Home > Health & Research > Health Education Campaigns & Programs > Cochrane Neonatal Review > Radiant warmers versus incubators for regulating body temperature in newborn infants

Radiant warmers versus incubators for regulating body temperature in newborn infants

Skip sharing on social media links
Share this:

Authors

Flenady VJ, Woodgate PG

Background - Methods - Results - Characteristics of Included Studies - References - Data Tables & Graphs


Dates

Date edited: 22/02/2005
Date of last substantive update: 17/06/2003
Date of last minor update: 28/01/2005
Date next stage expected 01/08/2006
Protocol first published: Issue 2, 1997
Review first published: Issue 1, 1998

Contact reviewer

Vicki Flenady

Centre for Clinical Studies-Women's and Children's Health
Women's and Children's Health Service
Mater Health Services Brisbane
Raymond Terrace
South Brisbane
Queensland AUSTRALIA
4101
Telephone 1: +61 7 38401591
Facsimile: +61 7 38401588

E-mail: vflenady@mater.org.au

Secondary contact person's name: Peter Gray

Contribution of reviewers

The reviewers worked collaboratively and equally on each stage of the review.

Sources of Support

Internal sources of support

Centre for Clinical Studies - Women's and Children's Health, Mater Hospital, South Brisbane, Queensland, AUSTRALIA
J P Kelly Research Foundation, Mater Hospital, South Brisbane, Queensland, AUSTRALIA
Department of Neonatology, Mater Mothers' Hospital, South Brisbane, Queensland, AUSTRALIA

External sources of support

Commonwealth Department of Health and Ageing, Canberra, Australian Capital Territory, AUSTRALIA

What's new

This review updates the existing review "Radiant warmers versus incubators for regulating body temperature in newborn infants" published in The Cochrane Library, Issue 1, 1998 and substantively updated in Issue 3, 2003 (Flenady 2003) with additional data for two of the included studies incorporated into the review and one trial which was awaiting assessment subsequently excluded.

No new trials were identified in the search updated to January 2005, and as a result no substantive changes were made to the review

Dates

Date review re-formatted: / /
Date new studies sought but none found: 13/01/2005
Date new studies found but not yet included/excluded: / /
Date new studies found and included/excluded: / /
Date reviewers' conclusions section amended: / /
Date comment/criticism added: / /
Date response to comment/criticisms added: / /

Synopsis

Not enough evidence to show the effects of radiant warmers versus incubators for regulating body temperature in newborn babies.

Low birthweight babies have a higher chance of survival if they are kept warm. Incubators have been used for some time to maintain body temperature. More recently, open cots with an overhead radiant warmer have also been used for babies needing intensive care.

The review of trials found that radiant warmers increase water loss in low birthweight babies in the newborn period when compared to incubators and that this water loss needs to be taken into account when daily fluid requirements are calculated. However, there was not enough information available for this review to enable assessment of other important effects of radiant warmers. Therefore, at the present time, it is not clear which method of maintaining body temperature is best for newborn babies - radiant warmers or incubators. More research is necessary.

Abstract

Background

The provision of a thermoneutral environment is an essential component of the immediate and longer term care of newborn infants. A variety of methods are currently employed including incubators and open-care systems, with or without modifications such as heat shields and plastic wrap. The system used must allow ready access to the infant but should also minimise alterations in the immediate environment.

Objectives

To assess the effects of radiant warmers versus incubators on neonatal fluid and electrolyte balance, morbidity and mortality.

Search strategy

The standard search strategy of the Cochrane Neonatal Review Group was used. This included searches of electronic databases: Oxford Database of Perinatal Trials, Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 1, 2005), MEDLINE (1966 -2005), and CINAHL (1982-2005), previous reviews including cross references, abstracts, conferences, symposia proceedings, expert informants and journal hand searching mainly in the English language.

Selection criteria

Randomised or quasi-randomised trials in which radiant warmers were compared to incubators in a neonatal population.

Data collection & analysis

Independent data extraction and quality assessment of included trials was conducted by the authors. Data were analysed using relative risk (RR) and weighted mean difference (WMD). Results are presented with 95% confidence intervals. Meta-analysis was undertaken using a fixed effect model.

Main results

Eight studies are included in this review; six employed a crossover design. In the overall comparison of radiant warmers vs incubators, radiant warmers caused a statistically significant increase in insensible water loss (IWL) [WMD 0.94g/Kg/day (95% CI 0.47, 1.41)] and a trend towards increased oxygen consumption which was not statistically significant [WMD 0.27mL/kg/min (95% CI -0.09, 0.63)]. Due to small numbers, effects on important clinical outcomes could not be adequately assessed. A comparison of radiant warmers with heat shields vs incubators without heat shields showed a trend for increased IWL in the radiant warmer group, which was not statistically significant. No difference was shown in oxygen consumption.

Reviewers' conclusions

Radiant warmers result in increased IWL compared to incubators. This needs to be taken into account when calculating daily fluid requirements. The results of this review do not provide sufficient evidence concerning effects on important outcomes to guide clinical practice. Further randomised controlled trials are required to assess the effects of radiant warmers versus incubators in neonatal care on important short and long term outcomes, with particular attention to extremely low birthweight infants in the early neonatal period.

