U.S. flag

An official website of the United States government

Skip to main content
Return to topReturn to top

Battle Casualties in Korea: Studies of the Surgical Research Team Volume III

Fat Embolism in Korean Battle Casualties:  An Analysis of Its Incidence, Clinical Significance, and Pathologic Aspects

Captain Robert E. Scully, MC,USAR

The significance of fat embolism following trauma has remained controversial despite the large volume of literature on the subject. Some authors in this field 4  5 1215  have regarded this phenomenon as a relatively common and serious complication of injury, while others 3  6 16 17  have felt that it is rarely severe enough to be of clinical importance. Surgeons who handled large numbers of battle casualties in the Korean conflict were seldom able to recognize the clinical syndromes described in the literature as characteristic of fat embolism.2  8

Because of such conflicting opinions, a study of this subject was initiated in the Pathology Section of the 406th Medical General Laboratory in Tokyo, Japan. Autopsy cases of all types of trauma, including that from battle injuries, were analyzed from a clinicopathologic viewpoint in an attempt to determine the specific significance of fat embolism among the many sequelae of serious injury. This chapter is concerned primarily with the 110 cases of fatal battle casualties reviewed at the laboratory during 1953. Whenever pertinent, data obtained from analysis of similar cases reviewed in 1952 and of cases of death following other types of trauma are presented.

Material and Methods

In 1953, 132 cases of fatal battle injury were studied at our laboratory. Of these, 110 were selected for investigation of the incidence, clinical aspects, and pathologic changes of fat embolism. The remaining 22 cases were eliminated either because clinical or pathologic data were inadequate or because death occurred later than 4 weeks following injury. This arbitrary time limit for exclusion of cases was selected because, in our experience, fat embolism has not been demonstrable beyond this time. Although a few of the soldiers in the series were killed instantaneously or died while being evacuated, the great majority expired at army hospitals where autopsy facilities were


available. Obviously, the 110 studied were primarily those who survived to reach the hospital and do not represent a fair, over-all sample of soldiers who died of battle wounds.
Forty-four of the autopsies were performed by two trained pathologists at the 11th Evacuation Hospital, where casualties were transferred who developed post-traumatic renal failure. A medical officer, who had previous training in pathology, performed 30 autopsies at the 46th Mobile Army Surgical Hospital. The remainder were carried out by various medical officers, some of whom had little experience in pathology, at numerous surgical, evacuation, and general hospital of the Army either in Korea or in Japan. Moderately detailed clinical abstracts were available on most of the patients autopsied at the 11th Evacuation Hospital and the 46th Surgical Hospital. On the remainder of the patients, the clinical data forwarded to the laboratory were variable, at times being detailed and at other times meager.

Routine autopsy sections of tissue fixed in formalin, embedded in paraffin and stained with hematoxylin and eosin, were available for study. In some instances, only several sections had been prepared; for the great majority of autopsies, however, there were 20 or more sections. In 89 cases, Oil-Red-O fat stains were made on frozen sections of formalin-fixed lung tissue cut at 30 to 40 micra in thickness. In 33 cases, samples of kidney, and in 24, samples of one or several portions of the brain were similarly stained.

All Oil-Red-O sections were graded according to their content of intravascular fat. This grading was attended by some difficulty since the fat was sometimes distributed unevenly in a section. Because it was not feasible to do large numbers of fat stains on any one organ, estimation of the amount of intravascular vacuolation in routine hematoxylin and eosin sections (in which the fat is dissolved out) was resorted to as an ancillary method of grading.

The amount of fat in the lungs was graded: 0 to 8, Grade 1 representing less than 10 small-sized emboli per section, and Grade 8 signifying the presence of emboli in the majority of high-power fields and the appearance of extensive beading of alveolar capillaries with globules of fat. The intermediate grades were difficult to define accurately, however, and selection was made on the basis of visual impressions. Since all sections were examined at least twice and were finally reviewed within a short period of time, it is believed that their grading was reasonably accurate. Grades of Oil-Red-O and of hematoxylin and eosin sections of lung, determined independently of one another, were identical in 47 percent of the cases; they differed by one in 33 percent of the cases, by two in 17 percent, and by three in 3 percent. When grades differed by one. the higher of the two was selected for statistical purposes; but when the difference was two or three, the


The renal content of fat was graded 0 to 8. Grade 1 corresponded to less than 5 emboli per section, and Grade 8 to the presence of multiple emboli in the great majority of glomeruli. Hematoxylin and eosin grading was less successful in the kidneys than in the lungs. However, it was of definite value, matching closely with Oil-Red-O grading in all but a few instances.

