U.S. Army Medical Department, Office of Medical History
Skip Navigation, go to content

HISTORY OF THE OFFICE OF MEDICAL HISTORY

AMEDD BIOGRAPHIES

AMEDD CORPS HISTORY

BOOKS AND DOCUMENTS

HISTORICAL ART WORK & IMAGES

MEDICAL MEMOIRS

AMEDD MEDAL OF HONOR RECIPIENTS External Link, Opens in New Window

ORGANIZATIONAL HISTORIES

THE SURGEONS GENERAL

ANNUAL REPORTS OF THE SURGEON GENERAL

AMEDD UNIT PATCHES AND LINEAGE

THE AMEDD HISTORIAN NEWSLETTER

The Bacterial Flora of Wounds in the Korean War

Medical Science Publication No. 4, Volume 1

WEDNESDAY AFTERNOON SESSION
21 April 1954

MODERATOR
LIEUTENANT COLONEL EDWIN J. PULASKI, MC


THE BACTERIAL FLORA OF WOUNDS IN THE KOREAN WAR*

LIEUTENANT COLONEL ROBERT B. LINDBERG, MSC

The relation of wound infections to the bacterial population of war wounds has varied extensively during the period from World War I to the present. The classic battle problem of gas gangrene was relatively inconspicuous in the Korean war, at least during the major part of the conflict, which was fought over a relatively stable front. In the two preceding wars, principal pathogenic species of anaerobes included Clostridium perfringens, novyi, septicum, and histolyticum, while in the Korean war only one of these species, C1. perfringens, was present in a high percentage of cases. Several conditions of contemporary warfare undoubtedly contributed to this altered bacteriologic picture. These included the extensive use of penicillin and the broad-spectrum antibiotics; rapid evacuation of the wounded; forward location of the Mobile Army Surgical Hospital, with consequent shortened time from wound to definitive treatment; prompt and more effective resuscitation; and radical débridement, frequently within 6 hours of wounding. The major part of data on bacteriology of wounds in the Korean war was obtained after the war had passed from its initial fluid phase to the situation of a static front. Hence the cases observed tended to have a more optimal time interval between wounding and treatment, with a consequent reduction in the bacterial population of the wounds as compared with those incurred during the initial phase of retreat and defense of the Pusan perimeter, as well as under the circumstances of retreat from the Yalu River after the Chinese intervention.

Studies of war wounds in World Wars I and II, including the studies of Weinberg and Seguin (1), Stock (2), MacLennan (3), and Smith (4), showed presence of a complex bacterial flora, with multiple species of Clostridia and aerobic organisms typically present. Their reports are chiefly of cases which had progressed to the stage of wound infection, in contrast to the situation in recent wounds in the Korean war. Here we were concerned mainly with contamination and early bacterial proliferation, before clinical evidence of infection


*Presented 21 April 1954, to the Course on Recent Advances in Medicine and Surgery, Army Medical Service Graduate School, Walter Reed Army Medical Center, Washington, D. C.


312

appeared. Comparable studies of later wounds were made during the course of evacuation and treatment in Korea and Japan.

This report presents briefly the wound flora of battle injuries at three levels: first, those seen at the time of débridement at Army Surgical Hospitals, usually within 6 to 8 hours of wounding; second, those seen at the evacuation hospital, 24 to 72 hours post-wounding; and third, those seen at the general hospital level in Japan, 4 to 21 days after wounding. The bacteriology of wounds in cases necessitating amputation one week or more after wounding was also determined. The flora of local soil, clothing and the skin surface of line troops was observed for comparison with the flora of wounds. The aerobic flora of the wounds also was studied, and will be briefly summarized. Antibiotic sensitivity of Clostridia was determined, and the relationship of this sensitivity to blood and tissue levels of antibiotics in wounded men was observed.

Observations on Flora of Recent War Wounds

Wounded personnel from the east-central front (Mundung-ni) were studied during August and September, 1952, and from December 1952 to February 1953, with cultures of tissue samples collected during débridement at the 46th MASH.