[top]

Background

An understanding of neonatal thermoregulation led to significant decreases in morbidity and mortality in neonates, especially sick preterm infants requiring intensive care. In a series of randomised trials published in the late 1950's and early 1960's Silverman and colleagues demonstrated improved survival of small infants nursed in warmer environments (Silverman 1957; Silverman 1958; Silverman 1963). Maintaining a thermoneutral environment in the case of very low birthweight infants is now routine practice in neonatal nurseries.

Since the first incubators were used, many advances in design have been made. Nursing in closed incubators was initially standard practice and provided reasonable control over the neonate's immediate environment. Subsequently, open care cots incorporating radiant warmers were introduced to provide easier access to the baby receiving intensive care (LeBlanc 1991; Sinclair 1992).

Evaporative heat and water losses may be higher in infants nursed under radiant warmers than in incubators. These effects are of major importance for the very preterm infant especially in the first week of life (LeBlanc 1991). In a study observing hand washing practices in neonatal intensive care, although no difference was shown in handwashing practice, neonates nursed under radiant warmers were shown to have twice as many nurse interactions (mostly non emergency) than neonates nursed in an incubator (Davenport 1992). This suggests neonates under radiant warmers may be exposed to frequent handling which has been shown to increase oxygen and energy consumption and cause bradycardic episodes (Long 1980; Gorski 1990; Hutchinson 1994). Other noxious stimuli such as noise and light may be increased, possibly disrupting sleep patterns. However this exposure may have the advantage of providing the primary caregiver immediate unobstructed access and may also promote parent-infant attachment. These effects may have an impact on both short and long term outcomes.

Because of the sensitivity of sick neonates to alterations in their environment, the influence of different methods of controlling the environment on morbidity and mortality needs to be considered.

Objectives

To assess the effects of radiant warmers versus incubators on neonatal fluid and electrolyte balance, morbidity and mortality.

Planned sub-group analyses:

  • With or without added humidification
  • With or without the use of additional methods eg heat shields, plastic wrap
  • Duration of treatment - from birth to two weeks of age
  • Preterm < 28 weeks gestation or birthweight < 1000 gms
  • Preterm < 32 weeks gestation or birthweight < 1500 gms

Criteria for considering studies for this review

Types of studies

Randomised or quasi-randomised trials in which radiant warmers were compared to incubators.

Types of participants

Neonates nursed under radiant warmers vs incubators in neonatal nurseries.

Types of interventions

  • Radiant warmers
  • Incubators

Types of outcome measures

Primary:
  • Mortality
  • Oxygen consumption
  • Metabolic rate
  • Insensible water loss
  • Fluid and electrolyte balance
  • Patent ductus arteriosus
  • Time to regain birthweight
  • Bacterial colonisation
  • Infection
  • Necrotising enterocolitis*
  • Intraventricular haemorrhage (All grades)
  • Intraventricular haemorrhage (Grades 3 and 4)
  • Cerebral cystic lesions (Periventricular leukomalacia, porencephalic cysts)]
  • Chronic lung disease [infant receiving any respiratory support (supplemental oxygen or any form of assisted ventilation) for a chronic pulmonary disorder i) on the day they reached 36 weeks' post menstrual age; and ii) at 28 days postnatal age].
Secondary:
  • Length of hospital stay
  • Parent satisfaction
  • Staff satisfaction
  • Retinopathy of prematurity (All stages) *
  • Retinopathy of prematurity (Stages III and IV)*
  • Long term neurosensory impairment

NB: In the first update of this review, the reviewers changed the criteria for considering studies - Types of outcome measures - by adding three eligible outcomes (indicated by *).

Search strategy for identification of studies

The standard search strategy of the Neonatal Review Group was used. Searches were made of the Oxford Database of Perinatal Trials, Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library, Issue 1, 2005), MEDLINE (1966 -2005), and CINAHL (1982-2005). Text terms used included: incubator, isolette, radiant*, open warmer and MeSH term infant-newborn. The search strategy also included previous reviews including cross references, abstracts in conferences and symposia proceedings, journal hand searching in the English language.

[top]

Methods of the review

Standard methods of the Cochrane Collaboration were used as described in the Cochrane Handbook.

Quality assessment:

Included trials were assessed for blinding of randomization, blinding of intervention, completeness of follow-up, and blinding of outcome assessment. The authors independently undertook this assessment and assigned a rating of either Yes (Adequate), Can't Tell (Unclear), or No (Inadequate) for each. Differences were resolved by discussion.

Methods used to collect and synthesise data from included studies:

Each reviewer independently extracted data, then compared and resolved differences. Additional information concerning outcomes and method of randomisation was sought and received from the authors of three included trials (Bell 1980; Schnabel 1999; Meyer 2001).
For further details, please see table, Characteristics of Included Studies.

Three crossover studies employed more than one crossover (Bell 1979; Bell 1980; Marks 1986). (For details please see section, Description of studies). In the overall analysis, only one crossover (two exposures) were included from these studies as follows: infants nursed naked in radiant warmers and incubators in Bell 1979; radiant warmers and incubators without the use of heat shields in Bell 1980; and radiant warmers and incubators without phototherapy in Marks 1986.

Subgroup analyses were conducted to assess the effects of heat shields in each environment. Two trials were included in the comparisons on the use of heat shields (Bell 1980; Meyer 2001).

Categorical data were analysed with the use of relative risk (RR), and weighted mean differences (WMD) were used for outcome data measured on a continuous scale. All outcomes were analysed with the use of 95% confidence intervals (CI). Meta-analysis was conducted using a fixed effect model.

