Journal of Antimicrobial Chemotherapy (2001) 47, 87–91

JAC
High rates of resistance to cephalosporins among viridans-group
streptococci causing bacteraemia in neutropenic cancer patients

Anna Marrona*, Jordi Carratalàa, Fernando Alcaideb, Alberto Fernández-Sevillac and

Francesc Gudiola

Services of aInfectious Diseases, bMicrobiology and cClinical Haematology, Hospital de Bellvitge,

Institut Català d’Oncologia, Ciutat Sanitària i Universitària de Bellvitge, University of Barcelona, Spain

The prevalence of resistance to cephalosporins among viridans-group streptococci causing

88 (18%) cases among 485 bacteraemias in neutropenic cancer patients was studied. Rates of
resistance to ceftriaxone, ceftazidime, cefpirome and cefepime were 22, 53, 14 and 34%, respec-
tively. Previous administration of β-lactam therapy was the only factor significantly associated
with bacteraemia due to cephalosporin-resistant strains; only 11 (16%) of 68 patients infected
with cephalosporin-susceptible bacteria had received these antibiotics compared with 10 (50%)
of 20 patients infected with cephalosporin-resistant bacteria (P  0.0052).

Introduction an outpatient basis, because of its favourable pharmaco-

In recent years, viridans-group streptococci (VGS) have
become a major cause of bacteraemia in neutropenic can-
cer patients. There has also been an increase in the severity
of clinical manifestations and mortality caused by these
organisms.1 Fulminant infections, adult respiratory distress
syndrome (ARDS) and -streptococcal shock syndrome
are being observed in many centres. The emergence of
strains resistant to multiple antibiotics in this group of
organisms, which have long been considered very suscept-
ible to penicillin (MIC  0.12 mg/L), constitutes another
cause of great concern.2,3
Combination of an aminoglycoside with an antipseudo-
monal β-lactam drug has been one of the most frequently
used empirical treatments for febrile episodes in neutro-
penic patients with cancer. In many regimens the β-lactam
compound is ceftazidime. Recently, some institutions have
reported a decline in the frequency of infections caused by
Pseudomonas aeruginosa, especially among patients with
solid tumours and lymphomas. This fact, along with the
increase in Gram-positive bacteraemia, has prompted
evaluation in empirical regimens of third-generation
cephalosporins, with higher activity against Gram-positive
organisms (e.g. ceftriaxone) instead of the commonly used
ceftazidime.4 Recently, ceftriaxone has also been increas-
ingly used to treat low-risk febrile patients with cancer on
kinetic characteristics.5
The aims of the present study were to determine the
prevalence of resistance to cephalosporins among VGS
causing bacteraemia in neutropenic patients and to identify
risk factors associated with the development of bacter-
aemia due to these resistant strains.
Materials and methods
The study was carried out in a 1000 bed university hospital
for adults, where prospective surveillance of all cases of
bacteraemia in neutropenic (500 granulocytes/mm3)
patients with cancer is performed regularly. For the pur-
poses of the present study, we examined all episodes of viri-
dans streptococcal bacteraemia in neutropenic patients
with cancer documented from January 1986 to December
1996. The diagnosis of viridans streptococcal bacteraemia
was established by the presence of two or more sets of
blood cultures positive for VGS. Cases in which a single
blood culture yielded these bacteria were included only
when signs of septicaemia were present.
To identify factors associated with cephalosporin-
resistant viridans streptococcal bacteraemia, a comparison
between cases due to cephalosporin-susceptible bacteria

*Correspondence address. Feixa Llarga s/n, 08907 l’Hospitalet, Barcelona, Spain. Tel: 34-93-3357011 (ext. 2487);
Fax: 34-93-2607637; E-mail: jcarratala@csub.scs.es

