ORIGINAL INVESTIGATION

Outpatient Urine Culture

Does Collection Technique Matter?
Edward Lifshitz, MD; Liane Kramer, RNC, BSN

Background: Dysuria is one of the most common pre-
senting complaints of young women, and urinalysis is one
of the most common laboratory tests performed. De-
spite the fact that the midstream clean-catch technique
is commonly used for urine collection, contaminated urine
cultures occur with distressing regularity. The mid-
stream clean-catch technique is time-consuming to ex-
plain, frequently not performed correctly by patients,
costly for supplies, often embarrassing for patients and
staff, and of unproven benefit. Therefore, we designed a
study to compare various methods of obtaining speci-
mens for culture from acutely dysuric young women.
neal cleansing and spreading of the labia. In an attempt
to decrease contamination from the vagina, the third group
(n=81) was given the same instructions as group 2, with
the addition of using a vaginal tampon. Contamination
rates were calculated for all 3 groups.
Results: Contamination rates for the 3 groups were nearly
identical (29%, 32%, and 31%, respectively). Compar-
ing the no-cleansing group with the combined cleans-
ing, midstream groups also showed no difference in con-
tamination rates (28.6% and 31.5%, respectively, with
P=.65).

Methods: A total of 242 consecutive female patients who
presented with symptoms suggestive of a urinary tract
infection were randomized into 3 groups. The first group
(n=77) was instructed to urinate into a clean container.
No cleansing was done, and the specimen was not ob-
tained midstream. The second group (n = 84) was in-
structed to collect a midstream urine sample with peri-
Conclusions: In young, outpatient women with symp-
toms suggestive of a urinary tract infection, the mid-
stream clean-catch technique does not decrease contami-
nation rates.
Arch Intern Med. 2000;160:2537-2540

From Rutgers University Health
Services (Dr Lifshitz and
Ms Kramer) and the
Department of Medicine,
University of Medicine and
Dentistry of New Jersey–Robert
Wood Johnson Medical School
(Dr Lifshitz),
New Brunswick, NJ.
D
YSURIA IS one of the most
common complaints for
which young women
seek medical care, and
cystitis is one of the most
common diagnoses. In fact, in 1992, there
were an estimated 7 million episodes of
acute cystitis in the United States,1 with
the cost of treatment exceeding $1 bil-
lion.2 Among sexually active young women
at a university setting, the incidence of uri-
nary tract infections was found to be 0.7
per person-years.3
To minimize contamination from the
colon, skin, or vaginal areas, urethral cath-
eterization was routinely used prior to
1958 to collect urine specimens for cul-
ture. In that year, 2 articles arguing against
catheterization and for the midstream
clean-catch (MSCC) technique were pub-
lished.4,5 Other studies have confirmed that
the MSCC technique is equivalent to cath-
eterization in symptomatic women,6 and
for the past 40 years, this technique has
remained the de facto standard for outpa-
tient urine collection.
The MSCC technique in women in-
volves cleansing the perineum with either
saline or a bactericidal wash, spreading the
labia, discarding the first urine sample, and
collecting a midstream sample in a sterile
container. Yet the value of this involved and
difficult-to-teach technique has not been
proved. At least 3 studies in asymptom-
atic women found that cleansing did not
decrease the number of contaminated speci-
mens.7-9 Another study involving asymp-
tomatic females found no difference be-
tween catheterized specimens and speci-
mens collected in a nonsterile container
without any precautions,10 and the only
study we could find that involved symp-
tomatic women found no difference in
contamination rates between the MSCC
method and urine collection without any
precautions or instructions.11
Further evidence that the method of
urine collection may not be as important