Brain sections containing fat were seldom found; hence grading was not carried out. It was considered more appropriate to describe the occasional, positive sections as having minimal, slight or moderate quantities of fat. Recognition of fat vacuoles in hematoxylin and eosin sections of brain was difficult; therefore, quantification of the fat without examination of Oil-Red-O stains was unreliable.

Independently of the microscopic study, the abstracts of the clinical records were combed, giving special attention to duration of life, type and distribution of wounds, presence or absence of fractures, degree of shock, and prominence of cardiovascular, pulmonary, renal and cerebral symptoms. Other pathologic findings at autopsy, especially those referable to the cardiovascular system and lungs, were tabulated in an attempt to discover if their presence could be related statistically to the amount of fat embolism observed microscopically.

Results of Analysis

General Aspects of Cases. The ages of the patients ranged from 18 to 32 years. The average age was 22; and less than 10 percent of the patients were over 25. Shell fragments caused wounds in 85 percent of the injured soldiers, while 15 percent were caused by bullets. Ninety percent of the shell-fragment wounds and 33 percent of the bullet wounds were multiple. Seventy-five percent of the patients had fractures. In a few instances, significant burns complicated the injuries. The anatomic distribution of the wounds is presented in Table 1.

Table1.  Distribution of Wounds in 110 Cases

The duration of life after injury ranged from seconds to 28 days. Approximately 25 percent of the patients expired in less than 24 hours; 40 percent survived between 1 day and 1week; and 33 percent died later than 1 week following injury.


Incidence. Of the 89 patients of whom Oil-Red-O sections of lung were examined, 83 (93 per cent) showed pulmonary fat emboli. Using intravascular vacuolation in routine hematoxylin and eosin sections as the criterion for fat embolism, findings were positive in 87 out of 110 cases, an incidence of 79 percent. It is apparent that minimal degrees of embolism occasionally escaped detection when fat stains were not used. Analysis of both routine and fat stains, as presented in the section on Materials and Methods, revealed that 50 percent of the soldiers showed no fat embolism or only minimal degrees of it (Grades 0 to 2); 31 percent showed slight degrees (Grades 3 and 4); and 19 percent showed moderate to marked degrees (Grades 5 to 7).

Extensive experience with both routine and fat stains has shown that systemic fat emboli are most abundant and easiest to recognize microscopically in the kidneys. For this reason, the presence of fat in the kidneys has been regarded as the most sensitive pathologic index of the entrance of emboli into the systemic circulation. Previous studies by workers in this laboratory,9  13 and by others," have revealed that (1) renal involvement rises in incidence and grade with increasing degrees of pulmonary fat-embolism and (2) minimal to slight degrees (Grades 0 to 4) of the pulmonary fat-embolism are only rarely accompanied by the appearance of fat in the kidneys.

For the two reasons stated, renal fat stains were performed routinely in the present series only in cases in which the grade of pulmonary fat-embolism was 4 or greater. In 26 such instances, the kidneys were positive in 17. The grade of pulmonary fat-embolism in the remaining cases being 3 or less and hematoxylin and eosin sections of the kidneys disclosing no evidence of fat embolism, it was considered justifiable and conservative to grade the renal fat-embolism as "none to minimal" (Grades 0 to 2). Combining the results of grading both routine and fat stains, it was found that 92 percent of the 110 patients showed none or minimal renal fat-embolism. Four per cent exhibited a slight degree (Grades 3 and 4), and an additional 4 percent showed moderate to marked degrees (Grades 5 to 7).

Because of the importance of cerebral fat-embolism as a potential cause of death, sections of the central nervous system were studied with fat stains whenever the pulmonary grade was 4 or greater and brain tissue was available. Usually one section of cortex was examined. Often additional sections from cortex and other parts of the brain were studied as well. In the 24 cases in which the brain was stained with Oil-Red-O, findings were positive in five. Cerebral fat-embolism was less common and almost always less marked in degree than renal fat-embolism. Of the 24 cases mentioned previously, the kidney was positive for fat in 15, an incidence three


times that of the brain. Similarly, in a previous series dealing with various types of trauma, the kidney was positive for emboli in 18 and the brain in only 9 of 42 cases in which both renal and cerebral fat stains were made.13  In neither the present nor a previous series was the brain positive when the kidney was negative. In three of four cases of the present series in which renal fat- embolism was slight, cerebral embolism was minimal in two and absent in the third (in the fourth case no Oil-Red-O section of brain was available). In three of four cases in which renal fat- embolism was moderate to marked, brain embolism was minimal in two and moderate in the third (in the fourth case no Oil-Red-O section of the brain was available).

