SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE

2019, VOL. 37, NO. 1, 113–119

https://doi.org/10.1080/02813432.2019.1568708

RESEARCH ARTICLE

Sampling of urine for diagnosing urinary tract infection in general practice –

First-void or mid-stream urine?
Hoelmkjaer Pernillea, Bjerrum Larsa, M€akel€a Marjukkaa,b, Siersma Volkerta and Holm Annea
a
Research Unit for General practice and Department of General Practice, University of Copenhagen, Copenhagen, Denmark;
b
Department of Health and Social Care Systems, National Institute for Health and Welfare, Helsinki, Finland

ABSTRACT ARTICLE HISTORY

Objective: To study 1) whether the accuracy of point-of-care (POC) urine tests (dipstick, phase- Received 9 July 2018
contrast microscopy and culture) differs when performed on first-void urine (FVU) compared to Accepted 17 December 2018
mid-stream urine (MSU), and 2) if a delay of analysis up to four hours affects the accuracy of
KEYWORDS
POC tests when performed on urine from symptomatic of urinary tract infection (UTI), non-preg-
“Urinary tract infections”
nant women in general practice. [Mesh] “Urine” [Mesh];
Design: Prospective diagnostic study using paired samples. “Specimen handling” [Mesh]
Setting/Intervention: Three general practices in Copenhagen. Each woman delivered FVU and “Urine specimen collection”
MSU samples from the same void. As a reference standard, 8 ml of MSU was sent for culture at [Mesh] “Primary health
the Microbiology Department. care” [Mesh] “General
Patients: 117 women with one or more symptoms of UTI. practice” [Mesh]
Main outcome measures: Sensitivity, specificity and agreement with the reference standard of
FVU and MSU with different time delays (zero vs. one vs. four hours) as compared to reference
standard (MSU at time zero in boric acid tubes).
Results: All three POC tests performed on MSU were significantly more in agreement with the
reference than when performed on FVU when analysis was done immediately. The error rate
was 16% for MSU vs. 23% for FVU with POC culture, 27% vs. 40% with microscopy and 25% vs.
33% with dipstick testing. Delay of analysis up to four hours did not decrease agreement with
the reference.
Conclusion/Implication: MSU samples should be used in general practice for optimal accuracy
of POC tests. Analysis can be delayed up to four hours.

KEY POINTS

Point-of-care tests (dipstick testing, microscopy and culture) for diagnosing urinary tract infec-
tion performed on mid-stream urine samples are significantly more accurate than when per-
formed on first-void urine samples.

Delay of analysis up to four hours did not decrease the accuracy of any of the point-of-
care tests.

Point-of-care culture was more accurate than dipstick and microscopy both when performed
on mid-stream urine and first-void urine

The main contaminant in first-void urine samples was Enterococci spp., which contributed to
the majority of false positives.

Introduction samples may result in diagnostic misclassification,

Urinary tract infection (UTI) is a common reason for overtreatment, unnecessary side-effects and antibiotic
encounter in general practice and urine samples are resistance [3,4].
handled daily by both general practitioners (GPs) and In Denmark, most patients suspected for UTI are
practice personnel [1]. Typical uropathogens, e.g. instructed to deliver a mid-stream urine (MSU) sample
E. Coli and Enterococci spp. may act as contaminants by voiding firstly into the toilet and secondly into the
[2]. Features like high BMI, days since last shower or urine container. Practice personnel are trained to ana-
the degree of intimal hair removal could theoretically lyze the sample shortly after urination by point-of-care
contribute to this contamination. Contaminated urine (POC) testing, which may include urine dipstick,

CONTACT Pernille Hoelmkjaer pernillebrunsgaard@hotmail.com Research Unit for General practice and Department of General Practice, University
of Copenhagen, P.O.Box 2099, Copenhagen C, 1014, Denmark
ß 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
114 H. PERNILLE ET AL.