(REPRINTED) ARCH INTERN MED/ VOL 160, SEP 11, 2000 WWW.ARCHINTERNMED.COM
2537
Downloaded from www.archinternmed.com on May 4, 2010
©2000 American Medical Association. All rights reserved.
 SUBJECTS AND METHODS
A total of 242 consecutive female patients (mean age, 21.2
years; median age, 20 years; and age range, 17-50 years) who
presented to our university clinic and had symptoms sug-
gestive of cystitis were randomized into 1 of 3 groups. Rut-
gers University is the largest state university in New Jersey,
and all full-time students are eligible to be seen at the health
centers without charge. While specific socioeconomic data
were not collected for our subjects, patients at the health
centers are representative of the general university pop-
ulation: a highly educated, economically diverse group.
While the majority of our subjects were undergraduates, a
significant minority were graduate students.
Seventy-seven patients (mean age, 21.4 years; age
range, 17-50 years) randomized to the first of our groups
(“nothing”) were told to urinate into a clean, nonsterile
container without any preparation (no cleansing, no
midstream collection). The second group of patients
(n=84; mean age, 21.3 years; age range, 17-41 years) (“mid-
stream”) were told to cleanse the perineum with a bacte-
ricidal wipe of witch hazel, benzalkonium chloride,
solubilized lanolin, and methyl p-hydroxybenzoate
(Rantex; Clinipad Corp, Guilford, Conn) by wiping from
front to rear, to spread the labia, to discard the first urine
output, and to collect the midstream specimen in a clean,
nonsterile container. The third group (“everything”) was
identical to the second, except that the patients, who ranged
in age from 17 to 37 years (mean age, 21.0 years), were in-
structed to insert a vaginal tampon prior to collection. If
the patients in this group were unable or unwilling to use
a tampon, they were automatically relegated to the mid-
stream group. There were only a few patients who needed
to be reassigned this way.
The only exclusion criterion was antibiotic usage or
urethral instrumentation in the previous 7 days, or known
urologic abnormality or nephrolithiasis. According to the
current standard of care, the patients were treated based
on their symptoms and on the results of a “dipstick” analy-
sis (including tests to detect the presence of nitrites, leu-
kocyte esterase, hemoglobin, and protein).
Within 5 minutes of collection, all specimens were in-
jected into a receptacle containing boric acid and sodium
formate (Vacutainer Urine Transport Kit; Becton Dickin-
son & Co, Franklin Lanes, NJ) and transported to the labo-
ratory. Microbiologists at the laboratory were blinded as
to grouping. The specimens were plated within 24 hours
as is usually thought was found in a study of hospital-
ized patients that showed a 92.3% correlation between
catheterized specimens and those collected in dispos-
able diapers.12 Attempts to decrease contamination rates
via an educational campaign aimed at teaching male pa-
tients how to provide a MSCC specimen backfired when
contamination rates actually increased.13
There is some literature to support performing at
least some components of the MSCC technique, al-
though in each of 3 studies it was a different component
(spreading of the labia,8 cleansing of the perineum,14 and
midstream catch in men15).
The MSCC technique is difficult and time-
consuming to teach. Explaining the concept to patients
(REPRINTED) ARCH INTERN MED/ VOL 160, SEP 11, 2000
2538
Downloaded from www.archinternmed.com on May 4, 2010
©2000 American Medical Association. All rights reserved.
using a 1-µm loop, streaked on both blood and MacCon-
key agar, and incubated for 24 hours at 37°C. At this time,
a preliminary reading was done and the specimen was re-
incubated for another 24 hours, when a final reading
was done. Final culture reports were classified as no
growth, mixed, or pure. Mixed was defined as at least 2 or-
ganisms, and in most cases, specific identification of those
organisms was not made. Those that were pure were fur-
ther categorized according to species and colony-forming
units per milliliter using a standard technique.
One major decision concerned what constituted a“con-
taminated specimen. While there are several studies that
suggest that mixed cultures in hospitalized patients may
not be contaminated,16-19 there is a paucity of data on out-
patients. For example, while Stamm et al20 showed that ap-
proximately 13% of cultures obtained from symptomatic
female outpatients by catheterization or suprapubic aspi-
ration were mixed, they also found an unusually high rate
of mixed cultures obtained by the MSSC technique (64%).
In general, the prevailing view among clinicians is that since
Escherichia coli is the causative agent in the overwhelming
majority of cystitis cases, and since without manipulation
or anatomical defect the chance of a second infection de-
creases, a mixed culture in an uncomplicated outpatient
population likely indicates contamination. Because our
population was exclusively outpatient, had not under-
gone urethral instrumentation, and had no history of re-
nal disease or nephrolithiasis, and because we did not in-
vasively obtain urine specimens for comparison, we classified
all mixed cultures as contaminated.
A decision needed to be made as to what other re-
sults should be considered contaminated. While it is not
possible to know with absolute certainty, given the above-
mentioned characteristics of our population, low levels
(,104/mL) of organisms commonly found on the skin and
external and internal genitalia were considered to be con-
taminants. This group included all our cultures that yielded
Enterococcus and S epidermis (a total of 6 specimens). Other
organisms commonly found on the skin were also deemed
to be contaminants. This group included S viridans (2 speci-
mens) and S aureus (1 specimen). In keeping with the cur-
rent literature on dysuric women,20 coliform organisms of
102 colony-forming units per milliliter or more were con-
sidered significant.
This study was presented to the institutional review
board, which concluded that informed consent was not nec-
essary, as treatment was determined by standard clinical his-
tory, physical examination, and urinary dipstick results.
often involves embarrassment for both patient and staff.
The added costs of sterile containers and bactericidal wipes
are significant. If the MSCC technique were shown not
to improve contamination rates, then there would be a
substantial benefit from abandoning the practice.
The opposite argument can also be made. The
normal vaginal flora may include Staphylococcus
aureus, Staphylococcus epidermidis, Streptococcus viri-
dans, enterococci, peptostreptococci, group B strepto-
cocci, low-temperature–tolerant neisseriae, corynebac-
teria, some actinomycetes, members of the family
Enterobacteriaceae, acinetobacters, chlamydiae, Gard-
nerella vaginalis, occasionally clostridia and other an-
aerobic rods, mycoplasmata, Candida albicans, other
WWW.ARCHINTERNMED.COM
 Table 1. Culture Results for Each Group*

