STUDIES I N AUTOPSY BACTERIOLOGY
JOSEPH J. KURTIN, M.D.
WITH THE TECHNICAL ASSISTANCE OF ROY J. MAIKE
Departments of Pathology, Veterans Administration Center, Wood, Wisconsin, and
Marquette University School of Medicine, Milwaukee, Wisconsin
Bacteriologic study of suspected infections
during the course of necropsy is a common
procedure. Ascribing etiologic significance to
microorganisms often considered contami-
nants is problematic. There is often a
tendency to discount bacteriologic findings
when more than a few hours have elapsed
since death. The medical literature con-
cerning these problems has been sparse
during the past 15 years, a period in which
the concept of "opportunist" pathogens has
become most important. These considera-
tions served as a stimulus to study the
postmortem bacterial flora from several
sources and attempt correlation with the
clinical and morphologic aspects of autopsy
cases.
MATERIALS AND METHODS
Fifty consecutive unselected necropsies,
performed at the Veterans Administration
Center Hospital, Wood, Wisconsin, in 1957,
served as the basis for study. All of the nec-
ropsies were performed on unembalmed
bodies which had been refrigerated at 40 to
45 F. All patients were men, varying in age
from 31 to 85 years. The seventh decade
constituted, by far, the largest age group,
representing 32 of the 50 cases. The distribu-
tion of cases by major diagnosis, excluding
infections, is seen in Table 1. Of the 50
cases, 10 were free of infection. The 68 in-
fectious processes encountered in 40 cases
are noted in Table 2.
Specimens of heart blood, splenic pulp,
noninfected lung, peritoneum, and presum-
ably infectious processes, were microscop-
ically examined by direct smear preparations
stained by the Gram method, and by cultur-
Received, February 21, 1958; revision received,
March 31; accepted for publication April 21.
This paper was presented (in part) at the
Quarterly Meeting of the Wisconsin Society of
Pathologists, November 30, 1957.
239
ing in Fluid Thioglycollate Medium (BBL).
When culturing heart blood, in addition to
inoculation of thioglycollate broth with 2
ml. of the specimen, 10 ml. of blood was
placed in an Erlenmeyer flask contain-
ing approximately 50 ml. of Bacto-
Brain-Heart-Infusion-Broth (Difco). Pri-
mary cultures were incubated at 37 C. and
observed daily for evidence of growth. All
thioglycollate cultures were incubated for a
period of 14 days before being reported as
negative and discarded. Brain-heart-in-
fusion broth blood cultures were incubated
for a period of 7 days. Blood and splenic
pulp cultures were subcultured on blood
agar plates and/or examined by means of
Gram stained smear preparation, prior to
discarding as negative. Whenever possible,
all microorganisms isolated were identified as
to genera and species on the basis of mor-
phologic, tinctorial, colonial and biochemical
characteristics. The ability of strains of
Staphylococcus aureus to produce coagulase
was determined by the biochemical test
using Bacto-Mannitol Salt Agar (Difco).
Specimens for bacteriologic study were
obtained in the following manner:
Peritoneum. Immediately upon exposure
of the abdominal viscera, sterile cotton-
tipped applicator sticks were rubbed briskly
on the intestinal serosa.
Spleen. The splenic capsule was seared
in situ with a hot spatula. Splenic pulp was
forcibly aspirated by means of a sterile 20
ml. syringe with 15 gage hypodermic needle.
Noninfected lung. As soon as the thorax
was entered, an area of grossly normal lung
was chosen in situ, the pleura seared with a
hot spatula, and this area deeply incised
with a sterile knife. Sterile cotton-tipped ap-
plicator sticks were rubbed within the depth
of the incision.
Heart blood. On opening the parietal peri-
cardium, the surface of the right atrium was
240 KURTIN Vol. SO

TABLE 1 TABLE 2
MAJOR DIAGNOSES (EXCLUDING INFECTIONS) INFECTIOUS PROCESSES (40 CASES)
Number of Number of
Cases Cases
Malignancies Respiratory
Respiratory system 10 Tracheobronchitis 19
Gastrointestinal 5 Bronchopneumonia 21
Lymphomatous 4 Lobar pneumonia 2
Genitourinary 3 Empyema 1
Cerebral 1
Breast 1 Total 43

