Professional Documents
Culture Documents
-234, 1996
Copyright 0 1995 Elsevim Science Ltd
0@21-929o(!n)ooo39-9 Printed in Great Britain. All rights r%wed
M)Zl-9290/96 $15.00 + 00
Abstract-Despitedecades of animalexperiments, data on blast injury to the lung cover only a limited number of
circumstances and are in a fragmented form. This paper develops a mathematical model of the chest wall dynamics
and the subsequent generation of strong pressure waves within the lung, which havebeen hypothesized as the
mediator of injury. The model has been compared to an extensive database of observed pathologies from animal
tests. The incidence of injury and lethality is found to follow a log-normal correlation with the computed total
energy in these waves and, when the energy is normalized by the lung volume, the lethality correlation applies to all
large animal species. Small animals also correlate with the normalized energy, but at a different value, and it is
speculated that structural differences, other than lung volume, may be involved. This relatively simple model
allows the potential for blast injury to the lung to be determined from measured or computed pressure traces
without additional animal testing. Improved occupational exposure criteria should follow from this methodology.
Keywords: Lung; Blast injury; Injury.
growing military and civilian concernsabout primary lated to the piston velocity time history, has beencon-
blast injury involve so many possiblecombinationsof firmed in tests using a foam lung simulant (Yu et al..
factorsthat the costsof directly evaluatingnewsituations 1990a)and in computational analysesof the lung re-
have becomeprohibitive. Uncertainty about the extra- sponse(Vander Vorst et al., 1989).
polation betweenspeciesfurther cloudsthe applicability Usingthe relationshipconnectingthe chestwall veio-
of animal test results.This paper developsa mathemat- city to the pressurewave generatedin the lung paren-
ical modelthat connectsexternal blast conditionsof the chyma, we can construct a single degree of freedom
chestwall dynamicsand the generationof compression modelfor the pleural surfacedynamicsand derive a for-
waveswithin the lung for a variety of species.A correla- mula for the amount of energydeliveredto the lung. For
tion is madebetweenthe energy density of thosewaves simplicity, we assumethat the thoracic responseto blast
and observedinjury. The resultingformulation can pro- resultsis dominated by the chest wall inertia and the
vide a rational criterion for blast overpressurehazard to external pressureloading, Pioad,so we can ignore the
man. stresses arisingfrom the rib structure. As a corollary, we
assumethat different sectionsof the thorax act indepen-
dently to the external loading. Finally, we ignore the
METHODS effectsof internal reflectionsof the wave on the motion of
the thorax.
Stuhmilleret al. (1991)gave a review of the physicsof
If we apply Newton’slaw to the local thorax surface,
blast and its injury mechanisms. A completeunderstand- imaginingthat the chestwall andlung form a rectangular
ing mustincorporate the complexity of the blast loading
region,seeFig. 1, then weobtain the equationof motion
to the body, the biomechanicalresponseof the body, and
the mechanicalfailure of the tissues.The bioengineering dv
m - = Plead @)--PO 1
approach replacesempirical correlations, that can be dt (
situation and speciesdependent,with modelsthat are
(4)
anatomically and mechanicallycorrect and derive their
parametervalues from direct measurementof material where v is the velocity, x is the displacement,
m is the
properties.An application to the tympanic membraneis mass/chestwall area,and L is the ratio of the volume of
found in Stuhmiller (1989). the lung/chestwall area.The motion of the chestwall can
The motion of the chestwall into the underlying lung be determinedoncethe external blast loading, &,,d(t), is
material resultsin a compressionwave. Clemedsonand given.
Pettersson(1970)were among the first to observelow If the velocity and displacementaresmall,the equation
speedpressurewavesin animalsexposedto blast. The can be linearizedto the form
role of rate-dependent injury mechanismshas been
studied by many researcherssince that time and the
readeris referredto the work of Jonssonet al. (1979)and
other contained in the review by Stuhmiller (1991).Yen Finally, the normalized work, W*, definedastotal work
et al. (1986)quantified the phenomenaby measuringthe done to produce the wave divided by the volume of the
pressurewave through isolated rabbit and goat lung lung and the ambient pressure,can be computed using
and found that the wave speedvariedfrom 31to 64m s-l the velocity found from equation (5):
as the transpulmonary pressure varied from 0 to
20 cm HzO. W* = I= -!m. l; po~,v2&.
Sincethe lung behavesasa compressible material, we POV POL
can use the derivation of Landau and Lifshitz (1959), Injury is causedby a local excessstrain of the tissue,
which relatesthe pressurewave in a compressible gasto whosedetailsare not describedby this model.A correlate
the motion of a piston. The pertinent result is of grosspathology, however,may be related to the aver-
1); >$i 9
age energy dissipatedin the lung tissue, that is, the
P(t)=Po
( l+&- (1) normalizedwork.