87
© 2001 The British Society for Antimicrobial Chemotherapy
 A. Marron et al.
and those due to cephalosporin-resistant strains was car-
ried out. For this purpose, susceptibility to cephalosporins
was considered to be the same as that to ceftriaxone.
During the study period, norfloxacin was used as pro-
phylaxis. Ceftazidime or imipenem plus amikacin were the
empirical antibiotic regimens most commonly used for
febrile episodes. Vancomycin was added to the initial regi-
men of patients in whom infection due to Gram-positive
bacteria was initially suspected and also to those who had
not improved after initial therapy for 48 h or who worsened
before that time.
Severe mucositis was defined as the presence of multiple
ulcerations covering 25% of the oral mucosa. Serious
complications included the following: septic shock (defined
as a systolic blood pressure of 90 mmHg and evidence of
peripheral hypoperfusion) and ARDS (adult respiratory
distress syndrome; defined as respiratory failure with bilat-
eral pulmonary infiltrates and neither evidence of cardiac
failure nor isolation of organisms in respiratory specimens).
In the study of risk factors, prophylactic norfloxacin and
prior β-lactam therapy were considered to be present when
they were given during the 2 weeks before viridans strepto-
coccal bacteraemia developed. Attributable mortality was
considered when it occurred within the first 7 days from
the bacteraemia and after excluding other causes of death.
Overall mortality was defined as any death occurring
within 30 days from the bacteraemia.
VGS recovered from blood cultures were classified
according to the taxonomy and nomenclature proposed by
Bruckner & Colonna6 which includes five species/groups,
namely the Streptococcus mitis, Streptococcus sanguis,
Streptococcus milleri, Streptococcus salivarius and Strepto-
coccus mutans groups. They were identified by standard
methods, including colony morphology and the production
of acid from trehalose, sorbitol, lactose, mannitol, sucrose,
inulin, raffinose, glycerol, arabinose, maltose and sorbose.
The isolates were additionally tested for reactions on aes-
culin agar and bile–aesculin agar, growth in 6.5% NaCl
broth, ammonia production from arginine, pyruvate util-
ization, sodium hippurate hydrolysis and hydrolysis of
starch. The antimicrobial agents tested were penicillin,
ceftriaxone, ceftazidime, cefepime, cefpirome, imipenem,
erythromycin, ciprofloxacin and vancomycin. The MICs
were determined by the microdilution method.7 Strains
were classified for penicillin susceptibility according to the
NCCLS criteria,8 as follows: susceptible (MIC  0.12 mg/L),
intermediately resistant (MIC 0.25–2 mg/L) and highly
resistant (MIC  4 mg/L). VGS were considered to be
resistant to cephalosporins when the MIC was 2 mg/L.8
Streptococcus pneumoniae ATCC 49619 and Staphylo-
coccus aureus ATCC 29213 were used for the quality con-
trol.
We used 2 or Fisher’s exact test when appropriate for
analysis of categorical variables and Student’s t-test for
analysis of continuous variables.
88
Results
Over the study period, 485 episodes of bacteraemia were
documented among neutropenic cancer patients. Eighty-
eight (18%) of these episodes were caused by VGS. Strains
were identified as Streptococcus mitis group (n 72),
Streptococcus salivarius group (n 7), Streptococcus
sanguis group (n 5), Streptococcus milleri group (n 3)
and Streptococcus mutans group (n 2). One episode was
caused by S. mitis and S. salivarius groups concomitantly.
Table I shows the in vitro activities of nine antimicrobial
agents against the 89 VGS isolated from blood cultures.
Thirty-five (39%) isolates showed some degree of resist-
ance to penicillin with an MIC range of 0.25–16 mg/L, with
20 isolates (57%) being highly resistant. Rates of resistance
to erythromycin, imipenem and ciprofloxacin were 36, 7
and 38%, respectively (MIC ranges 1–32 mg/L, 2 mg/L
and 2–32 mg/L, respectively). All isolates were susceptible
to vancomycin. Twenty isolates (22%) were resistant to
ceftriaxone (MIC range 2–8 mg/L) and 47 (53%) to cef-
tazidime (MIC range 2–256 mg/L). Susceptibility to
fourth-generation cephalosporins was tested in 35 of 89
isolates. Cefpirome was more active than cefepime; five
isolates (14%) strains were resistant to cefpirome and 12
(34%) to cefepime (MIC range 2–4 mg/L).
A comparison of the clinical findings in patients with
cephalosporin-susceptible and cephalosporin-resistant viri-
dans streptococcal bacteraemia is shown in Table II. Prior
administration of β-lactam antibiotics was the only factor
significantly associated with cephalosporin-resistant cases;
11 (16%) of 68 patients with cephalosporin-susceptible
bacteraemia had received β-lactam antibiotics as compared
with 10 (50%) of 20 patients with cephalosporin-resistant
bacteraemia (P 0.0052). Ceftazidime was the prior β-
lactam most frequently given to patients with cephalo-
sporin-resistant bacteraemia (eight of 10 patients). In fact,
four patients developed breakthrough bacteraemia due to
resistant strains while receiving ceftazidime. The other
β-lactams given to patients with bacteraemia due to
cephalosporin-resistant isolates were imipenem in two
cases (one patient received both imipenem and ceftazi-
dime) and amoxycillin–clavulanate in one case. In patients
with cephalosporin-susceptible bacteraemia, the β-lactams
used were ceftazidime in nine cases and imipenem in two
cases.
Ten of the 88 patients (11%) developed serious compli-
cations: ARDS plus septic shock in five patients, ARDS in
three patients and septic shock in two patients. The VGS
isolated from five of these 10 patients were resistant to
ceftazidime (MIC range 2–32 mg/L), with three of them
also being resistant to ceftriaxone (MICs 2–8 mg/L) and
cefepime (MIC range 2–4 mg/L). Attributable mortality
was 9% (eight of 88 patients) and overall mortality was
22% (19 of 88 patients). The comparative study did not
show significant differences between cephalosporin-sus-
ceptible and cephalosporin-resistant bacteraemic episodes
 Cephalosporin-resistant viridans streptococci