Description of studies

Eight studies were included in this review. Six employed a crossover design (Darnall 1978; Bell 1979; Bell 1980; Marks 1980; LeBlanc 1982; Marks 1986) and two did not (Schnabel 1999; Meyer 2001). A further seven studies were identified. Six of these were excluded as neither random nor quasi-random allocation to the exposure was employed (Levison 1966; Wu 1974; Williams 1974; Jones 1976; Wheldon 1982; Kjartansson 1995) and one trial (Merenstein 1979) was excluded following correspondence with the author as data were not available in a format for inclusion in this review.

Participants in the included trials were similar. All studies enrolled preterm neonates (mean gestational ages 28-32 weeks, mean birthweights 1.1 - 1.6 kg across the trials). Darnall 1978 also included term neonates. The two more recent trials studied a less mature population (Schnabel 1999; Meyer 2001). The age at which the neonates were studied varied; however, the majority of infants were greater than seven days of age. One study enrolled infants shorty after birth on admission to the nursery (Meyer 2001). The duration of the exposure in the crossover trials ranged from one hour to three days (most were between one and three hours). Two non-crossover trials studied infants in the two environments for longer periods - one study from seven to 35 days of life (Schnabel 1999) and the other from admission to a weight of 1800g (Meyer 2001).

The number and types of exposures studied differed in the included crossover studies. Three crossover trials studied infants in each of the two environments (radiant warmers and incubators) without additional thermal measures (Darnall 1978; LeBlanc 1982; Marks 1980). Three trials reported additional exposures. One study (Bell 1979) assessed the effects of clothing and included three exposures: naked in incubators, naked in radiant warmers and clothed in incubators. Another (Marks 1980) studied infants in four environments: radiant warmers and incubators with and without phototherapy. Bell 1980 compared the effects of heat shields in the two environments (four exposures).

One non-crossover trial (Meyer 2001) assessed the use of heat shields with radiant warmers compared to incubators without heat shields.

Double-walled incubators were used in two trials (Schnabel 1999; Meyer 2001), while single-wall incubators were used in all six crossover trials. In all studies, attempts were made to achieve a thermal state associated with minimal resting oxygen consumption in each environment. Added humidification was mentioned in four trials (LeBlanc 1982; Marks 1986; Schnabel 1999; Meyer 2001). Five studies reported a relative humidity of approximately 30%. Two trials reported much higher relative humidity of 50 - 80% for both radiant warmer and incubator groups (LeBlanc 1982; Meyer 2001). The infant's temperature was maintained with the use of servo-control in the majority of the included trials.

The most frequently reported outcomes in the crossover trials were insensible water loss (three trials) and oxygen consumption (five trials). One crossover trial also reported weight gain, and skinfold fat gain (Marks 1986). The two non-crossover trials assessed temperature control, fluid and electrolyte balance, weight gain, major neonatal morbidity and mortality (Schnabel 1999; Meyer 2001). In these trials, the outcome of chronic lung disease was reported by different definitions: oxygen use at 28 days postnatal age (Schnabel 1999) and at 36 weeks corrected age (Meyer 2001). The reviewers decided due to small numbers of infants studied to report this outcome according to one definition only and chose the definition of oxygen use at 36 weeks' post menstrual age.

(For further details on included studies see table, Characteristics of Included Studies).

Methodological quality of included studies

Six of the eight included studies used a crossover design. Random allocation of the order of exposure was undertaken in four crossover studies (Bell 1979; Bell 1980; LeBlanc 1982; Marks 1986). Whether blinding of randomisation was undertaken is unknown. Quasi-random allocation was used in two studies (Darnall 1978; Marks 1980 ). A blinded method of random allocation was undertaken in the two non-crossover trials (Schnabel 1999; Meyer 2001). Meyer 2001 used a numbered card with the allocation concealed beneath an adhesive label; numbered cards were kept in a locked cupboard until the point of randomisation and Schnabel 1999 used sealed envelopes. Blinding of the intervention was not possible. Blinding of outcome measure was reported by one trial Schnabel 1999 following correspondence with the author. All neonates enrolled in the included studies were included in the analysis apart from Marks 1986 (four of the 17 enrolled infants were excluded due to illness), Schnabel 1999 (four of 34 excluded - one death and three transfers), and Meyer 2001 (one exclusion of 61 enrolled due to no parental consent).

[top]

Results

The results of six crossover trials with a total of 66 infants and two non-crossover trials in 90 infants are included in this review.

Radiant warmers versus incubators

In the overall analysis comparing radiant warmers and incubators, a statistically significant increase in insensible water loss (IWL) was shown for neonates nursed under radiant warmers [WMD 0.94 g/kg/day, (95% CI 0.47, 1.41)] when compared to incubators. There was a trend towards increased oxygen consumption which was not statistically significant for infants nursed under radiant warmers [WMD 0.27 mL/kg/min (95% CI -0.09, 0.63)].

The results of one cross-over trial (Marks 1986) showed no difference in metabolic rate [MD 0.13 kcal/kg/h (95% CI -0.26, 0.52)]. However, this trial showed a statistically significant increase in skinfold fat gain for infants nursed under radiant warmers [MD 0.04 mm/d (95% CI 0.01, 0.07)]. No statistically significant difference was shown in weight gain over the study period [WMD 1.06 g/kg/d (95% CI -0.94, 3.06)] (Marks 1986; Meyer 2001) or time to regain birthweight [WMD 0.86 days (95% CI -1.49, 3.21)] (Meyer 2001; Schnabel 1999).

