MAJOR ARTICLE

Antimicrobial Resistance Among Uropathogens

that Cause Community-Acquired Urinary Tract
Infections in Women: A Nationwide Analysis
Kalpana Gupta1, Daniel F. Sahm2, David Mayfield2, and Walter E. Stamm1
1

Department of Medicine/Division of Allergy and Infectious Diseases, University of Washington School of Medicine, Seattle;
and 2MRL Pharmaceutical Services, Herndon, Virginia

The Infectious Diseases Society of America (IDSA) recently published guidelines for the treatment of women
with community-acquired urinary tract infections
(UTIs) that are based in part on the rate of resistance
to trimethoprim-sulfamethoxazole (TMP-SMZ) in the
geographic region of the practitioner [1]. However, as

Received 24 July 2000; revised 28 November 2000; electronically published 5

June 2001.
Presented in part: 5th Annual International Infectious Diseases Society of
Obstetrics and Gynecology USA Meeting, San Francisco, May 2000.
Financial support: National Institutes of Health, Bethesda, Maryland (DK 53369),
and Procter & Gamble, Cincinnati.
K. G. has served as a consultant for and received research support and/or speaking
honoraria from Procter & Gamble and Bayer Pharmaceutical. W E. S. has served as
a consultant for and received research support from Procter & Gamble, Ortho McNeil,
and Bayer Pharmaceutical.
Reprints or correspondence: Dr. Kalpana Gupta, Division of Allergy and Infectious
Diseases, 1959 NE Pacific St., University of Washington, BB1225, Box 356523,
Seattle, WA 98195 (kalg@u.washington.edu).
Clinical Infectious Diseases 2001; 33:8994
 2001 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2001/3301-0014$03.00

the authors of the guidelines acknowledge, such data

are not always readily available in many areas. Although
we and others have recently reported increasing antimicrobial resistance rates among uropathogens that
cause acute cystitis in women [24], these studies have
been conducted within a single institution or a defined
geographic area. Furthermore, many studies do not distinguish between urinary isolates that have been recovered from adult female outpatients (most of whom
have uncomplicated UTIs) and those that have been
recovered from men, children, or inpatients (most of
whom have complicated UTIs). To address these issues,
we used a national laboratory database to assess the
rates of resistance among uropathogens that were recovered from female outpatients both nationally and
within 9 geographic regions in the United States.

METHODS
Data collection.

The data presented here were ob-

Resistance in Community-Acquired UTI

CID 2001:33 (1 July) 89

Downloaded from http://cid.oxfordjournals.org/ by guest on May 11, 2014

Current recommendations for empirical therapy for community-acquired urinary tract infection (UTI) in
women hinge on knowledge of antimicrobial susceptibility patterns in the geographic region of the practitioner.
We conducted a survey of antimicrobial susceptibilities of 103,223 isolates recovered from urine samples that
were obtained in 1998 from female outpatients nationally and within 9 geographic regions in the United
States. Resistance of Escherichia coli isolates to trimethoprim-sulfamethoxazole varied significantly according
to geographic region, ranging from a high of 22% in the western United States to a low of 10% in the Northeast
(P ! .001). There were no clinically significant age-related differences in the susceptibility of E. coli to any of
the study drugs, but the susceptibility to fluoroquinolones of nonE. coli isolates that were recovered from
women who were aged 150 years was significantly lower than that of isolates recovered from younger women
(P ! .001). The in vitro susceptibility of uropathogens in female outpatients varies according to age and
geographic region.