regarding the development of serious complications or in

256

_
_
100
_
_
_
100
_
_
mortality rates.

98.1

86.1
Table I. In vitro activities of nine antimicrobial agents against 89 strains of VGS isolated from blood of neutropenic patients with cancer

128

_
_

_
_
_

_
-
Discussion

98.1 Over recent years, penicillin-resistant VGS have been

86.1
increasingly isolated worldwide. In South Africa, 38% of
64

_
_

_
_
_

_
_
strains isolated from blood cultures between 1988 and 1991
were found to be resistant to penicillin.2 A recent study
conducted in the USA, which included mostly non-
84.9

86.1
neutropenic patients, showed that 56% of VGS recovered
_
_

_
_
_

100
_
32

from blood cultures were resistant to penicillin.3 In addi-

Cumulative percentages of isolates for which the MICs (mg/L) were

tion, some studies dealing with neutropenic patients with

cancer9 have also shown an increasing incidence of viridans
97.7
75.5

86.1
100
_

_
_
_

_
16

streptococcal bacteraemia caused by penicillin-resistant

strains. Nevertheless, experiences regarding the activities
of cephalosporins against VGS isolated from the blood of
96.6
98.8

60.4

82.3

neutropenic cancer patients are limited.10 In the present

100

_
_
_

_
8

study, a substantial proportion of isolates were resistant to

ceftriaxone. Importantly, ceftazidime was found to be the
95.4
95.2
92.4
58.5

81.0

least active of the cephalosporins tested. Regarding the

100
100
_

_
4

activities of the fourth-generation cephalosporins, we

found that the rate of resistance to cefepime was higher
than that to cefpirome.
84.7
78.3
87.3
45.3
94.3
80.6

