Biol Neonate 1990;57:10-20
Radiant Heat Loss versus Radiant Heat Gain in Premature Neonates under Radiant Warmers' Stephen Baumgart Department of Pediatrics, University of Pennsylvania School of Medicine. Division of Neonatology. The Children's Hospital of Philadelphia. Pa.. USA
Key Words. Radiant warmers • Premature neonates • Convection • Evaporation • Metabolic rate Abstract. Premature infants nursed on open radiant warmer beds are exposed to shortwavelength infrared power density distributed evenly over the bed surface. Additionally, infants’ sides are exposed to relatively cooler nursery' walls, and to the radiant warmer bed platform which may heat and reradiate to the baby. Therefore, infants may not only gain heat from the warmer (Qradiam warmer) hut lose or gain radiant heat to the sides as well ( ± Qradiam loss)- In order to quantitate these parameters, ten premature newborn infants nursed under radiant warmers servocontrolled to 36.5 °C skin temperature (weight 1.27 ± 0.24 SD kg. gestation 31 ± 3 weeks) were investigated, and partitional calorimetry pre viously reported. In the present study, calculation of net rate of radiant heat transfer (Q nctradiant) w-as made from these data (-2.63 ± -1.52 kcal/kg/h), and compared to direct measurements of Qradiam warmer (-2.49 ± -0.90 kcal/kg/h). The present report further parti tions net radiant heat transfer to evaluate Qradiam lossi -0.13 ± 1.82 kcal/kg/h (range -3.16 to 1.93). From these calculations mean radiant temperature of this environment was estimated (45.3 ± 4.3 °C) and compared to the radiant warmer temperature received (45.0 ± 2.9 °C). This information suggests other strategies to reduce radiant heat loss as well as convective and evaporative losses in premature neonates nursed on open radiant warmer beds.
Powerful, nonionizing radiant heat sources are now being used routinely to con trol body temperature in critically ill. very low birth weight infants. The configuration of these heaters generally employs an over head radiant heat source which uniformly
distributes radiant power over a mattress surface on an open bed platform [1], The surface of an infant directly facing the ra diant warming element receives heat suffi cient to maintain anterior abdominal wall 1 This work was supported by McCabe Fund. NIH RR00240.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Introduction
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
mean radiant temperature experienced by premature infants maintained under these devices. Such information would be useful in devising a more homogenous radiant envi ronment for these infants. The purpose of the present report is to calculate the mean radiant temperature, and the net radiant heat gain experienced by neonates under ra diant warmers from a previously reported heat partition study of these infants, and to compare this calculated balance to direct measurements of radiant heat emitted by the warmers.
Methods Study Population Ten premature newborn infants (mean study weight 1.27 ± 0.24 SD kg and mean gestation 31 ± 3 weeks) recovering from respiratory disease were orig inally studied while nursed naked and supine under a radiant warmer [5], Warmers were servocontrolled to maintain anterior abdominal wall skin temperature at 36.5 °C which has recently been determined as ther mally neutral for these infants [3], All infants were less than 15 days old (mean age 8 ± 3 days). Every effort was made to study low weight infants as soon as possible during the critical phase of their hospital course, since this is the population most often requir ing care under radiant heaters. Each infant was stud ied for 90 min to determine rates of heat exchange through radiation, convection, conduction, evapora tion and metabolism. This protocol was approved by our Committee for the Protection of Human Subjects and informed parental consent was obtained.
Radiant Wanner The configuration of the radiant warmer used in this study has been described (Air Shields, Infant Care System. Hatboro. Pa.) [1]. Distribution of ra diant power density over the small area of the bed platform occupied by the infant is reasonably uni form and the warmer may be considered as a finite plane source of heat [ 1. 7], Side walls are located cir cumferentially around the bed platform and form a shallow well 11 cm deep wherein the infant lies. The
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
skin temperature at a preselected level, usually between 36 and 37 °C. This tempera ture is maintained by a proportional servocontroller which receives signals from an electronic thermistor taped to the infant's abdomen. Most clinicians agree it is sufficent to control this one site since empirical data demonstrate normal and stable infant core temperature and minimum oxygen con sumption using this approach [2, 3], The radiant temperature experienced by the low birth weight premature neonate un der a radiant warmer, however, may not be uniform, and is actually a complex summa tion of all radiant temperatures the infant experiences from all directions in which the infant lays exposed to surrounding solid surfaces [4], Under a warmer, the infant’s surface facing the heating element receives radiant heat, but the infant's sides are left exposed to the nursery walls and windows which are cool relative to skin temperature, and to the side walls and mattress of the warmer bed platform which are only heated indirectly by the radiant heating element. One can imagine a radiant hemisphere sur rounding the infant with a measurable mean radiant temperature which is deter mined to a large degree by the relatively hot radiant heating element located overhead. The rate of net radiant heat transfer would be determined by the gradient between the infant’s mean skin temperature and the mean hemispheric radiant temperature ac cording to the infant’s complex surface geometry and the geometry of his environ ment. Radiant heat gained from the warmer should accurately balance all heat losses from babies nursed in this open environ ment [5. 6], There are no data currently reported quantifying net radiant heat gain and the
11
Baumgart
12
Direct Meaurement o f Radiant Heat ini’ ( Q radiant warmer)
In vitro. Direct measurement of radiant heat re ceived at bed level from the radiant warmer was made in vitro using an Eppley Model E6 Thermopile (Eppley Laboratories Inc., Newport. R.I.). This device comprises a Coblentz-type thermopile which mea sures heat as radiant power density in mW/cm2 from radiant bodies relative to the ambient environment. The transducer port is shielded such that radiation admitted to the sensor plate subtends an arc of 65°. The range of wavelengths the thermopile senses is from 300 to 50,000 nm which encompasses the range emitted by the radiant warmer used in a clinical set ting. Radiant power density received by the thermo pile may be measured from zero to 50 mW/cm2 on a linear scale with this device. The absolute accuracy of this instrument compared to a standard radiant cali bration source (5,000 W tungsten filament at 3,000K) is ± 0.5%. The thermopile reference stan dard for calibration is room temperature at 50% rela tive humidity and atmospheric pressure. Tempera ture dependence of the Coblentz-type thermopile is only -0.15% /°C rise in ambient temperature. Resul tant reference error was, therefore, considered less than 1%. After equilibrating the radiant warmer bed for 1 h at one half the maximum heater power output, the thermopile was positioned at bed level, centered and directly facing the radiant warming element over head. This location was designated 90° perpendicular to the direction of radiant heat delivery. A reading of radiant power density was made after 15 s sufficient for the thermopile’s response to remain stable. The thermopile was then moved through an arc in 22.5° increments from 0°, facing the left warmer bed side wall, to 180° facing the right side wall. A similar power density reading was made at each position. In vivo. Heat received by infants from the radiant warmer (O radiam warmer) was monitored by a previously reported technique [8]. A wattmeter and the thermo pile were used during the study to quantify radiant heat delivered to the bed level surface, and a geomet ric model of the infant was employed to calculate the amount of incident energy absorbed by the infant (keal/kg/h). This model employs a series of cylinders
representing an infant's torso and extremities, and provides a measured body surface area for study infants [9]. The infant’s cross-sectional area projected at bed-level represents the portion of incident radiant heat received from the warmer, and may be deter mined from the cylindrical model. The 65° viewingangle of the thermopile aperture, oriented 90° per pendicular to the direction of radiant heat delivery, limits this measurement of Q radiant warmer to energyreceived by infants from the overhead heater and a small portion of surrounding nursery walls. The ther mopile reference temperature used to determine Q radiam » arac r for each infant was corrected for the dif ference between ambient environmental temperature and the infant’s mean skin temperature according to the Stefan-Boltzman relationship described below [10]. Details of this technique and treatment for sources of error have been previously reported [7], Calculation o f Net Radiant Heat Transfer and Radiant Ileal Loss (Q m i,a„t loss) The general equation for metabolic heat balance may be written as follows: (Qnet radiant)
Qmetabolic
Q siorcd
Qcvaporalion * Qconvcclion
Qradiaiion + Qconduction
where expressions on the right side of this equation designate heat losses and a positive sign convention is adopted. LeBlanc [11] has recently reported that the bed surface and walls, as well as nursery walls may be important radiant surfaces related to the infant’s ex posed sides and this mode of heat transfer will be considered in the analysis of Qradiaiion [11]. By rear ranging the general heat equation: Q n cl radiant = Qmctabolic ” Qcvaporalion — Qconvcclion ‘ Qconduciii
where QncI radiant represents the net radiant heat ex change with the infant’s environment required to bal ance the partition. Q radiam warmer * Q radiam I
A negative number for Q ncl radiant would indicate that the radiant warmer represents a heat gain (Q rad iam warmer) sufficient to balance all other radiant losses to the nursery walls as well as major heat losses represented by convection and evaporation. Q radiant loss, then, is simply the difference between Q net radiant I^SS Q radiant warmer measured directly.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
sides are completely open above these walls, freely permitting radiant, convective and evaporative heat exchange.
