How to Cool a Burn: A Heat Transfer Point of ViewAnne Baldwin,* Jie Xu, PhD,† Daniel Attinger, PhD‡
The objective of this work is to develop and validate a numerical model that can conduct atransient analysis of heat transfer and the corresponding damage in skin burns. Once thismodel is developed, an examination of the effect of cooling on reducing damage from skinburns is carried out. A finite element numerical model is used to simulate the conduction ofheat and the transient progress of irreversible injury in the skin. The damage function ofHenriques and Moritz is used to model the damage that occurs in the skin during the burnand cooling periods. Numerical results are presented that describe the heat transfer during askin burn. Comparison is made between different burns: a high-temperature, short-duration burn (99°C for 1 second) and a medium-temperature, long-duration burn (80°Cfor 15 seconds). Cooling parameters such as the nature of the cooling fluid, the duration ofthe cooling period, the temperature of the coolant fluid, and the delay between the termina-tion of the burn and the initiation of the cooling therapy are examined. The authors findthat the most influential way to significantly reduce the damage from a burn is to immedi-ately cool the burn. In addition, it was found that cooling a burn for a prolonged period oftime or with very cold water cannot be justified from purely a heat transfer point of view. (J BurnCare Res 2012;33:176–187)
Physical and Psychological Impact of BurnsSkin plays various important roles in the human body,eg, providing sensory feeling, protecting the bodyfrom harmful external conditions, and providingthermoregulation. Burns, therefore, are one of themost painful injuries humans can sustain and survive.Burns are not only potentially lethal but also poselong-lasting physical and mental damage to burn vic-tims.1 The greatest causes of nonfatal burns are scaldor thermal burns, with a 68% incidence rate for malesand 76.9% incidence rate for females.2 In this study,we will focus on thermal burns. Thermal burns areprogressive, potentially resulting in edema formation,liberation of vasoselective substances, and inducing aprofound impact on all organ systems. A first-degreeburn is characterized by damage to the epidermal
layer of the skin, in which the skin around the burnbecomes red and tender without any blister formation.A second-degree burn can be a superficialpartial-thickness burn or a deep partial-thickness burn.Superficial partial-thickness burns involve injury to theepidermis and papillary layer, with good remaining per-fusion in the dermis. Deep partial-thickness burns ex-tend to the reticular layer of the dermis, causing damageto the hair follicles and sweat and sebaceous glands.Third-degree burns cause damage to all epidermal anddermal structures, causing the skin to be pale, leathery,and charred.3 Thermal injury can result in changes tocell membrane functions, tissue acid-base balance, he-modynamics, and hematologic derangement.4 The painassociated with burns often worsens with time, as a re-sult of depression, nerve regeneration, skin contraction,or tissue damage from the burn and subsequent proce-dures.5 Burns are serious injuries that have long-lasting,sometimes permanent, effects. A thorough understand-ing of the physics of burns will help their treatment andprevention.
Emergency Treatment for Burns and Needfor Further ClarificationThe emergency treatment for burns varies widely, de-pending on the protocol used by the personnel at-tending the injury and on the initial conditions of the
From the *Department of Chemistry and Chemical Biology,Harvard University, Cambridge, Massachusetts; †Departmentof Mechanical Engineering, Washington State University,Vancouver; and ‡Department of Mechanical Engineering, IowaState University, Ames, Iowa.
Address correspondence to Daniel Attinger, PhD, AssociateProfessor, Mechanical Engineering, Iowa State University, 2025Black Engineering, Ames, IA 50011-2161.