The object was to determine the extent and nature of wound contamination during the first few hours postwounding. Infection was not a major problem at this brief time interval. Specimens were collected, placed in a modified Robertson's cooked meat medium,* and shipped to Japan for study. This holding medium offered an optimal means for transporting clostridial cultures with maximum survival of multiple species in a single sample. However, an occasional series of tissue samples were found to be negative for Clostridia despite gross contamination. Such results may have been due to undetected variations in the transport and enrichment medium, and these tissue cultures hence show fewer Clostridia than were undoubtedly present in the tissues at the time of debridement.

Principal emphasis was placed on study of wounds of the extremities. The object was to determine the flora present, without losing sight of the fact that presence of Clostridia in a wound does not of necessity connote anaerobic infection. In World War II over 50 percent of wounds were contaminated with Clostridia, yet only 1 percent to 2 percent of cases showed any clinical evidence of clostridial infection (5).


*The classical formula of an aqueous meat infusion with meat particles added was modified by the incorporation of a heart infusion in place of distilled water. pH was 7.4.


313

Findings in the two series of early wound cases were sufficiently disparate to make separate presentation of the results desirable.

During August and September, 1952, 94 tissue samples were collected from 33 patients, of whom 31 (94 percent) yielded Clostridia. Of the 94 tissue blocks, Clostridia were recovered from 63 (67 percent); 125 strains, including 19 species, were recovered. The total number of species of Clostridia recovered per patient was 4. An average of 2 strains of Clostridia were recovered from each positive tissue block. The average time elapsing between injury and obtaining of tissue sample was 41/2 hours. The relative extent of tissue destruction was related to the extent of clostridial involvement: Artillery wounds were 88 percent positive and harbored an average 3 species per patient. Mortar and grenade wounds were 65 percent positive for Clostridia and harbored an average of 6 species per patient. Small arms fire produced wounds of which 55 percent showed Clostridia with an average of 1.3 species per patient. The incidence of Clostridia in tissues and in cases of 28 patients of this series is shown in table 1. It will be seen that Cl. sporogenes was the predominant strain in terms of cases and of tissue samples studied. The principal pathogenic species was Cl. perfringens, in 46.4 percent of cases, and

Table 1. Incidence of Clostridia in Tissues of 28 Patients with Recent Wounds


Clostridium species

Number of cases in which found

Percent of 28 cases positive for Clostridia

Total number of tissues examined in cases positive for sporogenes

Number of tissues positive of all positive cases

Percent of total tissues positive from positive cases

Sporogenes

18

64. 2

78

40

51. 2

Perfringens

13

46. 4

58

29

50

Bifermentans

10

35. 7

57

16

28

Novyi

7

25

24

10

41. 6

Multifermentans

7

25

37

11

29. 7

Paraputrificum

5

17. 8

16

5

31. 2

Sordelli

3

10. 7

9

4

44. 4

Lentoputrescens

3

10. 7

8

3

37. 5

Tertium

3

10. 7

6

3

50

Tetanomorphum

3

10. 7

15

4

26. 6

Butyricum

2

7. 1

7

3

42. 8

Tetani

2

7. 1

14

3

21. 4

Aerofoetidum

2

7. 1

5

2

40

Carnis

1

3. 9

2

1

50

Capitovale

1

3. 9

4

1

25

Cochlearium

1

3. 9

3

1

33

Putrefaciens

1

3. 9

3

1

33

Sphenoides

1

3. 9

4

1

25

Histolyticum

1

3. 9

2

1

50


314

in 50 percent of the tissues. Cl. novyi was present in 25 percent of this brief series of cases. No other pathogenic species appeared in large numbers.

Blood cultures taken on 15 severely wounded men yielded an unexpectedly high proportion of anaerobes: 11 were positive, with 4 species and 19 strains of Clostridia recovered. Although bacteremia due to Clostridia had been observed previously, this proportion was deemed unusually high. Subsequent studies failed to yield a comparable proportion of positives, but 3 out of 30 later cultures on wound cases were also positive for Clostridia. There was no indication that this blood stream invasion was persistent, and it may have been benign. However, it had not been thought that blood stream involvement was so widespread in war wound cases.