No. (%) of Total

Nothing (n = 77) Midstream (n = 84) Everything (n = 81) Total (N = 242), No. (%)
Escherichia coli 37 (48) 41 (49) 41 (51) 119 (49)
Staphylococcus saprophyticus 1 (1) 0 (0) 3 (4) 4 (2)
Enterobacter 3 (4) 0 (0) 1 (1) 4 (2)
Klebsiella 1 (1) 0 (0) 1 (1) 2 (1)
Proteus 2 (3) 1 (1) 0 (0) 3 (1)
Enterococcus † 1 (1) 2 (2) 1 (1) 4 (2)
Streptococcus viridans † 0 (0) 1 (1) 1 (1) 2 (1)
Staphylococcus aureus† 0 (0) 0 (0) 1 (1) 1 (0)
Staphylococcus epidermidis † 1 (1) 1 (1) 0 (0) 2 (1)
No growth 11 (14) 15 (18) 10 (12) 36 (15)
Mixed 20 (26) 23 (27) 22 (27) 65 (27)
Total contaminated† 22 (29) 27 (32) 25 (31) 74 (31)

*See “Subjects and Methods” section for definitions of groups.

†These organisms were labeled contaminants; thus, Total contaminated = Mixed + ( Enterococcus + S viridans + S aureus + S epidermis).

yeasts, and Trichomonas vaginalis. Since vaginal secre-

tions may potentially contaminate a urine specimen, it Table 2. Culture Results for 2 Groups*
is reasonable to postulate that adding a vaginal tampon
to the MSCC technique would decrease urine sample con- No. (%)
tamination rates. Nothing Midstream + Everything Total
We devised a study to compare 3 methods of collec- Uncontaminated 55 (71.4) 113 (68.5) 168 (69.4)
tion: standard midstream collection with bactericidal wipe, Contaminated 22 (28.6) 52 (31.5) 74 (30.6)
midstream collection with bactericidal wipe and the inser- Total 77 (100) 165 (100) 242 (100)
tion of a vaginal tampon, and no precautions (no-wipe,
nonmidstream). All specimens were collected in clean, non- *See “Subjects and Methods” section for definitions of groups.
sterile containers. Unlike most previous studies,7-11 which
involved asymptomatic women, our study comprised symp-
Table 3. Colony Counts Greater Than 105 Colony-Forming
tomatic, premenopausal women, a population for which
Units per Milliliter for Uncontaminated Specimens
urine cultures are frequently ordered in clinical practice.
No. (%)
RESULTS
No Cleansing Bactericidal Wipe
Table 1 contains a summary of our results. As ex- Escherichia coli 27 (73) 64 (78)
pected in young, outpatient women, E coli was over- Staphylococcus 1 (100) 3 (100)
whelmingly the most common pathogen, accounting for saprophyticus
Enterobacter 3 (100) 1 (100)
90% of documented infections. The other 10% were al-
Klebsiella 1 (100) 1 (100)
most evenly divided among Staphylococcus saprophyti- Proteus 2 (100) 2 (100)
cus, Klebsiella, Proteus, and Enterobacter. Total 34 (77.3) 71 (79.8)
The contamination rates in each of our 3 groups
(29%, 32%, and 31%, respectively) were essentially iden-
tical. The null hypothesis of equality between our 3 groups 105 colony-forming units per milliliter were seen in 77.3%
could not be rejected, as a x2 analysis of the last 3 rows of the pure specimens obtained without cleansing and
of Table 1 revealed a P value of .82. Also, if we state that in 79.8% of those obtained with cleansing.
no-growth urine cultures are not contaminated, then we
can further categorize results as either contaminated or COMMENT
uncontaminated. And since groups 2 and 3 used both mid-
stream collection techniques with cleansing, and since Our 3 groups—no precautions, midstream, midstream
we could see that these groups were nearly identical (Table plus vaginal tampon—had essentially identical contami-
1), we were able to combine them into 1 larger group, nation rates of 29%, 32%, and 31%, respectively (Table
which allowed us to simplify our results (Table 2). 1). These data suggest that the method of urine collec-
As the data in Table 2 show, there was actually a tion does not affect the results of the urine culture and
slightly lower contamination rate among our nothing that, at least in our study population, collecting a speci-
group than among our midstream plus everything group men in a nonsterile container, with no special instruc-
(28.6% vs 31.5%), but this trend did not reach signifi- tions, may save time, money, and embarrassment.
cance (P=.65). We also found virtually identical colony counts in-
Finally, we found that the use of a bactericidal wipe dependent of the use of bactericidal wipes (Table 3). This
did not decrease colony counts (Table 3). More than finding suggests that, at least for symptomatic young

