Cot-nursing versus incubator care for preterm infants

An optimal thermal environment is desirable for preterm infants. When an infant is challenged by cold, the baby attempts to conserve body heat by vasoconstriction and to maintain body temperature by increasing the rate of heat production. This involves thermogenesis by the metabolism of brown adipose tissue and an increase in oxygen consumption. The increase in energy expenditure may reduce weight gain (Glass 1969). Traditionally for preterm infants in the intensive care unit, temperature has been maintained by the use of air heated incubators or thermo controlled radiant heaters. The more stable preterm infants have usually been nursed in incubators, being moved to open cots at 1700 to 1800 g (Sutter 1988), though practice varies widely between neonatal units.

Given the complexities of caring for a preterm infant in an incubator, there is a need to assess the effectiveness of nursing in an open cot. This method of nursing does require a number of extra measures to ensure body temperature is maintained. These have included extra clothing/bedding, warming up the nursery and heating the cot mattress.

Metabolic effects were studied in infants in an open nursery cot provided with a small electric heating element below the mattress (Hey 1970). Oxygen consumption and heat loss fell compared to when the infants were studied in an unheated cot. In more recent years, a heated water-filled mattress (HWM) has been developed to maintain the temperature of a preterm infant, who would otherwise be nursed in an air heated incubator (Tunell 1986). The mattress can be placed in a standard baby cot. With the HWM there is conductive heat transfer to the baby and indirect heating of the air around the baby (limiting convective heat loss), thus tending to keep the baby at a normal temperature. Preliminary clinical results indicated that during treatment in the heated cot, the skin temperature was generally found to be 35 to 37 degrees C in all parts of the body covered by a blanket, while the skin of the face exposed to room air was 3 to 4 degrees C lower (Tunell 1986). This may be beneficial, as it has been suggested that a lower skin face temperature may have a positive effect on thermal comfort and behavioural state (Bruck 1968).

Parents experience a feeling of loss following the birth of a preterm infant. This is compounded by the separation of an infant in an incubator from its mother. Stress and anxiety are frequently experienced by new mothers (Shields-Poe 1997). Maternal perceptions of their infants may influence infant development (Watt 1989). Maternal perceptions and feelings may be more positive in the case of infants receiving cot-nursing because of greater maternal access to babies in open cots. Additionally, nursing staff may perceive that better care may be provided to infants nursed in open cots due to increased accessibility.

The cost of an air-heated incubator is substantial compared to the relatively cheap alternative of an open cot. If it could be demonstrated that nursing an infant in an open cot instead of an incubator could be achieved without adverse effect, then considerable benefit could accrue in economic terms both in developing and developed countries.

Although there may be benefits of nursing infants in open cots, there may be potential risks. It is possible that Infants nursed in cots may be handled more due to easier open access, which could may result in an increased risk of nosocomial infection. Conversely, there may be risks of nursing infants in incubators, such as exposure to excessive noise, resulting in negative effects on behavioural state and longer term development (Benini 1996).

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Electronic searches

Searching other resources

The systematic review followed the method described of the Cochrane Collaboration, documented in the Cochrane Handbook for Systemic Reviews of Interventions (Higgins 2008).

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Results of the search

Included studies

Excluded studies

Allocation (selection bias)

Blinding (performance bias and detection bias)

Incomplete outcome data (attrition bias)

Other potential sources of bias

The results of this review come from five studies involving a total of 247 babies. Three comparisons were undertaken: the overall comparison of cot-nursing vs incubator care, and two subgroup comparisons: cot-nursing with heated water-filled mattress vs incubator care, and cot-nursing using warming of the nursery versus incubator care.

Cot-nursing (any type) vs. incubator (COMPARISON 1):

Five trials of 247 infants contributed to this comparison. Data were available for pre-specified primary outcome measures, body temperature, episodes of cold stress, episodes of hyperthermia and weight gain, and for the secondary outcomes of oxygen consumption, not breast feeding at hospital discharge, episodes of nosocomial sepsis, maternal perceptions of infant's condition, maternal stress and anxiety and death prior to hospital discharge.

Primary outcomes:

Body temperature (Outcomes 1.1 - 1.3):

There was no difference between the cot- nursing and incubator groups in mean body temperature [mean difference (MD) 0.02 degrees C; 95% confidence interval (CI) - 0.20 to 0.07]. There was, however, statistical heterogeneity (I² 83%) due to the inclusion of the trial (Sarman 1989b) in which most of the babies on recruitment had a temperature < 36 degrees C. As the Green-Abate 1994 trial only reported the core-skin temperature gradient, it was not possible to use this data in the meta-analysis. Only the Gray 2004 trial reported the number of infants with episodes of cold stress and hypothermia. No differences between the cot-nursing and incubator groups were found. Similarly only the Gray 2004 trial reported on episodes of hyperthermia, with this outcome being found to occur more frequently in the cot-nursing group (RR 1.48; 95% CI 1.04 to 2.09). Green-Abate 1994 reported on the number of temperature measurements in the hypothermic and hyperthermic range per person-hours and hence these data were not used in this analysis.