A non-statistically significant reduction in chronic lung disease (defined as oxygen use at 36 weeks postmenstrual age) [RR 0.20 (95% CI 0.01, 4.00)], and an increase in infants with serum sodium >150mmol/L which was also not statistically significant [RR 3.00 (95% CI 0.66, 13.69)], were shown for infants nursed under radiant warmers (data from one trial, Meyer 2001).

No differences were shown in the rates of other reported neonatal morbidities (infection, necrotising enterocolitis, patent ductus arteriosus, retinopathy of prematurity, intraventricular haemorrhage, cerebral cystic lesions, or death). However, due to the small numbers of infants studied, all measures of effect for these outcomes are imprecise.

Subgroup analyses by use of heat shields

  1. Radiant warmers vs incubators - without the use of heat shields
    In this comparison, as shown in the overall analysis, a statistically significant increase in insensible water loss (IWL) [WMD 0.94 g/kg/day (95% CI 0.47, 1.41)] and a trend towards increased oxygen consumption [WMD 0.27 mL/kg/min (95% CI -0.09, 0.63)] which was not statistically significant was shown for infants nursed under radiant warmers when compared to infants nursed in incubators without the use of heat shields. No statistically significant differences were shown for the outcomes of time to regain birthweight, infection, necrotising enterocolitis, patent ductus arteriosus, retinopathy of prematurity, cerebral cystic lesions, or neonatal death (data from one small trial of 34 infants, Schnabel 1999).
  2. Radiant warmers with heat shields vs incubator without heat shields
    This comparison showed a similar trend to that of the overall analysis for increased IWL in the radiant warmer group which was not statistically significant [WMD 1.00 g/kg/day (95% CI -0.10, 2.10)]. No difference was shown in the rate of oxygen consumption when radiant warmers with heat shields were compared to incubators [WMD -0.05 mL/kg/min (95% CI -0.84, 0.74)]. Data from one trial of sixty infants (Meyer 2001) showed non-statistically significant reductions for infants nursed under radiant warmers in chronic lung disease, necrotising enterocolitis and intraventricular haemorrhage (Grades 3 and 4), an increase in infants with retinopathy of prematurity and serum sodium > 150 mmol/L, and no difference between groups in time to regain birthweight or weight gain over the study period. However, due to the small numbers of infants studied, adequate assessment of these outcomes is not possible.
  3. Radiant warmers without heat shields vs incubators with heat shields
    Only two outcomes were able to be assessed in this comparison: IWL and oxygen consumption. A statistically significant increase in IWL [WMD 1.27 g/kg/day (95% CI 0.02, 2.52)] was shown for the infants nursed under radiant warmers. Oxygen consumption was not statistically significantly different [WMD -0.05 mL/kg/min (95% CI -0.89, 0.79).
  4. Radiant warmers with heat shields vs incubators with heat shield
    Once again, only two outcomes were able to be assessed in this comparison. A statistically significant increase in IWL for radiant warmers [WMD 1.24 g/kg/day (95% CI 0.30, 2.18)] and a trend towards a reduction in oxygen consumption was shown for the radiant warmer group [WMD -0.45 mL/kg/min (95% CI -1.33, 0.43)], again with no statistical significance.

Due to insufficient data, planned sub-group analyses by duration of treatment (birth to two weeks of age), added humidification and lower gestational age could not be conducted.

Discussion

This review includes eight trials which assessed short term effects of radiant warmers vs incubators. A statistically significant increase was shown in IWL for the infants nursed under radiant warmers. The mean increase in IWL was 0.94 mL/kg/h or 22.6 mL/kg /day. This effect needs to be considered when calculating daily fluid requirements. A non significant trend for increased oxygen consumption in neonates nursed under radiant warmers was shown. The crossover design used by most studies included in this review has the limitation of being able to assess immediate effects of the intervention only. There is limited information available in this review on more meaningful clinically important outcomes, as only two trials did not use a cross-over design.

One trial (Marks 1986) reported no statistically significant differences in urine output, specific gravity and osmolality in low birthweight neonates aged 28 days (mean) nursed under radiant warmers compared to incubators for a three day period with similar fluid intake. Meyer 2001 reported an increase in the number of infants with serum sodium levels > 150 mmol/L in the radiant warmer group, although this was not statistically significant. Data from two trials showed no difference in weight gain over the study period in infants nursed under radiant warmers compared to those nursed in incubators (Marks 1986; Meyer 2001) and no difference in time to regain birthweight in two trials (Schnabel 1999; Meyer 2001). One trial also showed no difference in metabolic rate over a three day period although a statistically significant increase in skinfold fat gain was shown for infants nursed under radiant warmers was shown (Marks 1986).

The results of the two non-crossover trials in this review showed no difference in important clinical outcomes of major neonatal morbidity and mortality (Schnabel 1999; Meyer 2001). However, Meyer 2001 reported an overall trend to less neonatal morbidity for infants under radiant warmers with heat shields when compared to infants in an incubator without heat shields and postulated that this may have been due to less hypothermia demonstrated in the immediate newborn period for infants nursed under radiant warmers. Due to small numbers of infants studied, it is not possible to adequately assess these effects.