Table 1. Distribution of uropathogens that cause urinary tract infections

in women.
Frequency among women,
a
by age group, %
Uropathogen
Escherichia coli

1550 years
of age

150 years

of age

72

53

Klebsiella species

12

Proteus species

Enterobacter species

Pseudomonas aeruginosa

Other gram-negative rod

Enterococcus species

12

Staphylococcus aureus

Staphylococcus saprophyticus

0.2

Coagulase-negative staphylococci

Other

A total of 46,768 isolates were recovered from women aged 1550 years and 56,455
isolates were recovered from women aged 150 years.

tained from The Surveillance Network (TSN) DatabaseUSA,

a national surveillance system that collects antimicrobial susceptibility results from clinical microbiology laboratories in 9
geographic regions of the United States. The methods used for
data collection, verification, and quality control for this database have been described in detail elsewhere [5]. In brief, all
laboratories must pass an inspection prior to enrollment. Data
from enrolled laboratories are sent electronically to the central
database. Unusual antimicrobial susceptibility profiles are verified by repeated testing, and duplicate data received within 5
days with regard to the same specimen source in a given patient
are excluded.
In order to limit our analyses to episodes of communityacquired UTI in women, including those UTIs that are most
amenable to empirical therapy, the database was searched for
isolates recovered from urine samples that were obtained from
women aged 1550 years and 150 years in an outpatient setting
in 1998. Data from patients in nursing facilities were not included. Organisms that accounted for 2% of the total isolates
within each age category were included in the study. Susceptibility patterns of other infrequently isolated organisms are not
reported.
Laboratory methods. Antimicrobial susceptibility test results collected through TSN were generated by each of the TSN
participant laboratories. All TSN laboratories use standard
methods that comply with those published by the National
Committee for Clinical Laboratory Standards [6] or with commercial methods approved by the US Food and Drug Admin-

90 CID 2001:33 (1 July) Gupta et al.

istration. For this study, results regarding resistant isolates and

intermediate susceptible isolates were combined.
Statistical methods. The proportion of resistant organisms
was calculated by dividing the number of isolates that were resistant to each antimicrobial agent by the number of organisms
that were tested against that antimicrobial agent. The main outcome of interest was the proportion of organisms that were
resistant to each antimicrobial agent nationally and within the 9
geographic regions for patients in each age group. The significance of differences in resistance according to geographic region
or age was determined by means of the x2 test for independence
[7]. Because of the extremely large number of isolates in the
database, most differences in the prevalence of resistance that we
tested were significant (P ! .05 ), even if they varied by only 2%.
Therefore, to make our results more meaningful for clinicians,
we defined a variance of 10% as clinically significant and set
our level of statistical significance at P ! .01.
RESULTS
A total of 103,223 isolates were recovered from urine samples
obtained from female outpatients in 1998. The distribution of
the pathogens according to age group is shown in table 1.
Escherichia coli accounted for 33,462 (72%) of 46,768 isolates
recovered from women who were 1550 years of age and 29,734
(53%) of 56,455 isolates recovered from women who were 150
years of age (P ! .001). Staphylococcus saprophyticus was also
present in significantly higher proportions in the younger

Downloaded from http://cid.oxfordjournals.org/ by guest on May 11, 2014

NOTE. There are significant differences in prevalence between age groups (P !

.0001) for all organisms with the exception of S. aureus, Enterobacter species, and
coagulase-negative staphylococci.

women (P ! .001). Klebsiella pneumoniae, Enterococcus species,

and Proteus mirabilis were the most common nonE. coli isolates in both age groups, but they were present in significantly
higher proportions in women who were 150 years of age
(P ! .001 for each organism). Pseudomonas species, other gramnegative rods, and other organisms were also more frequently
isolated from women who were 150 years of age (P ! .001 for
each organism). S. aureus, Enterobacter species, and coagulasenegative staphylococci accounted for 2%3% of isolates recovered from women in each of the age groups.
The overall national susceptibilities of E. coli, nonE. coli,
and all isolates combined, stratified by age group, are shown
in table 2. Almost all E. coli isolates that were recovered from
female outpatients were susceptible to nitrofurantoin and to
the fluoroquinolones that were tested, regardless of age group.
In contrast, 33%40% of E. coli isolates were resistant to ampicillin and 16%18% were resistant to TMP-SMZ in each age
group. There were no clinically significant age-related differences in the susceptibility of E. coli to any of the drugs that
were studied.
The nonE. coli isolates demonstrated higher resistance rates
to each of the drugs that were studied with the exception of
TMP-SMZ, which demonstrated in vitro activity against

