Interventions to prevent hypothermia at birth in preterm and/or low birthweight babies.

BACKGROUND

Hypothermia incurred during routine postnatal resuscitation is a world-wide issue (across all climates), with associated morbidity and mortality. Keeping vulnerable preterm infants warm is problematic even when recommended routine thermal care guidelines are followed in the delivery suite.

OBJECTIVES

To assess efficacy and safety of interventions, designed for prevention of hypothermia in preterm and/or low birthweight infants, applied within 10 minutes after birth in the delivery suite compared with routine thermal care.

SEARCH STRATEGY

The standard search strategy of The Cochrane Collaboration was followed. Electronic databases were searched: MEDLINE (1966 to May Week 4 2004 ), CINAHL (1982 to May Week 4 2004), EMBASE (1974 to 09/07/04), the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 3, 2004), Database of Abstracts of Reviews of Effects (DARE 1994 to July 2004), conference/symposia proceedings using ZETOC (1993 to July 2004), ISI proceedings (1990 to 09/07/2004) and OCLC WorldCat (July 2004). Identified articles were cross-referenced. No language restrictions were imposed.

SELECTION CRITERIA

All trials using randomised or quasi-randomised allocations to test a specific intervention designed to prevent hypothermia, (apart from 'routine' thermal care) applied within 10 minutes after birth in the delivery suite to infants of < 37 weeks' gestational age or birthweight </=2500 g.

DATA COLLECTION AND ANALYSIS

Methodological quality was assessed and data were extracted for important clinical outcomes including adverse effects of the intervention by at least three independent reviewers. Authors were contacted for missing data. Data were analysed using RevMan 4.2.5. Relative risk (RR), risk difference (RD) and number needed to treat (NNT) with 95% confidence limits were calculated for each dichotomous outcome and mean differences (MD) with 95% confidence limits for continuous outcomes.

MAIN RESULTS

Six studies giving a total of 304 infants randomised and 295 completing the studies were included. Four comparisons to 'routine care' were undertaken within two categories: 1) barriers to heat loss (four studies): plastic wrap or bag (three), stockinet caps (one) and 2) external heat sources (two studies): skin-to-skin (one), transwarmer mattress (one). Plastic barriers were effective in reducing heat losses in infants < 28 weeks' gestation (three studies, n = 159; WMD 0.76 degrees C; 95% CI 0.49, 1.03) but not in the 28 to 31 week group. There was insufficient evidence to suggest that plastic wrap reduces the risk of death within hospital stay (three studies, n = 161; typical RR 0.63; 95% CI 0.32, 1.22; typical RD -0.09; 95% CI -0.20, 0.03). There was no evidence of a significant difference in major brain injury, mean duration of oxygen therapy or hospitalisation for infants < 29 weeks' gestation. Stockinet caps were not effective (borderline significant for infants < 2000 g birthweight) in reducing heat losses.Skin-to-skin care was shown to be effective in reducing the risk of hypothermia when compared to conventional incubator care for infants 1200 to 2199 g birthweight (one study, n = 31; RR 0.09; 95% CI 0.01, 0.64; NNT 2; 2 to 4). The transwarmer mattress significantly kept infants </=1500 g warmer and reduced the incidence of hypothermia on admission to NICU (one study, n = 24; RR 0.30; 95% CI 0.11, 0.83; NNT 2 range 2 to 4).

AUTHORS' CONCLUSIONS: Plastic wraps or bags, skin-to-skin care and transwarmer mattresses all keep preterm infants warmer, leading to higher temperatures on admission to neonatal units and less hypothermia. Given the low NNT, consideration should be given to using these interventions in the delivery suite. However, the small numbers of infants and studies and the absence of long term follow-up mean that firm recommendations for clinical practice cannot be given. There is a need to conduct large, high quality randomised controlled trials looking at long-term outcomes.