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
Qconvcclion
= (a + bVn) ( rm eanskin — r.i:r)
where (a) is a constant expressing heat loss through natural convection and bVn is an expression for forced convection when air velocity V is in m/s. The assumption these investigators made was that air ve locity within conventional incubators was negligible, hence the expression (bVn) was dropped from the equation. There is empirical evidence validating this assumption, since forced convection does not appar ently alter heat loss from adults when (V) is less than 0.20m/s. Bell and Rios [ 12] reported air velocities less than this in their study conducted in incubators. We have previously reported velocities for convective air currents under radiant warmers in this range. For the present study, air velocity was measured with an Alnor 8630 hot-wire anemometer as previously re ported [1], Measurements were made 6 cm above the mattress at the infant’s umbilicus and recorded by computer at I Hz. Mean air velocity was calculated during each 90-min study and reported. In each sub ject mean air velocity was less than or equal to 0.105 m/s, well below' the forced convective threshold [5], In the present study, therefore, the simplified formula of Bell and Rios [12] and empiric assumptions were used to recalculate (Qconvcclion) for study infants: Qconvcclion
= 0.34
( T mean skin
—3
air),
where a = 0.34 kcal/kg/h/ °C, T mcan skin is mean skin temperature in “C calculated from a weighted mean of cheek, abdomen and heel temperature and Talr is the temperature of the ambient nursery' air over the infant’s abdomen [13]. Yellow' Springs Series 400 thermistors were used to measure skin temperatures, and an Amer ican Instruments shielded temperature/hygrometer probe was used to measure air temperature and relative humidity 9 cm above the infant’s mattress.
Conductive Ileal Loss (Qconduction) Conductive heat loss occurs from infants depend ing on the area of skin in contact with the bed. the temperature gradient between the skin and bed sur face, and the insulating properties of bedding between the skin and bed surface. Hey and O’Connell [14] have estimated the thermal insulation of a double layer of flannel cotton to be 0.11 °C -m—h/kcal. Four such blankets folded twice (total of 16 layers) were
bunched together on the Potter Baby Scale’s load plat form with the baby naked and supine centered on top. Estimated total insulation of this technique was 0.88°C 'm 2-h/kcal. Potter Scale temperature mea sured during studies in 7 infants and averaged (30.3 °C). was subtracted from each infant’s mean skin temperature to obtain the conductive gradient. A fractional body surface area in contact with the blan kets was assumed to be 0.3. Conductive heat loss (Qconduction* kcal/kg/min) was thus calculated and nor malized to study weight according to a standard pub lished method [ 15].
Evaporative Heat Loss (Qnaporation)■ Metabolic Rate (Q metabolic)• and Heat Stored (Qstored) Values for these parameters were measured as pre viously reported [5, 12] and are summarized in the Appendix.
Calculation of Radiant Temperlures (Tradiant) Radiant temperatures of the infant’s environment (3*radiant warmer» 3'mcan radiam) may now be calculated from the above data using the Stefan-Boltzmann rela tionship: Q radtant = O * Bskin * ^environment (3'm can skin*1 ~ 3'radiant** )
A• F•W
1
where a is the Stefan-Boltzmann constant 4.93 X 10"8 kcal/h/m2/K -4; eskin is the emissivity of the infant’s skin which is assumed to be one since infants are con sidered black bodies [12]. The emissivity of the com plex radiant environment, environment» will be assumed to be one for the purpose of this calculation. For the overhead radiant plane dominated by the radiant warming element this is probably true: however, emissivities for the various walls and other surfaces near to the infant w'ere not determined and their rela tive importances weighted. Overestimation of emis sivity in the calculation of Tra(jjam may result in modest overestimation of this value. Temperatures are con verted to Kby adding 273 to °C. The quantity A is the body surface area calculated from Haycock et al. [9]. The fraction of effective radiating body surface ex posed (F) was assumed to be 0.5 after the estimation of Bell and Rios [12], and W is weight in kg. Rearranging these values and solving for Tradiam, this expression becomes: -r 4 / t* j * radiant ~ v ■ * moan skin *“
V radiant * G
”
* 6$kin * ^environment * A • r
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Convective Heal Loss (Ocomection) Recently. Bell and Rios [12] have simplified the general calculation of convective heat loss:
13
14
Fig. 1. Radiant heat density measured in an arc simulating a baby's sides exposed to room wall tem perature and the overhead warmer element. Note thermopile used was referenced to room temperature approximately 29 °C in the nursery. Infants with skin temperature reference would experience radiant heat density corrected by - 7 mW/cm-.
Results The results for direct radiant power den sity measurement made by the thermopile at bed level in vitro are shown in figure 1. This figure demonstrates that the incident radiant heat is greatest from the warmer on the side of an object directly facing the warmer (8.7 mW/cm2 at 90°). When moved through an arc from one side of the warmer bed plat form to the other, the radiant power density falls rapidly to 0.0 mW/cm2 on the thermo pile when facing the walls of the nursery. Since the thermopile’s reference standard is the ambient nursery temperature, the zero reading implies no radiant heat gradient rel ative to normal room wall temperature. An interesting observation made with the ther mopile facing the side walls of the warmer
bed platform was that radiant power density again increased to 0.9 mW/cm2, suggesting that the side walls and bed surface were sec ondary irradiators assimilating and reradiat ing heat from the warming element. Table 1 presents the partitional calorime try derived from environmental assessment, and measurements of infant metabolic rate, water loss, and temperature. Mean skin tem perature for these infants was 36.3 ± 0.3 SD °C, and mean air temperature was 29.1 ± 2.0 °C yielding a skin to air temperature transfer gradiant of 7.3 ± 2.0 °C. The pres ence of forced convection was negligible with respect to convective heat loss as pre dicted, in every infant being less than 0.105 m/s air velocity. The rate of convective heat loss from infants recalculated from these re sults according to the assumptions described was 2.92 ± 0.80 kcal/kg/h. Evaporative heat loss, calculated from weight loss and the difference in weights of metabolic gases exchanged was 1.39 ± 1.05 kcal/kg/h as reported [5]. Insensible water loss concomitantly ranged from 1.14 to 7.33 ml/kg/h, the largest losses tending to occur in the smallest and least mature infants. Con ductive heat loss was calculated at 0.19 ± 0.03 kcal/kg/h. Mean metabolic rate of heat production was 1.90 ± 0.46 kcal/kg/h and heat storage was negligible. From these partitional data, net radiant heat transfer and mean radiant temperature for the infants’ complex environment was calculated (table 2): mean Qnet radiant was -2.63 ± -1.52 kcal/kg/h, indicating that for every infant net heat transfer was toward the infant, and Tmcan radiam was 45.3 ± 4.3 °C, a relatively warm environmental hemisphere. Table 2 also displays thermopile measure ments of radiant heat delivered to infants directly by the warmer: mean Q radiant warmer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Baumgart
15
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
Table 1. Partitional calorimetry Qconvcction
A 1 Vcvapo ration
Qcondu:tion
keal/kg/h
keal/kg/h
Q metabolic* keal/kg/h
O sto red 1
keal/kg/h 1
3.90
4.25
0.26
2.56
0.23
2
2.14
1.51
0.20
2.13
0.01
3
2.41
0.85
0.21
2.09
-0.05
4
2.31
0.66
0.19
2.40
0.05
Patient No.
