International Journal of

Environmental Research
and Public Health

Review
Diagnostic Test Accuracy of First-Void Urine Human
Papillomaviruses for Presence Cervical HPV in Women:
Systematic Review and Meta-Analysis
Peter Bober 1, * , Peter Firment 2 and Ján Sabo 1

1 Department of Medical and Clinical Biophysics, Faculty of Medicine, University of P.J. Šafárik in Košice,
Trieda SNP 1, 04011 Košice, Slovakia; jan.sabo@upjs.sk
2 Department of Anaesthesiology and Intensive Medicine, FNsP J. A. Reimana Prešov, Jána Hollého 5898/14,
08181 Prešov, Slovakia; firment@fnsppresov.sk
* Correspondence: peter.bober@upjs.sk

Abstract: First-void urine usually contains exfoliated cells of the debris and mucus from the female
genital organs and cervix, i.e., high concentration of human papillomavirus deoxyribonucleic acid
(HPV DNA). We conducted a meta-analysis of published data and determined an accuracy of HPV
detection in first-void urine compared to the women’s cervix. According to Preferred Reporting Items
for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we carried out a comprehensive
literature search. Eligible articles published from 2011 until 2021 were gathered by searching Embase,
PubMed and Cochrane Library Central databases. The patient selection, index test, standard test,



and patient flow were the factors involved in quality evaluation. A meta-analysis of 15 studies


(3412 women) based on 5054 potential records was conducted. Pooled sensitivity for high-risk
Citation: Bober, P.; Firment, P.; Sabo,
HPV detection in urine of 78% (70–84%) and specificity of 89% (81–94%) were calculated. Any
J. Diagnostic Test Accuracy of
HPV detection in urine of 87% (74–94%) and 91% (83–96%) were pooled sensitivity and specificity,
First-Void Urine Human
respectively. HPV 16 and 18 had a pooled sensitivity of 77% (76–77%) and specificity of 98%
Papillomaviruses for Presence
(98–98%). Meta-analysis indicated variations between the pooled specificities and sensitivities. In
Cervical HPV in Women: Systematic
Review and Meta-Analysis. Int. J.
meta-regression analysis, a heterogeneity in accuracy by using covariates (bias in patient selection,
Environ. Res. Public Health 2021, 18, purpose, sample timing, storage temperature and HPV detection method) were not detected. Our
13314. https://doi.org/10.3390/ meta-analysis demonstrates the accuracy of detection of HPV in urine for the presence of cervical
ijerph182413314 HPV. Although progress is continuously made in urinary HPV detection, further studies are needed
to evaluate and to improve the accuracy of the first-void urine test in order to be comparable with
Academic Editor: Jan Y. Verbakel other screening methods.

Received: 13 November 2021 Keywords: human papillomavirus; HPV DNA; cervical cancer; CIN; first-void urine
Accepted: 15 December 2021
Published: 17 December 2021

Publisher’s Note: MDPI stays neutral 1. Introduction

with regard to jurisdictional claims in
Is widely known that HPV is the primary cause of cervical cancer [1]. Cervical cancer
published maps and institutional affil-
iations.
presents the fourth-most cause of cancer deaths in women worldwide [2]. HPV is detected
in almost all cervical cancer biopsies with more than 90% presence in high-grade squamous
intraepithelial lesions (HSIL) [3]. More than 200 genotypes of HPV have been identified to
date [4]. Of them, HPV16 and HPV18 represent the high-risk oncogenic genotypes, as they
cause approximately 70% of nearly all cervical cancer [5–7].
Copyright: © 2021 by the authors.
A major impediment to controlling cervical cancer is lack of attendance for screening,
Licensee MDPI, Basel, Switzerland.
i.e., in those countries without well-developed screening programs, from 50% to more than
This article is an open access article
80% of women are not screened [8]. In addition, in countries with well-organised screening
distributed under the terms and
conditions of the Creative Commons
programmes, half of all potentially detectable carcinomas are found in women who have
Attribution (CC BY) license (https://
not attended screening programmes [9].
creativecommons.org/licenses/by/
There has been a drastic decline in the incidence, as well as the mortality, of cervical
4.0/). cancer worldwide since the introduction of the Pap test [10,11]. However, screening