In view of the unexpectedly high proportion of early wounds harboring Clostridia and the continuing problem of wound infections observed at the general hospital level, a further study of similar wounds was made under conditions permitting more extensive sampling and a closer liaison between the base laboratory and the MASH. These cultures were taken during December and January, 1952-53. Winter conditions altered the tactical situation: A higher proportion of patients were evacuated by ambulance because of decreased flying time available, so that the mean time from wounding to taking of cultures rose to 7.9 hours. Of the 87 patients who were studied, 43 showed Clostridia in débrided tissues. This striking drop in incidence of clostridial contamination could not be explained by any specific factor observed. It is possible that the fact that these patients' extremities were inevitably colder during the period between wounding and treatment may have retarded bacterial proliferation.

Of the 285 tissue samples cultured from patients who harbored Clostridia, 84, or 29 percent, were positive. This lowered proportion of positive tissues may well reflect the fact that more extensive sampling was done, so that samples of uncontaminated tissue were more frequently included. Among patients harboring Clostridia, an average of 2.9 species per patient were recovered, and among tissue samples "positive" for Clostridia, 1.5 species per sample were recovered.

Table 2 shows the predominant species recovered in the two series of observations. Seventeen species were recovered during the summer and 19 during the winter study. It will be seen that the predominant species in both studies was Cl. sporogenes, while the principal pathogenic species was C1. perfringens. The incidence of C1. novyi is noteworthy; 5 to 7 percent of the total strains recovered is less than one-third the incidence of this potent pathogen observed in other recent wars (3, 4). The minor incidence of C1. septicum in contrast to its


315

importance in World War II is also notable. Of the remaining species, only C1. sordelli is generally regarded as pathogenic. (Cl. tetani will not be discussed here since it does not fall within the scope of this study.) Multifermentans, lentoputrescens and paraputrificum were the other species of principal interest.

Table 2. Clostridia Flora in Two Series of Recent War Wounds


Species

Summer 1952

Winter 1952-53

Number

Percent of total strains

Number

Percent of total strains

Sporogenes

40

29

25

20. 3

Perfringens

29

21. 2

22

17. 8

Bifermentans

15

11

3

2. 4

Sordelli

4

2. 8

4

3. 2

Multifermentans

11

8. 0

3

2. 4

Novyi

10

7. 2

7

5. 6

Paraputrificum

5

3. 6

7

5. 6

Tetanomorphum

4

2. 9

2

1. 6

Butyricum

3

2. 2

1

0. 8

Tertium

3

2. 2

3

2. 4

Lentoputrescens

3

2. 2

6

4. 8

Tetani

3

2. 2

4

3. 2

Carnis

1

0. 7

4

3. 2

Aerofoetidum

2

1. 4

2

1. 6

Histolyticum

1

0. 7

0

0

Cochlearium

1

0. 7

2

1. 6

Putrifaciens

1

0. 7

1

0. 8

Sphenoides

1

0. 7

2

1. 6

Difficile

0

0

1

0. 8

Capitovale

0

0

3

2. 4

Fallax

0

0

2

1. 6

Unclassified and unidentified

0

0

19

15. 3

 

139

 

123

 
 

Percent of total

 

Percent of total

Number of patients with Clostridia

30

90

43

57

Number of tissues with Clostridia

63 

67

84

29. 6

The flora of the environment, including the skin surface, clothing and soil of the combat area, was cultured during this investigation to obtain an indication of the actual source of wound contaminations. The skin surface harbored, among total strains recovered, 25 percent of C1. sporogenes and 23 percent of Cl. perfringens, with small numbers of bifermentans, novyi and paraputrificum. A total of 17 species


316

were found. This pattern most closely resembled that seen in the fresh wound. Only 0.7 strain per skin swab was recovered, which proportion may reflect the difficulties of survival of Clostridia on swabs even in an optimal environment. Clothing and soil showed an essentially similar population, predominantly Cl. perfringens (32 percent to 37 percent), with 12 percent to 20 percent of Cl. sporogenes. This reversed ratio of predominant species between wounds and soil, with a closer correspondence between skin and wound flora, lends credence to an hypothesis that wound contamination in this type of warfare is primarily a seeding of tissues with skin flora, rather than that of soil or clothing. An additional observation bearing on this point is the low incidence of C1. buytyricum in wounds at the MASH level, and its virtual absence in wounds cultured later. It appears that even when soil bacteria are driven into a wound, nonadapted species are promptly eliminated. Of exceptional interest is the fact that Cl. tetani was never recovered from Korean soil, although there was no difficulty in recovering it from tissues.