keal/kg/h
5
1.78
1.60
0.17
1.48
0.11
6
3.82
0.97
0.20
1.18
-0.03
7
3.73
0.86
0.19
2.22
-0.06
8
2.89
0.82
0.19
1.81
-0.02
9
3.73
1.07
0.17
1.43
0.04
10
2.51
1.30
0.17
1.65
-0.13
Mean
2.92
1.39
0.19
1.90
0.02
SD
0.80
1.05
0.03
0.46
0.10
1 Previously reported data [5].
Patient No.
O nct radiant
keal/kg/h
Tm can radiant
°c
Q radiant warmer
keal/kg/h
Tradiant warmer
Q radiant loss
°c
keal/kg/h
-2.92
45.1
-3.16
42.1
-3.66
48.1
1.93
40.6
-2 .7 6
45.0
1.43
1
-6.08
53.4
2
-1.73
3
-1.33
4
-0.81
39.3
-2.30
44.3
1.49
5
-2.18
44.8
-2.66
46.6
0.48
6
-3.78
49.6
-1.96
43.5
-1.82
7
-2.5 0
45.6
-0.82
39.6
-1.68
8
-2.07
44.2
-2.19
44.6
0.12
9
-3.58
49.0
-1.78
42.6
-1.80
10
-2.20
44.3
-3.87
50.2
1.67
Mean
-2.63
45.3
-2.49
45.0
-0.13
SD
-1.52
4.3
-0.90
2.9
1.82
Q radiant loss " Q nel radiant ~ Q radiant warmer*
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Table 2. Calculation of mean environmental radiant temperatures from partitional calorimetry vs. direct measurement
16
was -2.49 ± -0.90 kcal/kg/h. The radiant temperature of the warmer experienced by infants was 45.0 ± 2.9 °C. Therefore, mean Qradiam loss calculated from the partition is -0.13 ± 1.82 (range-3.16 to 1.93) kcal/kg/ h.
Discussion These data demonstrate that low birth weight infants nursed on open beds under radiant warmers experience an environment which differs from the convection warmed incubator in two important ways. First, air temperature is much cooler (29.1 °C). Sec ond, cooler air temperature results in a re quirement for considerably higher mean ra diant temperature (45.3 °C) to maintain both a normal core temperature, and skin temperature within the predicted thermal neutral range. The radiant warmers in the present study provided a similar radiant temperature to meet this demand (45.0 °C). Moreover, in vitro studies indicate that the mean radiant temperature of the environ ment is not homogenous in all directions from the infant's position on the bed. A warmer temperature is experienced by the surface of the infant exposed perpendicu larly to the warming device, while the in fant’s sides exposed to the nursery walls may perceive the cooler nursery wall tempera ture. It is interesting to note that the bed's side walls and perhaps mattress surface con stitute secondary irradiators presenting ra diant surfaces somewhat warmer than the nursery walls. Also, radiant heat loss to the sides of individual infants ranged from posi tive to negative values, indicating a complex summation of net heat losses and gains from bed platform, nursery walls, and the radiant heater.
Bell and Rios [ 12] have recently described the partition of heat losses for single- and doublewalled incubators. These authors described relatively temperate environments wherein air and wall temperatures were within a few degrees of the infant's skin and core temper atures. Using their model to evaluate convec tive heat loss, we have described a different environment under the radiant warmer. Heat losses via convection and evaporation are of a magnitude sufficient to require a mean ra diant temperature much higher than the com parable mean wall temperature in the incuba tor. This higher mean radiant temperature is generated by the servocontrolled radiant warmer assigned to maintain the anterior ab dominal wall temperature at 36.5 °C. This surface of the infants in the present study was directly facing the heater and probably con stituted the warmest side of these infants. Even though the exact radiant and convective geometry of an infant’s body is complex, the model we cited is relatively conservative in estimating the infant's exposure to convec tive heat loss which may have resulted in an underestimation of Tmcan radiant in the present calculation. Bell and Rios [12] and LeBlanc [11] have previously described similar results regard ing radiant heat transfer in the relatively more temperate incubator environment. An important part of radiant heat transfer de scribed in LeBlanc’s study was the relatively large contribution of the mattress surface immediately adjacent to the baby’s sides to radiant heat loss. As in the present study, emissivity of the surrounding surfaces was assumed to be near unity. The results of our study suggest that the sides of an infant nursed under a radiant warmer which face room walls are exposed to a radiant temper ature which is less than for infants in incuba
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Baumgarl
tors, and to the side walls of the warmer bed and mattress which are probably secondarily heated by the warmer. LeBlanc [16] specu lates that such environmental heterogeneity under the radiant warmer may result in a slightly increased metabolic rate compared to the standard incubator. Information regarding radiant tempera tures may be useful, then, in designing ra diantly warmed environments wherein the overhead heater not only warms the infant’s exposed surface area, but also a series of stra tegically placed secondary irradiating sur faces. This tactic would significantly reduce radiant heat loss from the infant's sides and moderate the environmental temperature re quired to maintain thermal equilibrium. Several types of clear plastic heat shields have been advocated in this capacity for cov ering infants under radiant warmers. These shields have ranged from saran blankets to rigid plastic hoods as much as 3 mm in thick ness. Such devices may also serve as second arily heated walls since they are partially opaque to the infrared heat from the warmer [17], Certain types of shields may function to reduce the radiant temperature gradient be tween the infant’s skin and the surrounding walls while still permitting the radiant heat from the warmer through from above. One shield tested, (a 3-mm-thick plastic hood) absorbed almost all of the infrared energy from the servocontrolled warmer and re sulted in an increased power demand from the warmer to maintain infant skin tempera ture [ 17], The reasons for heterogeneity observed in the calculation of radiant heat loss as the dif ference between net radiant balance (parti tion), and radiant gain from the warmer (thermopile), may be inherent in measure ment errors encountered with partitional
17
calorimetry (Appendix). However, it is impor tant to note that in the present study when the gradient between ambient air and skin tem peratures is high (i.c. high Q COnvcction)> radiant heat loss is low or even negative (indicating heat gain). Either high convective losses may predispose to higher servocontrolled heater outputs, or higher heater outputs correspond to warmer secondary irradiators in the envi ronment. This latter hypothesis would require experimental trial to justify the use of second ary or passive radiant heat sources in an infant’s environment. In conclusion, mean environmental ra diant temperature is greater under radiant warmers than inside convectively heated in cubators. This mean radiant temperature is several degrees warmer than the infant’s skin suggesting that despite probable radiant losses from the infant’s sides to the nursery walls, net radiant heat transfer is dominated by the radiant heating element overhead and is toward the infant. The primary source of this radiant heat is from the radiant warming element, however, the mattress and sidewalls of the bed platform may be secondary irradiators receiving heat from the warmer and reradiating it to the infant. The rela tively higher mean radiant temperature the infant experiences on the radiant warmer bed also suggests that heat losses via evapo ration, convection and perhaps radiation to the sides are increased over the incubator environment. Heated walls, mattresses or other shielding interventions must be care fully evaluated to determine their effect on the complete thermal environment before being routinely applied to infant care. The goal of these devices should be to reduce radiant, evaporative and convective losses from infants who must be nursed in an open and accessible fashion.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
18
Baumgatt
Scale of equilibration
Appendix Metabolic Rate (& m etabolic) Metabolic rate (kcal/kg/h) was calculated from oxygen consumption (V0,. ml/kg/min) and carbon dioxide production (Vc0,, ml/kg/min) according to the relationship described by Lusk [18]. An open cir cuit technique was used to monitor the rates of gas exchange continuously [19]. The details of our com puterized metabolic monitoring system have been re ported previously [3, 5]. The relative accuracy of this apparatus has been determined in vitro to be ± 3.0% for V0:, VCo;. and respiratory quotient by previously established techniques [20]. Evaporative Heat Loss ((¿evaporation) Infant weight loss (IL. g/kg/h) was monitored con tinuously by a Potter baby scale (absolute accuracy ± 0.5 g) and a Hewlett-Packard strip-chart recorder during a 90-min sampling period. Insensible weight loss thus determined was corrected for the weight of respiratory gas exchange to determine insensible wa ter loss (IWL, ml/kg/h): IWL = IL -(0.1179 VC0: -0.0857 V0>). This result was multiplied by 0.58 kcal/ ml of water evaporated at 36.5 °C to determine the rate of evaporative heat loss in kcal/kg/h. Five sources of potential error have been de scribed previously during validation of these proce
Fig. A2. Thermal equilibration of infant scale.