The findings from the comparisons assessing the effects of heat shields come from one small crossover trial (Bell 1980) and one small non-crossover trial (Meyer 2001). In these comparisons there is a suggestion that the use of heat shields with radiant warmers may reduce oxygen consumption when compared to incubators; no differential effect was shown for IWL. There is some evidence to suggest that heat shields may diminish transmission of radiant heat to the infant's skin (Baumgart 1982). There is insufficient evidence in this review of the benefits and possible harm of heat shields to guide clinical practice. This review was unable to carry out the planned sub-group analyses limited to high risk infants of very short gestation or very low birth weight, or according to use (or not) of humidification. The effects of radiant warmers and incubators in the early newborn period, when infants (particularly those of extremely low birthweight) are most sensitive to environmental conditions, was unable to be assessed.

Further randomised controlled trials in which infants are allocated to radiant warmers or incubators in the period immediately following admission to the neonatal nursery are necessary. Specific issues to be addressed by further research include:

  • important thermal and non thermal effects, ie neonatal mortality, neonatal morbidity including chronic lung disease, nosocomial infection, sleep disturbance and parental perceptions
  • the effects in neonates less than seven days of age, particularly extremely low birthweight neonates
  • the effects of extra measures to modify the environment, eg humidification, 'bubble plastic' and plastic heat shields

Reviewers' conclusions

Implications for practice

Radiant warmers result in an increased insensible water loss when compared to incubators. The results of this review do not provide sufficient evidence regarding clinically important effects to guide clinical practice.

Implications for research

Further randomised controlled trials are required to assess the effects of radiant warmers versus incubators in neonatal care with particular attention to the extremely low birthweight population in the early neonatal period.

Acknowledgements

The reviewers would like to thank Dr Merenstein, Dr Bell, Dr Myer and Karin Schnabel for providing additional information on their trials and also Prof David Henderson-Smart (Cochrane Neonatal Group Regional Coordinator, Australasia) for advice on this review and Zsuzsoka Kecskes for assistance with translation of an included trial.

We would also like to acknowledge Katie Welsh for assistance with literature searching and formatting the review.

Potential conflict of interest

  • None noted.

[top]

Characteristics of studies

Characteristics of Included Studies

Study Methods Participants Interventions Outcomes Notes Allocation concealment
Bell 1979 Blinding of randomization - can't tell
Blinding of intervention -no
Complete followup - yes
Blinding of outcome measure -no
10 infants studied. Mean birthweight: 1.24 kgs, gestational age: 30 weeks, mean age at entry: 12 days, mean weight at entry: 1.15 kg

Crossover design. Each infant was studied for a 2 hour period in four conditions with a one hour pause in between.

  1. Incubator, no phototherapy
  2. Radiant warmer, no phototherapy
  3. Radiant warmer with low-dose phototherapy
  4. Radiant warmer with high-dose phototherapy
Insensible water loss, heart rate, respiratory rate Temperature control: servo controlled in both environments to maintain rectal temperature 36.8 - 37.1C. Infants were studied naked.
Mean skin temperatures for both RW and INC groups - 36.4C
Relative humidity: RW - 29.4%, INC - 26.4%. (p value < 0.025)
B
Bell 1980 Blinding of randomization - can't tell (coin toss)
Blinding of intervention -no
Complete followup - yes
Blinding of outcome measure -no
8 infants studied. Mean birthweight: 1.57 kg, gestational age: 32 weeks, weight on study entry:1.47 kg, age when studied: 12.2 days

Crossover design. Each infant was studied during four consecutive two-hour periods in four different conditions.

  1. single wall incubator
  2. incubator with heat shield
  3. radiant warmer
  4. radiant warmer with a heat shield
Insensible water loss, oxygen consumption, heart rate, respiratory rate Temperature control: servo controlled to keep rectal temperature between 36.8 and 37C in all groups. Infants nursed naked. Mean skin temperature: RW - 36.9, INC - 37C
Relative humidity: RW - 39%, INC - 31.4% ( p value < 0.01)
B
Darnall 1978 Blinding of randomization - can't tell
Blinding of intervention -no
Complete followup - yes
Blinding of outcome measure -no
10 infants studied. Birthweight range: 820-3800 gm, gestational age range: 28 - 40 weeks. Age range when studied: 4 hrs - 30 days

Crossover design. Infants were studied from one to three hours in each of two environments.