burn. Some experts recommend pouring water di-rectly on the burned area, as early cooling can reducethe depth and degree of burn injury.1 However, 58%of the ambulances surveyed in a UK study stated thatthey had no emergency treatment policy for burnpatients.6 Also, cooling a burn too greatly could in-crease risk of hypothermia.2 A clear understanding ofthe physics of heat burns and their cooling is neededto clarify the advantages and disadvantages of eachform of treatment. Understanding cooling from botha heat transfer and clinical point of view is importantto tackle the burn treatment issue. Examining the roleof heat transfer in the damage caused during skinburns and their cooling informs the clinical perspec-tive of the issue and is also relevant to novel medicaltechnologies such as laser resurfacing and lasersurgery.7
Importance of Cooling the BurnThis study focuses on one of the most practical (andwidely available) responses to burn injuries: coolingthe burn area. Studies have shown the clear analgesicbenefits of cooling the burn.2,5,8 There is significantevidence that cool water can limit the devastationfrom a burn. The zone of ischemia surrounding theburn has the potential to progress to full tissue necro-sis in the days after the burn unless the ischemia isreversed.3 A study by Davies,8 which reviews multiplemethods of cooling, concludes that prompt coolingof small burned areas reduces the severity of injuryand reveals that cooling controls edema formation,thus limiting cell death. In addition to preventingnonthermal necrosis, cooling by water has also beensaid to decrease pain, lower the infection rate, prom-ise a shorter recovery period, and lessen the need forgrafting.8 Davies8 concluded in his examination ofthe prompt cooling of burns that rapid reduction ofskin temperature by cooling reduced skin loss andenhanced healing. As evidence points to the impor-tance of cooling a burn wound by water, a morequantitative analysis of different cooling techniques isnecessary for the development of emergency responseand rehabilitative measures.
Previous StudiesStudies Done on Simulating Burn Damage.
Many studies simulating the burning of the skin havebeen conducted to determine the extent of damagefrom a burn. Henriques and Moritz9 created thebuilding blocks for investigation of cell damage fromburns. He solved the Fourier heat conduction equa-tion and was thus able to model the transfer of heatthrough skin. He created a general theory of heat flowthat gave the time-temperature relationships within
the skin during exposure to heat. Henriques andMoritz went further to calculate the extension of tis-sue damage by using an Arrhenius first-order rateequation, which is used in this study.10 Ng andChua11 were able to create a mesh-independent pre-diction of a skin burn injury using a finite elementnumerical model. Using both the bioheat transferequation and the Henriques damage function, theydetermined that the greatest damage occurred veryquickly and mostly in the epidermal layer. They de-termined that the burn process keeps expanding inthe deeper layers of the skin for about 1 second afterthe removal of the heat source, in the event of a 90°C,15-second thermal insult. Ng and Chua11 used a one-dimensional model to investigate various parametersthat influence the damage caused by a burn. Theydiscovered that the initial tissue temperature was themost influential factor and that thermal conductivityranked second in importance. They also found thatblood perfusion had minimal effect.12 Their resultsconcur with those of Jiang et al,13 who concludedthat the initial temperature distribution and theblood perfusion had little effect on the transient tem-perature distribution and the resulting thermal dam-age in the skin.
Studies Conducted on Measuring DamageDuring Burning and Cooling Process. Henriquesand Moritz10 determined experimentally that tissuenecrosis continued 2 to 3 minutes after the burnwhen the skin was subject to free air convection. Thisfinding might signify that cooling the burn wouldreduce the tissue temperature below the damagethreshold, lessening tissue necrosis. However, evi-dence that water cooling reduces thermal damage intissues is scant. In fact, Diller and Hayes tested differ-ent cooling durations and water temperatures for theability to limit thermal damage.14 Using the Hen-riques damage function, they discovered that immer-sion in cool water has no effect on limiting thermaldamage of tissues.14 The greatest temperature gradi-ent caused by the cold water occurred in the epider-mal and dermal layers, where the thermal damage wasalready the largest, so that cooling would have littleeffect. In the subcutaneous layer, immersing the burnin cold water did little to change the tissue tempera-ture because of the low thermal conductivity of theskin.14 A recent study conducted by Sompel et al alsoused the Arrhenius damage model to compute thepostburn accrued damage after cooling therapy. Theyalso concluded that the therapeutic benefits of cool-ing cannot be explained by thermal analysis.15 Todate, no study has examined how long or at whattemperature a burn should be cooled. Also, no dataexist to assess the importance of cooling a burn im-
Journal of Burn Care & ResearchVolume 33, Number 2 Baldwin, Xu, and Attinger 177
mediately. These are the questions we propose toconsider in this study from a heat transfer point ofview.
