Vol. 12, No. 6
JOURNAL OF CLINICAL MICROBIOLOGY, Dec. 1980, p. 772-775
Comparison of Media and Culture Techniques for Detection of
Streptococcus pneumoniae in Respiratory Secretions
TERESA C. WU,t* LINDA M. TRASK, AND ROLANDO E. PHEE
Department ofPathology and Clinical Laboratories, Mount Sinai Hospital, Hartford, Connecticut 06112
We compared the relative efficacy of three methods for the isolation of Strep-
tococcus pneumoniae in lower respiratory secretions. Based on results from 294
clinical specimens, we found that S. pneumoniae was isolated at a frequency of
65% with 5% sheep blood agar or 5% sheep blood agar containing 5 yg of
gentamicin per ml, both incubated in 5% CO2. Anaerobic incubation of 5% sheep
blood agar enhanced the recovery rate of S. pneumoniae to 93%. The improved
efficacy with anaerobic incubation is due to the greater ease ofrecognition of the
larger and more mucoid colonies of S. pneumoniae, and to the suppression of the
growth of other oral bacteria present in the respiratory sections.
The isolation rate of Streptococcus pneumo-
niae in sputum cultures from patients with
pneumococcal pneumonia by
method, i.e., 5% sheep blood agar incubated in
5% C02, is low. Several studies have shown that
approximately 55% of patients with bacteremic
pneumococcal pneumonia yield positive culture
of S. pneumoniae from the sputum (1, 4). This
low yield cannot be attributed to a decreased
viability of the organism in the presence of pha-
ryngeal flora when plating of the specimen is
delayed (12). The recognition of S. pneumoniae
is hindered by the overgrowth of other less fas-
tidious pharyngeal organisms when nonselective
media are used. Both Dilworth et al. (2) and
Sondag et al. (11) reported significant improve-
ments in the recovery of S.pneumoniae with 5%
sheep blood agar containing 5[igof gentamicin
per ml. On the other hand, Schmid et al. (9)
have concluded that gentamicin-blood agar is
generally less efficient than the standard method
for the isolation of S. pneumoniae. Howden has
compared the results of aerobic and anaerobic
cultivations for the primary isolation of S. pneu-
moniae from the respiratory tract of children
(6). His results indicated that about half of the
S. pneumoniae grew bothaerobically and anaer-
obically; the remaining half grew only anaero-
bically. Since CO2 was present in the anaerobic
incubation only, it is not clear whether the im-
proved recovery of S. pneumoniae was due to
the presence of C02, the anaerobic environment,
or both. An improved recovery rate of S. pneu-
moniae with anaerobic culture from the sputum
of bacteremic pneumococcal pneumonia pa-
tients was further demonstrated by Drew (3).
t Present address: Microbiology Service, Clinical Center,
National Institutes of Health, Bethesda, MD 20205.
The present report describes the results of a
comparative study of the yields of S. pneumo-
niae from clinical specimens with various media
and culture conditions. Our results indicate that
anaerobic culture with sheep blood agar is the
simplest procedure with the highest recovery of
MATERIALS AND METHODS
During a 3-month period, all sputa, bronchial wash-
ings, and transtracheal aspirates submitted for routine
culture to the Clinical Microbiology Laboratory of the
Mount Sinai Hospital were included in the study.
Specimens were first screened microscopically for the
suitability for culture by the criteria and procedure
outlined by Murray et al. (8). Acceptable specimens
were processed for culture by the conventional proto-
col suggested by Isenberg et al. (7). In addition to the
standard culture media, additional plates of 5% sheep
blood agar and gentamicin-containing sheep blood
agar were used for each specimen. Throughout the
study, all platings were carried out by the same tech-
Media were prepared weekly with the aid ofa Bench
Top Agarmatic Sterilizer (New Brunswick Scientific
Co., New Brunswick, N.J.) so that uniform depth (6
mm) was maintained for each batch of plates. Genta-
micin powder was supplied by Schering Corp., Bloom-
The culture methods used in the present study are
designated as follows: SBA-CO2 for sheep blood agar
incubated in 5% C02; GBA-CO2 for sheep blood agar
supplemented with 5 jig of gentamicin per ml incu-
bated in 5% C02; and SBA-ANA for sheep blood agar
incubated in a GasPak jar (BBL Microbiology Sys-
tems, Cockeysville, Md.). After overnight incubation
(18 h) at 35°C, the plates were examined by three
technicians working independently. The GBA-C02
plates were re-examined at the end of 48 h of incuba-
tion. To minimize the variation due to individual ex-
perience, the three examiners rotated during the 3-
month study so that they worked with all three culture
TECHNIQUES FOR S. PNEUMONIAE DETECTION
Initial identification ofS.pneumoniae was based on
the typical colonial morphology seen under a dissect-
ing microscope. Each suspect was further checked by
the bile solubility and Optochin disk sensitivity tests
of subcultures. In addition, representatives of each
morphological type of alpha-hemolytic colonies which
did not resemble those of typical pneumococci were
indiscriminately subcultured and tested as described
above. All subcultures were carried out on 5% sheep
blood agar incubated in 5% CO2.
