Use of erythromycin for preterm feeding intolerance needs better assessment of risk
Article Outline
- Nuntnarumit P, Kiatchoosakun P, Tantiprapa W, Boonkasidecha S. Efficacy of oral erythromycin for treatment of feeding intolerance in preterm infants. J Pediatr 2006;148:600-5
- Reference
- Copyright
Nuntnarumit P, Kiatchoosakun P, Tantiprapa W, Boonkasidecha S. Efficacy of oral erythromycin for treatment of feeding intolerance in preterm infants. J Pediatr 2006;148:600-5
Question Among preterm infants <35 weeks gestation with feeding intolerance, is oral erythromycin (EM) safe and efficacious?
Design Randomized, double-blinded, placebo-controlled trial.
Setting Neonatal care nurseries of 3 university hospitals in Thailand.
Participants 46 preterm infants with feeding intolerance.
Intervention EM ethyl succinate 10 mg/kg every 6 hours for 2 days, followed by 4 mg/kg every 6 hours for another 5 days.
Outcome Time to full feeding (150 mL/kg/d).
Main Results Times to full feeding were significantly shorter in the intervention group, with a median (interquartile ranges) of 7 days (6 to 9 days) versus 13 days (9 to 15 days) (P < .001). In addition, the number of withheld feeds was significantly less in the EM group, with a median of 1 episode (0 to 2 episodes) versus 9 episodes (2 to 13 episodes) (P < .001). No significant differences in episodes of sepsis, necrotizing enterocolitis, and cholestasis were observed.
Conclusions Oral EM was effective and safe for treatment of feeding intolerance in preterm infants.
Commentary This is a well designed and reported trial. The authors clearly define their subject population and give clear definitions of important terms and protocols. They also include a discussion of sample size estimate and randomization. The authors found that preterm infants with feeding intolerance treated with erythromycin were able to attain full feeds sooner than control infants. What is a practicing clinician to do when presented with a statistically significant (P< .05) difference in Kaplan-Meier survival curves, as well as the hazard ratio from the Cox regression analysis as estimates of risk? Should practice change? How can a clinician translate this information into clinically meaningful terms? Calculating the number to treat (NNT) is one option.1 Using the Kaplan-Meier curves and text, the proportion of erythromycin-treated and placebo control–treated infants at full feeds at 14 days is 1 and 0.5, respectively. This gives an absolute risk reduction of 0.5 and an NNT of 2. This means that treating 2 premature infants who have feeding intolerance with erythromycin will result in 1 more infant who will attain full feeds by 14 days. This seems impressive. However, it is not clear whether the reduction in time to full feeds has a clinical impact on the number of days of parenteral nutrition and its associated complications or length of hospital stay. Hence, one must ask whether this potential benefit is worth the risk. The authors state in their discussion that the study was not powered to detect important secondary outcomes, such as NEC. Decisions to change practice are based on evaluations of both benefits and risks. Thus for fear of doing harm (in this case as yet undefined—which is scary indeed), practice is not likely to change. We agree with the authors that the recommendation for the routine use of erythromycin for feeding intolerance should await confirmation from larger clinical trials powered to detect important secondary outcomes.
Reference
PII: S0022-3476(06)00810-9
doi:10.1016/j.jpeds.2006.08.037
© 2006 Mosby, Inc. All rights reserved.