dures [5, 7], Briefly summarized, these sources are: (1) scale hysteresis; (2) load drag from wires and tub ing on the scale platform: (3) condensation of respira tory gas humidification; (4) fluid intake and urine and stool weights excreted, and (5) thermal expan sion of the scale under the warmer. An example of scale calibration is shown in the figure A l. Numeral I indicates removal of a 5-gram weight from one exper imental apparatus with an infant and tubing, etc. in place. The polygraph record subtends 25 mm within 30 s and represents less than 0.5 mm distortion with a light tap from below the platform (Numeral 2). Hys teresis and load drag combined produce less than 0.1 g error in absolute weight determination. Conden sation was not detected in our respiratory support apparatus both in in vitro and in vivo experiments [21]. Respiratory tubing is passively warmed under the radiant heater and unless super-saturated gas mix tures are used, the 'rain-out’ phenomenon with infant water aspiration is not routinely observed. Fluid in take and urine and stool excretion are measured dur ing our studies, and the polygraph rezeroed to ac count for weight adjustment. Thermal expansion of the scale apparatus under the radiant heater was taken into account by providing adequate time for the scale to reach thermal equilibrium before commenc ing data acquisition. The time constant for scale equilibration was estimated empirically as 57 min.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Fig. A l. Scale calibration.
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
First hour
Second hour
0.51 2.63 3.52 3.50 2.45 1.58 2.15 1.07
0.51 2.17 2.54 3.94 2.45 1.58 2.12 1.42
2.18 0.38
2.09 0.35
(fig. A2), therefore a minimum of 3 h was alotted before an experiment (less than 0.1 g error in weight incurred). Replicability of infant weight change evaluations has been previously tested in eight premature subjects [7]. Table AI demonstrates weight loss during 2 con secutive hours in this study. Disregarding possible changes in infant physiology and/or environment, mean weight change was the same for both hours (dif ference less than 0.1 g absolute, p = 0.87). Standard error was less than 0.4 g/kg/h). These data suggest that cumulative error in measuring insensible water loss using these techniques is less than the ± 0 .5 g re ported for the Potter scale. Error in calculating evapo rative heat loss from these data is probably less than 0.3 keal/kg/h. Heal Stored (Quoted) Q siorcd (keal/kg/h) should be nearly zero if the infant is at thermal equilibrium and neither gains nor loses heat. Q slored may be calculated over the 90-min study for each infant: Q stored
= 0.84 (0.6 ATrectai + 0.4 AT mean skin)*
where 0.84 kcal/kg/°C is the specific heat constant for human tissue and AT (°C ) is the difference in temperature between the end and the beginning of the study. Rectal temperatures were measured with a Yel low Springs rectal thermistor inserted 5 cm past the anus.
1 Baumgart S. Engle WD. Fox WW. et al: Effect of heat shielding on convection and evaporation, and radiant heat transfer in the premature infant. J Pediatr 1981:99:948-956. 2 Robinson RO. Jones R: Advantages and disad vantages of overhead radiant heaters. Proc R Soc Med 1977:70:209. 3 Malin SW. Baumgart S: Optimal thermal manage ment for low birth weight infants nursed under high-powered radiant warmers. Pediatrics 1987; 79:47. 4 Hey EN. Katz G: The optimum thermal environ ment for naked babies. Arch Dis Child 1970:45: 328. 5 Baumgart S: Partitioning of heat losses and gains in premature newborn infants under radiant warmers. Pediatrics 1985:75:89. 6 Wheldon AC, Rutter N: The heat balance of small babies nursed in incubators and under radiant warmers. Early Human Dev 1982:6:131. 7 Baumgart S, Engle WD. Fox WW. et al: Radiant warmer power and body size as determinants of insensible water loss in the critically ill neonate. Pediatr Res 1981:15:1495. 8 Baumgart S, Engle WD. Langman CB, et al: Mon itoring radiant power in the critically ill newborn under a radiant warmer. Crit Care Med 1980;8: 721. 9 Haycock GB. Schwartz GJ. Wisotsky DH: Geo metric method for measuring body surface areas: a height-weight formula validated in infants, chil dren and adults. J Pediatr 1978:93:62. 10 LeBlanc MH: Personal communication. 1988. 11 LeBlanc MH: Oxygen consumption in premature infants in an incubator of proven clinical efficacy. Biol Neonate 1983,44:76. 12 Bell EF, Rios GR: A double-walled incubator alters the partition of body heat loss of premature infants. Pediatr Res 1983:17:135. 13 Silverman WA, Sinclair JC, Agate FJ: The oxygen cost of minor changes in heat balance of small newborn infants. Acta Paediatr Scand 1966:55: 294. 14 Hey EN, O’Connell B: Oxygen consumption and heal balance in the colnursed baby. Arch Dis Child 1970:45:335. 15 Swyer PR: Heat loss after birth; in Sinclair JC (ed): Temperature Regulation and Energy Metab-
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
References
Table AI. Weight loss, g/kg/h
Mean SEM
19
16
17
18
19
20
Baumgart
olisrn in the Newborn, New York, Gruñe & Strat ton. 1978, p 92. LeBlanc MH: Relative efficacy of an incubator and an open warmer in producing thermoneutral ity for the small premature infant. Pediatrics 1982;69:439. Baumgart S, Fox WW, Polin RA: Physiologic im plications of two different heat shields for infants under radiant warmers. J Pediatr 1982:100:787. Lusk G: The Elements of the Science of Nutrition, ed 4.. Philadelphia, WB Saunders, 1928. pp 6468 . Lister G. Hoffman JIE. Rudolph AM: Oxygen uptake in infants and children: a simple method for measurement. Pediatrics 1974;53:656. Marks K.H. Coen P, Kerrigan JR. et al: The accu racy and precision of an open-circuit system to measure oxygen consumption and carbon dioxide production in neonates. Pediatr Res 1987;21:58.
21 Sosulski R. Polin RA, Baumgart S: Respiratory water loss and heat balance in intubated infants receiving humidified air. J Pediatr 1983:103: 307.