Total 24 Genitourinary
Prostatitis 7
Vascular disease (degenerative) Epididymitis 3
Coronary 9 Cystitis 6
Cerebral 3 Pyelonephritis 4
Aorta (aneurysm) 1
Total 20
Total 13
Others
Physical agents Peritonitis 3
Trauma 2 Endocarditis 1
Thermal 1 Burns, infected 1

Total 3 Total 5
Grand total 68
Chronic pulmonary emphysema 5

Acute pancreatitis 2 little variation from the results obtained

during the first 12 hr. postmortem. During
Others the first 48 hr. after death, the factor of time
Cardiac arrest (during operative pro- 1
cedure) apparently is not an important consideration
Fatty cirrhosis 1 in contamination rate. These results imply
Incarcerated hernia 1 that, contrary to the general impression, 2 ' 3
the validity of postmortem bacteriologic
Total 3 findings need not be disregarded beyond the
Grand total 50 first 6-hr. period.
Regarding the cultures of heart blood, 11
brain-heart-infusion broth cultures were
seared with a hot spatula. Approximately 12 positive whereas their corresponding thio-
ml. of blood was withdrawn with a sterile glycoUate cultures were negative. This
syringe and 15-gage hypodermic needle. discrepancy in results may be explained on
the basis of external contamination, because,
RESULTS AND DISCUSSION in some instances, growth did not appear
1. Sterility of sources related to the post- prior to opening the flasks for subculture.
mortem interval. These data are presented in Furthermore, the bacteria isolated in 8 of
Table 3, which includes all of the cases re- the 11 cultures were of environmental
gardless of the presence of infection. The variety. These included Staphylococcus al-
proportion of sterile cultures from all sources bus, diphtheroids, and biochemically inert
remains quite constant at different postmor- Gram-negative bacilli. In subsequent tables
tem intervals. With the exception of the and discussion, only the results obtained
heart blood, the 2 groups (the second 12-hr. from heart blood in thioglycoUate medium
period and the 24- to 48-hr. period) show are considered.
Sept. 1958 AUTOPSY BACTERIOLOGY

TABLE 3
STERILE CULTURES RELATED TO TIME OF DEATH

Source of Material Cultured

Time Between Heart blood
Death and Autopsy Noninfected All
Peritoneum lung Spleen
Thio Brain-heart- sources
infusion flask
per cent per cent per cent per cent per cent per cent
Less than 6 hr. (9 cases) 56 56 89 56 78 67
6 to 12 hr. (5 cases) 20 60 80 so 60 60
12 to 24 hr. (23 cases) 39 52 83 61 78 63
24 to 48 hr. (13 cases) 38 62 69 46 77 58
Irrespective of time 40 56 80 58 76

TABLE 4
BACTERIOLOGIC FINDINGS IN CASES OF RESPIRATORY INFECTION
Types of Microorganisms Isolated
Source Sterile
Cultures Virulent Streptococcus Proteus-
D. pneu- Coliform Pseudomo- Sapro-
micrococcus moniae bacilli nas phytic*
Respiratory Fecal
Lung
Pneumonia (14 cases) 1 8 3 3 1 10 3 1
Tracheobronchitis (8 0 3 1 2 0 7 4 0
cases)
Noninfected (17 9 4 2 1 0 5 0 2
cases)
Heart blood 14 0 0 1 0 3 1 1
Peritoneum 6 5 1 4 0 5 2 5
* Avirulent Micrococcus species, Bacillus sublilis, Alcaligenes fecalis, and biochemically inert
Gram-negative bacilli.