82%84% of E. coli isolates, versus 89%94% of nonE. coli

isolates. There was also a significant reduction in the percentage
of nonE. coli isolates that were susceptible to the fluoroquinolones (range, 12%16% reduction) in women who were 150
years of age, compared with data for the younger age group.
When all isolates were considered together, susceptibility to the
fluoroquinolones and nitrofurantoin was 190% in women who
were 1550 years of age but was somewhat lower (range,
8%9% reduction) in women who were 150 years of age. Rates
of susceptibility to TMP-SMZ and to ampicillin were similar
between age groups for all isolates combined.
The overall prevalence of antimicrobial resistance among the
nonE. coli species by age group is shown in table 3. Nitrofurantoin demonstrated the highest and most consistent activity
against the gram-positive isolates, whereas the fluoroquinolones
demonstrated the best activity against the nonE. coli, gramnegative isolates. There was a clinically significant, age-related
variation in susceptibility of S. aureus and enterococci to the
fluoroquinolones but for none of the other drugs that were
studied.
The variation in rates of resistance of E. coli to ampicillin
and TMP-SMZ, according to geographic region and age, is
shown in figure 1. There was not any clinically significant age-

Resistance in Community-Acquired UTI CID 2001:33 (1 July) 91

Downloaded from http://cid.oxfordjournals.org/ by guest on May 11, 2014

Figure 1. Map showing that there is clinically significant (10%) and statistically significant (P ! .001 ) variation in the percentage of Escherichia
coli isolates that were susceptible to trimethoprim-sulfamethoxazole (TMP-SMZ; white boxes) and ampicillin (dark boxes), in relation to geographic
region and age group (1550 years old vs. 150 years old), among strains that cause urinary tract infection in women who were treated on an outpatient
basis.

related variation in resistance within each region for either of

the antimicrobial agents. However, there was significant variation in susceptibility to both TMP-SMZ and ampicillin in
relation to geographic region within each age group (P ! .001).
Susceptibility of E. coli isolates to TMP-SMZ was lowest in the
western and southern regions of the United States, and it tended
to increase in a northeasterly direction. This pattern was not
observed with regard to susceptibility to ampicillin, although
the lowest rates of susceptibility were observed in the southern
regions. Overall, the 2 northeastern regions had the highest
rates of susceptibility to both TMP-SMZ and ampicillin in comparison with the other regions. Susceptibility of E. coli to nitrofurantoin and to the fluoroquinolones did not vary according to geographic region (data not shown).
DISCUSSION

Table 2. National rates of antimicrobial susceptibility for isolates that cause urinary tract
infection in female outpatients in 1998.
Percentage of isolates susceptible, by patient age group

Antimicrobial
agent

E. coli
(n p 63,196)

NonE. coli
(n p 31,602)

All isolates
(n p 94,798)

1550 years

150 years

1550 years

150 years

1550 years

150 years

Ampicillin

60

67

51

51

58

61

Nitrofurantoina

99

98

60

61

91

84

Ciprofloxacin

99

98

90

78

98

90

Levofloxacin

99

97

95

79

98

89

a,b

82

84

94

89

84

86

TMP-SMZ

NOTE. The susceptibilities of coagulase-negative staphylococci that were not specifically identified as Staphylococcus saprophyticus and other rare organisms are not included. The reduction in susceptibility to nitrofurantoin
and the fluoroquinolones among all isolates recovered from women who were 150 years of age was statistically
significant (P ! .001) but not clinically significant (see the Statistical Methods section). E. coli, Escherichia coli; TMPSMZ, trimethoprim-sulfamethoxazole.
a
b

Pseudomonas species were not tested.

Enterococcus species were not tested.