73.4

According to our data, prior β-lactam therapy appears to

100

_
2

be a major predisposing factor for the occurrence of bac-

teraemia caused by resistant strains. It should be noted,
62.3
71.1
73.4
32.1
80.0
58.1
93.7
67.1

however, that these resistant strains are particularly preva-

100
1

lent in countries with a high incidence of penicillin-resistant

pneumococci, such as Spain and South Africa. Interest-
ingly, several in vitro studies have demonstrated the poten-
49.7
68.7
68.4
15.1
57.1
48.4
79.7
63.3

77.6
0.5

tial for transfer of resistance determinants to penicillin and

cephalosporins between these two streptococci species.
Thus, the selection pressure of exposure to β-lactams can
Activities of cefpirome and ceftazidime were tested in 35 of the 89 strains.
0.25

10.3
65.1
60.8
9.4
48.6
45.2
73.4
59.5

7.9

promote the enrichment and spread of resistant VGS as

well as resistant pneumococci.
In our experience, the outcome in patients with cephalo-
sporin-resistant viridans streptococcal bacteraemia was not
0.12

4.6
60.2
50.6
2.0
48.6
32.3
65.8
46.8

3.9

worse than that in patients with cephalosporin-susceptible

viridans streptococcal bacteraemia. However, it should be
pointed out that our study was not designed to address this
0.06

2.3
39.8
39.2

34.3
25.8
57.0
41.8

issue. Further studies specifically analysing the clinical

_

significance of resistance to cephalosporins are needed,

both for bacteraemia and for other serious infections
caused by VGS that are occasionally treated with cephalo-
0.03

24.1
29.1

25.7
12.9
45.6
30.4

sporins, such as endocarditis or, less frequently, meningitis.

_
_

A recent study11 of the outcome of bacteraemia in

neutropenic cancer patients found that six of nine patients
with viridans streptococcal bacteraemia who died had been
empirically treated with ceftazidime. In contrast, all patients
Erythromycin
Ciprofloxacin

who survived had received vancomycin in the initial anti-

Vancomycin
Ceftazidime
Ceftriaxone

Cefpiromea
Cefepimea
Antibiotic

Imipenem

biotic regimen. Thus, in light of our findings of a high per-

Penicillin

centage of VGS showing increased MICs to ceftazidime,

we believe that, at least at institutions where penicillin-
resistant VGS are prevalent, ceftazidime should not be
a

89
 A. Marron et al.

Table II. Comparison between cephalosporin-susceptible and cephalosporin-resistant VGS

bacteraemia in 88 neutropenic patients with cancera

Cephalosporin-susceptible Cephalosporin-resistant
Variable bacteraemia (n 68) bacteraemia (n 20) P

Age, years: mean (range) 43.3 (16–80) 36.9 (16–63) NS

Gender
male 38 (56) 11(55) NS
female 30 (44) 9 (45)
Underlying disease
acute leukaemia 33 (49) 12 (60) NS
other malignancies 35 (51) 8 (40)
Bone marrow transplantation
yes 24 (35) 7 (35) NS
no 44 (65) 13 (65)
Severe neutropeniab
yes 64 (94) 18 (90) NS
no 4 (6) 2 (10)
Severe mucositis
yes 30 (44) 8 (40) NS
no 38 (56) 12 (60)
Prophylaxis with quinolones
yes 56 (82) 13 (65) NS
no 12 (18) 7 (35)
Prior β-lactam therapy
yes 11(16) 10 (50) 0.0052
no 57 (84) 10 (50)
Species of VGSc
S. mitis group 54 (78) 19 (95) NS
other 15 (22) 1 (5)
Serious complications
yes 8 (12) 2 (10) NS
no 60 (88) 18 (90)
Overall mortality 15 (22) 4 (20) NS
Attributable mortality 6 (9) 2 (10) NS

Data are presented as number (%) of patients. NS, not significant.

a
Cephalosporin susceptibility was determined on the basis of ceftriaxone MICs.
b
Fewer than 100 granulocytes/mm3.
c
One episode was caused by S. mitis group and S. salivarius group concomitantly.