Stephen Baumgart, MD Associate Professor of Pediatrics Department of Pediatrics University of Pennsylvania School of Medicine Division of Neonatology The Children's Hospital of Philadelphia Philadelphia. PA 19104 (USA)
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
20
Radiant Heat Loss versus Radiant Heat Gain in Premature Neonates under Radiant Warmers' Stephen Baumgart Department of Pediatrics, University of Pennsylvania School of Medicine. Division of Neonatology. The Children's Hospital of Philadelphia. Pa.. USA
Key Words. Radiant warmers • Premature neonates • Convection • Evaporation • Metabolic rate Abstract. Premature infants nursed on open radiant warmer beds are exposed to shortwavelength infrared power density distributed evenly over the bed surface. Additionally, infants’ sides are exposed to relatively cooler nursery' walls, and to the radiant warmer bed platform which may heat and reradiate to the baby. Therefore, infants may not only gain heat from the warmer (Qradiam warmer) hut lose or gain radiant heat to the sides as well ( ± Qradiam loss)- In order to quantitate these parameters, ten premature newborn infants nursed under radiant warmers servocontrolled to 36.5 °C skin temperature (weight 1.27 ± 0.24 SD kg. gestation 31 ± 3 weeks) were investigated, and partitional calorimetry pre viously reported. In the present study, calculation of net rate of radiant heat transfer (Q nctradiant) w-as made from these data (-2.63 ± -1.52 kcal/kg/h), and compared to direct measurements of Qradiam warmer (-2.49 ± -0.90 kcal/kg/h). The present report further parti tions net radiant heat transfer to evaluate Qradiam lossi -0.13 ± 1.82 kcal/kg/h (range -3.16 to 1.93). From these calculations mean radiant temperature of this environment was estimated (45.3 ± 4.3 °C) and compared to the radiant warmer temperature received (45.0 ± 2.9 °C). This information suggests other strategies to reduce radiant heat loss as well as convective and evaporative losses in premature neonates nursed on open radiant warmer beds.
Powerful, nonionizing radiant heat sources are now being used routinely to con trol body temperature in critically ill. very low birth weight infants. The configuration of these heaters generally employs an over head radiant heat source which uniformly
distributes radiant power over a mattress surface on an open bed platform [1], The surface of an infant directly facing the ra diant warming element receives heat suffi cient to maintain anterior abdominal wall 1 This work was supported by McCabe Fund. NIH RR00240.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Introduction
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
mean radiant temperature experienced by premature infants maintained under these devices. Such information would be useful in devising a more homogenous radiant envi ronment for these infants. The purpose of the present report is to calculate the mean radiant temperature, and the net radiant heat gain experienced by neonates under ra diant warmers from a previously reported heat partition study of these infants, and to compare this calculated balance to direct measurements of radiant heat emitted by the warmers.
Methods Study Population Ten premature newborn infants (mean study weight 1.27 ± 0.24 SD kg and mean gestation 31 ± 3 weeks) recovering from respiratory disease were orig inally studied while nursed naked and supine under a radiant warmer [5], Warmers were servocontrolled to maintain anterior abdominal wall skin temperature at 36.5 °C which has recently been determined as ther mally neutral for these infants [3], All infants were less than 15 days old (mean age 8 ± 3 days). Every effort was made to study low weight infants as soon as possible during the critical phase of their hospital course, since this is the population most often requir ing care under radiant heaters. Each infant was stud ied for 90 min to determine rates of heat exchange through radiation, convection, conduction, evapora tion and metabolism. This protocol was approved by our Committee for the Protection of Human Subjects and informed parental consent was obtained.
Radiant Wanner The configuration of the radiant warmer used in this study has been described (Air Shields, Infant Care System. Hatboro. Pa.) [1]. Distribution of ra diant power density over the small area of the bed platform occupied by the infant is reasonably uni form and the warmer may be considered as a finite plane source of heat [ 1. 7], Side walls are located cir cumferentially around the bed platform and form a shallow well 11 cm deep wherein the infant lies. The
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
skin temperature at a preselected level, usually between 36 and 37 °C. This tempera ture is maintained by a proportional servocontroller which receives signals from an electronic thermistor taped to the infant's abdomen. Most clinicians agree it is sufficent to control this one site since empirical data demonstrate normal and stable infant core temperature and minimum oxygen con sumption using this approach [2, 3], The radiant temperature experienced by the low birth weight premature neonate un der a radiant warmer, however, may not be uniform, and is actually a complex summa tion of all radiant temperatures the infant experiences from all directions in which the infant lays exposed to surrounding solid surfaces [4], Under a warmer, the infant’s surface facing the heating element receives radiant heat, but the infant's sides are left exposed to the nursery walls and windows which are cool relative to skin temperature, and to the side walls and mattress of the warmer bed platform which are only heated indirectly by the radiant heating element. One can imagine a radiant hemisphere sur rounding the infant with a measurable mean radiant temperature which is deter mined to a large degree by the relatively hot radiant heating element located overhead. The rate of net radiant heat transfer would be determined by the gradient between the infant’s mean skin temperature and the mean hemispheric radiant temperature ac cording to the infant’s complex surface geometry and the geometry of his environ ment. Radiant heat gained from the warmer should accurately balance all heat losses from babies nursed in this open environ ment [5. 6], There are no data currently reported quantifying net radiant heat gain and the
11
Baumgart
12
Direct Meaurement o f Radiant Heat ini’ ( Q radiant warmer)
In vitro. Direct measurement of radiant heat re ceived at bed level from the radiant warmer was made in vitro using an Eppley Model E6 Thermopile (Eppley Laboratories Inc., Newport. R.I.). This device comprises a Coblentz-type thermopile which mea sures heat as radiant power density in mW/cm2 from radiant bodies relative to the ambient environment. The transducer port is shielded such that radiation admitted to the sensor plate subtends an arc of 65°. The range of wavelengths the thermopile senses is from 300 to 50,000 nm which encompasses the range emitted by the radiant warmer used in a clinical set ting. Radiant power density received by the thermo pile may be measured from zero to 50 mW/cm2 on a linear scale with this device. The absolute accuracy of this instrument compared to a standard radiant cali bration source (5,000 W tungsten filament at 3,000K) is ± 0.5%. The thermopile reference stan dard for calibration is room temperature at 50% rela tive humidity and atmospheric pressure. Tempera ture dependence of the Coblentz-type thermopile is only -0.15% /°C rise in ambient temperature. Resul tant reference error was, therefore, considered less than 1%. After equilibrating the radiant warmer bed for 1 h at one half the maximum heater power output, the thermopile was positioned at bed level, centered and directly facing the radiant warming element over head. This location was designated 90° perpendicular to the direction of radiant heat delivery. A reading of radiant power density was made after 15 s sufficient for the thermopile’s response to remain stable. The thermopile was then moved through an arc in 22.5° increments from 0°, facing the left warmer bed side wall, to 180° facing the right side wall. A similar power density reading was made at each position. In vivo. Heat received by infants from the radiant warmer (O radiam warmer) was monitored by a previously reported technique [8]. A wattmeter and the thermo pile were used during the study to quantify radiant heat delivered to the bed level surface, and a geomet ric model of the infant was employed to calculate the amount of incident energy absorbed by the infant (keal/kg/h). This model employs a series of cylinders
representing an infant's torso and extremities, and provides a measured body surface area for study infants [9]. The infant’s cross-sectional area projected at bed-level represents the portion of incident radiant heat received from the warmer, and may be deter mined from the cylindrical model. The 65° viewingangle of the thermopile aperture, oriented 90° per pendicular to the direction of radiant heat delivery, limits this measurement of Q radiant warmer to energyreceived by infants from the overhead heater and a small portion of surrounding nursery walls. The ther mopile reference temperature used to determine Q radiam » arac r for each infant was corrected for the dif ference between ambient environmental temperature and the infant’s mean skin temperature according to the Stefan-Boltzman relationship described below [10]. Details of this technique and treatment for sources of error have been previously reported [7], Calculation o f Net Radiant Heat Transfer and Radiant Ileal Loss (Q m i,a„t loss) The general equation for metabolic heat balance may be written as follows: (Qnet radiant)
Qmetabolic
Q siorcd
Qcvaporalion * Qconvcclion
Qradiaiion + Qconduction
where expressions on the right side of this equation designate heat losses and a positive sign convention is adopted. LeBlanc [11] has recently reported that the bed surface and walls, as well as nursery walls may be important radiant surfaces related to the infant’s ex posed sides and this mode of heat transfer will be considered in the analysis of Qradiaiion [11]. By rear ranging the general heat equation: Q n cl radiant = Qmctabolic ” Qcvaporalion — Qconvcclion ‘ Qconduciii
where QncI radiant represents the net radiant heat ex change with the infant’s environment required to bal ance the partition. Q radiam warmer * Q radiam I
A negative number for Q ncl radiant would indicate that the radiant warmer represents a heat gain (Q rad iam warmer) sufficient to balance all other radiant losses to the nursery walls as well as major heat losses represented by convection and evaporation. Q radiant loss, then, is simply the difference between Q net radiant I^SS Q radiant warmer measured directly.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
sides are completely open above these walls, freely permitting radiant, convective and evaporative heat exchange.