METHODS
Heat Transfer EquationThe heat transfer in this study is modeled with thebioheat transfer equation used by Ng and Chua11:
pCp
�T�t � k��2T
�r2 �1r
�T�r �
�2T�z2� � �b�bcb(Ta � T)
� qm (1)
Their equation represents the conduction of heat inbodily tissues. Their equation is derived from thePennes bioheat transfer equation, which is based onexperimental evidence on human tissue. The equa-tion is expressed in two-dimensional, axisymmetriccoordinates. The left-hand side term refers to the rateof heat storage in the tissues, where � is the volumicmass (kg/m3), C is the heat capacity (J/(kg � K)), andT is the temperature. The first right-hand side termrepresents the thermal diffusion of heat through thetissue, where k is the thermal conductivity (W/mK).The second right-hand side term describes theheat transport due to microcirculatory blood perfusion.This term has been shown to be negligible duringthermal burns by Ng and Chua11 and will be ne-glected in our numerical implementation. The lastterm, qm, represents the internal heat generated dueto metabolism in the skin. Because the contributionof this term is negligible in comparison to the largeheat flux from the burn, it will not be accounted for inthis model, as done in the study conducted by Ng andChua,11 Diller et al,13 and Jiang et al.14 In addition,we determined that our simulation results were inde-pendent of the value of the metabolism in the skin(see Mesh, Domain, Metabolism, and Time Indepen-dence section).
The values of the symbols used in this study arespecified in Table 1.
Geometric and Computational DomainThe computational domain created in this study isshown in Figure 1. It is representative of the typicalthickness and arrangement of skin layers. Each layerhas its own density, heat capacity, and thermal con-ductivity values, as defined in Table 1.
Initial and Boundary ConditionsThe initial temperature in all the skin layers (sub-domains) is defined by the initial temperature profile
of the skin determined by Ng and Chua.11 This dis-tribution accounts for the fact that the outer skinsubject to free convection is cooler than the inner skintissue:
T � 76568z3 � 5503.5z2 � 200z
� 305.93 (2)
where the temperature (T, in degree Kelvin) is plot-ted as a function of the depth in the skin (z). The rightand bottom boundaries are thermally insulated, asshown in Figure 1, and axial symmetry is applied tothe left boundary. The typical extension of the do-main has a total radius of 0.012 m and a depth of0.005 m, where the heating pad has a radius of 0.005m. The epidermal surface surrounding the heatingpad where the heat burn is applied is subject to freeconvection with ambient air at 27°C during and afterthe burn. A free convection coefficient h � 7 W/m2Kis used to account for the free convection of airaround the skin. Different boundary conditions forthe burnt epidermal surface are created for each post-burn treatment technique. When running water isapplied in the simulation, it is applied across the entirecalculation domain, as one would flush the entire skinwith water. For instance, for cases involving coolingwith running water, a convection coefficient h �1000 W/m2K is used on the top surface of the epi-dermis. For cases involving cooling with a water bath,a free convection coefficient h � 500 W/m2K is used.Perfect thermal contact between the heat source andthe skin is assumed.
Thermal Damage and Cooling EfficacyTo compare different cooling therapies, the followingapproach is taken. The Henriques and Moritz equa-tion is used to compute the thermal damage function� from the temperature history of the burn:
��r,z,����0
�
Pexp� ��ERT �r,z,���d�
� �f �r,z,��d� (3)
where � denotes the severity of tissue injury, P rep-resents the frequency factor in the damage integral(1/s), �� represents the activation energy (J/kmol),R represents the gas constant (J/kmol � K), T (K)represents temperature, z is depth, and � is time, withvalues in Table 1.10 The standard classification ofburn injury depends nonlinearly on the value of �.10
Journal of Burn Care & Research178 Baldwin, Xu, and Attinger March/April 2012