92 CID 2001:33 (1 July) Gupta et al.

Downloaded from http://cid.oxfordjournals.org/ by guest on May 11, 2014

This study provides current information regarding the etiologic

agents that cause community-acquired UTI in women in the
outpatient setting and their antimicrobial susceptibility patterns
in relation to age and geographic region on a national basis.
By using this large national database and by restricting the
analyses to female outpatients within these 2 age groups, we
have defined the population that is most amenable to empirical
therapy as recommended in the recently published IDSA guidelines. The safety and efficacy of such empirical therapy depends
upon periodic assessment of antimicrobial resistance profiles.
The frequency distribution of etiologic organisms within
each age group is quite consistent with the findings of previous,
smaller studies of acute cystitis in women that were conducted
in individual centers [8, 9]. E. coli and S. saprophyticus were
more prevalent in younger women, whereas other gram-negative rods and enterococci were more prevalent in older women.

Therefore, the agents that cause acute cystitis remain highly

predictable, although they vary by age, even as the susceptibility
patterns of these organisms change.
These data illustrate several important points regarding the
drugs that are most frequently considered for empirical therapy
for community-acquired UTI in women. First, rates of resistance to ampicillin range from 30% to 40% among E. coli and
nonE. coli isolates nationwide. This resistance does not vary
greatly by age group, either nationally or within each of the
geographic regions that we studied. Although there is significant
geographic variability in resistance to ampicillin, it is unacceptably high (25%) throughout the United States. Therefore,
as has been recommended by the IDSA and others [8], ampicillin should be avoided in first-line therapy for patients with
community-acquired UTIs.
Resistance to TMP-SMZ, the drug that was recommended
for use as first-line therapy in the IDSA treatment guidelines,
is 15% nationally in women in both age groups. However,
rates of resistance of E. coli to this drug vary significantly according to geographic region, ranging from a high of 22% in
the western United States to a low of 10% in the Northeast.
Within regions, there is not much variation in rates of resistance
to TMP-SMZ according to age group. Therefore, TMP-SMZ
remains a reasonable first-line choice for empirical therapy in
the northern and eastern states, although it may no longer be
appropriate in western and southwestern states. Of note, a similar geographic distribution in TMP-SMZ resistance was recently reported by Talan et al. [10] among E. coli strains that
cause acute pyelonephritis in a nationwide study.
How much variation there may be from locale to locale
within each region cannot be readily ascertained from the database. The cause of this geographic variation in resistance to
TMP-SMZ is also not apparent. One possibility could be re-

Table 3. National rates of antimicrobial susceptibility for nonEscherichia coli isolates that cause urinary tract infection recovered
from female outpatients, according to age group.
Percentages of isolates susceptible
Antimicrobial
agent

Staphylococcus Staphylococcus Enterococcus

Enterobacter
Klebsiella
Proteus
Pseudomonas
saprophyticus
aureus
species
species
species
species
aeruginosa
(n p 763/92) (n p 753/1009) (n p 2563/6933) (n p 1031/1376) (n p 2885/6643) (n p 1756/3333) (n p 457/2008)

Ampicillin

29/27

16/7

98/91

3/5

1/2

92/86

NT

Nitrofurantoin

99/94

99/98

98/94

51/55

57/58

1/2

NT

Ciprofloxacin

99/97

90/38

67/44

96/92

99/97

99/92

74/71

Levofloxacin

100/100

95/50

83/56

99/96

99/98

98/84

72/71

93/95

97/91

95/87

92/92

94/85

NT

TMP-SMZ

NOTE. Data are % of isolates recovered from women who were 1550 years of age/% recovered from women who were 150 years of age. NT, not tested;
TMP-SMZ, trimethoprim-sulfamethoxazole.