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
Qconvcclion
= (a + bVn) ( rm eanskin — r.i:r)
where (a) is a constant expressing heat loss through natural convection and bVn is an expression for forced convection when air velocity V is in m/s. The assumption these investigators made was that air ve locity within conventional incubators was negligible, hence the expression (bVn) was dropped from the equation. There is empirical evidence validating this assumption, since forced convection does not appar ently alter heat loss from adults when (V) is less than 0.20m/s. Bell and Rios [ 12] reported air velocities less than this in their study conducted in incubators. We have previously reported velocities for convective air currents under radiant warmers in this range. For the present study, air velocity was measured with an Alnor 8630 hot-wire anemometer as previously re ported [1], Measurements were made 6 cm above the mattress at the infant’s umbilicus and recorded by computer at I Hz. Mean air velocity was calculated during each 90-min study and reported. In each sub ject mean air velocity was less than or equal to 0.105 m/s, well below' the forced convective threshold [5], In the present study, therefore, the simplified formula of Bell and Rios [12] and empiric assumptions were used to recalculate (Qconvcclion) for study infants: Qconvcclion
= 0.34
( T mean skin
—3
air),
where a = 0.34 kcal/kg/h/ °C, T mcan skin is mean skin temperature in “C calculated from a weighted mean of cheek, abdomen and heel temperature and Talr is the temperature of the ambient nursery' air over the infant’s abdomen [13]. Yellow' Springs Series 400 thermistors were used to measure skin temperatures, and an Amer ican Instruments shielded temperature/hygrometer probe was used to measure air temperature and relative humidity 9 cm above the infant’s mattress.
Conductive Ileal Loss (Qconduction) Conductive heat loss occurs from infants depend ing on the area of skin in contact with the bed. the temperature gradient between the skin and bed sur face, and the insulating properties of bedding between the skin and bed surface. Hey and O’Connell [14] have estimated the thermal insulation of a double layer of flannel cotton to be 0.11 °C -m—h/kcal. Four such blankets folded twice (total of 16 layers) were
bunched together on the Potter Baby Scale’s load plat form with the baby naked and supine centered on top. Estimated total insulation of this technique was 0.88°C 'm 2-h/kcal. Potter Scale temperature mea sured during studies in 7 infants and averaged (30.3 °C). was subtracted from each infant’s mean skin temperature to obtain the conductive gradient. A fractional body surface area in contact with the blan kets was assumed to be 0.3. Conductive heat loss (Qconduction* kcal/kg/min) was thus calculated and nor malized to study weight according to a standard pub lished method [ 15].
Evaporative Heat Loss (Qnaporation)■ Metabolic Rate (Q metabolic)• and Heat Stored (Qstored) Values for these parameters were measured as pre viously reported [5, 12] and are summarized in the Appendix.
Calculation of Radiant Temperlures (Tradiant) Radiant temperatures of the infant’s environment (3*radiant warmer» 3'mcan radiam) may now be calculated from the above data using the Stefan-Boltzmann rela tionship: Q radtant = O * Bskin * ^environment (3'm can skin*1 ~ 3'radiant** )
A• F•W
1
where a is the Stefan-Boltzmann constant 4.93 X 10"8 kcal/h/m2/K -4; eskin is the emissivity of the infant’s skin which is assumed to be one since infants are con sidered black bodies [12]. The emissivity of the com plex radiant environment, environment» will be assumed to be one for the purpose of this calculation. For the overhead radiant plane dominated by the radiant warming element this is probably true: however, emissivities for the various walls and other surfaces near to the infant w'ere not determined and their rela tive importances weighted. Overestimation of emis sivity in the calculation of Tra(jjam may result in modest overestimation of this value. Temperatures are con verted to Kby adding 273 to °C. The quantity A is the body surface area calculated from Haycock et al. [9]. The fraction of effective radiating body surface ex posed (F) was assumed to be 0.5 after the estimation of Bell and Rios [12], and W is weight in kg. Rearranging these values and solving for Tradiam, this expression becomes: -r 4 / t* j * radiant ~ v ■ * moan skin *“
V radiant * G
”
* 6$kin * ^environment * A • r
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Convective Heal Loss (Ocomection) Recently. Bell and Rios [12] have simplified the general calculation of convective heat loss:
13
14
Fig. 1. Radiant heat density measured in an arc simulating a baby's sides exposed to room wall tem perature and the overhead warmer element. Note thermopile used was referenced to room temperature approximately 29 °C in the nursery. Infants with skin temperature reference would experience radiant heat density corrected by - 7 mW/cm-.
Results The results for direct radiant power den sity measurement made by the thermopile at bed level in vitro are shown in figure 1. This figure demonstrates that the incident radiant heat is greatest from the warmer on the side of an object directly facing the warmer (8.7 mW/cm2 at 90°). When moved through an arc from one side of the warmer bed plat form to the other, the radiant power density falls rapidly to 0.0 mW/cm2 on the thermo pile when facing the walls of the nursery. Since the thermopile’s reference standard is the ambient nursery temperature, the zero reading implies no radiant heat gradient rel ative to normal room wall temperature. An interesting observation made with the ther mopile facing the side walls of the warmer
bed platform was that radiant power density again increased to 0.9 mW/cm2, suggesting that the side walls and bed surface were sec ondary irradiators assimilating and reradiat ing heat from the warming element. Table 1 presents the partitional calorime try derived from environmental assessment, and measurements of infant metabolic rate, water loss, and temperature. Mean skin tem perature for these infants was 36.3 ± 0.3 SD °C, and mean air temperature was 29.1 ± 2.0 °C yielding a skin to air temperature transfer gradiant of 7.3 ± 2.0 °C. The pres ence of forced convection was negligible with respect to convective heat loss as pre dicted, in every infant being less than 0.105 m/s air velocity. The rate of convective heat loss from infants recalculated from these re sults according to the assumptions described was 2.92 ± 0.80 kcal/kg/h. Evaporative heat loss, calculated from weight loss and the difference in weights of metabolic gases exchanged was 1.39 ± 1.05 kcal/kg/h as reported [5]. Insensible water loss concomitantly ranged from 1.14 to 7.33 ml/kg/h, the largest losses tending to occur in the smallest and least mature infants. Con ductive heat loss was calculated at 0.19 ± 0.03 kcal/kg/h. Mean metabolic rate of heat production was 1.90 ± 0.46 kcal/kg/h and heat storage was negligible. From these partitional data, net radiant heat transfer and mean radiant temperature for the infants’ complex environment was calculated (table 2): mean Qnet radiant was -2.63 ± -1.52 kcal/kg/h, indicating that for every infant net heat transfer was toward the infant, and Tmcan radiam was 45.3 ± 4.3 °C, a relatively warm environmental hemisphere. Table 2 also displays thermopile measure ments of radiant heat delivered to infants directly by the warmer: mean Q radiant warmer
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Baumgart
15
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
Table 1. Partitional calorimetry Qconvcction
A 1 Vcvapo ration
Qcondu:tion
keal/kg/h
keal/kg/h
Q metabolic* keal/kg/h
O sto red 1
keal/kg/h 1
3.90
4.25
0.26
2.56
0.23
2
2.14
1.51
0.20
2.13
0.01
3
2.41
0.85
0.21
2.09
-0.05
4
2.31
0.66
0.19
2.40
0.05
Patient No.