microscopy and culture. The recommendation to use or not being able to deliver enough urine to make
MSU is based on research done at hospitals with a two urine samples from the same portion of urine.
much higher prevalence of UTI than in primary care. The women were orally informed about the study,
Instructing patients in MSU can be time-consuming presented with written material and asked to sign a
and challenging, leaving interpretation of POC tests consent form. The involved clinics were located in
difficult. A systematic review suggested that sampling Copenhagen and covered a total of approximately
technique may not affect accuracy of urine culture 18,000 listed patients.
performed on urine from symptomatic patients in pri-
mary care, but none of the included studies compared
Data collection
the accuracy based on paired urine samples techni-
ques [5]. If the urine sample is left at room tempera- Upon inclusion, clinical information was collected
ture beyond four hours before analysis it may lead to using a structured case-report form. See Table 1 for
increased number of colony forming units (CFU) inter- list of covariates.
preted as significant growth. However, no studies
have investigated delays of up to four hours before
Urine samples
analysis. Furthermore, studies of the consequences of
delay are quite old and conducted before the thresh- The included women were instructed to deliver a first-
olds for significant growth were lowered to the cur- void urine (FVU) sample in one cup, squeeze off, and
rent cut-offs [6–8]. Because the threshold for a mid-stream urine (MSU) sample in a second cup.
significant growth has been lowered, delayed urine Hence, each woman delivered two urine samples.
analysis could affect the accuracy of POC testing Both samples were placed at room temperature and
as well. analyzed immediately after urination (time zero), after
In Danish general practice, antibiotic treatment is one hour and after four hours. First author (PB) proc-
usually initiated based on clinical history combined essed all samples.
with POC diagnostics. Waiting on results from the hos-
pital may cause unnecessary delay of up to four days, Reference standard
which is unacceptable for most patients [9].
The aims of the study were to investigate the influ- A few milliliters of all MSU-samples were transferred
ence of 1) sampling technique (First-void urine(FVU) to a boric acid tube and sent to Hvidovre Hospital
vs. MSU) and 2) delay of analysis (zero vs. one vs. four Microbiology Department for culture. Urine samples
hours) on the accuracy of POC tests (culture, phase- were analyzed on Inoqul ATM Bi-plate (CHROMagar
contrast microscopy and urine dipstick) for UTI in and blood agar) with 10 mL on each half of the agar
symptomatic non-pregnant women with suspected [10,11]. All samples were quantified. Significant growth
UTI in general practice. The reference standard for was defined as growth of 103 CFU/mL for E. coli and
both aims was urine culture performed on MSU at the S. saprophyticus, 104 CFU/mL for other typical uro-
microbiological department where bacterial multiplica- pathogens and 105 CFU/mL for possible uropatho-
tion was stabilized by adding boric acid immediately gens in accordance with European consensus [8].
after sampling. Growth of more than two different colonies (mixed
cultures) was considered as non-significant growth.

Material and methods

POC tests
Study design
POC culture
Prospective diagnostic study using paired samples.
10 ml of urine from each urine sample (FVU and MSU)
was inoculated on two separate agars (Flexicult IDV R
Setting and recruitment of patients
SSI Diagnostica) and the number of colonies was
Patients presenting with symptoms of UTI in three counted the following day. The Flexicult IDV R is a

general practices in Copenhagen were included. chromogenic agar allowing identification and quantifi-
Women were eligible if they were 18 or older, non- cation of most primary and secondary uropathogens.
pregnant and presented one or more symptoms of The culture was interpreted according to European
UTI (dysuria, frequency or urge). Exclusion criteria guidelines [8]. More than five colonies corresponded
were recent bladder surgery, urinary tract abnormality to 103 CFU/mL and was interpreted as a positive result
 SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 115

for a primary uropathogen. For secondary uropatho- Table 1. Baseline data (N ¼ 117).
gens, more than 50 colonies corresponded to 104 n / (%)
CFU/ml was interpreted as a positive result. Age
30 or below 58 (50)
31 to 60 46 (39)
Microscopy 61 or older 13 (11)
History of vaginal deliveries
A drop of uncentrifuged urine from each sample was Nulliparous 75 (64)
Body-mass index (self-reported)
analyzed by an Olympus phase-contrast microscopy at Below 25 93 (79)
400 times magnification. Urine samples were classified 25–30 16 (14)
30 or above 8 (7)
as positive if there were one or more bacteria or four Duration of symptoms
or more leucocytes per field of vision unless there Less than 4 days 55 (47)
Between 4 and 6 days 15 (13)
were 3 squamous cells or 3 different microorgan- 7 days or more 47 (40)
isms present. In this case the sample was labeled as Urine incubation time in bladder
1 hour or less 69 (59)
contaminated. 1–4 hours 41 (35)
4 hours or more 7 (6)
Days since latest shower
Urine dipstick 0 65 (56)
1 or more 52 (44)
The first (n ¼ 25) were analyzed by Combur 5 V R and
Genital hair removal
interpreted visually. The rest (n ¼ 92) were analyzed by All removed 45 (38)
Some removed 18 (15)
a semi-automated urine analyzer Urisys 1100 V R.
Nothing removed 54 (46)
Dipstick analyses were considered positive if there was Reference culture result
Significant growth (positive sample) 52 (44)
a positive reaction for nitrite (þ) or a positive reac- E. coli – out of total  41 (35)
tion (þþ) for leucocytes [12]. Of the 52 positive samples, 79% were due to E.coli.