  1. incubator
  2. radiant warmer
Oxygen consumption, carbon dioxide production, skin temperature Temperature control: Servo control in both environments to maintain skin temperature 36.1 36.8C. Mean skin temperatures: RW - 36.4, INC - 36.5C
Relative humidity or infant clothing in either environment was not mentioned
C
LeBlanc 1982 Random allocation not described "order randomised".
Blinding of randomization - can't tell
Blinding of intervention -no
Complete followup - yes
Blinding of outcome measure -no
Entry criteria: low birthweight, feeding and gaining weight
16 infants studied. Mean birthweight 1.212 kg, mean gestational age: 31 weeks, mean age when studied: 22 days
Crossover design. Infants were studied on a single day in four environments. Two environments are reported in this publication - radiant warmer and incubator groups Oxygen consumption, body temperature Temperature control:servo controlled in both environments to maintain skin temperature 36C. Mean skin temperatures: RW - 35.99, INC - 36.04C (not significant). Infants were nursed naked
Relative humidity: RW - 55.2%, INC (set at maximum humidification) - 75.9%. (p value < 0.001)
B
Marks 1980 Quasi-random allocation: "order alternated".
Blinding of randomization - can't tell
Blinding of intervention -no
Complete followup - yes
Blinding of outcome measure -no
Inclusion criteria: low birthweight
9 infants studied. Mean birthweight:1.812 kg, mean gestational age: 34 weeks, mean age when studied: 7 days
Crossover design. Infants were studied for 2 consecutive three hour periods in a single wall incubator and under a radiant warmer Heart rate, respiration rate, oxygen consumption, insensible water loss, carbon dioxide production Temperature control: servo controlled to maintain skin temperature 36.5C. Infants were nursed naked. Mean skin temperatures: RW - 36.3, INC - 36.3C
Relative humidity:RW - 30.6%, INC - 29.9% (not significant)
C
Marks 1986 Blinding of randomization - can't tell
Blinding of intervention -no
Complete followup - no (4 infants out of 17 enrolled were excluded)
Blinding of outcome measure -no
Inclusion criteria: in an incubator, gaining weight, tolerating three hourly gavage feeds, free of complicating disease
17 infants enrolled, 4 were excluded. Mean gestational age: 34 weeks, mean weight at study entry:1.395 kg, mean age at study entry: 28 days

Crossover design. Infants were studied for a total of nine days in three different conditions (3 days in each)

  1. Single-walled forced air incubator, infant naked
  2. incubator, infant clothed
  3. radiant warmer, infant naked
Heart rate, respiration rate, oxygen consumption, carbon dioxide production, metabolic rate, body temperature, dietary intake, growth, urine - volume, specific gravity, osmolality Temperature control: RW - servo controlled to maintain skin temperature 36.5, INC - manual control "in the higher portion of the NTE". Infants nursed in diaper only
Mean skin temperature: RW - 36.2, INC - 35.8C ( p value 0.03)
Relative humidity : RW - not stated, INC - "maintained at < 40%"
B
Meyer 2001 Blinding of randomization - yes
Blinding of intervention -no
Complete followup - no ( one post randomisation exclusion)
Blinding of outcome measure - can't tell
61 preterm infants studied from admission to nursery up to 1800g.
Inclusion criteria:
< 33 wks gestation.
Exclusion criteria:
Major congenital abnormality, congenital infections, outborn infants.
Mean BW of study groups: 1182 and 1211 g.
Mean GA of study groups: 28 and 29 wks
Radiant warmer:
Fisher and Paykel "Cosy Cot" with polythene "tent-like" cover to deliver humidification.

Incubator: Air Shields double-walled (C550 or Isolette, Vickers Medical.)
Temp control - Air servocontrol. Humidification via incubator mechanism.

Body temperatures, Weight gain, caloric intake, fluid intake.
Major neonatal morbidity: infection, PDA, CLD, ROP, IVH, other major cerebral abnormalities.
Mortality, use of phototherapy

Unpublished data received from author for outcome: time to regain birthweight

Target body temp - Abdominal wall - 36.8 and axilla 36.8 to 37.3.
All infants nursed in diapers and bonnets and booties after 1 week, incubator infants also wore woolen tops after 1 week.
Added humidity (relative humidity 70-80%) for infants < 1250 (< 1000 for 5 days, 1000-1249 for 3 days).
Sample size calculation based on % time body temp < 36 in the first week and weight gain.
A
Schnabel 1999 Blinding of randomization - yes
Blinding of intervention - no
Complete followup - no (4 post randomisation exclusions)
Blinding of outcome measure - yes
34 preterm infants < 1500 g at birth. Studied from 7 to 35 days of age.
Mean BW in study groups 1125 and 894 g.
Mean GA in study groups: 29 and 28 wks.
Radiant warmer: Drager Babytherm 8010 or the Air Shield Open Care System or the Heinen-Lowenstein Open Care System. Skin temp control as well as the Heinen-Lowenstein open Care System.

Incubator: Drager Isolette 8000 or Air Shields Incubator.

Bacterial colonisation of infant and cot,
infection, body temperatures, intravenous fluid intake, weight gain, use of phototherapy, NEC, ROP, duration and mode of oxygen therapy.
Unpublished data received from the author for outcomes of neonatal death, ROP, IVH, cerebral cystic lesions, time to regain birthweight.
Four babies lost to followup due to: death - 1, transfer to other unit - 3.

Relative humidity : RW - not stated, INC - mean 55%(Unpublished information from the author)

A

RW = Radiant warmer
INC = Incubator
NTE = Neutral thermal environment
NEC= Necrotising enterocolitis
IVH= Intraventricular haemorrhage
CLD= Chronic lung disease
PDA= Patent ductus arteriosus
ROP= Retinopathy of prematurity
BW = Birthweight
GA = Gestational age

Characteristics of excluded studies

Study Reason for exclusion
Jones 1976 Random allocation to exposure was conducted for a subgroup only.
Kjartansson 1995 Neither random or quasi-random allocation to study groups was performed.
Levison 1966 Neither random or quasi-random allocation to study groups was performed.
Merenstein 1979 Data were not available in a format for inclusion in this review.
Wheldon 1982 Neither random or quasi-random allocation to study groups was performed.
Williams 1974 Neither random or quasi-random allocation to study groups was performed.
Wu 1974 Neither random or quasi-random allocation to study groups was performed.