In summary, about 90 percent of the soldiers dying in hospitals less than 1 month after battle trauma showed pulmonary fat-embolism at autopsy; however, only 19 percent of them showed more than a slight degree. Less than 10 percent of the soldiers had more than minimal renal fat- embolism; and a much smaller percentage had more than minimal cerebral involvement.

With regard to the comparative incidence and severity of fat embolism in various types of trauma, our experience has been that the highest incidence and severest grades were observed in individuals who had been severely beaten and had extensive contusions of the extremities. There were 13 such cases. A study of more than 50 deaths following vehicular accidents revealed an incidence and severity of fat embolism roughly paralleling that complicating battle trauma. The lowest incidence and mildest degrees of pulmonary fat-embolism among the large groups of trauma cases studied were in patients who died of single bullet wounds of the head or chest (over 40 cases examined).

Although an organized study was not made of the incidence of pulmonary fat-embolism in nontrauma cases, it has been impressive whenever present because of its rarity and its mildness.

In 18 cases of patients dying after extensive burns, findings showed that pulmonary fat-embolism was absent or minimal in 15, with slight embolism in two cases and moderate embolism in the other case.

Pathogenesis. Little new information on the pathogenesis of fat embolism was obtained from an analysis of the present series of cases. It was not possible to make an accurate estimate of the amount of soft-tissue and bone injury from the autopsy records; hence these injuries could not be related to the degree of pulmonary fat-embolism. However, it was feasible to make a correlation between the topography of the wounds, the presence or absence of fractures, and the amount of fat in the lungs.

It was found that only one patient showed moderate pulmonary fat-embolism among the 31 patients having wounds confined to the


head, neck, trunk, or a combination of these areas. In contrast, 20 of the 79 patients whose extremities were involved showed moderate to marked embolism. One explanation for this difference is probably the larger amount of osseous damage suffered in the extremity cases. Of the 79 extremity cases, 63 were associated with fractures and 19 of these showed moderate to marked pulmonary fat-embolism. On the other hand, only one of the 16 patients not having fractures exhibited moderate embolism.

Although the wounded-in-action cases provided no opportunity for correlation between the amount of soft-tissue damage and the degree of fat embolism, study of the 13 cases of death due to beating suggested the importance of trauma to adipose tissue in the production of fat embolism. At autopsy, these patients commonly showed marked pulmonary fat-embolism and occasionally a marked systemic fat-embolism, despite the absence of obvious fractures. It is possible, of course, that jarring of the bones with microscopic fractures provided a source of embolic fat in these cases.

Clinical Aspects. An attempt to delineate a pulmonary symptom complex in the present series of cases was unsuccessful. Dyspnea and cyanosis were often mentioned as prominent features in the courses of these patients. However, there appeared to be no correlation between these symptoms and the amount of fat in the lungs. An occasional patient had a more or less acute onset of respiratory distress at some interval after injury and died, showing moderate to marked pulmonary fat-embolism. On the other hand, such histories were also obtained in association with only minor degrees of embolism, and were usually absent in cases showing high grades.

In this series, there was no evidence to support a cardiovascular symptom complex. No signs were observed of right-sided heart failure which could be correlated with moderate or marked pulmonary fat-embolism. In a small percentage of the cases in which it was possible to estimate the degree of shock, no correlation was demonstrable between it and the pulmonary fat content. Furthermore, the occasional occurrence of unexpected sudden death bore no consistent relationship to the quantity of embolic fat in the lungs.

With reference to systemic fat-embolism, opthalmoscopic examinations and laboratory examinations of the urine for fat were not recorded in our abstracts. Cutaneous petechiae were noted in occasional cases of the series, but could not be correlated statistically with the degree of fat embolism at autopsy.

Only two cases presented a clinical picture suggestive of cerebral fat-embolism. At autopsy, one had a moderate degree, while the other showed cerebral edema but no fat-embolism.