Int. J. Environ. Res. Public Health 2021, 18, 13314. https://doi.org/10.3390/ijerph182413314 https://www.mdpi.com/journal/ijerph
Int. J. Environ. Res. Public Health 2021, 18, 13314 2 of 15

strategies for cervical cytology or Papanicolaou (Pap) tests requires uncomfortable and
invasive pelvic examinations. Moreover, healthcare providers find it time-consuming
and it cannot be carried out easily in resource-poor settings [12,13]. Additionally, cervical
cytology can be susceptible as a result of technical or subjective errors, due to low sensitivity
and false negative results [14,15].
There has been a great deal of interest lately in using urine as a liquid biopsy for HPV
DNA testing, and this has increased due to observation of high correlations between urine
and cervical HPV infections [16–20]. Urine samples are a good option for self-sampling
screening since they are cheap, noninvasive and simple to collect [21,22]. The HPV test
using urine appears to be an effective method for detecting HPV infection, so there is a
possibility that it could be used for women who do not participate in routine screenings [23].
Urine voiding in the first part (first-void urine) usually contains exfoliated cells of
the debris and mucus from the female genital organs and cervix, i.e., the first-void urine
contains higher concentrations of HPV DNA than midstream urine. According to this
theory, the identification of biomarkers in first-void urine, as well as HPV DNA, can be
used to screen for (pre)cervical cancer [24].
Therefore, we conducted a systematic review and meta-analysis to determine the
accuracy of detection of HPV in first-void urine compared with the cervix in women.

2. Materials and Methods

According to recommended methods, a meta-analysis and systematic review was
conducted in compliance with Preferred Reporting Items for Systematic Reviews and
Meta-Analyses (PRISMA) guidelines [25].

2.1. Criteria for Search and Eligibility

A literature review for the past 10 years (from January 2011 up to May 2021) in the
three databases: Embase and Cochrane library (Title/Keywords/Abstracts) and PubMed
(Title/Abstracts) was conducted. In each database, using Boolean logic, we searched for
the following terms: (HPV or hrHPV or human papillomavir *) OR (HPV or hrHPV or
human papillomavir *) AND (test * or assay * or genotyping or typing or detection or
amplification) OR (HPV or hrHPV or human papillomavir *) AND (deoxyribonucleic or
ribonucleic or nucleic or DNA or RNA or mRNA) OR (cervical or cervix or cervixes or
cervico *) AND (precancer * or cancer * or neoplas * or dysplas * or dyskaryos * or tumor *
or tumour * or malignanc * or carcinoma * or adenocarcinoma * or lesion * or squamous or
small cell or large cell) OR (cervical intraepithelial neoplasia or CIN or CINII * or CIN2 *
or CINIII * or CIN3 * or SIL or HSIL or LSIL or ASCUS or AS-CUS) AND (urin *). We
manually searched the relevant publications.
The eligibility criteria included any test-of-accuracy study comparing HPV DNA
detection in urine and cervix samples, in women with concern about infection with HPV or
development of cervical cancer. If the reference standard was different or not available, we
excluded the study. Meta-analysis included studies with data that could be converted into
2 × 2 table. A test’s diagnostic value can be overestimated by certain factors. Therefore,
we excluded case-control studies, i.e., studies testing only cervical cancer patients or
non-infected patients from the meta-analysis.

2.2. Study Extraction, Quality and Selection

For relevant studies, we screened all titles and abstracts. Two reviewers (P.B. and J.S.)
independently performed a systematic literature search. In addition, P.B. screened the full
texts of the included papers and extracted the subsequent data: characteristics of the study
(authors, publication year, country, and purpose), characteristics of the patients (median
age and range, cytology and histology results), index test characteristics (volume of sample,
storage temperature, DNA extraction and amplification method, test timing as compared
to the reference standard). To all studies the quality assessment of diagnostic accuracy
Int. J. Environ. Res. Public Health 2021, 18, 13314 3 of 15

studies-2 (QUADAS-2) was applied [26]. The patient selection, index test, standard test,
and patient flow were the factors involved in quality evaluation.

