduction vs 24.7%) and an overall symptom score (53.7%reduction vs 33.4%). The findings of this study suggest thathoney is better than no treatment for reducing coughfrequency and improving combined symptom scores. A keystrength of this study is that the investigators were partic-ularly thorough in their attempts to maintain blinding andwere diligent in treating groups in a similar fashion asidefrom the interventions. However, one limitation is the lackof information on the validity and reliability of the coughquestionnaire. Despite this, the study offers an interestingalternative to traditional over-the-counter remedies forcough in children. Further randomized controlled trialscould explore whether improvement is seen with varietiesof honey other than buckwheat and whether consumptionof other sweet liquids results in the same improvement asthat seen after administration of honey. Many pediatriciansare reluctant to prescribe remedies such as dextromethor-phan, given the lack of evidence for its efficacy and thepotential for adverse effects. Some may now view honey asa possible alternative for the treatment of cough for chil-dren over 1 year of age. Other clinicians may not prescribeany therapy for the treatment of symptoms of upper respi-ratory tract infections; for those clinicians, the results ofthis article may be less applicable.

Michael D. Warren, MDWilliam O. Cooper, MD, MPH

Vanderbilt University School of MedicineNashville, Tennessee

Use of caffeine for apnea of prematurity alsohas long-term neurodevelopmental benefitsSchmidt B, Roberts RS, Davis P, Doyle LW, Barrington KJ,Ohlsson A, et al, for the Caffeine for Apnea of PrematurityTrial Group. Long-term effects of caffeine therapy for apneaof prematurity. N Engl J Med 2007;357:1893-902

Question Among preterm infants, does treatment with caf-feine for apnea of prematurity have any long-term effects onneurodevelopment and growth?

Design Randomized controlled trial.

Setting Multiple centers in the United States, Canada,Europe, Australia, and Israel.

Participants 2006 infants with birth weights of 500 to 1250 g.

Intervention Infants received either caffeine or placebo untiltherapy for apnea of prematurity was no longer needed. Par-ticipants were followed to a corrected age of 18 to 21 months.

Outcome A composite of death, cerebral palsy, cognitivedelay (defined as a Mental Development Index score of �85on the Bayley Scales of Infant Development), deafness, orblindness.

Main results Of the 937 infants assigned to caffeine forwhom adequate data on the primary outcome were available,377 (40.2%) died or survived with a neurodevelopmentaldisability, as compared with 431 of the 932 infants (46.2%)assigned to placebo for whom adequate data on the primaryoutcome were available (odds ratio adjusted for center, 0.77;95% confidence interval [CI], 0.64 to 0.93; P � .008, numberneeded to treat [NNT] � 17). Treatment with caffeine ascompared with placebo reduced the incidence of cerebral palsy(4.4% vs 7.3%; adjusted odds ratio, 0.58; 95% CI, 0.39 to0.87; P � .009, NNT � 35) and of cognitive delay (33.8% vs38.3%; adjusted odds ratio, 0.81; 95% CI, 0.66 to 0.99; P �.04, NNT � 23). The rates of death, deafness, and blindnessand the mean percentiles for height, weight, and head cir-cumference at follow-up did not differ significantly betweenthe 2 groups.

Conclusions Caffeine therapy for apnea of prematurity im-proves the rate of survival without neurodevelopmental disabilityat 18 to 21 months in infants with very low birth weight.

Comment Methylxanthines have been a mainstay oftreatment for apnea of prematurity for over 3 decades.They reduce the frequency of apnea episodes, but evidenceof favorable effects on long-term outcomes has been lack-ing. Central nervous system effects might actually adverselyimpact neurological development. This report, along withthe short-term outcomes from the same trial, lays thoseconcerns to rest. More than 2000 infants were enrolled inthis multicenter, randomized, placebo-controlled trial. In-fants treated with caffeine were extubated, weaned fromcontinuous positive air pressure, and weaned from supple-mental oxygen sooner, and fewer required oxygen at 36

740 The Journal of Pediatrics • May 2008

weeks gestation. At 18 to 21 months of age, fewer treatedinfants had died or survived with neurodevelopmental im-pairment, with an estimated number needed to treat of 17to avoid one such outcome. In particular, rates of cerebralpalsy and cognitive delay were reduced. Because of thesedata, we now know that our time-honored reliance onmethylxanthines for management of apnea of prematuritywas not misguided, and in fact appears to be beneficial.Moreover, we know who (infants with birth weights of 500to 1250 g), when (within 10 days after birth), how (loadingdose of 20 mg/kg of caffeine citrate followed by daily

maintenance doses of 5 mg/kg, increased to 10 mg/kg ifapneas persisted), and how long (until apneas resolve, or tomedian gestational age of 35 weeks) to treat, and thatmonitoring of caffeine levels is unnecessary. Criteria forinitiation of treatment (which were “treatment or preven-tion of apnea or to facilitate endotracheal tube removal” inthis trial) remain imprecise.

William E. Benitz, MDStanford University School of Medicine

Palo Alto, California

Clinical Research Abstracts for Pediatricians 741