Hence, a first-degree burn corresponds to � � 0.53,a second-degree burn corresponds to � � 1.0, and athird-degree burn corresponds to � � 104.10,11 Notethat Henriques and Moritz examined the tempera-ture history only at specific locations such as the sur-face or epidermal-dermal layer. Their method, there-fore, could only determine the degree and not thespatial extension of a burn. Numerical simulations,however, provide the temperature at any point ofspace in time. It makes sense therefore to use transientnumerical results to estimate the spatial extent of thedamage by calculating locally the value of a damagefunction similar to the one developed by Henriques
and Moritz.10 This approach to estimate the spatialextent of the burn damage has been used by Ng andChua11 and others,14,15 and this is the procedure wefollow in this article. For the typical case where theskin is burnt by contact with an external heat source,1) any injury deeper than the � � 0.53 location isreversible; 2) any injury more superficial than the � �1.0 location is irreversible, and 3) that any injurybetween the � � 0.53 and 1.0 locations is indeter-minate. Because of this uncertainty, there is a contro-versy over the exact value of � that should be consid-ered as the dividing line between reversible andirreversible injury zones. Indeed, the dividing line
Table 1. Nomenclature
Symbol Property Meaning Value Range Reference
k (W/mK) Thermal conductivity of skin 9Epidermis 0.21Dermis 0.37Subcutaneous 0.16
h (W/m2K) Convective heat transfer between 9Air skin and air 7.0Water skin immersed in water 500Running water skin flushed with running water 1000
�b (kg/m3) Density of skin 11Epidermis 1200Dermis 1200Subcutaneous 1000
Cp (J/kg � s) Specific heat of skin 11Epidermis 3590Dermis 3300Subcutaneous 2500
rs (m) Radius of heating disk 0.005 9rmax (m) Radius of skin measured 0.007 9ts (sec) Duration of heat exposure 15 9Ti Initial tissue temperature 10
EpidermisDermisSubcutaneous
Tamb (°C) Temperature of ambient air 27°C 9Theating pad (°C) Temperature of heating disk 9
Case 1 90°CCase 2 80°C
P (1/sec) Frequency factor in damage integral 3.1 � 1098 9R (J/(kmol � K)) Ideal gas constant 8.314 9�E (J/kmol) Activation energy 6.27 � 108 9�z (m) Depth of skin tissue 9
Epidermis 7.5 � 10�5
Dermis 1.5 � 10�3
Subcutaneous 3.425 � 10�3
�x (m) Element size (m) 0.05 � 10�3
�t (sec) Time step (sec) 0.01
Journal of Burn Care & ResearchVolume 33, Number 2 Baldwin, Xu, and Attinger 179
could be anywhere between 0.53 and 1. However, itdoes not change the validity of our results because forall cases considered in this article, our simulationsshow that the � � 1.0 and � � 0.53 locations arealmost contiguous. In this study, the results are basedon � � 0.53 as the separation between a reversibleand an irreversible damage.
Therefore, at a given time, the irreversible damagevolume is given by:
� � �� � �0�dV (4)
where �0 is a threshold value equal to 0.53. To eval-uate how efficient a cooling therapy is, we comparethe final irreversible damage volume cooled by ambi-ent air with the final irreversible damage volumewhen more efficient methods are used. The relativereduction of burn volume is expressed by the coolingefficacy :
�%� �amb � cool
amb 100 (5)
where represents the efficacy of the cooling therapy,amb represents the final volume affected by irre-versible damage after cooling a burn by free con-
vection with ambient temperature air, and coolrepresents the final volume affected by irreversibledamage after cooling a burn by the investigatedcooling therapy. The fraction is multiplied by 100to give a percentage of the damage caused by theburn that is reduced by the investigated coolingtherapy. Both cooling therapies (free convectionwith ambient air and investigated) are applied forthe same amount of time.
Description of the Burn and InvestigatedCooling MethodsIn this article, two main factors, the duration of theheat exposure and the temperature of the burn, areused to characterize the burn.11 The burn is causedthrough direct contact with a 5-mm-radius heatingpad, maintained at constant temperature, for a spe-cific time. In the cooling process, the duration ofthe exposure to a coolant (seconds), the tempera-ture of the cooling agent (K), and the delay be-tween the termination of the burn and the initia-tion of cooling therapy (seconds) are definingfactors. Table 2 describes all the cases consideredin this study, with respect to heating and coolingconditions.
