After approval by the Institutional Review Board, a multiple case, within-subject design was used to explore body temperature and peripheral vasoconstriction in 10 ELBW infants over their first 12 hours in the NICU in a North Carolina hospital. Sample size was set at 10 in order to have a sample large enough to have representation of various weights, genders, and races of ELBW infants while maintaining a sample size small enough to permit detailed within-subject analyses. Infants were included if they had a birth weight between 500–1000 grams and a gestational age between 23–28 6/7 weeks by obstetrical dates determined by ultrasound or clinical dating. Consent was obtained from mothers admitted for preterm labor.

Skin surface thermistor probes were attached once the infant was on the Giraffe warmer (GE Healthcare) and weight was verified. Central (abdominal) and peripheral (foot) temperatures were measured with a Mini-Logger (Mini Mitter, Oregon) monitor using two thermistor probes (Steri-Probe, Skin surface probe #499B, Cincinnati Sub-Zero). Abdominal surface temperature was used as central body temperature because we were unable to use esophageal temperature due to the critical illness in this extremely preterm population and because rectal temperature measurement cannot be done continuously for 12 hours due to risk of rectal perforation.14 Abdominal surface temperature provides the closest temperature to deep tissue core temperature in this population because the probe is covered which provides near zero heat flux and the skin is thin lacking much body fat.15 The central temperature study probe was positioned adjacent to the standard care incubator temperature probe in the key-hole opening of the duoderm on the skin surface of the abdomen and covered with reflective temperature probe cover. The peripheral temperature probe was secured to the sole of each infant’s foot underneath the tape which held the pulse oximeter probe. The sole of the foot was used secondary to the practical ease of probe attachment and because most ELBW preterm infants do not have the ability to sweat when born,16 which makes the choice of sole of the foot versus other areas of the foot acceptable. Temperature readings were sampled at one-minute intervals and recorded using Mini-Logger software on a laptop computer. These infants were followed during their first 12 hours of life and infants showed very little motor activity due to their immaturity and illness severity; therefore, motion should not have impacted thermistor measurement.

Preterm infants at this hospital were placed in polyurethane occlusive wrapping upon delivery, resuscitated according to NRP standards and then transported to the NICU under warm blankets. Each infant was placed on a Giraffe radiant warmer (GE Healthcare) equipped with a computerized heat control system that was set on manual at maximum heat output prior to admission, then the infant was controlled at a body temperature of 36.5° C. The warmer converted to an incubator after placement of umbilical lines and humidity was then added and maintained between 60–80%. A data collector sat at each bedside and recorded procedures, manipulation to the infant and incubator environmental temperature at five minute intervals.

Temperature data were exported into Mini-Logger software (Mini-Mitter), converted into Excel files, then exported into SAS^®, Version 8 (Cary, NC) for analyses. Analyses were conducted within each infant, using approximately 720 measurements (one each minute for 12 hours) for each variable, although these data were not consistently analyzable for all minutes due to instances when both temperature probes were not properly secured. Temperature measurements were plotted to display any trend over time by visual inspection. Peripheral vasoconstriction was assessed by taking the difference between the abdominal and peripheral temperatures.15 A decrease in peripheral temperature that occurs before an alteration in the abdominal temperature is an early sign of thermal cold stress due to peripheral constriction.15 Larger differences (> 2° C) have been used to indicate vasoconstriction.11 Temperature data were analyzed through descriptive statistics, two sample t-tests; and Pearson correlation coefficients within subjects. Alpha level was set at 0.05 level of significance, since we treated the within-infant observations as conditionally independent.

Only one infant exhibited peripheral vasoconstriction (abdominal temperature minus peripheral temperature [Δ T] by at least 2° C) with 9% of his measurements exhibiting peripheral vasoconstriction (see Figure 1). This infant was an 880 gram, 26-week male. The largest infant in the study (960 gram, 26-week GA female) showed an attempt at peripheral vasoconstriction; 63% of her observations had a ΔT of + 1° C. These two infants were the only subjects that had at least 8% their abdominal temperatures greater than their peripheral temperatures by at least 1° C; the remaining eight infants exhibited generally higher peripheral temperatures rather than higher abdominal temperatures (−ΔT) (see Table 2).

Seven infants (500–710 grams, 24–25 weeks GA) showed a pattern in which at least 50% of their peripheral temperature measurements were higher (by any amount) than their matched central temperature observations (see Table 2). This pattern can be seen in Figure 2 for a 590 gram, 24 week gestational aged male infant (I). This infant had an abdominal temperature range of 31.24° C to 36.91° C. Study monitors were attached and began recording at 1 hour of age for this infant. Stabilization of the infant (attachment of the ventilator, monitors and umbilical line placement) in the NICU took place for approximately 1 hour prior to beginning data collection. Even though admission temperature was 35.4° C, rectal temperature at 1 hour of age and post stabilization was recorded as too low to register. In contrast, the larger infant that exhibited peripheral vasoconstriction kept his abdominal temperature higher than his peripheral temperature. This infant was an 880-gram African American male (D) delivered by vaginal delivery at 26-weeks gestation. The temperature monitors were attached at 1 hour of age and the infant was cold upon admission (34.8° C axillary); however, his temperature increased fairly quickly. Umbilical line placement took place over the first 3 hours of age. It is interesting to note that when the infant became cold around 6 hours of study time, his peripheral temperature dropped representing peripheral vasoconstriction for approximately 1 ½ hours, then both temperatures increased together to a warmer temperature level.