Weight gain (Outcome 1.4):

The Sarman 1989a, Green-Abate 1994 and Gray 2004 trials provided data on weight gain (g/kg body weight/day) subgrouped by week. The results showed no significant differences between the cot-nursing and incubator groups in either weeks one, two or three, either in the individual trials or in the meta-analysis (week one MD 0.19 g/kg/day; 95% CI -2.39 to 2.77; week two MD 0.91 g/kg/day; 95% CI -0.51 to 2.33; week three MD 0.35 g/kg/day; 95% CI -4.41, 5.11). The results for week one indicated statistical heterogeneity (I² 59%), which related to the trial of Green-Abate 1994 which included babies cot-nursed in a space heated room as well as those on a heated water-filled mattress. Interestingly, the Green-Abate 1994 trial did not report any statistically significant differences in caloric intake between the groups during the trial period, these data were not provided by either Sarman 1989a or Gray 2004.

Secondary outcome measures:

Oxygen consumption (ml/kg/min) (Outcome 1.5):

Two trials contributed data for the outcome of oxygen consumption (Sarman 1989a; Sarman 1992). Neither trial found evidence of effect, and no significant difference between the two groups was found in the meta-analysis (MD -0.17 ml/kg/min; 95% CI -0.47 to 0.13).

Not breast feeding at hospital discharge (Outcome 1.6):

The Sarman 1989a, Sarman 1989b and Gray 2004 trials collected data on breast feeding on discharge. Sarman 1989b and Gray 2004 found no evidence of effect, but Sarman 1989a found a reduction in the number of infants not breast feeding at discharge that was of borderline statistical significance. In the meta-analysis, cot-nursing resulted in a reduction in the number of infants not breast feeding at discharge, but this was not statistically significant (typical RR 0.74; 95% CI 0.48 to 1.14).

Episodes of nosocomial sepsis (Outcome 1.7):

Two trials (Gray 2004; Sarman 1989a) reported data on nosocomial sepsis. No sepsis occurred during the Sarman 1989a trial, while no significant differences between the two groups were reported in the Gray 2004 trial (typical RR 0.27; 95% CI 0.01 to 6.41)

Maternal perceptions of infants condition (Outcome 1.8):

In the Gray 2004 trial, mothers completed questionnaires on the perceptions of their preterm infants using the Neonatal Perception Inventory (Broussard 1970; Broussard 1971). At the time of randomisation there were no differences between the cot-nursing and incubator groups of mothers. On completion of the trial, again no differences between the groups were reported (MD -1.0; 95% CI -2.96 to 0.96).

Maternal stress and anxiety (Outcome 1.9):

The Gray 2004 trial collected data on maternal anxiety and stress levels using the anxiety scale of the Delusions Symptoms States Inventory (DSSI) (Bedford 1978) and the Los Angeles Stress Scale (Neader 1973). In these questionnaires, no differences between the two groups either at randomisation or during follow-up were reported (MD -2.9; 95% CI -6.54 to 0.74).

Death prior to hospital discharge (Outcome 1.10):

No significant difference was shown for the outcome of death prior to discharge either in the individual trials (Sarman 1989a, Sarman 1989b; Green-Abate 1994; Gray 2004) or in the meta-analysis, though there was a strong trend towards less death in the cot-nursing group (RR 0.59; 95% CI 0.28 to1.25).

Cot-nursing with heated water-filled mattress and incubator care (Comparison 2):

Five trials with 223 infants were included in this comparison.

Primary outcomes:

Body temperature (Outcomes 2.1 - 2.3):

Data on mean body temperature were obtained from the four trials (Sarman 1989a; Sarman 1989b; Sarman 1992; Gray 2004) with the results being the same as for the overall comparison. The one trial that reported number of infants with episodes of cold stress and hyperthermia (Gray 2004) indicated that there were no differences between the cot-nursing and incubator groups for episodes of cold stress, while there was an increase in episodes of hyperthermia in the cot-nursing group (RR 1.48; 95% CI 1.04 to 2.09).

Weight gain (Outcome 2.4):

No statistically significant difference was shown in weight gain in either weeks one, two or three of treatment, either in the individual trials, (Sarman 1989a; Green-Abate 1994; Gray 2004) or in the meta-analyses: mean difference (MD) and 95% CI for weeks one, two and three were 1.30 g/kg/day (-1.48 to 4.08), 0.90 g/kg/day (-0.57 to 2.37) and 1.21g/kg/day (-3.77 to 6.18) respectively.

Secondary outcomes (Outcomes 2.5 - 2.10):

For the secondary outcome measures of oxygen consumption, breast feeding at hospital discharge, episodes of nosocomial sepsis, maternal perceptions of infant's condition and maternal stress and anxiety the same trials were included as for the overall comparison and hence the same results were obtained. No statistically significant difference was shown for death, the only other reported secondary outcome, either in the individual trials (Sarman 1989a; Sarman 1989b; Green-Abate 1994; Gray 2004) or in the meta-analysis: (typical RR 0.63; 95% CI 0.30 to 1.34).

Cot-nursing using warming of the nursery vs incubator care (Comparison 3):

Weight gain (g/kg/day) (Outcome 3.1); Death prior to hospital discharge (Outcome 3.2):

One trial involving 45 infants was included in this analysis (Green-Abate 1994). Infants receiving cot-nursing with warming of the nursery had less weight gain than those in the incubator group during week one (MD -5.90 g/kg/day; 95% CI -11.13 to -0.67. No statistically significant differences were shown for weight gain in weeks two (MD 1.40 g/kg/day; 95% CI -2.70, 5.50) or three (MD -6.00 g/kg/day; 95% CI -18.81 to 6.81). Data on temperature control were not presented in a way that could be included in the review. The only secondary outcome that could be analysed was death and there was no significant difference between the groups (RR 0.44; 95% CI 0.04 to 4.49).

Data were not available for the secondary outcome measures of length of hospital stay, nosocomial and nursing perceptions and cost. Due to insufficient data, other planned sub-group analyses could not be performed (for detail please refer to the secondary objectives of the review included under "Objectives").

The results of the review showed no evidence of effect of cot-nursing versus incubator care on weight gain in the overall analysis, or in the subgroup analysis comparing cot-nursing using a heated water-filled mattress with incubator care. However, cot-nursing with warming of the nursery during week one when compared to incubator care revealed poorer weight gain which was on average 6 g/kg/day less (Green-Abate 1994). The primary outcomes related to temperature control - mean body temperature and episodes of cold stress indicated on overall analysis no effect of cot-nursing compared to incubator care. Episodes of hyperthermia in the cot-nursing group were reported more frequently in one trial (Gray 2004). It was reported that it was considered concerns raised by staff providing nursing care at the beginning of the trial of the possibility of cold stress/hypothermia, may have led to the temperature of the mattress being set higher than appropriate. The secondary outcomes of oxygen consumption, breast feeding at hospital discharge, episodes of nosocomial sepsis, maternal perceptions of infant's condition, maternal stress and anxiety and death prior to hospital discharge revealed there was no effect of cot-nursing compared to incubator care. There was, however, a strong trend towards less death prior to hospital discharge. This was largely related to the results were obtained from the trials undertaken in Turkey (Sarman 1989b) and Ethiopia (Green-Abate 1994) and thus may not be applicable to neonatal nurseries in developed countries. Nevertheless the implications of these findings deserve consideration, particularly in the context of a developing country.

The protocol did not place strict a priori definitions for the primary or secondary outcomes. Hence we used the operational definitions as reported by the investigators in the eligible trials.

This review has a number of limitations. The numbers of babies enrolled in the included trials are relatively small and data were either not available or only available for very few infants for many of the pre-specified important clinical outcomes for this review. Importantly, the outcome measure related to safety - episodes of nosocomial sepsis was only addressed in two studies (Sarman 1989a; Gray 2004).

No trials were identified that compared care in newborn cots with additional warmth provided by extra clothing/bedding or a heated element below the mattress. Furthermore, comparisons of cot nursing interventions with varying methods for incubator care e.g. servo-controlled incubators, naked babies in incubators, could not be performed due to lack of available trials. Two trials (Sarman 1989b; Green-Abate 1994) took place in developing countries (Ethiopia and Turkey) where major differences in medical and nursing care were evident in comparison to developed countries. In the study carried out in Turkey (Sarman 1989b), a limited number of nurses were available in the nursery, with care largely being provided by the mothers. Accordingly, one may not be able to generalize from these the results.