Figure 1. A model of the mesh-independent solution grid approximating an axisymmetrical half cross-section of the model aswell as the boundary conditions used during the burn simulation.10
Table 2. Cases examined in simulations
CaseBoundary and Transient Conditions
During HeatingBoundary and Transient Conditions During Cooling
(Entire Epidermis Surface)
1 15-sec, 90°C burn by 0.005-m heating disk 20 sec natural convection with 27°C air2 5-sec, 80°C burn by 0.005-m heating disk 2A:10 sec natural convection with 27°C air
2B:10 sec flushed by 10°C water2C:10 sec cooling by immersing in 10°C water
3 1-sec, 99°C burn by 0.005-m heating disk 2A:10 sec natural convection with 27°C air2B:10 sec flushed by 10°C water
4 15-sec, 80°C burn by 0.005-m heating disk N/A
Journal of Burn Care & Research180 Baldwin, Xu, and Attinger March/April 2012
NUMERICAL IMPLEMENTATION
Modeling, Time, and Space DiscretizationIn this work, we solve the partial differential Eq. (1)with the commercial finite element packageCOMSOL Multiphysics. The finite element methodhandles complex geometries and boundary condi-tions that usually make an analytical solution difficultor impossible to establish. A backward Euler timediscretization is used, and the Courant-Friedrich-Levy criterion is used to determine the maximumallowable time step.16 The Henriques and Moritzequation [Eq. (3)] is numerically integrated in timeusing a nonlinear solver and quadratic Lagrangescheme.10 In the model used in this study, a meshwith 1442 elements is selected to perform the simula-tions (Figure 2). The independence of our simulationmethod to the spatial and temporal discretization is re-ported in the Results section. Nondimensional temper-atures were used to solve the heat equation, so that onesimulation could express a range of ambient and heat-ing pad temperatures. The initial temperatures werescaled between 0 and 1, with 0 being ambient tem-perature, 1 being the temperature of the heating disk,and body temperature ranging between 0 and 1. Thenondimensional temperature � is expressed by thefollowing equation:
��r,z� �T�r,z� � Tamb
Theatingpad � Tamb(6)
where �(r,z) is equal to the nondimensional temper-ature, T(r,z)is equal to the temperature of the skin indegrees Celsius, Tamb is equal to the ambient temper-ature, and Theating pad is equal to the temperature ofthe heating pad used to burn the skin.
Limitations of the Numerical SimulationThe model developed to simulate the heat transferand damage that occurs in the skin during a burn islimited in its ability to model a realistic burn. Thereare some biological changes in the burnt tissues—such as edema formation—that the model in thisstudy does not account for. Also, the model does notaccount for the boiling of body fluids and can there-fore only simulate burns at temperatures up to100°C.
RESULTS
Mesh, Domain, Metabolism, andTime IndependenceResults with three meshes, refined (1590 elements),regular (1480 elements), and coarse (1083 elements)mesh, are shown in Figure 2, for a 90°C, 15-secondburn. The temporal evolution of the temperature atone location, the dermis-subcutaneous interface (r �0.0025 m, z � 0.001575 m), was examined, with z �0 at the surface of the skin (Figure 1). The time step,0.01 seconds, was kept constant for all three experi-ments, as was the size of the computational domain.
Figure 2. Comparison of the temperature change at the dermis-subcutaneous interface (r � 0.0025, z � 0.001575) during a90°C, 15-second burn using normal mesh (1480 elements), fine mesh (1590 elements), extra coarse mesh (1083 elements),as well as an enlarged domain (subcutaneous region twice as large), and a smaller time step (0.001 seconds).
Journal of Burn Care & ResearchVolume 33, Number 2 Baldwin, Xu, and Attinger 181
The coarse mesh varied slightly from the normal andfine mesh curves and so was deemed unsuitable forthis study. There was negligible temperature differ-ence between the fine mesh and normal mesh, and soa normal mesh was used for the simulation analyzedin this study. Figure 2 confirms that the mesh used inthe simulation creates a mesh-independent solution.
It is also important to verify that the results are notinfluenced by the size of the computational domain.To do so, the temperature at a given point (r �0.0025 m, z � 0.001575 m) was simulated for dif-ferent substrate sizes, for a 90°C, 15-second burn.Figure 2 shows that similar results were obtained forthe domain shown in Figure 1 (1480 elements) and alarge domain with the subcutaneous layer twice aslarge and twice as deep (with the same mesh density).Therefore, in all simulations used in this study, thecomputational domain has the size given in Figure 1.
In addition, it is necessary to determine that theresults are not influenced by the value chosen for themetabolic heat generation (qm). The simulation ofthe 99°C, 1-second burn was conducted for two val-ues qm of 0 and 368.1 W/m3, the value used in thestudy by Ng and Chua.11 Differences between bothtime-temperature curves were 1%, because of thelarger values of the heat flux due to the burn. Therefore,the contribution of qm is neglected for the simulationsconducted in this study, in accordance with other workreviewed in Heat Transfer Equation section.