rently recommended [1]. Further studies of the efficacy of 3-day

regimens of this drug would be of interest.
The extremely large database and the ability to define urinary
isolates with regard to patient age, sex, region, and community
versus nosocomial acquisition are major strengths of this study.
The recently published IDSA guidelines make these data particularly relevant, because the guidelines provide practitioners
with an estimate of susceptibility rates within well-defined geographic regions and patient groups. Local resistance patterns
are clearly best when they are available, but they are increasingly
not available because of the empirical treatment (without culture) of UTI in women.
The data do have certain limitations that are inherent to
laboratory-based surveys. For example, because cultures are not
performed for all empirically treated patients, unrecognized
biases may cause physicians to obtain specimens for culture
from some patients but not from others, thus influencing the
susceptibility data that they obtain. In addition, although we
have reported in vitro susceptibility data from those patients
who are most likely to represent the population amenable to
empirical therapy for cystitis (namely, adult women seen in the
outpatient setting), we do not have specific data regarding the
presence and nature of symptoms and signs or of complicating
factors such as diabetes, recent hospitalization, catheterization,
and urological or gynecologic surgery.
An unknown proportion of the women in the study undoubtedly had asymptomatic bacteriuria, complicated UTI, or
pyelonephritis that was evaluated in the outpatient setting. However, the fact that the distribution of uropathogens and their rates
of susceptibility are consistent with those from previous studies
of women known to have acute uncomplicated cystitis suggests
that the number of such cases is likely small. For example, the
overall national rates of susceptibility for all isolates that were
recovered from women who were 1550 years of age in this study
are very similar to the rates of susceptibility that we recently
reported following a population-based study in Seattle in which
we were able to specifically identify women with symptomatic
Resistance in Community-Acquired UTI CID 2001:33 (1 July) 93

Downloaded from http://cid.oxfordjournals.org/ by guest on May 11, 2014

gional differences in the use of TMP-SMZ as prophylaxis for

patients with HIV infection or AIDS or for other uses. An
alternative possibility could be introduction of TMP-SMZ
resistant strains into the western and southwestern United
States from sources in Mexico or Asia, where markedly increased rates of resistance to TMP-SMZ are seen.
Alternative drugs that have been recommended by the IDSA
for empirical treatment of patients with acute cystitis include
fluoroquinolones and nitrofurantoin [1]. In general, resistance
to the fluoroquinolones was minimal among women aged
1550 years and did not vary by geographic region. Susceptibility to this class of drug was somewhat reduced among older
women, mainly because of the resistance and higher prevalence
of enterococci among women in this age group. Overall, the
activity of fluoroquinolones was better than that of TMP-SMZ
against all isolates combined in each age group; therefore, fluoroquinolones would be reasonable choices for empirical therapy
in cases in which TMP-SMZ could not be used.
Nitrofurantoin also demonstrated activity against the majority of isolates that cause UTIs that were recovered from
women aged 1550 years. Older women were less likely to be
infected with a strain susceptible to nitrofurantoin, mainly owing to the decreased activity of this drug against Proteus species
and Klebsiella species, both of which are relatively more prevalent among women in this age group than they are among
younger women. In general, the rates of susceptibility of isolates
that were recovered from younger women to nitrofurantoin
were better than the rates of susceptibility to TMP-SMZ; however, the isolates recovered from older women had equivalent
rates of susceptibility to the 2 drugs.
In contrast to TMP-SMZ, however, susceptibility to nitrofurantoin did not differ by geographic region; therefore, this
drug would be a reasonable alternative choice for therapy in
areas where rates of resistance to TMP-SMZ are high. However,
studies to date have demonstrated that 3-day regimens of nitrofurantoin have lower cure rates than do regimens of TMPSMZ or fluoroquinolones [1, 11], so a 7-day regimen is cur-

94 CID 2001:33 (1 July) Gupta et al.

Acknowledgment

The authors thank Thomas M. Hooton, MD, for helpful

review of and comments on the manuscript.