Wounds at Evacuation Hospital Level

At the evacuation hospital level data are less available because of the difficulties of followup of this extremely mobile population. However, on the basis of 145 cultures identified from wounds at the 11th Evacuation Hospital during early 1953, some conclusions may be drawn: The incidence of Cl. sporogenes and of Cl. perfringens strains remained approximately that noted at the MASH. Cl. novyi, relatively infrequent at the MASH level, here made up 13 percent of the strains recovered. Cl. multifermentans also increased in incidence from approximately 6 percent to 12 percent, while Cl. bifermentans decreased in incidence from approximately 10 percent to 4 percent. The total of 19 species of Clostridia recovered included most of those observed throughout the chain of evacuation, although one rare species, Cl. histolyticum, was found twice. Table 3 summarizes this information.

Wound Flora in General Hospitals

From Japan most cultures obtained were taken on patients who were not responding well to treatment. Delayed primary closure of a clean wound, within 6 days of wounding, was the rule, and such wounds, which might have shown extensive tissue contamination at the MASH, would not usually be cultured again prior to closure.

Table 4 shows the results for 1951 and 1952 in specimens submitted for culture from general hospitals. Seventy-four percent of the cases yielded Clostridia, and 67 percent of the tissues cultured were positive.


317

Table 3. Flora of Wounds at 11th Evacuation (2 to 4 days average since wounding), January through April 1953


Species


Strains


Percent of total

Sporogenes

39

26. 8

Perfringens

22

15. 1

Multifermentans

17

11. 7

Bifermentans

6

4. 1

Novyi

19

13. 1

Tetani

4

2. 7

Sordelli

1

0. 7

Carnis

5

3. 4

Histolyticum

2

1. 4

Lentoputrescens

5

3. 4

Capitovale

2

1. 4

Paraputrificum

4

2. 7

Parabotulinum

3

2. 0

Tetanomorphum

1

0. 7

Aerofoetidum

4

2. 7

Butyricum

4

2. 7

Sphenoides

1

0. 7

Tertium

2

1. 4

Cochleanium

1

0. 7

Unidentified

3

2. 0

145

 

The predominant species was Cl. perfringens, which appeared more frequently than it did earlier in the course of wound treatment. Cl. sporogenes was the remaining dominant species. C1. lentoputrescens, a proteolytic nonpathogenic species, was more common than has been observed in other series, while Cl. novyi was seen in only 4.6 percent of the total. This important pathogen showed a higher incidence in previous wars. The 23 species identified indicate the diversity of flora present. The unusual observation of Cl. feseri was made; this pathogen has apparently not previously been observed in human wounds. Another unique species was Cl. difficile, not previously reported in human cases (6).

Table 4. Clostridial Flora of Wounds at General Hospitals in Japan, 1951-52


Species


Strains


Percent of total

Perfringens

127

36. 3

Sporogenes

76

21. 7

Multifermentans

17

4. 8

Bifermentans

18

7. 3

Sordelli

1

0. 4

Lentoputrescens

29

12. 3

Novyi

11

4. 6

Cochlearium

7

2. 9

Tertium

6

2. 7

Chauvoei (feseri)

3

1. 1

Tetani

7

2. 9

Paraputrificum

8

2. 9

Tetanomorphum

4

1. 6

Difficile

1

0. 4

Filiforme

2

0. 8

Innominatum

2

0. 8

Capitovale

1

0. 4

Histolyticum

4

1. 6

Carnis

2

0. 8

Sphenoides

2

0. 8

Septicum

4

1. 6

Butyricum

2

0. 8

Aerofoetidum

2

0. 8

Unclassified

13

3. 7

349

 