Finally, it is important to verify that the results arenot influenced by the size of the time step, which
governs how the numerical solver advances in time.The temperature is given at the same point as themesh and computational domain tests (r � 0.0025m, z � 0.003425 m), simulated when the tempera-ture solution to a 90°C, 15-second burn is solved attwo distinct time steps, 0.01 and 0.001 seconds. Fig-ure 2 demonstrates that a 0.01-second time step gen-erates a temperature evolution very similar to that ofa 0.001-second time step. Therefore, the time stepused in this study has no effect on the final solution,and so a 0.01-second time step is used.
Validation With Published ResultsAnother way to validate our code is to compare itsresults with published results. In Figure 3, 45°C iso-therms are plotted 1 and 14 seconds after the start ofa 90°C, 15 seconds burn: the isotherms locations arein very good agreement with the results of Diller etal.14 Also, the separation between reversible and irre-versible damage is plotted for the same burn case inFigure 4, with a good agreement with the data fromDiller et al.14 The purpose of Figure 4 is to examinethe validity of the model, which we have confirmedwith contemporary research. This agreement is a signthat our modeling is reliable and suited to the pro-posed study.
Cooling Effect on the Skin DamageThe volume of skin irreversibly damaged after usingvarious cooling methods is demonstrated in Figure 5.We see that the kind of postburn treatment does not
Figure 3. A comparison of 45°C isotherm for a 90°C burn between the current study and the results of Diller and Hayes at twodistinct times, 1 and 15 seconds.
Journal of Burn Care & Research182 Baldwin, Xu, and Attinger March/April 2012
significantly reduce the irreversible damage, and nosensible difference is seen between the air cooling andthe water flushing treatment. This probably occursbecause most of the damage occurs during the burn-ing process. Also, the mathematical structure of Eq.(3) makes it impossible to reverse the damage bycooling: once a given volume of skin has been burnt,
no cooling can reverse the process. This concurs withthe results of Diller et al.14
Progression of the Burn During CoolingA numerical analysis of the volume of irreversibledamage during cooling of the 80°C, 5-second burnwas performed to quantitatively compare the efficacy
Figure 4. A comparison of the location (� � 0.53) separating irreversible (above line) from reversible (below line) damagebetween the current model and the one of Diller and Hayes for a 90°C, 15-second burn.
Figure 5. A numerical comparison of the irreversible damage after 25 seconds of cooling the case 2 burn (80°C, 5 seconds) byfree convection with 27°C air, immediately flushing with 10°C water, and immediately immersing in 10°C water. The location(� � 0.53) is assumed to separate irreversible from reversible damage.
Journal of Burn Care & ResearchVolume 33, Number 2 Baldwin, Xu, and Attinger 183
of different cooling therapies [Eq. (5)]. A graph ofthe irreversible damage vs the cooling time is shownin Figure 5. Let us compare first the progression ofthe damage for the case A where no specific cooling isapplied (just ambient air convection): the irreversibledamage volume increases to its final value withinabout 5 seconds. This explains why cooling by im-mediately flushing the burn with water will lessenthe total damage. For instance, cooling an 80°C,5-second burn for 25 seconds by flushing with10°C water thus reduces the volume irreversiblydamaged by about 9% [Eq. (5)]. Figure 5 also
shows that immersing the wound with water (case2B) reduces the final damage almost as well asflushing the burn with water.
Parameters That Control the Efficacyof the Cooling
Duration of the Cooling Time. A numerical anal-ysis of the effect of the duration of the cooling therapyfor a case 2 (80°C, 5 seconds), case 3 (99°C, 1 sec-ond), and case 4 (80°C, 15 seconds) burn is pre-sented in Figure 6, in terms of cooling efficacy. All thedata show that cooling the burn for at most 20 sec-
Figure 6. A comparison of the effect of immediately cooling 80°C, 15-second burns and 99°C, 1-second burns with 10°Cwater for varying durations of cooling therapy.
Figure 7. A comparison of the effect of immediately cooling 80°C, 15-second burns and 99°C, 1-second burns with varyingtemperatures of water for 10 seconds.