Blinding and timing

external reference culture was done using a paired
PB was blinded to the result of the dipstick test when
logistic regression model taking into account paired
performing microscopy and to the result of the culture
samples. A Wald test was performed to evaluate the
performed at the microbiological department when
difference in agreement. These analyses were done
interpreting the results of urine dipstick, phase-con-
separately for each sampling modality and at each of
trast microscopy and culture performed in general
the time delays.
practice. All POC cultures were photographed and
To see whether third variables (listed in Table 1)
sent for a second evaluation by the last author (AH) influenced the effect of sampling technique on agree-
who was blinded to clinic information about patients ment with the reference, the interaction between
and results of urine analyses. Discrepancies between these third variables and the variable indicating the
the two evaluations done by PB and AH were dis- sampling technique was added to the above logistic
cussed, and the colonies were recounted in case regression models and tested with a Wald test. This
of doubt. model was also used to see if manual reading versus
automated reading of the dipsticks influenced the
Statistical analysis effect of sampling technique on agreement with the
reference. The accuracy (sensitivity and specificity) was
We assessed agreement of the POC tests performed calculated for both sampling techniques and all time-
on MSU at time zero to be 90% according to previous delays. A p-value of 0.05 was considered significant.
studies [13]. In a power calculation we determined The statistical analysis was done in SAS v.9.4
that we needed samples from 125 patients to detect a
drop in accuracy from 90% to 80% due to sampling
technique or delay of analysis with 80% probability, Results
assuming an intra-class correlation of 0.2 between the The study was conducted from September 2015 to
samples from the same patient; this with a signifi- June 2016 and a total 122 women were eligible for
cance level of 5% participation. Two were excluded due to not being
Agreement between the result of POC tests per- able to deliver a sufficient amount of urine and three
formed on urine from the two different sampling tech- were excluded due to already having participated in
niques and the three different time-delays with the the study leaving 117 women as eligible for inclusion.
116 H. PERNILLE ET AL.

Table 2. Overall agreement of urine POC test modalities as compared to reference (N ¼ 117).
Agreement with reference standard n (%)
p-value
Both FVU Neither FVU Overall Overall (sampling
and MSU Only MSU Only FVU nor MSU agreement, MSU agreement, FVU method)
POC culture
Immediate analysis 89 (76) 9 (8) 1 (1) 18 (15) 98 (84) 90 (77) .03
Analysis after 1 hour 87 (74) 11 (9) 1 (1) 19 (16) 98 (84) 88 (75) .003
Analysis after 4 hours 91 (78) 7 (6) 4 (3) 15 (13) 98 (84) 95 (81) NS
p-value (time delay) NS NS
Urine dipstick
Immediate analysis 73 (63) 14 (12) 5 (4) 25 (21) 87 (74) 78 (67) .04
Analysis after 1 hour 74 (63) 13 (11) 6 (5) 25 (21) 87 (74) 80 (68) NS
Analysis after 4 hours 74 (63) 11 (9) 9 (8) 23 (20) 85 (73) 83 (71) NS
p-value (time delay) NS NS
POC microscopy
Immediate analysis 63 (54) 22 (19) 7 (6) 25 (21) 85 (73) 70 (60) .005
Analysis after 1 hour 70 (60) 13 (11) 12 (10) 22 (19) 83 (71) 82 (70) NS
Analysis after 4 hours 83 (71) 13 (11) 4 (3) 18 (15) 96 (82) 87 (74) .03
p-value (time delay) 0.0045 0.0004
MSU: mid-stream urine; FVU: first-void urine; POC: Point of Care; NS: not significant.