[top]

References to studies

Included studies

Bell 1979

{published data only}

Bell EF, Neidich GA, Cashore WJ, Oh W. Combined effect of radiant warmer and phototherapy on insensible water loss in low-birth-weight infants. Journal of Pediatrics 1979;94:810-3.

Bell 1980

{published data only}

Bell EF, Weinstein MR, Oh W. Heat balance in premature infants: Comparative effects of convectively heated incubator and radiant warmer, with and without plastic heat shield. Journal of Pediatrics 1980;96:460-5.

Darnall 1978

{published data only}

Darnall RA, Ariagno RL. Minimal oxygen consumption in infants cared for under overhead radiant warmers compared with conventional incubators. Journal of Pediatrics 1978;93:283-7.

LeBlanc 1982

{published data only}

LeBlanc M. Relative efficacy of an incubator and an open warmer in producing thermoneutrality for the small premature infant. Pediatrics 1982;69:439-45.

Marks 1980

{published data only}

Marks KH, Gunther RC, Rossi JA, Maisels MJ. Oxygen consumption and insensible water loss in premature infants under radiant heaters. Pediatrics 1980;66:228-32.

Marks 1986

{published data only}

Marks KH, Nardis EE, Momin MN. Energy metabolism and substrate utilization in low birth weight neonates under radiant warmers. Pediatrics 1986;78:465-72.

Meyer 2001

{published data only}

Meyer MP, Payton MJ, Salmon A, Hutchinson C, de Klerk A. A clinical comparison of radiant warmer and incubator care for preterm infants from birth to 1800 grams. Pediatrics 2001;108:395-401.

Schnabel 1999

{published data only}

Schnabel K. Incubator care versus "open care" in the warming bed for very small premature infants. Kinderkrankenschwester 1999;18:195-200.

Excluded studies

Jones 1976

{published data only}

Jones RW, Rochefort MJ, Baum JD. Increased insensible water loss in newborn infants nursed under radiant heaters. British Medical Journal 1976;2:1347-50.

Kjartansson 1995

{published data only}

Kjartansson S, Arson S, Hammarlund K, Sjors G. Water loss from the skin of term and preterm infants under a radiant heater. Pediatric Research 1995;37:233-8.

Levison 1966

{published data only}

Levison H, Linsao L, Swyer PR. A comparison of infra-red and convective heating for newborn infants. Lancet 1966;2:1346-8.

Merenstein 1979

{published data only}

Merenstein G B, Koziol D F, Brown G L, Weisman LE. Radiant warmers vs incubators for neonatal care. American Journal of Diseases of Children 1979;133:857-8.

Wheldon 1982

{published data only}

Wheldon AE, Rutter N. The heat balance of small babies nursed in incubators and under radiant warmers. Early Human Development 1982;6:131-43.

Williams 1974

{published data only}

Williams PR, Oh W. Effects of radiant warmer on insensible water loss in newborn infants. American Journal of Diseases of Children 1974;128:511-4.

Wu 1974

{published data only}

Wu PYK, Hodgman JE. Insensible water loss in preterm infants: changes with postnatal development and non-ionizing radiant energy. Pediatrics 1974;54:704-12.

* indicates the primary reference for the study

Other references

Additional references

Baumgart 1982

Baumgart S, Fox WW, Polin RA. Physiological implications of two different heat shields for infants under radiant warmers. Journal of Pediatrics 1982:787-90.

Davenport 1992

Davenport SE. Frequency of hand washing by registered nurses caring for infants on radiant warmers and in incubators. Neonatal Network 1991:21-5.

Gorski 1990

Gorski PA, Huntington L, Lewkowicz DJ. Handling preterm infants in hospitals. Clinics in Perinatology 1990:103-12.

Hutchinson 1994

Hutchinson AA. Respiratory disorders of the neonate. Current Opinion in Pediatrics 1994:142-53.

LeBlanc 1991

LeBlanc MH. Thermoregulation: incubators, radiant warmers, artificial skins, and body hoods. Clinics in Perinatology 1991:403-21.

Long 1980

Long JG, Philip AGS, Lucey JF. Excessive handling as a cause of hypoxaemia. Pediatrics 1980:203-7.

Silverman 1957

Silverman WA. Blanc WA. Effect of humidity on survival of newly born premature infants. Pediatrics 1957:447.

Silverman 1958

Silverman WA, Fertig JW, Berger A. Influence of the thermal enviroment upon the survival of newly born premature infants. Pediatrics 1958:876-86.

Silverman 1963

Silverman WA, Agate FJ, Fertig JW. A sequential trial of the nonthermal effects of atmospheric humidity on survival of newborn infants of low birthweight. Pediatrics 1963:719-24.

Sinclair 1992

Sinclair JC. Management of the thermal environment. In: Sinclair JC, Bracken MB, editor(s). Effective Care of the Newborn Infant. Oxford: Oxford University Press, 1992:40-55.

Other published versions of this review

Flenady 1998

Flenady VJ, Woodgate PG. Radiant warmers versus incubators for regulating body temperature in newborn infants. In: The Cochrane Database of Systematic Reviews, Issue 1, 1998.