Using 36.4° C as the cut-point for hypothermia, we originally wanted to examine whether there was increased observations with peripheral vasoconstriction below that limit. Because only one infant displayed peripheral vasoconstriction for minimal observations, we looked at whether the temperature difference between abdominal and foot temperatures (Δ T) increased when abdominal temperatures were lower than or equal to 36.4° C compared to when abdominal temperatures were greater than 36.4° C (see Table 3). Nine of the ten infants had significantly increased mean ΔT values between abdominal and peripheral temperatures, when their abdominal temperatures were less than or equal to 36.4° C compared to when their temperatures were greater than 36.4° C. Seven of these nine infants had negative mean Δ T values, meaning most of their peripheral temperatures were greater than their abdominal temperatures and one infant had a ΔT value that was very close to 0, meaning the abdominal and peripheral temperature were almost the same during most of the study (see t tests in Table 3). The remaining infant (D) had a significant increase in mean Δ T when abdominal temperature was less than or equal to 36.4° C, however, most of the abdominal temperatures were higher than the peripheral temperatures (+ΔT). Infant D (880 grams) had more observations that appeared to be towards peripheral vasoconstriction (abdominal > peripheral temperature by 2° C) when his abdominal temperatures were less than or equal to 36.4° C (out of 422 observations).

In addition, the correlation between extent to which Δ T increased or decreased and the number of degrees the abdominal temperature fell below 36.4° C was significant in 7 of 10 infants (see Table 3). Two of the significant correlations (infants “C” and “I”) were negative, meaning ΔT increased as abdominal temperature decreased. Both of these infants weighed between 500–600 grams. The five remaining infants had a significant positive correlation between ΔT and the decrease in abdominal temperature, meaning ΔT decreased as abdominal temperature fell. This would make theoretical sense as one would expect peripheral vasoconstriction (decrease in foot temperature) when abdominal temperature is decreased.

Looking at the graphic trends (see Figure 1) from the infant that exhibited peripheral vasoconstriction, it is evident that once the infant exhibited peripheral vasoconstriction, the abdominal temperature climbed higher than 36.4° C as the peripheral temperature decreased (symbolizing vasoconstriction). These data would not have been in the analysis of temperatures less than 36.4° C.

Study data collection took place for each infant’s first 12-hours of life; therefore, concomitant infant conditions were largely unknown. It can be assumed that each infant’s ductus arteriosus was either open or closing; however, it was impossible to know patency status during the study period. Additionally, blood culture results were not available until at least 48–72 hours after birth and were not collected as part of this study. Maternal antibiotics were administered prior to delivery in eight of the ten infants (Table 1). Two infants manifested hypotension requiring inotropic drugs. Infant B received a Dopamine infusion at one and a half hours of age, progressing to a Dobutamine infusion at seven hours of age, and then required an Epinephrine infusion at ten hours of age. The only other infant (E) in this sample with persistent hypotension during the first 12-hours of life received a Dopamine infusion at two hours of age. Mortality and morbidity data were not collected past the 12 hours of data collection.

Limitations in examining peripheral vasoconstriction in this study included that peripheral vasoconstriction is altered not only by thermoregulation but also by other factors stimulating the autonomic nervous system including responses to stressful stabilization procedures, the dynamic nature of transitional circulatory changes, and fluid and medication administration. This was a small pilot study, limited to 12-hours of data collection. Our subsequent study will examine status of ductus arteriosus when peripheral temperatures are greater than abdominal temperatures, ongoing disease status, morbidity, and mortality information to give a more complete clinical case presentation in which to analyze each subject.

Because this was a small study, results cannot be generalized to the larger population of preterm ELBW infants. Future studies need to include a larger sample of infants with analyses within subjects as well as between subjects, perhaps using stratification of gestational age and weight classes. Another limitation is the use of thermistors for temperature measurement. A temperature measured on the skin under a reflective tape is not necessarily equal to an adjacent skin surface site that is not covered, due to local vasodilatation caused by the insulated covering. Skin surface temperature can also be measured by themal imaging with infrared technology, which allows measurement of multiple skin surface sites simultaneously. Our future studies will employ this technique to better examine skin surface temperature for central and peripheral measurements. This study was also limited in that environmental temperatures were documented by an observer, and not recorded using research data loggers continuously. Our future study will measure temperature inside the incubator continuously to use in analyses with body temperature.

Lastly, this study and future studies are limited by the obstetrical prenatal assessment of gestational age. Gestational age was recorded by obstetrical dates or fetal ultrasound when available; both estimates have large variation in exact gestational age assessment. For this reason, we use birth weights as an approximate measure of maturity when the preterm infant is less than 1000 grams at birth.

We did not track brain hemorrhage outcomes in this small sample; however, future research should examine brain injury in association with body temperature and vasomotor tone during the first week of life. Because it is essential to understand physiological responses to hypothermia in ELBW infants which may lead to increased morbidity and mortality in this vulnerable population, we will continue to study peripheral vasoconstriction as a mechanism to conserve heat. Studying this concept in extremely preterm infants may lead to new information about maturation of vasomotor tone and add evidence to factors which may affect risk for intraventricular hemorrhage during the first week of life.