Case 1 (Cerebral Fat-Embolism.). A 29-year-old Korean soldier was admitted to an army surgical hospital 7 hours after being treated at a battalion aid station. On arrival, he was in shock and had a


blood pressure of 76 systolic, 48 diastolic, lie had traumatic amputations of the right leg through its lowest third, of the left hand, and of several fingers of the right hand. There were open, comminuted fractures of the left tibia and fibula. Multiple shell-fragment wounds of the right hand and scalp were present. Both corneas showed abrasions.

Several pints of blood were administered, and 2 hours later the patient was taken to the operating room. A guillotine amputation of the right lower leg was performed through its lowest third. A similar amputation of the lower left forearm was performed. The first, second, third, and fourth fingers of the right hand were amputated through the proximal phalanges. Débridement of the left leg, foot, ankle, the right hand, and the scalp was carried out.

The patient`s immediate postoperative course was not unusual. However, 24 hours following operation he became stuporous, and an intracranial lesion was suspected. He was evacuated to another surgical hospital for consultation and; on arrival, he showed respiratory difficulty. His blood pressure was elevated. He was taken to the operating room 35 hours after admission to the battalion aid station and a cranial exploration revealed the brain to be under a considerable amount of pressure. No epidural or subdural hemorrhage was found. Postoperatively, the patient`s respirations were labored. His blood pressure dropped and he expired in 10 hours.

At autopsy, the principal gross findings were moderate dilatation of the left ventricle, slight to moderate pulmonary congestion and edema (lung weight: 600 gm. each), cerebral swelling, and a cerebellar pressure cone. Microscopically, the lungs showed moderate to marked congestion and edema and Grade 6 fat embolism. There were scattered foci of early bronchopneumonia. The kidneys showed Grade 6 fat embolism. Groups of capillaries here and there in the myocardium were plugged with fat. The brain exhibited groups of capillaries filled with fat in all sections, including cerebral cortex, basal ganglia, brain stem and cerebellum. Small foci of rarefaction surrounded some of the emboli in the cortex and basal ganglia. Occasional vacuoles produced by fat emboli were seen in the splenic and adrenal sinusoids in hematoxylin and eosin sections.

Case 2 (Cerebral Fat-Embolism). A young Korean soldier was injured by a land-mine explosion sustaining traumatic amputation of the lowest third of the right leg, multiple shell-fragment wounds of the left leg with avulsion of the skin of the knee, shell-fragment wounds of the hands and of the left upper inner arm with an axillary hematoma, and avulsion of the scrotum with fragmentation of the left testis
On arrival at a Marine medical battalion aid station, his blood pressure was 75 systolic and 65 diastolic, pulse 120, and respirations 20.  His chest was clear to percussion and ausculation. There was


generalized abdominal tenderness. Under gas-oxygen-ether anesthesia, a guillotine amputation of the right lower leg was made; the left testis was excised, and the multiple wounds were débrided. Six pints of blood were given before and during the operation.

Postoperatively, the patient sank into deep coma, On the first postoperative day, his blood pressure was 120 systolic and 60 diastolic, pulse 110, and respirations 20. His chest was clear; and his abdomen was slightly rigid. Catheter output of urine was 1 liter. Early during the second day his temperature was 102o F. He had noisy respirations due to retained secretions; otherwise there was no change in his condition. About 49 hours following operation and 57 hours after injury, he developed signs of an expanding intracranial lesion. The right pupil was larger than the left. He had weak, generalized "shaking" convulsions, and his blood pressure rose to 170 systolic, 100 diastolic. A tracheotomy was performed. Bilateral cranial exploration was negative. Skin petechiae appeared. Death occurred 6 hours after the onset of the final episode, i. e., 55 hours after the first operation.

The most striking finding at gross autopsy was the presence of petechiae disseminated throughout the white matter of the brain. These were most numerous in the region of the substantia nigra. The gray matter appeared uninvolved. Microscopically, the brain showed numerous round to oval areas of infarction; some of the larger ones exceeded in dimension the diameter of a low-power field. The infarcts were characterized by disintegration of the intercellular framework, decrease of its tinctorial properties, nuclear shrinkage and, in some instances, infiltration of polymorphonuclear leukocytes. Occasional petechiae were seen; but these were less numerous than the anemic infarcts. The latter were distributed chiefly in the cortical white matter, but they were also present in the deeper layers of the cortical gray matter and elsewhere. Oil-Red-O stains revealed that many of the capillaries and small arteries, especially in the gray matter of the cortex, were distended by fat globules. Only occasional infarcts in the white matter contained demonstrable fat at their centers.