80
Fig. 3. Correlation of incidence of injury with normalized work [(A) trace or greater; (A) slight or greater;
(0) moderate or greater; and (0) severe]. The occurrence of injury above a certain level follows
a log-normal distribution. The probability of the occurrence of a particular level of injury has a bell-shaped
distribution that depends on the range of pathologies defined by that level and can be found by taking the
difference between the bounding cumulative curves.
1.0
0.9
0.8 m colEplexwawsludiss(1990-1991)
I I I I/lll/ll I dYIIIllI
0.7
c
* 0.6
2
2
s 0.5
0.4
t
*
0.3
0.2
0.1
0.0
0.001 0.01 0.1 1 10 100
Normallad Work
Fig. 4. Correlation of lung are injured with normalized work. The area is determined by averaging the
injured areas of individual animals. Each data point represents a number of test animals, so the vertical
error bars represent one standard deviation of the areas and the horizontal bars represent one standard
deviation of the calculated work.
viscoelastic forces arising from the thorax itself have been from fluid leakage are correlated with the total energy
ignored.Without suchfeaturesit is not possibleto pre- deposited into the compression wave.
diet the specificlocations of injury, however, the gross Viano and Lau (1988)proposeda viscoelasticmodalof
pathology of hemorrhageand the increaseof lung weight thorax forces, whose parameters were selected to capture
Model of blast overpressure injury to the lung 231
3.5
q FresFiiSWdiis(1991-1991)
z 3.0
x CmplexWave~Sbdii1990&1991
E
s 2.5
-Injurycorreletll(
Mean
)
s
- -- lnjuryConelat~(Mean+3xS.D.)
0”
g 2.0 ----..lnjuryCorrelation(Mean-3xS.D.)
‘s
E
::
5 1.5
P
E
E, 1.0
5
L
3z 0.5
0.0
0.001 0.01
Fig. 5. Correlation of lung weight with normalized work. The increase in lung weight due to fluid leakage
correlates with the normalized work generated by the chest wall motion. The variability of lung weights
seen in control animals, shown as the dotted band, explains the variability seen under injury conditions.
Fig. 6. Normalized work associated with 50% lethality for various species. Vertical bars represent the
uncertainty in the work due to uncertainty in the pressure data. Large and small mammal results group
around values 2.08 and 0.85, respectively.
64 Lvne
pOO
L -Ftiedfnndar 71 10000
----I
-Fh6landw
wave
Wave
.-~~~Triarigle Wave
*---.Ttim@eWava
ODog l fhuse
A &A
A Rat
1000 I
l GuineaPi
0 RabbII
100
1 10
0.1 1 10 loo loo0 loo00 0.1 1 10 loo loo0 loo00
Fig. 7. Blast characteristics producing 50% lethality in the free field. The 50% lethality conditions reported
by Bowen et al. (1968) can be correlated with a single normalized work value for large animals (2.08) and
a single value for small animals (0.85). The differences between large and small animals may be related to
thoracic forces not considered in the current model.
There is little actual data collected,but White inferred siblefor the lung pathologiesand lethality seenand that
trends of lethality with animal orientation to the blast. a predictive correlation can be madebetweenquantitat-
Thosetrendscan be reproducedby the normalizedwork ive measures of injury and the body-mass-weightedtotal
criterion if the blast load distribution variations with wave energy. Di&rences in smal animal response. and
orientation are taken into account (Fig. 9). the need for predicting location of injury suggestthat
We ooncludethat the compressionwave generatedby inclusionof more of the physiologicalstructure may be
rapid chest wall motion under blast loading is respon- profitably added.
Model of blast overpressure injury to the lung 233
E
0 \
* 1.0 9.1 1.0
\
Fig. 8. Effect of pressure waveform shape on lethality. Confirming trends reported in White et al. (1971), the
injury model predicts that normalized work and lethality are reduced when the pressure wave rises in steps
or rises with a finite slope.
Parallel to Blast
Fig. 9. Effects of body orientation on lethality. Curves depict conditions that produce 50% lethality levels
in man ( W* = 2.08) for three orientations of the body to the blast. The curves are nearly identical to those
proposed by White et al. (1971).
Acknowledgements-The authors acknowledgethe sup- Bowen,1. G., Fletcher, E. R., Richmond, D. R., Hirsch, F. G. and
oort of the U.S. Army Medical Researchand Develon- White, C. S. (1968) Biophysical mechanisms and scaling-_ pro-
cedures applicable in assessing responses of the thorax ener-
kent Commandunder contract DAMD17-93-C-3005: gized by air-blast overpressures or by nonpenetratina missiles.
