Body Temperatures of Very Low Birth Weight Infants on Admission to a Neonatal Intensive Care Unit

Abstract

Objective

Hypothermia occurs frequently in the first minutes after birth in preterm infants. Hyperthermia also occurs, often as a consequence of efforts to provide thermal support. Both hypothermia and hyperthermia are potentially harmful. Our objective was to examine the distribution of admission temperatures of very low birth weight (VLBW) infants, the effect of gestational age on admission temperatures, and the time required for correction of low temperatures.

Methods

Admission axillary temperatures were retrieved from the medical records for all VLBW infants born in our hospital during a 5-year period. The temperatures were classified as severe (<35.0°C), moderate (35.0–35.9°C), or mild (36.0–36.4°C) hypothermia, normothermia (36.5–37.4°C), or hyperthermia (≥37.5°C). The relationship between gestational age and admission temperature was examined. In addition, we analyzed the time required for normalization of low temperatures.

Results

Overall, 12% of infants were severely hypothermic, 40% moderately hypothermic, 27% mildly hypothermic, 19% normothermic, and 2% hyperthermic. Gestational age was inversely related to hypothermia risk and to the time required for recovery to normothermia.

Conclusion

Admission hypothermia is common among VLBW infants and is affected by gestational age.

Introduction

Hypothermia in the first minutes after birth is a common and potentially harmful event. After leaving the protected thermal environment of the uterus, newborn infants typically experience a fall in body temperature, often to levels that meet the World Health Organization definition of moderate hypothermia (body temperature <36.0°C) [1]. Preterm infants have greater difficulty maintaining normal body temperature than term infants and are at greater risk of hypothermia, with its associated increased risks of mortality and morbidity [2–6]. The increased vulnerability of preterm infants to hypothermia is explained by their inability to generate sufficient metabolic heat production to offset heat losses to the environment. These large heat losses result from a large surface-area-to-mass ratio, large evaporative water loss, meager tissue insulation, and limited ability to regulate vasomotor tone [7]. Preterm infants are also limited in their ability to transfer excess heat to the environment by vasodilating or sweating, which places them at risk of hyperthermia as their caregivers try to prevent or correct hypothermia [8]. Elevated body temperature is potentially as harmful as hypothermia, as it increases the risks of mortality and, in term infants with hypoxic-ischemic encephalopathy, impaired neurodevelopmental outcome [9, 10]. In an analysis of risks associated with admission temperature of preterm infants in the Canadian Neonatal Network, both hypothermia and hyperthermia were associated with increased risks of adverse outcomes [11].

The aims of this study were to: (a) examine the frequency and degree of low and high body temperature among very low birth weight (VLBW) infants at initial recording after admission to the neonatal intensive care unit (NICU); (b) examine the effect of gestational age on admission temperature; and (c) examine the time required for an infant with moderate or severe hypothermia to reach normal body temperature.

Methods

We conducted a single-center retrospective analysis of infants with birth weight <1500 grams born from May 1, 2008 through April 30, 2013 at the University of Iowa Hospitals and Clinics. Infants were placed under a radiant warmer and in plastic bags at birth, and beginning in October 2011, their heads were covered with plastic-lined hats. Once stabilized, they were moved to the neonatal intensive care unit, which is directly adjacent to the delivery suite.

Medical records of patients were reviewed, and data were entered into a computerized database [12]. Our standard clinical site for monitoring deep body temperature is the axilla, and normal axillary temperature is considered to be 36.5–37.4°C [13]. Data collected from the patient records included patient’s gestational age, birth weight, first recorded axillary temperature, age at first temperature reading, age at first axillary temperature ≥36.5°C, survival to discharge, and enrollment in follow-up.

The frequencies of mild, moderate, and severe hypothermia and of hyperthermia were examined. Axillary temperature at admission was classified as follows¹: normothermia, 36.5–37.4°C; mild hypothermia, 36.0–36.4°C; moderate hypothermia, 35.0–35.9°C; severe hypothermia, <35.0°C; hyperthermia, ≥37.5°C.

Descriptive statistics were calculated using SAS 9.4. The relation of admission temperature to gestational age was examined using linear regression. The relation between admission temperature and gestational age was determined by linear regression analysis. The relation of time to normothermia, defined as the time from admission temperature below 36.0°C to temperature ≥36.5°C, to gestational age was also examined using linear regression. The impact of introducing hats as a standard practice in October 2011 was examined in three ways. (1) The impact of hats on admission temperature was examined using the Wilcoxon rank sum test. (2) The impact of hat introduction on the incidence of admission temperatures below 35.0°C and below 36.0°C was examined using chi-square tests. (3) The impact of introducing hats on the relation between gestational age and admission temperature was examined using linear regression analysis. The study was approved by the University of Iowa institutional review board.