Sections of the lungs revealed Grade 7 embolism and numerous foci of bronchopneumonia. The heart. exhibited fat emboli distending groups of smell vessels. The kidney showed a Grade 5 embolism and moderate numbers of tubular casts of hemoglobin type. In the sinusoids and portal space vessels of the liver, there were scattered vacuoles produced by fat emboli. The parenchymal cells in the central halves of the lobules contained numerous medium-sized fat. droplets. The remainder of the organs exhibited occasional intravascular vacuoles suggesting fat emboli.

Pathology. Results of this study did not reveal any constantly occurring gross pathologic. changes in lungs showing fat. embolism. Neither edema, focal hemorrhage, patchy emphysema, nor atelectasis


Table2. Average Combined Lung Weights in Fat Embolism

were described consistently in the lungs with the severer grades of embolism. Table 2 shows the average combined weights of 102 pairs of lungs correlated with the grades of fat embolism at various times at which death occurred following injury. This table obviously affords no support for the thesis that fat embolism favors the development of pulmonary edema.

After considerable experience, it became possible to recognize with ease the vacuoles produced by fat emboli in hematoxylin and eosin sections of the lung. In the larger vessels these appeared as large, round or oval spaces which often coalesced (Fig. 1). Unlike arte-

FIGURE1. Lung. The arteries contain multiple round and oval vacuoles. Several alveolar wall capillaries are distended by large vacuoles. (X100 H&E)


factual spaces which merely displaced intravascular red cells, the vacuoles of fat compressed them so that their surfaces became concave along the margins of the vacuoles. Arterioles and capillaries were plugged and distended by round, oval or reniform vacuoles. At the periphery, the vessel walls and their endothelial linings were tightly compressed suggesting great tension (see Fig. 1). In sections of tissues fixed in acid formalin, the deposition of delicate, golden-yellow crystals of formalin pigment in the vacuoles was of confirmatory value in identifying them as fat.

In 4 of the 110 cases, bone-marrow emboli were found in small numbers in the pulmonary arteries (Fig. 2). A tiny fragment of embolic skin was seen in one case. It was common for thrombi composed largely of fibrin and eventually penetrated by endothelial cells to lay in juxtaposition to the embolic fat globules. Since more often than not these thrombi were situated away from the fat emboli, it seemed probable that the association of the two was of a collision type.

The use of fat stains revealed a striking picture, often outlining the pulmonary arterial tree and alveolar wall capillaries (Fig. 3).

The topographic localization of fat emboli within the lungs was not worked out in detail. Sections taken from various regions of the lungs generally contained fairly similar amounts of fat. Occasionally, one section of lung differed by two grades from another, and rarely by three. In sections in which only a small amount of fat was present, it was often confined to a thin rim of lung beneath the pleural surface.

FIGURE2. Lung. Bone marrow embolus in artery.  (X200 H&E)

FIGURE3.Lung. Arteries and alveolar wall capillaries are plugged with fat. The alveolar walls present a beaded appearance.  (X100 Oil-Red-O)


Microscopically, just as grossly, it was not possible to correlate the presence of pulmonary fat- emboli with focal edema, hemorrhage, emphysema, or atelectasis. Table 3 shows the relations of  pulmonary fat-embolism to the time of death following injury.

Table3.  Relation of Fat Embolism to Duration of Life After Injury

The severest degrees of fat-embolism were seen most often in the first week after injury, the peak incidence being in the early part of the week after the first day. There was a definite tendency for the pulmonary fat-content to decrease during the second, third, and fourth weeks. At one extreme, fat embolism was seen in patients who presumably lived no longer than minutes; at the other, in the wounded-in-action series, it was demonstrated as late as the 24th day following injury.