All plates were reviewed by one of us (T.C.W.).
Culture media giving no growth of S. pneumoniae
were reinoculated, using colonies from plates which
yielded positive results.
Quantitation of bacterial growth based on the re-
sults of standard streaking on the original plates was
scored as follows: 1+, light growth in the primary
streaking zone only; 2+, heavy growth in the primary
streaking zone only; 3+, growth in the primary and
secondary streaking zones; and 4+, growth in the
primary, secondary, and tertiary streaking zones.
The differences in the recovery rates of S. pneu-
moniae by various culture media and conditions were
analyzed by using Yates' correction ofMcNemar's test
for significance (10).
An autolytic enzyme-deficient mutant, CWL-1, orig-
inally derived from nonencapsulated pneumococcal
strain R36NC, was kindly provided by P. S. Vanke-
teschwaren of the University of Pennsylvania.
Of 294 acceptable specimens, there were 268
expectorated sputa, 21 bronchial washings, and
5 transtrachael aspirates. Fifty-five specimens
yielded S. pneumoniae by one or more tech-
niques. The yield of positive results by either
SBA-CO2 or GBA-CO2 was similar (36 of 55
specimens, 65%); SBA-ANA gave the best yield
(51 of 55 specimens, 93%). Although there is no
significant difference between the results of
SBA-CO2 and GBA-CO2, the differences be-
tween SBA-ANA and GBA-CO2 and between
SBA-ANA and SBA-CO2 are highly significant
All isolates of S.pneumoniae werepositive by
the bile solubility test and produced a zone of
growth inhibition of -15 mm with theOptochin
disk sensitivity test.
TABLE 1. Comparison ofthe number ofspecimens
in SBA-ANA versus those in GBA-C02 and in SBA-
ANA versus those in SBA-CO2 showingthepresence
(+) or absence (-) ofS. pneumoniaea
ap < 0.001 for SBA-ANA and SBA-CO2 and for
SBA-ANA and GBA-C02.
The colonies of S. pneumoniae on the SBA-
CO2 and GBA-CO2 plates were quite similar in
morphology but different in size. Colonies were
umbilicated, with a central depression, and non-
mucoid in most isolates. Zones of alpha-hemol-
ysis surrounding these colonies were character-
istically large. Because of the small size of colo-
nies on GBA-CO2 plates, in about 2/3 of the
cases the recognition ofcolonymorphology char-
acteristics of S. pneumoniae was possible only
after 48 h of incubation. In contrast to the um-
bilicated appearance on both SBA-CO2 and
GBA-CO2, the colonies
domed in shape, and the green hemolysis was
virtually absent. A narrow green hemolysis be-
came visible after a 30-min exposure ofthe plates
in the air at room temperature. SBA-ANA col-
onies were consistently two to five times larger
than their counterparts on SBA-CO2 or GBA-
CO2 plates and had a characteristic grey, watery,
or mucoid appearance. Whereas most isolates
were correctly recognized by their characteristic
morphologies, the identification of S. pneumo-
niae in a few instances was accomplished by
indiscriminate subculturing. Oftotal positive re-
sults, 6 to 8% by any of these culture techniques
were obtained by subculturing alpha-hemolytic
colonies without the typical pneumococcal mor-
All 55 isolates grew well by the methods of
SBA-CO2, GBA-CO2, and SBA-ANA on subcul-
turing despite the fact that some of the isolates
failed to grow on all plates initially.
