Kangaroo mother care for low birthweight infants

This review compares the effectiveness of the Kangaroo Mother Care (KMC) method with conventional care in reducing neonatal mortality and severity of illness and improving infant growth and development in low-birth-weight (LBW) infants. Conventional care in this review included thermal care of the infant by incubator and use of thermal cots, bulb-heated cots or warm rooms with open cots.

Fourteen trials of KMC for LBW infants were identified. Eleven trials were excluded: seven because they were non-randomized, one because allocation was by alternation, one because the intervention KMC was a combination of skin-to-skin contact and warmer/incubator, one because the intervention KMC was only skin-to-skin contact, and one because the intervention KMC was intermittent skin-to-skin contact and 20 out of 34 enrolled infants had birth weights >2500g. Three randomized controlled trials conducted in developing countries involving 1362 infants were included for meta-analysis (1),(2),(3),(4). All of these trials included babies <2000 g in the study. The reviewers felt that the inadequacies in the methods used in these trials weakened the credibility of their results.

Meta-analysis of the three trials revealed that KMC was associated with a significant reduction in the risk of nosocomial infection at 41 weeks corrected age and severe illness, including lower respiratory infection at six months. Infants in the KMC group gained significantly more weight per day at discharge from hospital, although this was of low clinical significance (3.6 g/day, 95% CI 0.8-6.4). There were no differences in the rates of mortality in the two groups. KMC was also associated with reduced likelihood of not exclusively breastfeeding at discharge from hospital (Relative Risk: 0.41, 95% confidence interval 0.25 to 0.68), but there was no difference in exclusive breastfeeding rates at one month and at six months of follow-up. In the KMC group, fewer mothers were likely to be dissatisfied with the method of care of their newborn babies (RR: 0.41, 95% CI 0.22-0.75). Psychomotor development was similar at one year of age.

Since most of the outcomes summarized above are derived from single studies, the results of this meta-analysis are not robust. Moreover, the three studies included in the meta-analyses were not controlled for duration of intervention and for co-interventions. Stabilization was poorly defined, which may have been a critical variable in very-low-birth-weight- (<1500 g) infants. The reviewers also could not carry out a planned subgroup analysis according to birth weight and gestational age due to paucity of data.

The search and analysis strategies of the review are appropriate. The data were analysed appropriately. The limitations of the review are primarily due to the restricted data and information available from the trials. For example, babies with birth weights <2000 g are a heterogeneous population made up of both preterm and growth-impaired neonates whose prognosis is distinctly different. Therefore, stratification would have been most useful, as the reviewers have pointed out. The reviewers also mention co-interventions, but do not elaborate on them. The inability to control for “conventional care” could possibly have confounded the results, as conventional care was non-standardized unlike the KMC method. Lastly, for the control group the review does not provide detailed information on the nature and intensity of maternal involvement in neonatal care.

The full RHL commentary also includes sections on: