Main

Care in a heated incubator has improved the survival of preterm infants of<1500-g birth weight(1) but separates the infants from their parents for weeks. To interrupt this separation, skin-to-skin contact between preterm infants and their parents outside the incubator is increasingly used in neonatal intensive care units(2). During skin-to-skin contact the infant is partly exposed to the maternal skin temperature, large parts of the infant's head are exposed to room temperature, and the back sides of the infant's trunk and extremities are covered with a towel. In this thermal environment preterm infants of>34-wk GA or with a postnatal age of 3 wk have stable body temperatures(3–5). We previously reported that preterm infants less than 1 wk of age had an increase in Trectal and normal Vo2 during skin-to-skin contact(6). But these studies did not include extremely premature infants or longitudinal measurements at different postnatal ages. This study includes a sufficient number of preterm infants of <28-wk GA and longitudinal measurements to analyze the effects of both gestational and postnatal age on the thermoregulation during skin-to-skin contact. These physiologic data are necessary for a evidence-based decision about which preterm infants are eligible for skin-to-skin contact. We therefore measured body temperature and Vo2 before, during, and after skin-to-skin contact in infants of 25-30-wk GA to determine at which gestational and postnatal age they tolerated skin-to-skin contact without a decrease in body temperature, increase in Vo2, or unrest.

METHODS

Patients. Preterm infants of 25-30-wk GA and a body weight appropriate for GA were studied in their first week of life as soon as they met the following criteria: spontaneous respiration, no apnea within the last 24 h that required bag and mask ventilation, less than 1 apnea/h, and stable circulation. Infants still on mechanical ventilation at a postnatal age of 7 d were excluded. The study was approved by the hospital's institutional review board, and written parental consent for each infant was obtained.

Sample size calculation. We calculated that a sample of 10 infants was necessary to detect a clinically relevant increase of Vo2 by 0.9 mL/kg per min (=15%) and a change in Trectal of >0.2°C/h (α error of 0.05, β error of 0.2).

Indirect calorimetry. For the continuous measurement of Vo2 we used open system flow-through indirect respiratory calorimetry with a face mask for breath sampling and a commercially available calorimeter (Deltatrac II metabolic monitor, Datex Division Instrumentarium Corp., Helsinki Finland) with a sampling flow of 3 L/min as described previously(6). We developed this system especially for small preterm infants and have shown that this system is accurate (mean experimental error for the Vo2 measurement 2 ± 2%) without increasing the infants' respiratory rate, activity, or body temperature(7).

In 16 of the 52 Vo2 measurements, infants needed increased inspiratory oxygen concentrations (FIO2), and oxygen was introduced through a side hole into the breath-sampling face mask. This produces FIO2 fluctuations that make Vo2 measurements inaccurate. Then only Vco2 was measured and Vo2 was calculated from Vco2 and an RQ estimated from nutritional intake (food quotient)(8). We showed previously that Vo2 calculated from Vco2 and food quotient agrees closely with measured Vo2(9). An increased Fio2 was required by 1 of the 16 infants of 28-30-wk GA in wk 1 and by 3 of the 16 in wk 2. Six of the 11 infants of 25-27-wk GA required an increased Fio2 in wk 1 and 6 of the 9 studied in wk 2. Vo2 calculated from Vco2 and food quotient did not differ from measured Vo2 in the infants studied. Therefore calculated and measured Vo2 values were combined in the analysis.

Temperature measurements and cardiorespiratory variables. Trectal was measured with a temperature probe inserted 2 cm into the rectum. The maternal skin temperature was measured in the midclavicular line 2 cm below the right clavicle. Air temperature under the towel covering the infant's trunk and extremities during skin-to-skin contact was measured with a probe in the center of a paper tube that prevented direct contact with the skin. Temperatures, heart rate, and Hb oxygen saturation were registered every minute (patient monitor CMS 5000, Hewlett Packard GmbH, Bad Homburg, Germany). The infant's activity was scored every minute by a bedside observer using a modified Brueck score(10). Time spent in sleep was defined as the time with a Brueck score of -3 (eyes closed, facial movements) or -4 (eyes closed, no movements).

Study protocol. As soon as an infant was eligible, the study was scheduled for the afternoon of the same day. The infants were naked except for a diaper in a double-walled, air temperature-controlled incubator(model 8000, Draeger AG, Lübeck, Germany) at a thermoneutral temperature and a humidity of 80%. Measurements always started on the same time of day and did not interfere with the three hourly nursing routine. At 1500 h the infant was fed, the diaper changed, and the temperature probes were attached. To minimize the effects of postprandial thermogenesis, the measurements in the incubator were started at 1600 h. At 1700 h the infant was transferred to the mother, who was wearing a hospital gown and reclining in a comfortable chair. All mothers had normal axillary temperatures. The infant was placed prone on the mother's chest, his trunk and extremities were covered with a towel, and his head was left uncovered and exposed to room air (temperature 26°C, relative humidity 40%). At 1800 h the infant was transferred back to the incubator and fed, and measurements were continued for another hour.

Measurements were repeated 7 d later if the infants were still clinically stable and breathing spontaneously. Between these two measurements 1 h of skin-to-skin contact was scheduled every day.

Nutrition and medication. The infants received parenteral nutrition (glucose and amino acids) and bolus gavage feedings from d 1. Intravenous lipids were added on d 2 and gavage feedings were advanced according to feeding tolerance. On the first study day infants received 8 bolus feeds of 5 ± 3 mL each representing 40% of total caloric intake, on the second study day infants received 8 bolus feeds of 12 ± 4 mL each representing 60% of caloric intake. Twenty infants received antibiotic treatment, no infant received medication known to influence Vo2 or activity.

Statistical analysis. Data were analyzed with the SPSS statistical software (SPSS Inc., Chicago, IL). Mean values of the three study periods were compared by paired t test. Differences between the two study groups were analyzed with the unpaired t test. Not normally distributed data were analyzed with nonparametric tests (Wilcoxon). Correlations between parameters were determined by regression analysis. A p value of <0.05 was considered statistically significant.

RESULTS

Study groups. Eleven of the 40 infants of 25-27-wk GA born during an 18-mo observation period were eligible for the study. Twenty-six infants were not eligible because they were ventilated for more than 7 d, two were small for GA, and one died on d 1. Two of these 11 infants were studied only in wk 1 because in wk 2 they were mechanically ventilated due to late onset sepsis; the other nine infants were studied in wk 1 and 2.

Sixteen of the 18 infants of 28-30-wk GA born during a 7-mo observation period were eligible for the study and were measured in wk 1 and 2; two were not eligible because they were small for GA. Clinical data of the study infants are given in Table 1.

Table 1 Clinical data of the study infants
Full size table

Effects of GA. GA had an effect on the change in Trectal during skin-to-skin contact in wk 1:Trectal decreased in infants of 25-27-wk GA from the first to the last 5 min of skin-to-skin contact, but increased in infants of 28-30-wk GA (r = 0.585, p = 0.0027) (Fig. 1). Therefore in the following, we separately analyzed infants of 25-27-wk GA and compared them to infants of 28-30-wk GA.

Figure 1
figure 1

Relationship between GA and temperature difference between the first and last 5 min of skin-to-skin contact in 27 preterm infants studied in wk 1 of life.

Full size image

In the 11 infants of 25-27-wk GA Trectal was stable in the last 30 min before skin-to-skin contact, but decreased during the transfer out of the incubator, during skin-to-skin care, and during the back transfer (Fig. 2), so that the mean Trectal in the hour after skin-to-skin contact was 0.3°C lower than in the hour before (p = 0.007;Table 2). Three of the 11 infants had a mean Trectal below 36.5°C in the hour after skin-to-skin contact. In the 16 infants of 28-30-wk GA, Trectal was stable in the hour before skin-to-skin contact, but rose steadily from 36.9± 0.4°C in the first 5 min to 37.4 ± 0.4°C in the last 5 min of skin-to-skin contact (p = 0.001) (Fig. 2). Six of the 16 infants had a Trectal > 37.5°C at the end of skin-to-skin contact.

Figure 2
figure 2

Time course of Trectal and Vo2 in preterm infants between 25 and 27 wk and preterm infants between 28 and 30 wk of gestation in wk 1 and 2 of life (data points are mean values for every 5 min; bars represent SD).

Full size image
Table 2 Trectal, Vo2, and duration of sleep before, during and after skin-to-skin contact in wk 1 and 2 of life(mean ± SD or median (range) for each 60-min period)
Full size table

Vo2 did not increase during skin-to-skin contact or after the temperature drops associated with the transfers in the infants of 25-27-wk GA(Fig. 2). In the infants of 28-30-wk GA, mean Vo2 was also not significantly higher during skin-to-skin contact.

In wk 1 air and surface temperatures measured during skin-to-skin contact(air temperature under the towel covering the infant's trunk and extremities and maternal skin temperature) were the same as air and surface temperatures measured in the incubator (air and mattress temperatures)(Table 3).

Table 3 Air and surface temperatures before, during, and after skin-to-skin contact in wk 1 and 2 of life (mean ± SD for each 60-min period)
Full size table

Effects of postnatal age. In wk 2, Trectal in all infants was stable in the incubator before and after skin-to-skin contact at lower incubator air and mattress temperatures than in wk 1 (Table 3). In wk 2, GA had no influence on Trectal, which increased in both groups by 0.2°C from the beginning to the end of skin-to-skin contact (p< 0.05) (Fig. 2).

The Vo2 in all three study periods increased with postnatal age(r = 0.599 Vo2 before, r = 0.432 Vo2 skin-to-skin, r = 0.664 Vo2 after; p < 0.001). Consequently in wk 2 mean Vo2 in all three study periods was higher than in wk 1 (p < 0.001) (Table 2).

Skin-to-skin contact had no effect on activity in wk 1 when infants were sleeping >90% of the time in the incubator and during skin-to-skin contact. In wk 2, however, the infants spent more time sleeping during skin-to-skin contact than in the incubator (Table 2). In wk 1 and 2 skin-to-skin contact had no effect on heart rate and oxygen saturation.

DISCUSSION

This study demonstrates effects of gestational and postnatal age on body temperature, Vo2, and activity during skin-to-skin contact in preterm infants. Our principal findings were that GA had an effect on the change in body temperature during skin-to-skin contact in wk 1:Trectal decreased in preterm infants of 25-27-wk GA, whereas it increased in infants of 28-30-wk GA, both changes were not accompanied by an increase of Vo2. In wk 2 the infants had a higher Vo2 than in wk 1, and even very immature infants of 25-27-wk GA had only small fluctuations in Trectal. Nevertheless, skin-to-skin contact should be used with caution in infants of 25 and 26 wk of gestation, because we studied only four such infants. In wk 2 the infants were sleeping more during skin-to-skin contact than when in the incubator.

The importance of skin-to-skin contact in the first days of life for maternal infant bonding is known from controlled studies in term infants(11). For preterm infants longer breastfeeding(12) and shorter hospital stay(13) are reported benefits of skin-to-skin contact. We focused our study on preterm infants of >30-wk GA fro whom temporary skin-to-skin contact may be most important, because they need long incubator care, which separates them from their mothers for weeks. But they are also the most fragile infants who benefits from skin-to-skin contact only when the necessary handling and partial exposure to room temperature does not impair their thermal and cardiorespiratory stability.

Preterm infants <30-wk GA are so vulnerable to cold exposure because their capacity for heat production is limited, they have high evaporative heat losses across their extremely water permeable skin, a large surface area for heat exchange, and only minimal insulating s.c. fat(14). Therefore they need incubator care. The incubator is a carefully studied thermal environment for preterm infants, and incubator temperatures necessary to maintain normal body temperatures in preterm infants of different gestational or postnatal ages are known(15). In contrast, skin-to-skin contact is a poorly studied thermal situation and it is not known at which gestational and postnatal age preterm infants can maintain a normal body temperature during skin-to-skin contact. In the few earlier studies of skin-to-skin contact Vo2 was not measured, and body temperatures only sporadically(16,17), or no infants were measured in the first 2 wk of life(3), or only few infants below 28-wk GA were included(6). In no previous study were measurements repeated at different postnatal ages.

We continuously measured Vo2 in addition to body temperature because neonates produce heat to compensate heat losses during cold exposure, which increases their Vo2(18). To ensure accurate Vo2 measurements which are difficult at the low Vo2 values of very premature infants we used a face mask that we had validated in vitro at low Vo2(7). To minimize influences of other factors affecting Vo2 measurements(19), we started Vo2 measurements always on the same time of the day and more than 1 h after the last enteral feeding, the study groups had a homogeneous gestational and postnatal age, and we monitored activity. In 16 of the 52 studies infants needed increased inspiratory oxygen concentrations. Then Vo2 was not measured but calculated from measured Vco2 and an estimated RQ. We have combined studies with calculated and measured Vo2 because there is good agreement between Vo2 calculated from Vco2 and measured Vo2: the error was within ±2% in children and adults(8) and within ±3% in preterm infants in our hands(9). When we analyzed studies with measured or with calculated Vo2 separately we found no significant differences in Vo2.

This study includes a sufficient number of infants below 28-wk GA and continuous measurements of Vo2, body temperature, and environmental temperatures in wk 1 and repeated in wk 2. We found that GA had a significant effect on the infants' thermal reaction to skin-to-skin contact in the first week of life. The very immature infants of 25- and 26-wk GA lost heat during skin-to-skin contact. Their thermal stability was also disrupted by the transfers out of and back into the incubator and rewarming in the incubator after skin-to-skin contact was slow. Despite the heat losses Vo2 did not increase. Possible explanations for this finding are that the environmental temperature during skin-to-skin contact was not low enough to elicit an increase in heat production because a Vo2 increase was observed in preterm infants only when environmental temperature fell below 33°C(18) or that the capacity for heat production of extremely preterm infants is very limited because they have a low resting metabolic rate during the first days of life(20) and can barely increase heat production during cold stress(21).

The infants of 28-30-wk GA gained heat during skin-to-skin contact in wk 1. Because they had the same heat production, i.e. Vo2, than the more immature infants their heat losses during skin-to-skin contact must have been smaller. It is known that the evaporative heat loss of infants of 28-30-wk GA is only half that of infants ≤27-wk GA(22) and that their body surface area per kg of body weight is 12% lower. The Trectal of the infants of 28-30-wk GA increased during skin-to-skin contact to values above those in the incubator, in six infants to more than 37.5°C. Because their heat production did not increase, their heat losses during skin-to-skin contact must have been smaller than that in the incubator, but heat losses were not measured. We made regional air and surface temperature measurements suggesting that conductive and convective heat losses from the trunk and the extremities were similar during skin-to-skin contact and incubator care because the air temperature under the towel covering the infant's trunk and extremities during skin-to-skin contact was as high as the air temperature in the incubator and maternal skin temperature was as high as the temperature of the incubator mattress. Therefore we speculate that radiative heat losses were lower during skin-to-skin contact when the infants were covered with a towel than in the incubator when the infants were naked.

Contrary to wk 1, infants of 25-27-wk gestation had an increase in Trectal during skin-to-skin contact in wk 2 because they had a higher metabolic rate measured as higher Vo2. A postnatal increase in metabolic rate caused by the energy-consuming processing of nutrients is well known(23). Another reason for the positive heat balance is the postnatal decrease of heat losses, especially of evaporative heat loss(22). Our observations that in wk 2 the infants tolerated the transfers out of and back into the incubator with a much smaller drop in Trectal than they did in wk 1 and that they had stable normal Trectal at lower incubator air temperatures are also evidence for their increased thermal stability.

In wk 2 the infants spent significantly more time sleeping during skin-to-skin contact than they did during incubator care. This is a benefit of skin-to-skin contact that we could not find in wk 1 when the infants spent more than 90% of the time sleeping, both in the incubator as well as during skin-to-skin contact.

Our physiologic data have the following implications for clinical practice. Skin-to-skin contact for preterm infants <1500 g cannot be generally recommended without considering gestational and postnatal age. In wk 1 of life marked drops in Trectal occurred with skin-to-skin contact and the necessary transfers between incubator and mother in infants of 25 and 26 wk of gestation. However, in wk 2 of life even the most immature infants studied could be taken out of the incubator for temporary skin-to-skin contact without disrupting their thermal and cardiovascular stability and with the additional benefit of more sleep during skin-to-skin contact.