keal/kg/h
5
1.78
1.60
0.17
1.48
0.11
6
3.82
0.97
0.20
1.18
-0.03
7
3.73
0.86
0.19
2.22
-0.06
8
2.89
0.82
0.19
1.81
-0.02
9
3.73
1.07
0.17
1.43
0.04
10
2.51
1.30
0.17
1.65
-0.13
Mean
2.92
1.39
0.19
1.90
0.02
SD
0.80
1.05
0.03
0.46
0.10
1 Previously reported data [5].
Patient No.
O nct radiant
keal/kg/h
Tm can radiant
°c
Q radiant warmer
keal/kg/h
Tradiant warmer
Q radiant loss
°c
keal/kg/h
-2.92
45.1
-3.16
42.1
-3.66
48.1
1.93
40.6
-2 .7 6
45.0
1.43
1
-6.08
53.4
2
-1.73
3
-1.33
4
-0.81
39.3
-2.30
44.3
1.49
5
-2.18
44.8
-2.66
46.6
0.48
6
-3.78
49.6
-1.96
43.5
-1.82
7
-2.5 0
45.6
-0.82
39.6
-1.68
8
-2.07
44.2
-2.19
44.6
0.12
9
-3.58
49.0
-1.78
42.6
-1.80
10
-2.20
44.3
-3.87
50.2
1.67
Mean
-2.63
45.3
-2.49
45.0
-0.13
SD
-1.52
4.3
-0.90
2.9
1.82
Q radiant loss " Q nel radiant ~ Q radiant warmer*
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Table 2. Calculation of mean environmental radiant temperatures from partitional calorimetry vs. direct measurement
16
was -2.49 ± -0.90 kcal/kg/h. The radiant temperature of the warmer experienced by infants was 45.0 ± 2.9 °C. Therefore, mean Qradiam loss calculated from the partition is -0.13 ± 1.82 (range-3.16 to 1.93) kcal/kg/ h.
Discussion These data demonstrate that low birth weight infants nursed on open beds under radiant warmers experience an environment which differs from the convection warmed incubator in two important ways. First, air temperature is much cooler (29.1 °C). Sec ond, cooler air temperature results in a re quirement for considerably higher mean ra diant temperature (45.3 °C) to maintain both a normal core temperature, and skin temperature within the predicted thermal neutral range. The radiant warmers in the present study provided a similar radiant temperature to meet this demand (45.0 °C). Moreover, in vitro studies indicate that the mean radiant temperature of the environ ment is not homogenous in all directions from the infant's position on the bed. A warmer temperature is experienced by the surface of the infant exposed perpendicu larly to the warming device, while the in fant’s sides exposed to the nursery walls may perceive the cooler nursery wall tempera ture. It is interesting to note that the bed's side walls and perhaps mattress surface con stitute secondary irradiators presenting ra diant surfaces somewhat warmer than the nursery walls. Also, radiant heat loss to the sides of individual infants ranged from posi tive to negative values, indicating a complex summation of net heat losses and gains from bed platform, nursery walls, and the radiant heater.
Bell and Rios [ 12] have recently described the partition of heat losses for single- and doublewalled incubators. These authors described relatively temperate environments wherein air and wall temperatures were within a few degrees of the infant's skin and core temper atures. Using their model to evaluate convec tive heat loss, we have described a different environment under the radiant warmer. Heat losses via convection and evaporation are of a magnitude sufficient to require a mean ra diant temperature much higher than the com parable mean wall temperature in the incuba tor. This higher mean radiant temperature is generated by the servocontrolled radiant warmer assigned to maintain the anterior ab dominal wall temperature at 36.5 °C. This surface of the infants in the present study was directly facing the heater and probably con stituted the warmest side of these infants. Even though the exact radiant and convective geometry of an infant’s body is complex, the model we cited is relatively conservative in estimating the infant's exposure to convec tive heat loss which may have resulted in an underestimation of Tmcan radiant in the present calculation. Bell and Rios [12] and LeBlanc [11] have previously described similar results regard ing radiant heat transfer in the relatively more temperate incubator environment. An important part of radiant heat transfer de scribed in LeBlanc’s study was the relatively large contribution of the mattress surface immediately adjacent to the baby’s sides to radiant heat loss. As in the present study, emissivity of the surrounding surfaces was assumed to be near unity. The results of our study suggest that the sides of an infant nursed under a radiant warmer which face room walls are exposed to a radiant temper ature which is less than for infants in incuba
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Baumgarl
tors, and to the side walls of the warmer bed and mattress which are probably secondarily heated by the warmer. LeBlanc [16] specu lates that such environmental heterogeneity under the radiant warmer may result in a slightly increased metabolic rate compared to the standard incubator. Information regarding radiant tempera tures may be useful, then, in designing ra diantly warmed environments wherein the overhead heater not only warms the infant’s exposed surface area, but also a series of stra tegically placed secondary irradiating sur faces. This tactic would significantly reduce radiant heat loss from the infant's sides and moderate the environmental temperature re quired to maintain thermal equilibrium. Several types of clear plastic heat shields have been advocated in this capacity for cov ering infants under radiant warmers. These shields have ranged from saran blankets to rigid plastic hoods as much as 3 mm in thick ness. Such devices may also serve as second arily heated walls since they are partially opaque to the infrared heat from the warmer [17], Certain types of shields may function to reduce the radiant temperature gradient be tween the infant’s skin and the surrounding walls while still permitting the radiant heat from the warmer through from above. One shield tested, (a 3-mm-thick plastic hood) absorbed almost all of the infrared energy from the servocontrolled warmer and re sulted in an increased power demand from the warmer to maintain infant skin tempera ture [ 17], The reasons for heterogeneity observed in the calculation of radiant heat loss as the dif ference between net radiant balance (parti tion), and radiant gain from the warmer (thermopile), may be inherent in measure ment errors encountered with partitional
17
calorimetry (Appendix). However, it is impor tant to note that in the present study when the gradient between ambient air and skin tem peratures is high (i.c. high Q COnvcction)> radiant heat loss is low or even negative (indicating heat gain). Either high convective losses may predispose to higher servocontrolled heater outputs, or higher heater outputs correspond to warmer secondary irradiators in the envi ronment. This latter hypothesis would require experimental trial to justify the use of second ary or passive radiant heat sources in an infant’s environment. In conclusion, mean environmental ra diant temperature is greater under radiant warmers than inside convectively heated in cubators. This mean radiant temperature is several degrees warmer than the infant’s skin suggesting that despite probable radiant losses from the infant’s sides to the nursery walls, net radiant heat transfer is dominated by the radiant heating element overhead and is toward the infant. The primary source of this radiant heat is from the radiant warming element, however, the mattress and sidewalls of the bed platform may be secondary irradiators receiving heat from the warmer and reradiating it to the infant. The rela tively higher mean radiant temperature the infant experiences on the radiant warmer bed also suggests that heat losses via evapo ration, convection and perhaps radiation to the sides are increased over the incubator environment. Heated walls, mattresses or other shielding interventions must be care fully evaluated to determine their effect on the complete thermal environment before being routinely applied to infant care. The goal of these devices should be to reduce radiant, evaporative and convective losses from infants who must be nursed in an open and accessible fashion.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
18
Baumgatt
Scale of equilibration
Appendix Metabolic Rate (& m etabolic) Metabolic rate (kcal/kg/h) was calculated from oxygen consumption (V0,. ml/kg/min) and carbon dioxide production (Vc0,, ml/kg/min) according to the relationship described by Lusk [18]. An open cir cuit technique was used to monitor the rates of gas exchange continuously [19]. The details of our com puterized metabolic monitoring system have been re ported previously [3, 5]. The relative accuracy of this apparatus has been determined in vitro to be ± 3.0% for V0:, VCo;. and respiratory quotient by previously established techniques [20]. Evaporative Heat Loss ((¿evaporation) Infant weight loss (IL. g/kg/h) was monitored con tinuously by a Potter baby scale (absolute accuracy ± 0.5 g) and a Hewlett-Packard strip-chart recorder during a 90-min sampling period. Insensible weight loss thus determined was corrected for the weight of respiratory gas exchange to determine insensible wa ter loss (IWL, ml/kg/h): IWL = IL -(0.1179 VC0: -0.0857 V0>). This result was multiplied by 0.58 kcal/ ml of water evaporated at 36.5 °C to determine the rate of evaporative heat loss in kcal/kg/h. Five sources of potential error have been de scribed previously during validation of these proce
Fig. A2. Thermal equilibration of infant scale.