2.3. Data Synthesis and Statistical Analysis

Upon the detection of any HPV, high-risk HPV, HPV 16 and 18, the 2 × 2 table was
made. If the study included more than one method for testing urine HPV, we selected the
one with methods closest to those used by other studies. From the estimates, we derived
a summary receiver operating characteristic (SROC) curve and the summary accuracy
measures with 95% confidence interval (CI) (sensitivity, specificity, likelihood ratio positive
and negative (LR+ and LR−)). The shape of a receiver operating characteristic (ROC)
curve and the area under the curve (AUC) can help us get a sense of a test’s discriminative
power, i.e., AUC presents the measure of diagnostic accuracy. If the curve is located
as close as possible to the upper-lefthand corner, and the larger the area under curve,
then the test will discriminate better between diseased and healthy individuals. A good
indicator of the quality of the test is the area under the curve, which can range from 0 to
1. In a perfect diagnostic test, the AUC is 1, whereas in a nondiscriminating one, the
AUC is 0.5 [27]. The forest plots showing the sensitivity and specificity with 95% CI to
visualise heterogeneity of studies were generated. In addition, we included the subsequent
covariates in meta-regression in order to investigate possible sources of heterogeneity: bias
caused by patient selection (high risk versus low risk), purpose (surveillance of HPV versus
cervical intraepithelial neoplasia (CIN) and cervical cancer screening), sample timing (urine
before versus after cervical tissue collection), storage temperature (more than 0 ◦ C versus
less than 0 ◦ C), HPV detection method (conventional PCR versus real time, quantitative
polymerase chain reaction (qPCR), DNA microarray, multiplex PCR).
A meta-analysis of diagnostic test accuracy was conducted using an online, freely
available interactive web-based tool: MetaDTA, version 2.01 (https://crsu.shinyapps.io/
dta_ma/ (Accessed date: 13 December 2021)). The MetaDTA statistical tool pools the sensi-
tivity and specificity estimates for bivariate random-effects models. This model was fitted
as a generalized linear mixed-effect model using the glmer function from the package lme4
of the statistical software R with shiny [28]. This approach accounts for potential threshold
effects and covariance between sensitivity and specificity. Using the logit estimates of
sensitivity and specificity, the diagnostic odds ratios (DORs) were obtained directly. In
addition, using parameters estimated from the bivariate model through the equivalence
equations of Harbord et al. [29], the SROC plot was rendered.
Meta-regression was performed using Meta-DiSc software (version 1.4). To explore
sources of heterogeneity in the studies, we used the Moses–Shapiro–Littenberg method
by adding covariates to the model [30]. Meta-regression analysis included the threshold
effect, weighted least squares method, the inverse of variance of the log of the DOR, and
the random effects between studies using restricted maximum likelihood.
Publication bias was conducted using R Studio (version 1.3.959) with “metafor” pack-
age. A p value < 0.05 was considered statistically significant.

3. Results
Identifying and selecting studies is summarized in Figure 1. Of the 5054 potential
records, 15 studies (3675 women recruited, 3412 women analysed) were included in the
meta-analysis [3,16,17,31–42].
Int. Int. J. Environ.
J. Environ. Res.Res. Public
Public Health
Health 18,18,x 13314
2021,
2021, FOR PEER REVIEW 4 of
4 15
of 15