Journal of Burn Care & Research184 Baldwin, Xu, and Attinger March/April 2012
onds is enough to maximize the cooling efficacy. Forinstance, cooling a case 2 burn for 25 seconds yields acooling efficacy of 9.02%, while cooling a burn for 15seconds yields and efficacy of 9.00%, a negligiblechange. These data concur well with that of Diller andHayes.14 The case 2, 3, and 4 burns all have similartrends. Similarly, for the shortest burn (99°C, 1 sec-ond), cooling the skin for more than 1 second hasvirtually zero effect on reducing the irreversible dam-age. As a rule, the duration of cooling matters mostfor the long, low-temperature burn, 80°C, 15 sec-onds (case 4), probably because in this case, the iso-
therms have penetrated deep in the skin. However,the results concur in showing that from a heat transferpoint of view, there is no need to cool a burn for timeslonger than a few seconds.
Temperature of the Coolant. The effect of in-creasing the temperature of the water coolant in case2, 3, and 4 burns is examined in Figure 7. In eachcase, the wound was flushed with water for 10 sec-onds, with the water temperature varied between60°C and 0°C. While increasing the temperature ofthe coolant generally decreases the cooling efficacy,for the shortest and hottest burn examined in this
Figure 8. A comparison of the effect of the delay (seconds) between the termination of the burn and the initiation of the coolingtherapy in 80°C, 15-second burns and 80°C, 5-second burns. The burns are subject to exposure to ambient temperature air fora variable amount of time and are then flushed by 10°C water for 10 seconds.
Figure 9. A comparison of the effect of the delay (seconds) between the termination of the burn and the initiation of the coolingtherapy in a 99°C, 1-second burn. The burned skin is subject to exposure to ambient temperature air for variant amounts of timeand then flushed by 10°C water for 10 seconds.
Journal of Burn Care & ResearchVolume 33, Number 2 Baldwin, Xu, and Attinger 185
study, 99°C, 1 second, increasing the temperature ofthe water to as much as 55°C still provides the samecooling efficacy, which remains constant at 14.3%.Increasing the temperature of the coolant has thegreatest impact on the longest burn examined in thisstudy, the 80°C, 15-second burn: using 1°C wateryields a cooling efficacy of 12.4%, while using 25°Cwater yields a cooling efficacy of 10.6%. For all cases,the difference between flushing burn wounds withcold water and with lukewarm water was minimal.Eventually, using too warm of a coolant has no ben-eficial effect at all; using a coolant of a temperaturemore than 65°C increases damage in all three burncases. As a conclusion, our thermal analysis concludesthat room temperature water works almost as well ascold water to cool a burn, which is counterintuitive.
Delay Between the Termination of the Burnand the Initiation of the Cooling. As shown inFigure 5, a great portion of the damage occurs eitherduring the burn process or within the first few sec-onds after the termination of a burn. As shown inFigures 8 and 9, the delay between the end of theburn and the start of the cooling procedure has asignificant impact on the efficacy of cooling. For in-stance, even delaying to cool the case 2 and 4 burnsfor 10 seconds makes the cooling process almost use-less, as shown by the very low values of cooling effi-cacy reached (down from about 10 to 2%). We cantherefore conclude that the most influential coolingparameter of all the parameters examined in this studyis the immediacy of the initiation of the cooling ther-apy for a burn.
DISCUSSION
We described, set up, and used a numerical modelthat can simulate the heat transfer and related skindamage that occurs during a burn. The thermal phys-ical properties of the skin, the temperature of theheating pad, the duration of heat exposure, and thenature of the cooling process have been investigatednumerically. Very good agreement has been shownwith published results. Our results show that most ofthe damage from a burn occurs during the burn pro-cess and that little damage occurs after the removal ofthe heat source, as shown in Figure 5. Damage froma burn cannot be undone; rather, from a heat transferstandpoint, the cooling therapy only slightly reducesthe final volume that is irreversibly burnt, as demon-strated in Figure 5, and in good agreement with theresults of Diller et al.14 We also observed the effect ofdifferent cooling techniques on the cooling efficacy.As demonstrated in Figure 5, we found that flushinga wound with water is marginally more effective than
immersing the wound in a cold water bath and moreeffective than cooling the burn by natural convectionof air (ie, doing nothing).