PB handled all samples and there was no miss- one hour time delay (P ¼ .003). The difference was not
ing data. significant at four hours delay. There was no signifi-
Table 1 shows the baseline data. Participants were cant effect of delay of analysis on agreement with the
generally slim, with a low BMI, good hygiene and reference of POC culture neither when performed on
evenly divided whether they had any hair removal or MSU nor on FVU.
symptoms for less than six days. Only a small percent- For urine dipstick analysis 63% of the tests were in
age managed to have urine in the bladder for four agreement with the reference both when performed
hours prior to urinating. on MSU and FVU and 21% of test were not in agree-
The columns show overall agreement with the ref- ment with the reference neither when performed on
erence for each POC modality and sampling technique MSU nor on FVU. 11% of tests were only in agreement
at the three different time-points for delay of analysis. with the reference when performed on MSU compared
The right column shows the significance-level of the to 6% when performed on FVU. This resulted in a
effect of sampling technique on overall agreement significantly higher agreement with the reference of
with the reference. The row below each point-of-care dipstick analysis performed on MSU than on FVU at
modality shows the significance-level of the effect of immediate analysis (P ¼ .04), but the difference was not
delay of analysis on overall agreement with the refer- significant at one and four hours delay. There was no
ence. The reference for all analyses is culture per- significant effect of delay of analysis for urine dipstick.
formed in the microbiological department on MSU For microscopy there was a significant effect of
incubated in boric acid immediately after voiding. delay of analysis for both MSU (P ¼ .0045) and FVU
N ¼ 117. p-value obtained from logistic regres- (P ¼ .0004) with both modalities increasing their agree-
sion model. ment with the reference with increasing time delay.
Table 2 shows the agreement of the three POC The overall agreement with the reference was signifi-
tests with the reference standard at the three cantly better when microscopy was performed on
time-points. MSU than FVU at immediate analysis (P ¼ .005) and at
For POC culture, about 76% of tests were in agree- 4 hours time delay (P ¼ .03) but the difference was not
ment with the reference both when the test was per- significant at one hour time delay.
formed on MSU and FVU irrespective of time delay. The columns show sensitivity and specificity for
About 15% of the tests were not in agreement with each POC modality and sampling technique at the
the reference neither when performed on MSU nor on three different time-points for delay of analysis. The
FVU irrespective of time delay. In 8% of the samples, reference for all analyses is culture performed in the
only MSU agreed with the reference compared to 2% microbiological department on MSU incubated in
were only FVU agreed. This resulted in a significantly boric acid immediately after voiding. N ¼ 117
higher overall agreement with the reference of POC Table 3 shows sensitivity and specificity of the three
culture performed on MSU than POC culture per- POC tests when performed on MSU and FVU at all
formed on FVU at immediate analysis (P ¼ .03) and at time delays. Sensitivity of POC culture based on both
 SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 117

Table 3. Sensitivity and specificity of urine POC test modalities as compared to reference.
Sensitivity Specificity Overall agreement
proportion/(95% CI) proportion/(95% CI) proportion/(95% CI)
MSU FVU MSU FVU MSU FVU
POC culture
Immediate analysis 0.88 (0.80–0.97) 0.90 (0.82–0.98) 0.80 (0.70–0.90) 0.68 (0.56–0.79) 0.84 (0.77–0.90) 0.77 (0.70–0.85)
Analysis after 1 hour 0.88 (0.80–0.97) 0.88 (0.80–0.97) 0.78 (0.68–0.88) 0.63 (0.51–0.75) 0.84 (0.76–0.90) 0.75 (0.66–0.82)
Analysis after 4 hours 0.87 (0.77–0.96) 0.89 (0.80– 0.97) 0.82 (0.72–0.91) 0.75 (0.65– 0.86) 0.84 (0.77– 0.90) 0.81 (0.74–0.88)
Urine dipstick
Immediate analysis 0.73 (0.59–084) 0.81 (0.67–0.90) 0.75 (0.63–0.85) 0.55 (0.43–0.68) 0.74 (0.66–0.82) 0.67 (0.58–0.75)
Analysis after 1 hour 0.69 (0.57–0.82) 0.73 (0.61–0.85) 0.78 (0.68–0.88) 0.65 (0.53–0.76) 0.74 (0.66–0.82) 0.68 (0.60–0.77)
Analysis after 4 hours 0.65 (0.52–0.78) 0.79 (0.68–0.90) 0.78 (0.68–0.88) 0.65 (0.53–0.76) 0.73 (0.65–0.81) 0.71 (0.63–0.79)
POC microscopy
Immediate analysis 0.61 (0.47–0.75) 0.69 (0.55–0.81) 0.82 (0.70–0.90) 0.52 (0.50–0.65) 0.73 (0.65–0.81) 0.60 (0.51–0.69)
Analysis after 1 hour 0.58 (0.44–0.71) 0.77 (0.65–0.88) 0.81 (0.72–0.91) 0.65 (0.53–0.76) 0.71 (0.63–0.79) 0.70 (0.62–0.78)
Analysis after 4 hours 0.71 (0.59–0.83) 0.75 (0.63– 0.87) 0.91 (0.84–0.98) 0.74 (0.63– 0.85) 0.82 (0.75– 0.89) 0.74 (0.66–0.82)
MSU: mid-stream urine; FVU: first-void urine; POC: Point of Care; NS: not significant.