(REPRINTED) ARCH INTERN MED/ VOL 160, SEP 11, 2000 WWW.ARCHINTERNMED.COM
2539
Downloaded from www.archinternmed.com on May 4, 2010
©2000 American Medical Association. All rights reserved.
women, the concern that bactericidal wipes may lower
pathogenic colony counts is not warranted.
Why did midstream collection with cleansing (with
or without a vaginal tampon) not decrease contamination
rates? Several possibilities come to mind. Despite careful
instruction, patients may not have been able to perform the
somewhat complicated procedures correctly, and thus not
gain their benefits. If there were not a sufficient quantity
of contaminating organisms in the urethra, vagina, or peri-
neal region to grow in culture, then the bactericidal wipes
and midstream collections would not decrease the num-
ber of mixed cultures. Finally, despite all precautions, some
areas of the perineum may not have come in adequate con-
tact with the bactericidal wipe, thus allowing contaminat-
ing bacteria to reach our clean container.
One criticism that could considered is that we used
clean, but not sterile, containers to collect the urine. De-
spite the fact that others10 found no difference in colony
counts or contamination rates when clean containers were
used, the argument may be made that the number of con-
taminated specimens in all 3 of our groups may have been
elevated. If the contamination rate was elevated enough,
it may have obscured small differences between our
groups.
In an attempt to gauge the effect of culturing from
a “clean” container, we poured sterile saline into a clean
cup, and from there into a clean test tube, and finally into
our transport container to be cultured. This simulated
the process by which the urine samples were collected
from our subjects. For 10 specimens, we were careful not
to touch the inside of the container (simulating the ideal
urine collection technique), and for 10 specimens, we pur-
posely touched the inside of the cup with our fingers to
“contaminate” the container (simulating the effect of a
patient inadvertently doing the same). None of these 20
cultures showed any growth.
Also, our contamination rate was actually lower than
that reported in other studies.7,8,20 We also compared our
culture results with those of 446 urine cultures per-
formed in the same laboratory but from a different health
center, which were collected using a sterile container and
the MSSC technique. This “control” group was predomi-
nantly composed of symptomatic women, but as the symp-
toms were not always recorded on the requisition slip,
an exact percentage was not attainable. We found simi-
lar results, with 26.9% of our patients and 24% of those
in the control group having mixed cultures. We ob-
tained 54.5% pure cultures; the control group had 41%.
We did have a lower percentage of no-growth samples
(14.9% vs 35%), but this difference could largely be at-
tributed to the fact that all our patients had signs and
symptoms suggestive of cystitis.
Our study population included women, mostly
young, with symptoms of lower urinary tract infection
(cystitis). They were all premenopausal and without other
significant medical illnesses. None had a history of neph-
rolithiasis or renal disease, and several had had negative
results on renal ultrasonography and/or intravenous py-
elography. However, while this population represents a
sizable percentage of all persons with urinary tract in-
fections, it is also one in which current thinking sug-
gests that cultures need not always be obtained.21,22
(REPRINTED) ARCH INTERN MED/ VOL 160, SEP 11, 2000
2540
Downloaded from www.archinternmed.com on May 4, 2010
©2000 American Medical Association. All rights reserved.
Overall, our results suggest that it may not be nec-
essary for urine cultures to be collected by means of the