Positive specimens (34) yielded pure or nearly
pure cultures ofS.pneumoniae, varying between
1+ to 4+ in quantity, whereas the concomitant
pharyngeal organisms were present in quantities
of 1+ or less. The remaining 21 positive speci-
mens consisted of a mixture of 2+ to 4+ growth
ofpharyngeal organisms and 1+ to 4+ growth of
S. pneumoniae. The most common type of bac-
teria was viridans group streptococci followed
by Staphylococcus aureus. Gram-negative en-
terics were found in only 4 cases (Table 2).
To understand the mechanisms that explain
TABLE 2. Concomitantpharyngeal organisms with
2+ to 4+ growth from 55 S. pneumoniae positive
Viridans group streptococci
WU, TRASK, AND PHEE
the differences in the recovery rates and colony
morphologies among the three methods, we did
the following experiments. We compared the
plating efficiencies of pure cultures from five
clinical isolates of S. pneumoniae by the three
methods. Results indicate that SBA-CO2, GBA-
C02, and SBA-ANA give essentially the same
plating efficiencies in regard to the growth of S.
pneumoniae. The ratio of average diameters of
the colonies was 0.5 to 0.8:1:2 to 5 (GBA-CO2:
SBA-CO2:SBA-ANA). It was suggested by How-
den that the large mucoid colonies produced in
the absence of oxygen were probably the result
of the suppressed autolytic activity of S. pneu-
moniae. We repeated the plating efficiency ex-
periment by using strain CWL-1 and found that
not only were the plating efficiencies similar, but
the ratios of the sizes of colonies were also the
same as obtained with the clinical isolates.
Contrary to previous reports, the addition of
gentamicin to sheep blood agar did not enhance
the isolation rate of S. pneumoniae in the lower
respiratory secretions. The most striking effect
ofgentamicin in sheep blood agar is to eliminate
gentamicin-susceptible gram-negative enterics
and S. aureus in the pharyngeal flora. Other
species, such as viridans group streptococci,
diphtheroids, Neisseria sp., and yeasts, are not
affected by the presence of this antibiotic. Of
294 clinical specimens, 13% of the positive spec-
imens had significant numbers of S. aureus and
7% of the positive specimens contained gram-
negative enterics. The smaller colonies of S.
pneumoniae on GBA-CO2 plates presented dif-
ficulties in distinguishing S. pneumoniae from
other alpha-hemolytic streptococci even after 48
h of incubation; this difficulty could account for
all 19 false-negative results. In one specimen, the
presence of viridans group streptococci pre-
vented the recognition ofS.pneumoniae, in spite
of the complete elimination of Escherichia coli
on the GBA-CO2 plate. In view of the nature of
colonized organisms in our patient populations
and the difficulty with the recognition of S.
pneumoniae, we believe GBA-CO2 is of limited
value for the primary culturing ofS.pneumoniae
from lower respiratory tract secretions.
SBA-ANA enhanced the isolation rate of S.
pneumoniae by 27% as compared with the stand-
ard method. There were four false-negative spec-
imens. In one case, the failure to recognize S.
pneumoniae can be attributed to the growth of
a large quantity of mucoid Klebsiella pneumo-
niae. The remaining three were due to the dif-
ficulty in distinguishing S. pneumoniae from
other oral streptococci. In the majority of cases,
the colonies of S. pneumoniae could be easily
recognized by their larger sizes and mucoid ap-
pearance in the midst of normal flora, whereas
organisms such as S. aureus, gram-negative en-
terics, diphtheroids, Neisseria sp., and yeasts
were suppressed under the anaerobic incubation.
Strict anaerobes, which are prevalent in the
mouth and upper respiratory tract as normal
flora, did not interfere in the SBA-ANA method.
All 294 specimens were collected and handled
without taking special precautions against oxy-
gen toxicity. In addition, SBA-ANA plates were
examined after 18 h of incubation, a period too
short for most fastidious anaerobes to become
visible on the plates.