FVU and MSU was nearly 90% irrespective of delay. Table 4. Characteristic of false positive tests from first-void
Specificity of POC culture based on MSU was about urine at time zero (n ¼ 9).
80% regardless of time delay, while specificity of POC Uropathogens identified
in POC cultureat Quantity
culture based on FVU was lower, ranging from 63 Patient immediate analysis (cfu/mL)
to 75%. 1 Enterococcus spp. 104
For POC microscopy, the sensitivity was generally Skin bacteria 103
2 E.coli <103
lower (61–71% for microscopy based on MSU and Enterococcus spp. 104
69–77% for microscopy based on FVU). Specificity for Skin bacteria 103
3 Enterococcus spp. 104
microscopy based on MSU ranged from 81–91% and Skin bacteria 103
on FVU from 52 to 74%. Urine dipstick had a sensitiv- 4 E.coli 104
Enterococcus spp. 105
ity of 65–73% based on MSU and 73–81% based on Skin bacteria 103
FVU and a specificity of 75–78% based on MSU and 5 E. coli 103
55–65% based on FVU. Enterococcus spp. <103
6 E. coli 103
The result of the urine dipstick analysis of the 92 Enterococcus spp. 103
women whose urine dipstick analysis was performed Skin bacteria 104
7 Enterococcus spp. 104
on the Urilyzer compared to those whose dipstick was Skin bacteria 105
read manually. It showed that the accuracy of the test 8 Enterococcus spp. 104
Skin bacteria 103
was significantly worsened (odds ratio (OR) 3.30 9 Enterococcus spp. 104
[95%CI 1.44–7.60]; p ¼ .0049, i.e.) by manual reading. Skin bacteria 103
Especially specificity was lower when read manually. CFU: colony forming units. Bacteriological characteristics of nine patients
with false positive POC cultures from first-void urine at immediate ana-
However, manual reading did not significantly influ- lysis but negative POC culture from mid-stream urine.
ence the effect of sampling technique on agreement
with the reference.
None of the variables shown in Table 1 (age, history Discussion
of vaginal deliveries, body-mass index, duration of
symptoms, urine incubation time in bladder, days Statement of principal findings
since latest shower and degree of genital hair We found that POC tests performed on MSU were sig-
removal) significantly influenced the effect of sampling nificantly more accurate than when performed on FVU
technique on agreement with the reference. at immediate analysis. Delay of analysis up to four
Nine patients had a false-positive POC culture per- hours did not compromise the accuracy of POC tests
formed on FVU, but true-negative POC culture per- significantly. We did not identify any patient-related
formed on MSU at time zero. The bacteriological factors that influenced the effect of sampling tech-
characteristics of these patients are listed in Table 4. nique on agreement with the reference.
Three of these patients had a true-neagtive POC cul-
ture performed on FVU after four hours time-delay.
Enterococcus spp. were responsible for false-positive Strengths and weaknesses of the study
results at time zero in six cases. The study is the first to assess how urine sampling
technique from voided urine samples affects the
118 H. PERNILLE ET AL.