One included trial (Gray 2004) assessed the effect of cot-nursing on maternal perceptions and maternal stress and anxiety and found no difference in comparison to the incubator group of infants. Another study (Sarman 1993) reported mothers' perceptions of their preterm infants treated in a cot with a heated water-filled mattress versus in an incubator. This pilot study suggested that the babies treated in a cot on a heated water-filled mattress were healthier and more vigorous than those in an incubator and that the mothers felt that their babies 'belonged to them'. These data could not be included as the infants were subsets of two other trials.

The heated water-filled mattress was designed as a cheap and technically simple device that when placed in a cot could be used as an alternative to incubator care (Tunell 1986). No studies were identified which compared the costs of cot-nursing compared to nursing a baby in an incubator.

The results of the review provides evidence that cot nursing is comparable to incubator care, without any indication that there is an increased risk of adverse outcome. Indeed it has been recommended that when a preterm infant is stable without the requirement for respiratory support, then cot-nursing and in particular using the HWM may replace incubator care (Tunell 2004).

Despite there being a relatively small number of trials included and infants studied, it would seem overall that cot-nursing using a heated water-filled mattress has similar effects to incubator care with regard to temperature control and weight gain. While limited data are available for important secondary outcomes including episodes of nosocomial sepsis, not breast feeding at hospital discharge and death prior to hospital discharge, together with cost it is suggested that cot-nursing with a heated water-filled mattress may be employed for preterm infants. Even though little information is available on cot-nursing in a space-heated room, it is suggested that cot-nursing in this manner is not appropriate.

Further investigation of the role of cot-nursing for preterm infants using randomised controlled trials is necessary. Different forms of providing extra warmth to the infants while in cots need to be examined in comparison with varying methods of incubator care. Important clinical outcomes to be addressed in future research include thermoregulation, weight gain, breast feeding, mortality and nosocomial infection. The issue of cost is important, especially as it pertains to developing countries. A formal assessment of expenditure for equipment purchase, the cost of nursing care and length of hospital stay needs to be carried out.

Risk of bias table

Risk of bias table

Risk of bias table

Risk of bias table

Risk of bias table

* Gray PH, Paterson S, Finch, G, Hayes M. Cot-nursing using a heated, water filled mattress and incubator care: a randomised clinical trial. Acta Paediatrica 2004;93:450-5.

Paterson S, Gray PH, Finch G, Hayes M. Cot-nursing using a heated water-filled mattress and incubator care: A randomised controlled trial. In: Proceedings of the Perinatal Society of Australia and New Zealand 7th Annual Congress. Baulkham Hills, Perinatal Society of Australia and New Zealand. 2003:P24.

Green-Abate C, Tafari N, Rao MR, Yu KF, Clemens JD. Comparison of heated water-filled mattress and space-heated room with infant incubator in providing warmth to low birthweight newborns. International Journal of Epidemiology 1994;23:1226-33.

Sarman I, Tunell R. Providing warmth for preterm babies by a heated, water filled mattress. Archives of Disease in Childhood 1989;64:29-33.

Sarman I, Can G, Tunell R. Rewarming preterm infants on a heated, water filled mattress. Archives of Disease in Childhood 1989;64:687-92.

Sarman I. Thermal responses and heart rates of low-birth-weight premature babies during daily care on a heated, water-filled mattress. Acta Paediatrica 1992;81:15-20.

Darragh J, Piper M, Byrne P, Watt MJ. The use of waterbeds for very low-birthweight infants: effects on neuromotor development. Developmental Medicine and Child Neurology 1994;36:989-99.

Deiriggi PM. Effects of waterbed flotation on indicators of energy expenditure in preterm infants. Nursing Research 1990;39:140-6.

Marks KH, Calder PD, Nardis EE, Ultman JS. A warming mattress for premature infants. Pediatric Research 1984;18:333A.

Sarman I, Tunell R, Vastberg L, Carlquist U, Can G, Toparlak D. Mothers' perceptions of their preterm infants treated in an incubator or on a heated water-filled mattress: a pilot study. Acta Paediatrica 1993;82:930-3.

Weintraub V, Mimouni FB, Dollberg S. Changes in energy expenditure in preterm infants during weaning: a randomised comparison of two weaning methods from an incubator. Pediatric Research 2007;61:341-4.

West CR, Williams M, Weston PJ. Feasibility and safety of early transfer of premature infants from incubators to cots: a pilot study. Journal of Paediatrics and Child Health 2005;41:659-62.

Zecca E, Corsello M, Priolo F, Tiberi E, Barone G, Romagnoli C. Early weaning from incubator and early discharge of preterm infants: randomized clinical trial. Pediatrics 2010;126:e651-6.