Ann. N.Y. Acad. Sci. i52, 122-146.. - -
Chuong, C. J. (1985) Biomechanical modeling of thorax re-
sponse to blast loading. Final Report Contract DAMD-17-
REFERENCES 82-C-2062. JAYCOR, San Diego, California.
Clemedson, C. J. (1953) Respiratory and pulmonary gas ex-
Baker, W. E. (1973) Explosions in Air. University of Texas Press, change in blast injury. J. Appl. Physiol. 6, 213-220.
Austin, TX. Clemedson, C. J. (1956) Blast injury. Physiol. Rev. 36, 336-354.
234 J. H. Stuhmiller et al.
Clemedson, C. J. and H&man, H. (1954) Air embolism and the (1988) Computer modeling of thoracic response to blast.
cause of death in blast injury. Milit. Surg. 114, 424-437. J. Trauma 28, S132-142.
Clemedson, C.-J. and Pettersson, H. (1970) Propagation of high Stuhmiller, J. H.. Phillips, Y. Y. and Richmond, D. R. (1991) Tht:
explosive air shock wave through different parts of an animal physics and mechanisms of primary blast injury. In Textbook
body. Am J. Physioi. 100, 24-30. of Military Medicine, Conoentional Warfare: Ballistic, Blasr
Dodd, K. T., Yelverton, J. T., Richmond, D. R., Morris, J. R. and and Burn Injuries (Edited by Bellamy, R. F. and Zajtchuk, R.).
Ripple, G. R. (1990) Nonauditory injury threshold for repeat- Office of the Surgeon General. Department of the Army,
ed intense freefield impulse noise. J. Occup. Med. 32,260-266. Washington, DC.
Fung, Y. C., Yen, R. T., Tao, Z. L. and Liu, S. Q. (1988) Vander Vorst, M. J., Dodd, K. T., Stuhmiller J. H. and Phillips
A hypothesis on the mechanism of trauma of lung tissue Y. Y. (1989) Calculation of the internal mechanical response of
subjected to impact load. J. Biomech. Engng 110, 50-56. sheep to blast loading. In Proc. 10th Inr. Symp. on Military
Jonsson, A., Clemedson, C. J., Jonsson, A., Clemedson, C. J., Applications of Blast Simulation (Edited by Reichenbach, H.
Sundquist, A. B. and Avvebo, E. (1979). Dynamic factors and Ackerman, J. H.). Oberjettenber. Germany.
influencing the production of lung injury in rabbits subjected Viano. D. C. and Lau. I. V. (19881 A viscous tolerance criterion
to blunt chest wall impact. Auiat. Space Enu. Med. 50, for soft tissue injury assessment: J. Biomechanics 21, 387-399.
325-337. White, C. S., Jones, R. K., Damon, E. G.. Fletcher, E. R. and
Landau, L. D. and Lifshitz, E. M. (1959) Fluid Mechanics. Richmond, D. R. (1971) The biodynamics of airblast. Tech-
Pergamon Press, London. nical Report DNA2738T, Defense Nuclear Agency, Washing-
Military Standard 1474C (1990) Noise Limits for Army ton, DC.
Material. Yen, R. T., Fung, Y. C., Ho, H. H. and Butterman, G. (1986)
Phillips, Y. Y. and Richmond, D. R. (1991) Primary blast injury Speed of stress wave propagation in lung. J. Appl. Physiol. 61,
and basic research: a brief history. In Textbook of Military 701-70s.
Medicine, Conventional Warfare: Ballistic, Blast and Burn In- Yu, J. H.-Y., Ho, H. H. and Stuhmiller, J. H. (1990a) A surrogate
juries (Edited by Bellamy, R. F. and Zajtchuk, R.). Office of the model of thoracic responses to blast loading. Final report for
Surgeon General, Department of the Army, Washington, DC. USAMRDC Contract DAMD17-85-C-5238. JAYCOR. San
Stahl, W. R. 1967. Scaling of respiratory variables in mammals. Diego, CA.
J. Appl. Physiol. 22, 453-460. Yu, J. H.-Y., Vasel, E. J. and Stuhmiller, J. H. (1990b) Design and
Stuhmiller, J. H. (1989) Use of modeling in predicting tympanic field test of a blast overpressure test module. Final report for
membrane rupture. Ann. Otol. Rhinol. Laryngol. 98, 53-60. USAMRDC Contract DAMD17-85-C-5238, JAYCOR, San
Stuhmiller, J. H., Chuong, C. J. Phillips, Y. Y. and Dodd, K. T. Diego, CA.