Figure 6 demonstrates that there is no need to coola burn for more than 30 seconds, from a heat transferstandpoint. Figure 7 demonstrates that the tempera-ture of the water coolant has a weak influence on thefinal burn volume, so that for all three types of burnsexamined in this study, flushing with lukewarm waterworks almost as well as flushing with cool water. Ourstudy shows that the greatest factor in reducing thefinal volume of skin irreversibly burnt is how imme-diately the burn is immersed or flushed the burn withwater, as demonstrated in Figures 8 and 9. Severalheat transfer results seem to directly contradict thecommon experience where cooling a burn for longtimes with cold running water helps in reducing thepain. This is probably a sign that the damage causedfrom burns is also related to biochemical processes,such as swelling, that occur long after the heat fromthe burn has been dissipated. The effect of cooling onreducing partial-thickness injury around a skin burnshould also be examined. How cooling the burn for along time influences these biochemical processesneeds to be studied with models that better couplethe biology and heat transfer.
ACKNOWLEDGMENTSWe thank Rajneesh Bhardwaj for teaching the use of
COMSOL and MATLAB.
REFERENCES
1. Singer AJ, Brebbia J, Soroff HH. Management of local burnwounds in the ED. Am J Emerg Med 2007;25:666–71.
2. Herndon DN. Total burn care. Edinburgh: SaundersElsevier; 2007.
3. Demling RH. Burns & other thermal injuries (Chapter 14). In:Doherty GM, editor. Current diagnosis & treatment: surgery.13th ed.; available from http://www.accessmedicine.com.ezp-prod1.hul.harvard.edu/content.aspx?aID�5213073. AccessedJuly 8, 2008.
4. Schwartz LR, Balakrishnan C. Thermal burns (Chapter 199).In: Tintinalli JE, Kelen GD, Stapczynski JS, Ma OJ, Cline DM,editors. Tintinalli’s emergency medicine: a comprehensive studyguide. 6th ed.; available from http://www.accessmedicine.com.ezp-prod1.hul.harvard.edu/content.aspx?aID�605000. Ac-cessed July 8, 2008.
5. Engrav LH, Covey MH, Dutcher KD, Heimbach DM,Walkinshaw MD, Marvin JA. Impairment, time out of school,and time off from work after burns. Plast Reconstr Surg 1987;79:927–34.
6. Allison K. The UK pre-hospital management of burnpatients: current practice and the need for a standard ap-proach. Burns 2002;28:135–42.
7. Hokesema H, Van de Sijipe J, Tondu T, et al. Accuracy ofearly burn depth assessment by laser Doppler imaging ondifferent days post burn. Burns 2009;35:36–45.
8. Davies JWL. Prompt cooling of burned areas: a review ofbenefits and the effector mechanisms. Burns 1982;9:1–6.
Journal of Burn Care & Research186 Baldwin, Xu, and Attinger March/April 2012
9. Henriques FC Jr, Moritz AR. Studies of thermal injury II.The relative importance of time and surface temperature inthe causation of cutaneous burns. Am J Pathol 1947;23:695–720.
10. Henriques FC Jr. Studies of thermal injuries V. The predictabilityand the significance of thermally induced rate processes leading toirreversible epidermal injury. Arch Pathol 1947;43:489–502.
11. Ng EYK, Chua LT. Mesh-independent prediction of skinburns injury. J Med Eng Technol 2000;24:255–61.
12. Ng EYK, Chua LT. Prediction of skin burn injury. Part 2:parametric and sensitivity analysis. Proceedings of the Insti-tution of Mechanical Engineers—Part H. J Eng Med 2002;216:171–83.
13. Jiang SC, Ma N, Li HJ, Zhang XX. Effects of thermal prop-erties and geometrical dimensions on skin burn injuries.Burns 2002;28:713–7.
14. Diller KR, Hayes LJ, Baxter CR. A mathematical model forthe thermal efficacy of cooling therapy for burns. J Burn CareRes 1983;4:81–9.
15. Sompel D, Kong TY, Ventikos Y. Modeling of experimentallycreated partial-thickness human skin burns and subsequenttherapeutic cooling: a new measure for cooling effectiveness.J Med Eng Phys 2009;31:624–31.
16. Reddy JN. An introduction to the finite element method.New Delhi: Tata McGraw-Hill Publishing CompanyLimited, 1993.
Journal of Burn Care & ResearchVolume 33, Number 2 Baldwin, Xu, and Attinger 187