Flenady 2002

Flenady VJ, Woodgate PG. Radiant warmers versus incubators for regulating body temperature in newborn infants. In: The Cochrane Database of Systematic Reviews, Issue 2, 2002.

[top]

Data and analyses

01 Radiant warmers vs incubators

Comparison or outcome Studies Participants Statistical method Effect size
01.01 Insensible water loss(mL/kg/h) 3 53 WMD (fixed), 95% CI 0.94 [0.47, 1.41]
01.02 Oxygen consumption(mL/kg/min) 5 111 WMD (fixed), 95% CI 0.27 [-0.09, 0.63]
01.03 Skinfold fat gain(mm/d) 1 26 WMD (fixed), 95% CI 0.04 [0.01, 0.07]
01.04 Metabolic rate(kcal/kg/h) 1 26 WMD (fixed), 95% CI 0.13 [-0.26, 0.52]
01.05 Weight gain(g/kg/d) 2 86 WMD (fixed), 95% CI 1.06 [-0.94, 3.06]
01.06 Time to regain birthweight (days) 2 90 WMD (fixed), 95% CI 0.86 [-1.49, 3.21]
01.07 Infection - suspected or proven 2 90 RR (fixed), 95% CI 0.93 [0.66, 1.30]
01.08 Infection - positive blood culture 1 60 RR (fixed), 95% CI 0.60 [0.16, 2.29]
01.09 Necrotising enterocolitis 2 90 RR (fixed), 95% CI 0.78 [0.20, 3.00]
01.10 Patent ductus arteriosus 2 90 RR (fixed), 95% CI 1.00 [0.38, 2.62]
01.11 Retinopathy of prematurity- all stages 2 88 RR (fixed), 95% CI 0.97 [0.59, 1.59]
01.12 Retinopathy of prematurity - Stages III and IV 2 88 RR (fixed), 95% CI 1.27 [0.28, 5.83]
01.13 Intraventricular haemorrhage - all grades 2 90 RR (fixed), 95% CI 0.76 [0.31, 1.91]
01.14 Intraventricular haemorrhage - Grades 3 and 4 1 90 RR (fixed), 95% CI 0.33 [0.01, 7.87]
01.15 Cerebral cystic lesions 2 90 RR (fixed), 95% CI 0.33 [0.04, 3.08]
01.16 Chronic lung disease 1 60 RR (fixed), 95% CI 0.20 [0.01, 4.00]
01.17 Neonatal death 2 94 RR (fixed), 95% CI 0.27 [0.05, 1.59]
01.18 Serum sodium >150mmol/L 1 60 RR (fixed), 95% CI 3.00 [0.66, 13.69]

02 Radiant warmers vs incubators - subgrouped by use of heat shields

Comparison or outcome Studies Participants Statistical method Effect size
02.01 Insensible water loss WMD (fixed), 95% CI Subtotals only
02.02 Oxygen consumption (mL/kg/min) WMD (fixed), 95% CI Subtotals only
02.03 Time to regain birthweight (days) 2 90 WMD (fixed), 95% CI 0.86 [-1.49, 3.21]
02.04 Infection - suspected or proven 2 90 RR (fixed), 95% CI 0.93 [0.66, 1.30]
02.05 Infection - positive blood culture 1 60 RR (fixed), 95% CI 0.60 [0.16, 2.29]
02.06 Necrotising enterocolitis 2 90 RR (fixed), 95% CI 0.78 [0.20, 3.00]
02.07 Patent ductus arteriosus 2 90 RR (fixed), 95% CI 1.00 [0.38, 2.62]
02.08 Retinopathy of prematurity - all stages 2 88 RR (fixed), 95% CI 0.97 [0.59, 1.59]
02.09 Retinopathy of prematurity - Stages III and IV 2 88 RR (fixed), 95% CI 1.27 [0.28, 5.83]
02.10 Intraventricular haemorrhage - all grades 2 90 RR (fixed), 95% CI 0.76 [0.31, 1.91]
02.11 Intraventricular haemorrhage - Grades 3 and 4 1 90 RR (fixed), 95% CI 0.33 [0.01, 7.87]
02.12 Cerebral cystic lesions 2 90 RR (fixed), 95% CI 0.33 [0.04, 3.08]
02.13 Chronic lung disease 1 60 RR (fixed), 95% CI 0.20 [0.01, 4.00]
02.14 Neonatal death 2 94 RR (fixed), 95% CI 0.27 [0.05, 1.59]
02.15 Serum sodium >150mmol/L 1 60 RR (fixed), 95% CI 3.00 [0.66, 13.69]

Notes

  • None noted.

Additional tables

  • None noted.

Contact details for co-reviewers

Dr Paul Woodgate, MBBS MMedSci(Clin Epid) FRACP

Neonatal Paediatrician
Department of Neonatology
Mater Mothers Hospital
Raymond Terrace
South Brisbane
Queensland AUSTRALIA
4101
Telephone 1: +61-7-38401911
Facsimile: +61-7-38401949

E-mail: Paul_Woodgate@mater.org.au


The review is published as a Cochrane review in The Cochrane Library, Issue 2, 2005 (see http://www.thecochranelibrary.com External Web Site Policy for information). Cochrane reviews are regularly updated as new evidence emerges and in response to comments and criticisms, and The Cochrane Library should be consulted for the msot recent recent version of the review.