accuracy of POC test in general practice using paired included in the review that compared different voided
samples allowing for direct comparison. techniques, were using a randomized design instead
One person performed all the analyses and we regis- of paired samples [16,17]. When looking at sampling
tered no missing data. Blinding was ensured as much techniques, a paired sample results in stronger and
as possible since PB was blinded to the results of the more compareble results.
dipstick when performing microscopy and the dipstick Enterococcus spp. was the the primary reason for the
analysis was automated for the majority of tests. false positive FVU-samples. A study by Hooton et al.
Objectivity of the POC culture readings was ensured by found the same tendency, where the MSU contained
a second reading of photographs of the POC tests by Enterococcus spp., but the catherurine did not [18]. This
AH who was blinded to all other information. suggest that Enterococcus spp. is only found in the
A major limitation in this study is that the choice of urethra and not in the bladder, and the significance of
reference standard could be questioned since sending the Enterococcus as a uropathogen is questionable.
urine from the MSU sample inherently introduces a A systematic review from 2016 found that urine
bias giving MSU analysis a higher accuracy. The optimal stored at room temperature for more than four hours
reference standard would have been a catheter sample resulted in overgrowth of both contaminants and sig-
after obtaining the two other samples [14] or to send nificant growth [19]. In our study we found that there
both the MSU and FVU sample to the microbiological was no change in accuracy for POC culture and dipstick
laboratory. However, it is doubtful whether women analysis when analysis was delayed up to four hours.
would provide consent for the procedure and whether Accuracy measures of POC culture were quite high
enough urine could be obtained for two boric acid compared to a previous study performed in a number
containers. Since our main aim was to identify a differ- of practices in the Copenhagen area [20]. This differ-
ence between the two sampling techniques, using the ence could reflect that this study was conducted
MSU sample as reference proved most feasible. under optimal conditions and the previous study was
There were some limitations to blinding in the performed in a clinical setting.
study since PB was not blinded to clinical history, the
sampling technique or the time-delay when interpret-
Meaning of the study, mechanisms and
ing the POC test. The increasing accuracy with increas-
implication
ing delay of analysis for POC microscopy could
possibly be due to review bias since PB was not This study has shown that POC tests performed on
blinded to the result of the previous microscopies and MSU are more accurate than on FVU, immediately
the dipstick analysis when performing the subsequent after voiding. However, in the clinical setting, analysis
microscopies [15]. This could be part of the explan- is often delayed, and according to our results, this
ation for the surprising finding that microscopy does not result in a lower accuracy of the POC tests
became more accurate if the analysis was performed regardless of voiding technique. The main reason for
after four hours instead of immediately. the lower accuracy of POC tests performed on FVU
A significant difference was found when comparing was a small amount of false positive due to
the dipstick results from the visual reading of the first Enterococcus spp., but the clinical significance of this
25 women vs. the Urilyzers reading of the last 92 bacterium is not fully established. Our findings were
women. However, visual reading did not significantly consistent for both urine dipstick, POC microscopy
modify the effect of sampling technique on accuracy and POC culture.
of urine dipstick. Thus, our main results can be inter- Based on this study we would recommend practi-
preted without taking this limitation into account. ces that use urine dipstick with or without POC
microscopy, to instruct their patients to deliver a MSU
sample. If the practice uses POC culture, MSU still pro-
Findings in relation to other studies, principally
vides the most accurate result, but if findings of
concerning differences in results
Enterococcus spp. are interpreted with caution as it
We found a difference in the diagnostic accuracy of should, FVU can be used without compromising accur-
the POC test when using MSU compared to FVU. This acy. Urine samples do not need to be stored in a
adds to the findings in the systematic review from refrigerator if analyzed within four hours.
Holm and Aabenhus, where they did not find consist-
ent evidence when comparing the different sampling 1. Any necessary ethical approval: All procedures fol-
techniques in general practice [5]. The two studies lowed were in accordance with the Helsinki
 SCANDINAVIAN JOURNAL OF PRIMARY HEALTH CARE 119