No clues were apparent from the histologic sections of lung as to how the embolic fat was eliminated. It was unusual to be able to demonstrate fat droplets within the alveoli. Occasionally, they were extravasated with blood. In one case, there was focal massive outpouring of fat into the air spaces in the absence of obvious hemorrhage. It has not been possible to recognize phagocytosis of extravasated fat by alveolar phagocytes. The lipochrome granules, so commonly seen in these cells, stained an orange or rust color with Oil-Red-O in contrast to the bright red of the embolic fat. Likewise, the quantity of phagocytes bore no consistent relationship to the amount of fat within nearby vessels. Only one case showed a reaction of vascular endothelium to fat emboli. In this instance, the endothelial cells surrounding the fat were filled with numerous small, uniform vacuoles, presumably fatty in nature. No variation in the appearance of the emboli was detected in relation to time. The embolic fat seen in several cases in which death occurred less than 10 hours after injury appeared identical to that observed within pulmonary vessels as late as 24 days after injury. It is possible that local enzyme activity is responsible for the gradual elimination of the pulmonary fat or that the droplets are being propelled continually into the systemic circulation


and excreted by the kidneys. The failure to find fat regularly in the air spaces does not exclude the possibility of rapid elimination in the sputum.

When more than a minimal amount, usually renal fat-embolism was easy to recognize in hematoxylin and eosin sections by observing the vacuolation of the glomerular tuft capillaries (Fig. 4). When the degree of embolism was minor, the fat was most apt to appear in the glomeruli near the renal capsule. Emboli were seen also in the inter-tubular vessels, chiefly in the cortex (Fig. 5). Rarely was fat visible in the form of droplets or casts within the tubules (Fig. 6). In a single case of death due to beating, fibrin thrombi lay in juxtaposition

FIGURE4. Kidney. Many glomerular tuft capillaries are distended by fat vacuoles.(X100 H&E)

to the fat emboli in the glomeruli. Fat-embolism was seen in the kidneys as early as 10 to 18 hours after injury and as late as 17 days.

Emboli were scattered in groups in both gray and white matter of the central nervous system. Vacuolation in hematoxylin and eosin sections of brain was often difficult to recognize. Occasionally, an empty, dilated vessel with a tenuous wall was highly suggestive of fat-embolism (Fig. 7). Parenchymal lesions secondary to the emboli were observed in only two cases. In these two cases, petechiae and small areas of anemic necrosis appeared (Fig. 8), especially in the white matter of the cortex, but elsewhere as well. The infarcts were characterized by a loosening and loss of tinctorial properties of the intercellular framework and shrinkage of both nerve and glial cells.


FIGURE5. Kidney. The glomeruli contain pretzel-like masses of fat. The interlobular vessels are also plugged with fat. (X100 Oil-Red-O)

FIGURE6.  Kidney. The uniformly black intratubular casts are fat. The paler granular casts are of hemoglobin type. (X100 Oil-Red-O)

Affinity was decreased for both eosin and fat stains. Some lesions showed infiltration by polymorphonuclear leukocytes. In occasional infarcts, centrally situated capillaries or arterioles were plugged with fat.

FIGURE7. Posterior Pituitary. Three small vessels are distended by vacuoles suggestive of fat. (X650 H &E)


FIGURE8. Cerebral Cortex. A pale, irregular area of rarefaction is seen at the junction of gray and white matter. Small black streaks, representing fat emboli in the gray matter can barely be made out at this power. (X50 Oil-Red-O)

Using fat stains, emboli were seen most commonly in the gray matter of the cortex (Fig. 9). A confusing element in the interpretation of fat stains of the brain was the frequent presence of phagocytes containing lipochrome granules in a perivascular location. These granules, which had a golden color in hematoxylin and eosin sections, seemed to be entirely unrelated to fat emboli; and often they appeared in their absence.

Oil-Red-O stains were made on only a few heart sections. In Case 1, small groups of capillaries were distended with fat globules (Fig. 10). The myocardium showed a few scattered small foci of degeneration characterized by hyalinization of fiber cytoplasm and nuclear pyknosis. It was not certain that these changes were ante-mortem. Fatty degeneration of muscle fibers surrounding the emboli and petechiae were not observed.

Next to the lungs and kidneys, fat emboli were most easily recognized in hematoxylin and eosin sections of the adrenals and spleen. Here they produced characteristic round, oval, and sausage- shaped vacuoles which were most numerous in the subcapsular zones (Figs. 11 and 12). In occasional cases, vacuoles highly suggestive of fat emboli were seen in other organs. With the exception of the brain. degenerative parenchymal lesions secondary to fat embolism were not observed.