318

Bacterial Flora of Amputations

In a series of 41 cases in which amputation was performed at a general hospital, tissue blocks from the amputated specimen were removed and cultured. The results of clostridial cultures are summarized in table 5. In these cases the predominant organism was again Cl. sporogenes, with Cl. perfringens observed half as often. This represents a reversal of the ratio of these species noted at the general hospital level. Three other species, multifermentans, bifermentans and novyi, were more common than they were in the hospital cases. Cl. feseri, which has been regarded as only a pathogen of horses, was recovered here as it was at the general hospital level.

Table 5. Clostridial Flora of 109 Amputation Specimens from 41 Cases of Clinical Gangrene due to Wounds, 1952

Organism

Number of strains

Percent of total strains

Number of patients

Percent of total patients

Perfringens

18

16. 5

13

31

Sporogenes

38

34. 8

23

56

Multifermentans

14

12. 8

10

24

Bifermentans

10

9. 0

8

19

Novyi

9

8. 0

9

21

Lentoputrescens

4

3. 6

4

9. 7

Cochlearium

3

2. 7

3

7. 3

Histolyticum

2

1. 8

1

2. 5

Carnis

2

1. 8

2

5

aerofoetidum

2

1. 8

2

5

Sordelli

2

1. 8

2

5

Chauvoei (feseri)

1

0. 9

1

2. 5

Tetani

1

0. 9

1

2. 5

Pasteurianum

1

0. 9

1

2. 5

Tertium

1

0. 9

1

2. 5

Tetanomorphum

1

0. 9

1

2. 5


Total


109

     

Average number of Clostridia per patient=2.7.

Assay of tissues of 28 patients at amputation was performed for determination of penicillin level in the muscle and other tissues. Proximal, viable samples of muscle showed an average of 0.26 unit of penicillin per gram, while the more distal samples contained on an averag of 0.22 unit per gram of muscle. These levels are about one-fourth of the corresponding blood content of penicillin in these cases. No specific correlation of antibiotic content and of clostridial flora was observed in the tissues assayed.


319

Aerobic Wound Flora

The aerobic flora of wounds at the various intervals described was determined in random samples of tissues. Streptococci and Staphylococci comprised over 50 percent of the total flora at the MASH, with Bacillus strains numbering 15 percent of the total. Coliform bacilli were prominent. The remainder were chiefly Pseudomonas and Proteus forms. At the evacuation hospital, a marked rise in the proportion of Proteus and Pseudomonas occurred. When patients reached the general hospital, an increase in the proportiton of Staphylococci appeared and coliform incidence fell. Streptococci were chiefly alpha and non-hemolytic forms, with hemolytic Streptococci occurring more frequently later in the healing process. The vast preponderance of hemolytic Streptococci were group D. Group A Streptococci were particularly infrequent in early wounds.

Sensitivity of Wound Flora to Antibiotics

Sensitivity of wound bacteria to antibiotics was determined on a total of 520 strains of recently isolated Clostridia. Extreme care in maintaining the proportion of inoculum to antibiotic dilution was found to be essential in order to obtain consistent results in sensitivity determinations with this group.

Table 6 illustrates the sensitivity of the principal species of clostridial flora of wounds. Approximately 85 percent of strains were sensitive to penicillin, aureomycin and terramycin; chloromycetin was relatively less effective on most strains. A small but significant proportion of strains were relatively resistant. This suggests that antibiotic resistance may well appear with sufficient readiness to constitute a problem in this situation.

When results of antibiotic sensitivity in successive years were studied, it was found that the proportion of Clostridia strains sensitive to penicillin fell from 94 percent in 1952 to 84 percent in 1953. The percentage sensitive to aureomycin and to terramycin remained constant or rose slightly during this time. Among other antibiotics tested, 100 percent of the strains were inhibited by 20 units of bacitracin.

Penicillin-resistant strains included bifermentans, novyi and sporogenes, among others.