Declaration of 1975, as revised in 1983. The study [7] Wheldon DB, Slack M. Multiplication of contaminant
was presented to the Ethical Committee for the bacteria in urine and interpretation of delayed cul-
Capital Region of Denmark and did not need ethical ture. J Clin Pathol. 1977;30:615–619.
[8] Aspevall O, Hallander H, Gant V, et al. European
approval (Ref. No.: FSP 15008103). All patients gave
guidelines for urinalysis: a collaborative document
written informed consent prior to participating in produced by European clinical microbiologists and
the study. The study was approved by the Danish clinical chemists under ECLM in collaboration with
Data Protection Agency (Ref. No.: 2015-41-4260). ESCMID. Scand J Clin Lab Invest. 2000;60:1–96.
2. The source of funding for the study: University of [9] Knottnerus BJ, Geerlings SE, Moll van Charante EP,
Copenhagen, 2016 funds, and The PLU fond et al. Women with symptoms of uncomplicated urin-
(Praktiserende Laegers Undervisningsfond) ary tract infection are often willing to delay antibiotic
treatment: a prospective cohort study. BMC Fam
3. Any conflict of interest: None
Pract. 2013;14:71.
4. Registration number if clinical trial: Trial registered
[10] Lauer BA, Reller LB, Mirrett S. Evaluation of preserva-
at: https://clinicaltrials.gov/ Study ID: NCT02585115 tive fluid for urine collected for culture. J Clin
Microbiol. 1979;10:42–45.
[11] Iversen J, Stendal G, Gerdes CM, et al. Comparative
Acknowledgement evaluation of inoculation of urine samples with the
Thank you to the GP’s and practice personnel at Vesterbro Copan WASP and BD Kiestra inoqula instruments.
Laegehus, Sundhedshuset Vesterbro and Laegerne på J Clin Microbiol. 2016;54:328–332.
Valdemarsgade. [12] Little P, Moore MV, Turner S, et al. Effectiveness of
five different approaches in management of urinary
tract infection: randomised controlled trial. BMJ. 2010;
Disclosure statement 340:c199–c196.
No potential conflict of interest was reported by the authors. [13] Bjerrum L, Grinsted P, Søgaard P. Detection of bacteri-
uria by microscopy and dipslide culture in general
practice. Eur J Gen Pract. 2001;7:55–58.
References [14] Stamm WE, Counts GW, Running KR, et al. Diagnosis
of coliform infection in acutely dysuric women.
[1] Pace WD, Dickinson LM, Staton EW. Seasonal variation N Engl J Med. 1982;307:463–468.
in diagnoses and visits to family physicians. Ann Fam [15] Whiting P, Rutjes AWS, Reitsma JB, et al. Sources of
Med. 2004;2:411–417. variation and bias in studies of diagnostic accuracy: a
[2] Jellheden B, Norrby RS, Sandberg T. Symptomatic systematic review. Ann Intern Med. 2004;140:189–202.
urinary tract infection in women in primary health [16] Bradbury SM. Collection of urine specimens in general
care. Bacteriological, clinical and diagnostic aspects in
practice: to clean or not to clean? J R Coll Gen Pract.
relation to host response to infection. Scand J Prim
1988;38:363–365.
Health Care. 1996;14:122–128.
[17] Lifshitz E, Kramer L. Outpatient urine culture: does
[3] Costelloe C, Metcalfe C, Lovering A, et al. Effect of
collection technique matter? Arch Intern Med. 2000;
antibiotic prescribing in primary care on antimicrobial
160:2537–2540.
resistance in individual patients: systematic review
[18] Hooton TM, Roberts PL, Cox ME, et al. Voided mid-
and meta-analysis. BMJ. 2010;340:c2096.
[4] Goossens H, Ferech M, Vander Stichele R, et al. stream urine culture and acute cystitis in premeno-
Outpatient antibiotic use in Europe and association pausal women. N Engl J Med. 2013;369:1883–1891.
with resistance: a cross-national database study. [19] LaRocco MT, Franek J, Leibach EK, et al. Effectiveness
Lancet. 2005;365:579–587. of preanalytic practices on contamination and
[5] Holm A, Aabenhus R. Urine sampling techniques in diagnostic accuracy of urine cultures: a laboratory
symptomatic primary-care patients: a diagnostic medicine best practices systematic review and meta-
accuracy review. BMC Fam Pract. 2016;17:72. analysis. Clin Microbiol Rev. 2016;29:105–147.
[6] De la Cruz E, Cuadra C, Mora JA. Effects of glucose, [20] Holm A, Cordoba G, Sørensen TM, et al. Clinical accur-
time and temperature on bacterial growth in urine. acy of point-of-care urine culture in general practice.
Rev Biol Trop. 1971;19:153–158. Scand J Prim Health Care. 2017;35:170–177.