FIGURE9. Cerebral Cortex.  The gray matter contains numerous fat emboli. (X100 Oil-Red-O)

FIGURE10.Myocardium. A group of capillaries contain fat emboli.  (X108 Oil-Red-O)

FIGURE11. Adrenal Cortex. Many Sinosoids are distended by round and sausage-shaped vacuoles produced by fat emboli.  (X138 H & E)

FIGURE12. Spleen. Several arterioles are distended by vacuoles of fat.  (X100 H & E)


Because of the many variables existing in any large series of trauma cases, it is not reasonable to compare the high incidence of fat embolism in our series (about 90 percent) with incidences in most other reported series. Mallory`s collection of autopsies on World


War II wounded 10 is perhaps as comparable to our series as any in the literature. His incidence of pulmonary fat-embolism (67 percent) is lower than ours. However, his figure for "significant" embolism (20 percent) tallies closely with this series for moderate to marked embolism (19 percent). Likewise, our incidence of moderate to marked systemic fat-embolism (4 percent) is identical to that given by Mallory for "significant" systemic fat-embolism.

This series does not support the widely accepted concept of fat embolism as a clinicopathologic entity in regard to its role in producing significant pulmonary dysfunction. Analysis of our cases did not reveal statistical evidence that fat embolism causes pulmonary distress. Likewise, there was no pathologic correlation between the amount of pulmonary fat-embolism observed and the presence and degree of pulmonary edema, pulmonary hemorrhage, atelectasis, emphysema, or right-sided cardiac dilatation. Mallory was similarly unable to correlate fat embolism and pulmonary edema on a statistical basis.10 It might be possible that serious pulmonary embarrassment might result from the presence of embolic fat in an occasional individual case. However, this finding must be quite uncommon, or it would be detected in some of the large series investigated, such as this series.

In reviewing the literature on fat embolism, the investigators were impressed by the lack of "controls," i.e., injured patients who died showing none or, at most, slight fat embolism at autopsy. The failure to compare such cases with cases of moderate to marked fat embolism exposed one to the risk of "post hoc, ergo propter hoc" fallacious reasoning in implicating fat embolism as the cause of various clinical and pathologic findings. Furthermore, the frequent absence of pulmonary symptoms and signs antedating the onset of cerebral fat-embolism7 11  18  is difficult to explain if one holds that pulmonary fat-embolism produces respiratory embarrassment. Cerebral fat-embolism, with possible rare exceptions, is encountered only when the degree of pulmonary involvement is moderate to severe.

Our conclusions, based on clinical data, are identical to those of Armin and Grant who employed experimental methods.1 These authors produced pulmonary fat-embolism by injecting rabbit perirenal fat into ear veins of unanesthetized rabbits. They estimated the amount of fat necessary to inject in order to produce a degree of pulmonary fat-embolism equivalent to a severe degree in the human. This estimate was based on measurement of the extractable fat content of severely involved human and rabbit lungs, as well as on grading of histologic sections.  From their experiments, Armin and Grant reported four conclusions. (1) The production of pulmonary fat-embolism corresponding to a severe grade in the human caused no change in blood pressure.


pulse rate, or respiration in the normal rabbit. The dose given to produce such an embolism was much less than that required to kill the rabbit. (2) In animals subjected to severe hemorrhage (averaging about 40 percent of the blood volume), the additional production of severe, pulmonary fat-embolism did not alter the clinical picture or increase the mortality above that due to hemorrhage alone. (3) The production of severe, pulmonary fat-embolism had no effect on the response of the animals to transfusion after hemorrhage. (4) Unless the rabbit tolerates fat embolism better than man, human gross pulmonary fat-embolism is unlikely to provoke symptoms or to be solely or in part responsible for death after injury.

As a corollary to our inability to relate evidence of pulmonary distress to fat embolism, the percentage of trauma deaths attributable to fat embolism was exceedingly low as compared with rates given elsewhere in the literature (in the vicinity of 1 percent). Cases listed as fatal were only those in which death was associated with a typical cerebral syndrome clinically and a significant degree of cerebral fat-embolism pathologically. The high mortality rates given in the literature depend largely upon the assumption that moderate to severe grades of pulmonary fat-embolism, whenever demonstrated, represent a primary cause of death or are a major factor contributing to death.