dures [5, 7], Briefly summarized, these sources are: (1) scale hysteresis; (2) load drag from wires and tub ing on the scale platform: (3) condensation of respira tory gas humidification; (4) fluid intake and urine and stool weights excreted, and (5) thermal expan sion of the scale under the warmer. An example of scale calibration is shown in the figure A l. Numeral I indicates removal of a 5-gram weight from one exper imental apparatus with an infant and tubing, etc. in place. The polygraph record subtends 25 mm within 30 s and represents less than 0.5 mm distortion with a light tap from below the platform (Numeral 2). Hys teresis and load drag combined produce less than 0.1 g error in absolute weight determination. Conden sation was not detected in our respiratory support apparatus both in in vitro and in vivo experiments [21]. Respiratory tubing is passively warmed under the radiant heater and unless super-saturated gas mix tures are used, the 'rain-out’ phenomenon with infant water aspiration is not routinely observed. Fluid in take and urine and stool excretion are measured dur ing our studies, and the polygraph rezeroed to ac count for weight adjustment. Thermal expansion of the scale apparatus under the radiant heater was taken into account by providing adequate time for the scale to reach thermal equilibrium before commenc ing data acquisition. The time constant for scale equilibration was estimated empirically as 57 min.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
Fig. A l. Scale calibration.
Radiant Heat Loss vs. Gain in Neonates under Radiant Warmers
First hour
Second hour
0.51 2.63 3.52 3.50 2.45 1.58 2.15 1.07
0.51 2.17 2.54 3.94 2.45 1.58 2.12 1.42
2.18 0.38
2.09 0.35
(fig. A2), therefore a minimum of 3 h was alotted before an experiment (less than 0.1 g error in weight incurred). Replicability of infant weight change evaluations has been previously tested in eight premature subjects [7]. Table AI demonstrates weight loss during 2 con secutive hours in this study. Disregarding possible changes in infant physiology and/or environment, mean weight change was the same for both hours (dif ference less than 0.1 g absolute, p = 0.87). Standard error was less than 0.4 g/kg/h). These data suggest that cumulative error in measuring insensible water loss using these techniques is less than the ± 0 .5 g re ported for the Potter scale. Error in calculating evapo rative heat loss from these data is probably less than 0.3 keal/kg/h. Heal Stored (Quoted) Q siorcd (keal/kg/h) should be nearly zero if the infant is at thermal equilibrium and neither gains nor loses heat. Q slored may be calculated over the 90-min study for each infant: Q stored
= 0.84 (0.6 ATrectai + 0.4 AT mean skin)*
where 0.84 kcal/kg/°C is the specific heat constant for human tissue and AT (°C ) is the difference in temperature between the end and the beginning of the study. Rectal temperatures were measured with a Yel low Springs rectal thermistor inserted 5 cm past the anus.
1 Baumgart S. Engle WD. Fox WW. et al: Effect of heat shielding on convection and evaporation, and radiant heat transfer in the premature infant. J Pediatr 1981:99:948-956. 2 Robinson RO. Jones R: Advantages and disad vantages of overhead radiant heaters. Proc R Soc Med 1977:70:209. 3 Malin SW. Baumgart S: Optimal thermal manage ment for low birth weight infants nursed under high-powered radiant warmers. Pediatrics 1987; 79:47. 4 Hey EN. Katz G: The optimum thermal environ ment for naked babies. Arch Dis Child 1970:45: 328. 5 Baumgart S: Partitioning of heat losses and gains in premature newborn infants under radiant warmers. Pediatrics 1985:75:89. 6 Wheldon AC, Rutter N: The heat balance of small babies nursed in incubators and under radiant warmers. Early Human Dev 1982:6:131. 7 Baumgart S, Engle WD. Fox WW. et al: Radiant warmer power and body size as determinants of insensible water loss in the critically ill neonate. Pediatr Res 1981:15:1495. 8 Baumgart S, Engle WD. Langman CB, et al: Mon itoring radiant power in the critically ill newborn under a radiant warmer. Crit Care Med 1980;8: 721. 9 Haycock GB. Schwartz GJ. Wisotsky DH: Geo metric method for measuring body surface areas: a height-weight formula validated in infants, chil dren and adults. J Pediatr 1978:93:62. 10 LeBlanc MH: Personal communication. 1988. 11 LeBlanc MH: Oxygen consumption in premature infants in an incubator of proven clinical efficacy. Biol Neonate 1983,44:76. 12 Bell EF, Rios GR: A double-walled incubator alters the partition of body heat loss of premature infants. Pediatr Res 1983:17:135. 13 Silverman WA, Sinclair JC, Agate FJ: The oxygen cost of minor changes in heat balance of small newborn infants. Acta Paediatr Scand 1966:55: 294. 14 Hey EN, O’Connell B: Oxygen consumption and heal balance in the colnursed baby. Arch Dis Child 1970:45:335. 15 Swyer PR: Heat loss after birth; in Sinclair JC (ed): Temperature Regulation and Energy Metab-
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
References
Table AI. Weight loss, g/kg/h
Mean SEM
19
16
17
18
19
20
Baumgart
olisrn in the Newborn, New York, Gruñe & Strat ton. 1978, p 92. LeBlanc MH: Relative efficacy of an incubator and an open warmer in producing thermoneutral ity for the small premature infant. Pediatrics 1982;69:439. Baumgart S, Fox WW, Polin RA: Physiologic im plications of two different heat shields for infants under radiant warmers. J Pediatr 1982:100:787. Lusk G: The Elements of the Science of Nutrition, ed 4.. Philadelphia, WB Saunders, 1928. pp 6468 . Lister G. Hoffman JIE. Rudolph AM: Oxygen uptake in infants and children: a simple method for measurement. Pediatrics 1974;53:656. Marks K.H. Coen P, Kerrigan JR. et al: The accu racy and precision of an open-circuit system to measure oxygen consumption and carbon dioxide production in neonates. Pediatr Res 1987;21:58.
21 Sosulski R. Polin RA, Baumgart S: Respiratory water loss and heat balance in intubated infants receiving humidified air. J Pediatr 1983:103: 307.
Stephen Baumgart, MD Associate Professor of Pediatrics Department of Pediatrics University of Pennsylvania School of Medicine Division of Neonatology The Children's Hospital of Philadelphia Philadelphia. PA 19104 (USA)
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/12/2018 5:45:16 AM
20