In spite of the fact that all 55 isolates could
grow both aerobically and anaerobically on sub-
culturing, we cannot rule out the possibility that
some strains of S. pneumoniae may be non-
aerotolerant during initial isolation. Since our
initial screening focused on alpha-hemolytic col-
onies, it is likely that the beta-hemolytic oblig-
atory anaerobic S. pneumoniae described by
Yatabe et al. (13) might have been missed in our
S. pneumoniae is currently classified as a fa-
cultative anaerobe. This organism lacks cyto-
chromes and utilizes oxygen through the flavo-
protein system, with hydrogen peroxide as its
metabolic end product. The peroxide production
is not only responsible for the green discolora-
tion surrounding colonies on blood agar, but is
also in part related to the autolysis of S. pneu-
moniae (5). A reduction of hydrogen peroxide
production during anaerobic growth may explain
the absence of alpha-hemolysis and the absence
of a central depression in the colonies. However,
the reason for the formation of large mucoid
pneumococcal colonies grown anaerobically is
not clear. Howden emphasized the importance
of the reduction in the potency of the pneumo-
coccal autolysin in the absence of oxygen (6).
We tested this possibility by employing an
CWL-1. As with the wild-type S. pneumoniae,
larger colonies were produced by this mutant
when grown anaerobically; however, we cannot
prove that the reduction of autolysis is solely
responsible for the phenomenon, and it seems
more likely that a number of events could have
contributed to the formation of the larger mu-
coid colonies in the absence of oxygen. Further
studies are needed such as the rate of DNA
synthesis and the activity of autolysin of S.
pneumoniae under both aerobic and anaerobic
This study does not intend to establish the
J. CLIN. MICROBIOL.
TECHNIQUES FOR S. PNEUMONIAE DETECTION
etiological diagnosis in pneumococcal pneu-
monia patients, but to explore a simple, cost-
effective method to improve the rate ofisolation
of S. pneumoniae in clinical laboratories. We
conclude that SBA-ANA is a simple procedure
with better results than the standard method
and gentamicin-supplemented blood agar.
1. Barrett-Connor, E. 1971. The nonvalue of sputum cul-
ture in the diagnosis of pneumococcal pneumonia. Am.
Rev. Respir. Dis. 103:845-848.
2. Dilworth, J. A., P. Stewart, J. M. Gwaltney, Jr., J. 0.
Hendley, and M. A. Sande. 1975. Methods to improve
detection of pneumococci in respiratory secretions. J.
Clin. Microbiol. 2:453-455.
3. Drew, W. L 1977. Value of sputum culture in diagnosis
of pneumococcal pneumonia. J. Clin. Microbiol. 6:62-
4. Fiala, M. 1969. A study of the combined role of viruses,
mycoplasmas and bacteria in adult pneumonia. Am. J.
Med. Sci. 257:44-51.
5. Holt, L. B. 1962. The culture ofStretococcuspneumoniae.
J. Gen. Microbiol. 27:327-330.
6. Howden, R. 1976. Use of anaerobic culture for the im-
proved isolation of Streptococcus pneumoniae. J. Clin.
7. Isenberg, H. D., F. D. Schoenknecht, and A. von
Graevenitz. 1979. Cumitech 9, Collection and process-
ing of bacteriological specimens. Coordinating ed., S. J.
Rubin. American Society for Microbiology, Washing-
8. Murray, P. R., and J. A. Washington II. 1975. Micro-
scopic and bacteriologic analysis of expectorated spu-
tum. Mayo Clin. Proc. 50:339-344.
9. Schmid, R. E., J. A. Washington II, and J. P. Anhalt.
1978. Gentamicin-blood agar for isolation ofStreptococ-
cus pneumoniae from respiratory secretions. J. Clin.
10. Sokal R. R., and F. J. Rohlf. 1969. Biometry, the prin-
ciples & practice of statistics in biological research, p.
614. W. H. Freeman & Co., San Francisco.
11. Sondag, J. E., R. K. Morgens, J. E. Hoppe, and J. J.
Marr. 1977. Detection of pneumococci in respiratory
secretions: clinical evaluation ofgentamicin-blood agar.
J. Clin. Microbiol. 5:397-400.
12. Williams, S. G., and C. A. Kauffman. 1978. Survival of
Streptococcus pneumoniae in sputum from patients
with pneumonia. J. Clin. Microbiol. 7:3-5.
13. Yatabe, J. H., K. L. Baldwin, and W. J. Martin. 1977.
Isolation ofan obligately anaerobicStreptococcuspneu-
moniae from blood culture. J. Clin. Microbiol. 6:181-
VOL. 12, 1980