Results

We enrolled 532 infants in the study; 48.9% were male. Their median birth weight was 1028 g with interquartile range (IQR) 807–1304 g, and their median gestational age was 29 weeks (IQR 26 to 30 weeks). The median chronological age at first temperature was 0.3 h (IQR 0.2–0.4 h). Ninety-five percent of the patients survived to discharge.

The mean admission axillary temperature was 35.8°C with standard deviation (SD) 0.8°C. Among all infants, 12% had severe hypothermia (admission temperature <35.0°C), 40% had moderate hypothermia (35.0–35.9°C), 27% had mild hypothermia (36.0–36.4°C). Overall, 79% (421) of the infants had some degree of hypothermia; only 19% were normothermic (36.5–37.4°C); and 2% were hyperthermic (≥37.5°C). Of the 421 hypothermic infants, 66% (277) had moderate or severe hypothermia, i.e. temperature <36.0°C.

The frequency of admission temperatures <35.0°C and <36.0°C decreased with increasing gestational age (Table 1). For each completed week of gestation, admission temperature increased by 0.05°C (P <0.0001); this was true for the entire study and for the periods both before and after hats were introduced as standard care. The time required for the 277 infants with moderate or severe hypothermia to reach axillary temperature ≥36.5°C ranged from 4 to 1500 minutes (Table 1). The time to normothermia, defined as the time from the admission temperature to the first temperature reading ≥36.5°C, was inversely related to gestational age (P<0.05).

Table 1.

Incidence of hypothermia and time to reach normothermia by gestational age

Gestational age, wk	n	Birth weight, g, mean (SD)	Admission temperature <35.0°C, n (%)	Admission temperature <36.0°C, n (%)	Age (min) at first temperature of 36.5°C, median (IQR)*
22	9	497 (58)	5 (56)	9 (100)	90 (75–170)
23	26	557 (84)	14 (54)	22 (85)	150 (91–225)
24	42	663 (128)	7 (17)	28 (67)	105 (66–220)
25	31	802 (138)	5 (16)	16 (52)	88 (51–158)
26	50	837 (215)	6 (12)	27 (54)	90 (53–150)
27	56	929 (183)	2 (4)	24 (43)	69 (50–83)
28	73	1100 (225)	4 (5)	31 (42)	100 (58–139)
29	82	1193 (207)	7 (9)	35 (43)	65 (35–109)
30	65	1214 (247)	4 (5)	28 (43)	63 (29–117)
31	43	1271 (160)	2 (5)	22 (51)	30 (23–60)
32	27	1236 (207)	2 (7)	18 (67)	53 (31–66)
33	13	1308 (147)	0 (0)	9 (69)	49 (43–78)
34	6	1266 (208)	2 (33)	3 (50)	90 (45–99)
35	3	1454 (30)	1 (33)	3 (100)	30 (25–235)
36	5	1364 (98)	1 (20)	2 (40)	42 (38–45)
37	0	–	–	–	–
38	0	–	–	–	–
39	1	1424	0 (0)	0 (0)	–
All infants	532	1035 (299)	62 (12)	277 (52)	75 (44–125)

SD, standard deviation

^*For the 277 infants whose temperature was <36.0°C on admission

The admission temperatures, incidence of severe and moderate hypothermia, and the relation between gestational age and admission temperature were all unaffected by the introduction of hats in October 2011.

Discussion

In our 5-year survey of admission temperatures for VLBW infants, 52% had temperatures <36.0°C, quite similar to the incidence reported by others in the US and other developed countries [3,4,6,14–16]. Laptook et al found that 47% of all VLBW infants in 15 centers of the US NICHD Neonatal Research Network had initial temperature <36.0°C, and admission temperature was inversely related to mortality risk [3]. The incidence of hypothermia (<36.0°C) by hospital varied from approximately 15% to more than 80%. In the Canadian Neonatal Network, 12% of infants below 33 weeks of gestation had admission temperatures below 36.0°C [11]. Admission hypothermia (<36.0°C) has been reported in 51% of preterm infants of gestational ages 22–33 weeks in Brazil [15], and 40% of VLBW infants in Malaysia were found to have admission temperatures <36.0°C [16]. In low resource countries, the incidence of admission hypothermia among all infants, term and preterm, has been reported as ranging widely [17–20], from 22% [19] to 85% [20]. The current study adds to the previous work in this area by contributing data on the incidence of hypothermia by week of gestational age.

Our study confirms that the prevalence of admission hypothermia among VLBW infants has remained consistently high over the past decade, indicating that increased efforts are needed to protect these at-risk patients from the inevitable cold exposure associated with birth. Reducing hypothermia soon after birth can be accomplished using proven methods [21–23] and has the potential to reduce mortality and improve the outcomes of preterm infants.