administered empirically to neutropenic patients suspected
of having viridans streptococcal bacteraemia. In recent
years, some authors have advocated the use of ceftriaxone4
in order to provide a better coverage of Gram-positive
infections. This antibiotic has also been increasingly used
to facilitate outpatient therapy for low-risk patients with
cancer.5 Therefore, our results demonstrating high rates of
resistance to this agent are of great clinical relevance.
Although fourth-generation cephalosporins showed a
better activity than ceftazidime, a significant proportion of
strains of VGS were also resistant to these agents. Imi-
penem was the most active of the β-lactams, and no
strain was found to be resistant to vancomycin. These
findings should be considered when choosing empirical
antibiotic therapy for febrile episodes in this population of
patients.
Acknowledgement
This work was supported by a grant from Fondo de Investi-
gaciones Sanitarias (93/1081, 93/1196) and Fundació
August Pi i Sunyer.

90
 Cephalosporin-resistant viridans streptococci
References
1. Elting, L. S., Bodey, G. P. & Keefe, B. H. (1992). Septicemia and
shock syndrome due to viridans streptococci: a case–control study
of predisposing factors. Clinical Infectious Diseases 14, 1201–7.
2. Potgieter, E., Carmichael, M., Koornhof, H. J. & Chalkley, L. J.
(1992). In vitro antimicrobial susceptibility of viridans streptococci
isolated from blood cultures. European Journal of Clinical Micro-
biology and Infectious Diseases 11, 543–6.
3. Doern, G. V., Ferraro, M. J., Brueggemann, A. B. & Ruoff, K. L.
(1996). Emergence of high rates of antimicrobial resistance among
viridans group streptococci in the United States. Antimicrobial
Agents and Chemotherapy 40, 891–4.
4. Anonymous. (1993). Efficacy and toxicity of single daily doses of
amikacin and ceftriaxone versus multiple daily doses of amikacin
and ceftazidime for infection in patients with cancer and granulo-
cytopenia. The International Antimicrobial Therapy Cooperative
Group of the European Organization for Research and Treatment of
Cancer. Annals of Internal Medicine 119, 584–93.
5. Karthaus, M., Wolf, H. H., Kampfe, D., Egerer, G., Ritter, J.,
Peters, G. et al. (1998). Ceftriaxone monotherapy in the treatment of
low-risk febrile neutropenia. Chemotherapy 44, 343–54.
6. Bruckner, D. A. & Colonna, P. (1997). Nomenclature for aerobic
and facultative bacteria. Clinical Infectious Diseases 25, 1–10.
91
7. National Committee for Clinical Laboratory Standards. (1993).
Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria
that Grow Aerobically—Third Edition: Approved Standard M7-A3.
NCCLS, Villanova, PA.
8. National Committee for Clinical Laboratory Standards. (1997).
Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria
that Grow Aerobically—Fourth Edition: Approved Standard M100-
S7/M7-A4. NCCLS, Villanova, PA.
9. Carratalà, J., Alcaide, F., Fernández-Sevilla, A., Corbella, X.,
Liñares, J. & Gudiol, F. (1995). Bacteremia due to viridans strepto-
cocci that are highly resistant to penicillin: increase among neutro-
penic patients with cancer. Clinical Infectious Diseases 20,
1169–73.
10. Pfaller, M. A., Marshall, S. A. & Jones, R. N. (1997). In vitro
activity of cefepime and ceftazidime against 197 nosocomial blood-
stream isolates of streptococci: a multicenter sample. Diagnostic
Microbiology and Infectious Disease 29, 273–6.
11. Elting, L. S., Rubenstein, E. B., Rolston, K. V. & Bodey, G. P.
(1997). Outcomes of bacteremia in patients with cancer and neutro-
penia: observations from two decades of epidemiological and
clinical trials. Clinical Infectious Diseases 25, 247–59.
Received 27 April 2000; returned 24 July 2000; revised 30 August
2000; accepted 22 September 2000