References
1. Warren JW, Abrutyn E, Hebel R, Johnson JR, et al. Guidelines for
antimicrobial treatment of uncomplicated acute bacterial cystitis and
acute pyelonephritis in women. Clin Infect Dis 1999; 29:74558.
2. Gupta K, Scholes D, Stamm WE. Increasing prevalence of antimicrobial
resistance among uropathogens causing acute uncomplicated cystitis
in women. JAMA 1999; 281:7368.
3. Dyer IE, Sankary TM, Dawson JA. Antibiotic resistance in bacterial
urinary tract infections, 1991 to 1997. West J Med 1998; 169:2658.
4. Weber G, Riesenberg K, Schlaeffer F, et al. Changing trends in frequency
and antimicrobial resistance of urinary pathogens in outpatient clinics
and a hospital in southern Israel, 19911995. Eur J Clin Microbiol
Infect Dis 1997; 16:8348.
5. Sahm DF, Marsilio MK, Piazza G. Antimicrobial resistance in key
bloodstream bacterial isolates: electronic surveillance with the Surveillance Network DatabaseUSA. Clin Infect Dis 1999; 29:25963.
6. National Committee for Clinical Laboratory Standards (NCCLS). 1999
Performance standards for antimicrobial susceptibility testing. 9th informational supplement, M100-S9. Vol. 18, no. 1. Villanova, Pennsylvania: NCCLS, 1999.
7. Rosner B. Fundamentals of biostatistics. 4th ed. Belmont, California:
Wadsworth Publishing, 1995:394.
8. Hooton TM, Stamm WE. Diagnosis and treatment of acute uncomplicated urinary tract infection. Infect Dis Clin North Am 1997; 11:
55181.
9. Bacheller CD, Bernstein JM. Urinary tract infections. Med Clin North
Am 1997; 81:71930.
10. Talan DA, Stamm WE, Hooton TM, et al. Short-course ciprofloxacin
for acute uncomplicated pyelonephritis in women: a randomized, double-blind trial. JAMA 2000; 283:158390.
11. Hooton TM, Winter C, Tiu F, et al. Randomized comparative trial and
cost analysis of 3-day antimicrobial regimens for treatment of acute
cystitis in women. JAMA 1995; 273:415.
12. Echols RM, Tosiello RL, Haverstock DC, et al. Demographic, clinical
and treatment parameters influencing the outcome of acute cystitis.
Clin Infect Dis 1999; 29:1139.
13. McCarty JM, Richard G, Huck W, et al. A randomized trial of ciprofloxacin, ofloxacin, or trimethoprim/sulfamethoxazole for the treatment
of acute urinary tract infection in women. Am J Med 1999; 106:2929.

Downloaded from http://cid.oxfordjournals.org/ by guest on May 11, 2014

acute uncomplicated cystitis and to exclude those with complicated UTI or pyelonephritis [2]. Likewise, the distribution of
uropathogens is similar to that found in clinical trials of treatment
of patients with acute uncomplicated cystitis [12]. Therefore, it
is likely that the great majority of isolates included in our analyses
were indeed recovered from patients who experienced episodes
of acute uncomplicated cystitis.
Finally, it is important to note that the clinical significance
of in vitro resistance in patients with uncomplicated cystitis
has been little studied. McCarty et al. [13] found a 50% rate
of bacterial eradication and a 60% rate of clinical cure among
10 women who were infected with a TMP-SMZ-resistant uropathogen and who had been randomized to receive treatment
with TMP-SMZ. However, more studies are needed before definitive conclusions can be made regarding the implications of
in vitro resistance in patients with cystitis. Certainly, factors
such as clinical presentation, level of infection (upper vs. lower
urinary tract), and cost need to be weighed in addition to
prevalence of resistance when therapeutic agents are chosen.
In conclusion, the in vitro susceptibility of community-acquired uropathogens varies according to age and geographic
region, even among adult women who were seen in the outpatient setting. This information needs to be considered when
empirical therapy for acute cystitis is chosen. As the magnitude
and variation of antimicrobial resistance in the community
grow, so does the need for continuous large-scale surveillance
systems such as TSN, which has yielded a database that can
link epidemiological and clinical factors with laboratory results.
In addition, studies of the clinical outcomes associated with in
vitro resistance are needed. The more detailed the clinical and
epidemiological data, the more precise the analyses that result
can be. Clearly, continued surveillance is essential to maintaining the safety and cost-effectiveness of empirical therapy as
a management strategy for acute cystitis.