2. Results in cases of respiratory infection. in intercurrent respiratory infections is fur-

Within the total group of 50 cases, appro-
priate cultures were obtained in 17 cases of
proven pyogenic infection of the lower re-
spiratory tract. The bacteriologic findings
in these cases are summarized in Table 4.
Strains of coagulase-positive Staphylococcus
aureus, coliform bacilli, and species of the
genera Proteus and Pseudomonas constituted
the bacteria most frequently isolated from
respiratory infections in this study. These
were predominantly intercurrent respiratory
infections in persons hospitalized for a
variety of disease processes (see Tables 1 and
2). In addition to isolation from the infec-
tious process, their role as etiologic agent,
rather than postmortem contaminant, is
supported by other factors discussed sub-
sequently. The dominant role of mixed in-
fections caused by "opportunist" pathogens
ther emphasized by clinical experience at
this hospital.
This bacterial flora contrasts, in variety,
with the bacteriologic results reported by
de Vries and Pritchard. 4 In a study of similar
material, based on intercurrent broncho-
pneumonia, these authors isolated Staphylo-
coccus aureus from the lung in 87 of 213
autopsies. Other bacterial species were iso-
lated from only 8 cases: Diplococcus pneu-
moniae in 4; Klebsiella pneumoniae in 3;
and Streptococcus pyogenes in 1.
3. Possible effect of postmortem transmi-
gration on the bacterial flora of respiratory
infections. Cultures of noninfected lung were
sterile in 9 of 17 cases of lower respiratory
tract infection. In an additional 2 cases, the
bacteria isolated from noninfected lung were
considered environmental organisms (.41-
242 KURTIN
caligenes fecalis, and a biochemically inert
Gram-negative bacillus). These lacked simi-
larity to the flora obtained from infected
portions of the lung. In 6 cases, the bacterial
flora was essentially the same in the non-
infected and infected lung. Diffuse disease
(tracheobronchitis) was demonstrated in
histologic sections of the lung considered
"noninfected" at the time of necropsy in 3
of these instances. Only 3 of 17 cases remain
in which there is evidence for postmortem
transmigration of bacteria from an infectious
process to other portions of lung.
The heart blood was sterile in 14 of 17
cases with pyogenic respiratory infection.
This compares favorably with the 80 per
cent overall incidence of sterile heart blood
cultures, and is a strong argument against
the concept of rapid postmortem bacterial
invasion of the blood stream, and contiguous
organs, in an infectious process.
The peritoneal cavity was sterile in 6 of
the 17 cases of respiratory infection. In 5
additional cases the bacteria recovered from
the peritoneum were saprophytic species and
in no way appeared related to the flora in-
volved in the respiratory infections. Post-
mortem transmigration from the peritoneal
cavity to the lung remained a possibility in
only 6 of the 17 cases.
Failure to demonstrate frequent evidence
of bacteria] transmigration, and the rela-
tively high incidence of sterility of nonin-
fected lung beyond the 6-hr. postmortem
interval, help to establish the etiologic sig-
nificance of certain bacteria isolated from
intercurrent respiratory infections. These
bacteria are often considered postmortem
contaminants but are felt to assume the role
of "opportunist" pathogens.
4- Comparison of the heart blood and
splenic aspirate. Cultures were obtained of
both heart blood and splenic pulp in all but
1 of the 50 cases under study. Splenic aspira-
tion was unsuccessful in this single instance.
In 34 of the cases (69 per cent), both the
heart blood and splenic aspirate proved to
be sterile. The same bacterial species were
isolated from both sources in 2 cases. In 11
instances there were differences in sterility,
or differences in type of bacterial flora, be-
tween the 2 sources. Three of these con-
Vol. SO
cerned cases of peritonitis in which the
organism recovered from the spleen was the
same as that isolated from the peritoneal
cavity. Corresponding heart blood cultures
were sterile. Excluding these 3 cases, the
correlation between results of heart blood
and splenic cultures was 78 per cent. It is of
interest that during the first 24-hr. post-