Aureomycin- and terramycin-resistant strains occurred among perfrigens, capitovale, lentoputrescens and sporogenes. Carnis was particularly resistant to terramycin. No specific relationship between age of wound and resistant strains was noted. In tests on the aerobic flora, the pyogenic cocci tended to be more often resistant to penicillin at all levels. Aureomycin and terramycin were most often effective


320

against the Staphylococci. The hemolytic group D Streptococci were predominantly resistant to all antibiotics tested. Proteus and Pseudomonas were largely resistant to all antibiotics tested, while Coliforms were most effectively inhibited by aureomycin and terramycin.

Table 6. Range of Concentration of Antibiotics which Inhibit More than 80 percent of Clostridium perfringens, multifermentans, bifermentans, sporogenes and novyi

Clostridium species

Antibiotic

Perfringens

Multifermentans

Bifermentans

Sporogenes

Novyi

Terramycin

Minimum units inhib. conc

0. 1-1. 0

0. 05-0. 5

0. 05-0. 5

0. 1-1. 0

0. 25-1. 0

Number of strains tested

193

45

60

135

7

Percent of total inhibited

88. 7

88. 8

91. 7

89. 6

85. 7

Aureomycin

Minimum units inhib. conc

0. 05-0. 5

0. 05-0. 5

0. 05-1. 0

0. 1-1. 0

0. 1-1. 0

Number of strains tested

189

47

65

132

9

Percent of total inhibited

92. 8

89. 4

90. 8

91. 0

88. 8

Penicillin

Minimum units inhib. conc

0. 1-1. 0

0. 05-1. 0

0. 05-1. 0

0. 1-2. 0

0. 05-1. 0

Number of strains tested

184

36

57

132

9

Percent of total inhibited

89

86. 4

94. 7

91. 7

88. 8

Chloromycetin

Minimum units inhib. conc

2. 5-10. 0

2. 5-8. 0

2. 5-8. 0

2. 5-10. 0

2. 5-10. 0

Number of strains tested

186

43

63

126

8

Percent of total inhibited

81. 5

83. 7

85. 2

90. 0

75. 0

inhib.=inhibiting.
conc.=concentration.

Summary

A series of observations on war wounds from the MASH level to the general hospital showed a significantly high degree of contamination of tissues with a mixed clostridial and aerobic population. The initially predominant species was Cl. sporogenes, followed by Cl. perfringens. However, as the patients moved down the chain of evacuation, those wounds still contaminated to an extent which occasioned culture exhibited a higher and higher proportion of Cl. perfringens, with C1. sporogenes second in significance. C1. novyi was conspicuously lower in incidence than has been noted in previous wars, and C1. septicum was virtually a rarity, in contrast to findings in Europe and North Africa. In specimens from amputations, the Clostridia once more showed a lowered incidence of Cl. perfringens, with a predominance of C1. sporogenes and numerous proteolytic forms present. From 19 to 27 species were collected at the various levels


321

of study. The aerobic flora comprised primarily pyogenic cocci, gram-positive bacilli and coliform bacteria. Group A Streptococci were not encountered. The major part of the aerobic flora was penicillin-resistant, while penicillin, aureomycin and terramycin were inhibitory to the major portion of the clostridial strains.

References

1. Weinberg, M., and Seguin, P.: La Gangrene Gayeuse. Masson et cie, Paris 1918.

2. Stock, A. H.: Clostridia in Gas Gangrene and Local Anaerobic Infections during the Italian Campaign. J. Bact. 54: 169, 1947.

3. a. Mac Lennan, J. D.: Anaerobic Infections of War Wounds in the Middle East. Lancet 245: 63, 1943.
    b. Ibid.: Lancet 245: 94, 1943.
    c. Ibid.: Lancet 245: 123, 1943.

4. Smith, L. De S.: Clostridia in Gas Gangrene. Bact. Rev. 13: 233, 1949.

5. Lowry, K. F., and Curtis, G. M.: Diagnosis of Clostridial Myositis. Am. J. Surg. 74: 752-757, 1947.

6. Mac Lennan, J. D.: Personal communication, 1954.