Several objections may be raised to the nature of the material from which our conclusions have been drawn. (1) The clinical abstracts at our disposal were not always detailed. (2) In caring for large numbers of wounded soldiers most of whom were seriously ill, field surgeons may not have accurately or uniformly recorded any but the severest signs or symptoms. (3) Pathologically, it was impossible to judge the ages of pulmonary fat-emboli; i. e., it could not be determined in any one case whether the emboli had accumulated in successive crops over a period of time or whether they had involved the lungs in a single, massive onslaught. Such a determination, if possible, would undoubtedly have increased the value of our clinicopathologic correlation. Despite these possible objections (one or more of which applies equally to the other series in the literature), it is believed: (1) that little in the way of conclusive proof has been offered for the opinion that moderate to severe pulmonary fat-embolism causes significant pulmonary dysfunction or death, and (2) that other possible sequelae of injury have not been adequately excluded as causative agents of pulmonary signs and symptoms when such appear.

It was evident that cerebral fat-embolism was seldom present in significant amounts, since only 1 percent of the cases of battle trauma coming to autopsy showed this entity.


A clinicopathologic analysis of 110 cases of death occurring at military hospitals up to 4 weeks after battle trauma was made in order to determine the incidence and significance of fat embolism.


Fat embolism, as evidenced by the presence of fat droplets in the pulmonary vessels, was demonstrable in approximately 90 percent of the 110 patients. In only 19 percent was the degree of pulmonary fat-embolism more than slight.

Of the 110 patients, only 4 percent showed more than slight systemic fat-embolism, as evidenced by the appearance of fat in the kidneys: and only 1 percent showed fatal (cerebral) fat embolism.

From both clinical and pathologic viewpoints, there was no evidence from the analysis that the presence of fat in the lungs causes pulmonary dysfunction or death. This finding is at marked variance with the concepts of many writers concerning the significance of pulmonary fat-embolism.

Fat emboli may be recognized and quantitated with considerable accuracy without the use of fat stains by searching for and estimating the amount of intravascular vacuolation. This can be done with greatest success in the lungs, and with somewhat less success in the kidneys.


1. Armin, J., and Grant, R. T.: Some Observations on Gross Pulmonary Fat Embolism. Cited by Grant, R. T.: Brit. M. Bull. 10: 17, 1954.
2. Artz, C. P.: Personal communication.
3. Darrach, W.: "Discussion of a Case Report." Harris, R. I.; Perrett, T. S.: and MacLachlin, A.: Fat Embolism, p. 1113. Ann. Surg. 110: 1095, 1939.
4. Denman, F. R., and Gragg, L.: Fat Embolism: A Diagnostic Enigma. Arch. Surg. 57: 325, 1948.
5. Glas, W. W.; Grekin, T. C.; and Musselman, M. M.: Fat Embolism. Am. J. Surg. 85: 363, 1953.
6. Groendahl, N. B.: Cited by Scuderi, C. S.: Fat Embolism: A Clinical and Experimental Study. Surg. Gynec. & Obst. 72: 732, 1941.
7. Groskloss, H. H.: Fat Embolism. Yale J. Biol. & Med. 8: 59, 175, and 297, 1935-1936.
8. James, E. S.: Fat Embolism. Canada M. Assn. J. 62: 548, 1950.
9. Lukeman, J.: Fat Embolism. Annual Historical Report, 406th Medical General Laboratory, pp. 92-94. United States Army, Washington, D. C., 1951.
10. Mallory, T. B.: "General Pathology of Traumatic Shock"; The Physiologic Effects of Wounds, pp. 283-305. Board for the Study of the Severely Wounded, North African-Mediterranean Theater of Operations. Department of the Army, Office of The Surgeon General, Washington, D. C., 1952.
11. McArdie, M. J. F.: Cited by Newman, P. H.: The Clinical Diagnosis of Fat Embolism. J. Bone & Joint Surg. SOB: 290, 1948.
12. Robb-Smith, A. H. T.: Pulmonary Fat Embolism. Lancet 1: 135, 1941.
13. Scully. R. E.: Fat Embolism. Annual Historical Report, 406th Medical General Laboratory, pp. 197-201. United States Army, Washington, D. C., 1952.
14. Vance, B. M.: The Significance of Fat Embolism. Arch. Surg. 23: 426. 1931.
15. Warthin, A. S.: Traumatic Lipaemia and Fat Embolism. Internat. Clin.. 4: 171, 1913.
16. Whitson, R. 0.: A Critique of Fat Embolism. J. Bone & Joint Surg. 33: 447, 1951.
17. Wilson, J. V., and Salisbury, C. V.: Fat Embolism in War Surgery. Brit. J. Surg. 31: 384, 1943-1944.
18. Winkelman, N. IV.: Cerebral Fat Embolism. Arch. Neurol. & Psych. 47: 57, 1942.