MSSC technique in young women with symptoms sug-
gestive of cystitis. Abandoning the use of the MSSC tech-
nique in this population may be warranted, thereby re-
sulting in considerable savings in money, time, and
embarrassment. We believe that further studies should
be performed to see whether the same result would be
obtained in male patients, patients with more compli-
cated urologic histories, and patients with signs and symp-
toms suggestive of upper or complicated urinary tract in-
fections.
Accepted for publication March 17, 2000.
Reprints not available from the authors.
REFERENCES
1. Schappert SM. National Ambulatory Medical Care Survey: 1992 Summary. Hy-
attsville, Md: National Center for Health Statistics; 1994. US Dept of Health and
Human Services publication PHS 94-1250. Advance Data From Vital and Health
Statistics, No. 253.
2. Johnson JR, Stamm WE. Diagnosis and treatment of acute urinary tract infec-
tions [published correction appears in Infect Dis Clin North Am. 1990;4:xiia]. In-
fect Dis Clin North Am. 1987;1:773-791.
3. Hooton TM, Scholes D, Hughes JP, et al. A prospective study of risk factors for
symptomatic urinary tract infection in young women. N Engl J Med. 1996;335:
468-474.
4. Beeson PB. The case against the catheter. Am J Med. 1958;24:1-6.
5. Boswell BR, Sanford JP. A screening method for the evaluation of urinary tract
infections in female patients without catheterization. Ann Intern Med. 1958;48:
1040-1046.
6. Walter FG, Knopp RK. Urine sampling in ambulatory women: midstream clean-
catch versus catheterization. Ann Emerg Med. 1989;18:166-172.
7. Leisure MK, Dudley SM, Donowitz LG. Does a clean-catch urine sample reduce
bacterial contamination? N Engl J Med. 1993;328:289-290.
8. Baerheim A, Digranes A, Hunskaar S. Evaluation of urine sampling technique:
bacterial contamination of samples from women students. Br J Gen Pract. 1992;
42:241-243.
9. Holliday G, Strike PW, Masterton RG. Perineal cleansing and midstream urine
specimens in ambulatory women. J Hosp Infect. 1991;18:71-75.
10. Immergut MA, Gilbert EC, Frensilli FJ, Goble M. The myth of the clean catch urine
specimen. Urology. 1981;17:339-340.
11. Morris RW, Watts MR, Reeves DS. Perineal cleansing before midstream urine:
a necessary ritual? Lancet. 1979;2:158-159.
12. Belmin J, Hervias Y, Avellano E, Oudart O, Durand I. Reliability of sampling urine
from disposable diapers in elderly incontinent women. J Am Geriatr Soc. 1993;
41:1182-1186.
13. Lipsky BA, Inui TS. “The best laid plans . . . ”: an evaluation of a patient educa-
tion program. Med Care. 1983;21:655-660.
14. Curtis P, Kim-Foley S, Kebede M. Urine sampling technique. Br J Gen Pract. 1993;
43:260-261.
15. Lipsky BA, Inui TS, Plorde JJ, Berger RE. Is the clean-catch midstream void pro-
cedure necessary for obtaining urine culture specimens from men? Am J Med.
1984;76:257-262.
16. Flournoy DJ. A descriptive analysis of urine cultures: pure versus mixed iso-
lates. Milit Med. 1992;157:514-517.
17. Siegman-Igra Y, Kulka T, Schwartz D, Konforti N. The significance of polymi-
crobial growth in urine: contamination or true infection. Scand J Infect Dis. 1993;
25:58-91.
18. Bartlett RC, Treiber N. Clinical significance of mixed bacterial cultures of urine.
Am J Clin Pathol. 1984;82:319-322.
19. Gross PA, Flower M, Barden G. Polymicrobic bacteriuria: significant association
with bacteremia. J Clin Microbiol. 1976;3:246-250.
20. Stamm WE, Counts GW, Running KR, Fihn S, Turck M, Holmes KK. Diagnosis
of coliform infection in acutely dysuric women. N Engl J Med. 1982;307:463-
468.
21. Ronald AR, Conway B. An approach to urinary tract infections in ambulatory
women. Curr Clin Top Infect Dis. 1988;9:76-125.
22. Hooton TM, Stamm WE. Diagnosis and treatment of uncomplicated urinary tract
infection. Infect Dis Clin North Am. 1997;11:551-580.
WWW.ARCHINTERNMED.COM