mortem period, the heart blood yielded a
slightly higher incidence of sterile cultures
than the spleen, indicating that the heart
appears to be a more valid source for post-
mortem blood culture than the spleen, par-
ticularly in the presence of peritonitis. The
overall incidence of sterile cultures from
the spleen in this series was 76 per cent.
These results are in close accord with the
results obtained by McLellan and Wigles-
worth.1 Using an applicator stick technic
rather than aspiration, they obtained sterile
cultures from the spleen in 37 of 48 consecu-
tive autopsies.
5. Comparison of results obtained by smear
examination and by the culture method. In no
instance, regardless of source, were bacteria
seen in smear preparations of noninfected
tissues. In respiratory infections, positive
smears and positive cultures were obtained
in 7 of 17 cases of pneumonia, and in 4 of 5
cases of tracheobronchitis. Negative smears
and positive cultures were obtained in 9 of
17 cases of pneumonia and in 1 case of tra-
cheobronchitis. In no instance were smears
positive and corresponding cultures nega-
tive. Both smear and culture were negative
in 1 case of lobar pneumonia. These results
demonstrate the invalidity of denying bac-
terial etiology to an inflammatory process
at autopsy because of negative smears. How-
ever, the demonstration of microorganisms
in direct smears was always a significant
finding.
SUMMARY AND CONCLUSIONS
The bacterial flora from a number of
sources was studied in 50 consecutive autop-
sies. During the first 48-hr. postmortem
period, the factor of time does not appear
to be an important consideration in the rate
of bacterial contamination. "Opportunist"
pathogens (often considered postmortem
contaminants), as the coliform bacilli,
Sept. 1958 AUTOPSY BACTERIOLOGY
Staphylococcus aureus, and species of the
genera Proteus and Pseudomonas, play an
etiologic role in the majority of intercurrent
respiratory infections encountered at au-
topsy. Failure to demonstrate frequent evi-
dence of postmortem bacterial transmigra-
tion serves to support this thesis.
Postmortem blood cultures from the heart
and spleen give comparable results in most
instances. However, the heart appears to be
a more valid source of blood culture when
peritonitis exists.
Failure to find bacteria in direct smears
of a suspected infectious process at autopsy
need not rule out bacterial etiology, and does
not invalidate the significance of a positive
culture.
SUMMARIO IN INTERLIJSTGUA
Le flora bacterial ab varie focos esseva
studiate in 50 consecutive necropsias. Du-
rante le prime 48 horas post morte, il pare
que factor de tempore es sin grande im-
portantia pro le grado del contamination
bacterial. Pathogenos "opportunistic" (fre-
quentemente considerate como contami-
nantes post morte), como per exemplo le
bacillos coliforme, Staphylococcus aureus, e
species del generes Proteus e Pseudomonas,
ha un rolo etiologic in le majoritate del
intercurrente infectiones respiratori que es
243
incontrate al necropsia. Iste these es sup-
portate per le facto que transmigration bac-
terial post morte es infrequentemente de-
monstrabile.
In le majoritate del casos, simile resultatos
es obtenite per culturas post morte de san-
guine ab corde e splen. Tamen, le corde es
apparentemente un plus valide fonte pro
cultura de sanguine quando le patiente ha
habite peritonitis.
Le non-detection de bacterios in frottis
directe de un suspecte processo infectiose al
tempore del necropsia non suffice a excluder
un etiologia bacterial. Illo non rende invalide
le signification de un cultura positive.
Acknowledgment. Grateful acknowledgment is
made to Joseph M. Lubitz, M.D., and Marie L.
Koch, Ph.D., for their assistance.
REFERENCES
1. MCLELLAN, N. W., AND WIGLESWORTII, F. W.:
A simple method of taking bacteriologic cul-
tures of the organs at autopsy. J. Tech.
Methods, 19: 109-112, 1939.
2. SCHAUB, I. G., AND FOLEY, M. K.: Diagnostic
Bacteriology. Ed. 4. St. Louis: C. V. Mosby
Co., 1952, pp. 120-122.
3. SMITH, D. T., AND MARTIN, D. S.: In Zinsser's
Textbook of Bacteriology, Ed. 9. New York:
Appleton-Century-Crofts, Inc., 194S, pp.
947-948.
4. DE VRIES, J. A., AND PRITCHARD, J. E.: The in-
crease in serious staphylococcal infections as
shown by post-mortem investigation. Ca-
nad. M. A. J., 73: S27-S2S, 1955.