Flowdiagram
Figure1.1.Flow
Figure diagram of
of the
the studies
studies selected
selectedfor
forthis
thismeta-analysis.
meta-analysis.
3.1. Studies Description
3.1. Studies Description
The characteristics of included studies in this review and meta-analysis are shown
The characteristics
in Tables of included
1 and 2. We recruited 8 outstudies
of 15 in this review
populations ofand meta-analysis
studies are shown
from gynaecology or in
Tables 1 and clinics,
colposcopy 2. We recruited 8 outcentres,
3 from health of 15 populations of studies from
1 from genitourinary gynaecology
medicine and 1 from or acol-
poscopy
general clinics, 3 fromInhealth
practitioner. most centres, 1 from
populations of genitourinary
study, cervicalmedicine and 1 from
cancer screenings a general
were the
practitioner.
purpose of the In most populations
testing of study,
(10/15). Those cervical
remaining cancer
were screenings
for CIN were
follow-up the purpose
(3/15) or HPV of
surveillance
the (2/15).Those remaining were for CIN follow-up (3/15) or HPV surveillance
testing (10/15).
(2/15).Results of cytological analysis were recorded for 15 populations, i.e., 51% (1706/3360)
women hadof
Results normal conditions,
cytological 25%
analysis (848/3360)
were recorded had
foratypical squamous
15 populations, cells
i.e., 51% of (1706/3360)
undeter-
mined significance (ASCUS), 16% (542/3360) had low-grade squamous
women had normal conditions, 25% (848/3360) had atypical squamous cells of undeter- intraepithelial
lesionsignificance
mined (LSIL), 0.42% (14/3360)16%
(ASCUS), had(542/3360)
atypical squamous
had low-gradecells, possible
squamous high-grade lesion le-
intraepithelial
(ASC-H), and 7.4% (250/3360) had high-grade squamous intraepithelial lesion (HSIL).
sion (LSIL), 0.42% (14/3360) had atypical squamous cells, possible high-grade lesion (ASC-
From the 9 populations with reported histology results, 33.3% (304/912) of women had
H), and 7.4% (250/3360) had high-grade squamous intraepithelial lesion (HSIL). From the
normal conditions, 25% (229/912) had CIN1, 14.6% (133/912) had CIN2, 1.2% (11/912) had
9 CIN2+,
populations with reported
25% (229/912) histology
had CIN3, results,
and 0.66% 33.3%had
(6/912) (304/912) of women
histology had normal
proved cervical con-
cancer.
ditions, 25% (229/912)
Conventional PCRhad
wasCIN1,
used14.6%
in most (133/912)
studies,had
but CIN2, 1.2%methods
the testing (11/912) used
had CIN2+,
were not 25%
(229/912) had CIN3, and 0.66% (6/912) had histology proved cervical cancer.
uniform. Five of the 15 studies used real-time PCR [31,32,34,40,41], and there was only one
Conventional
PCR-based PCR was [37]
DNA microarray usedused
in most
out ofstudies, butstudy,
15. In one the testing methods
real time PCR was used were not
evaluated,
uniform.
in the lastFive of the 15
multiplex studies
PCR. used
Storage real-time PCR
temperatures [31,32,34,40,41],
of urine ranged from and ◦
−80 there was only
C [33,35,40]
one ◦ C [31,32,34,37].
to 4PCR-based DNAInmicroarray [37] used
13 and 11 studies out of 15.available
commercially In one study, real time
amplification PCR was
platforms
evaluated, in the DNA
and commercial last multiplex
extractionPCR. Storage temperatures
kits, respectively, were used. of In urine ranged
all studies, thefrom −80 °C
reference
standard to
[33,35,40] of 4cervical samples forInHPV
°C [31,32,34,37]. 13 andDNA 11testing
studieswere used.
commercially available amplification
platforms and commercial DNA extraction kits, respectively, were used. In all studies, the
reference standard of cervical samples for HPV DNA testing were used.
Int. J. Environ. Res. Public Health 2021, 18, 13314 5 of 15

Table 1. Qualitative characteristics of included studies.

Author, Year, Study Context Cytology HPV Detection DNA Extraction DNA Amplification
Country Timing Primers
[Ref] (Purpose) (Histology) Method Method Platform
abnormal (CIN2,
Hyun-Woong Cho, colposcopy QIAamp DNA
South Korea CIN3, urine after cervical real-time PCR Seegene L1
2020, [31] (follow-up of CIN) blood minikit
cervical cancer)
MagNA Pure LC
Mette Tranberg, 2020, general practitioner ASC-US (normal, another day, urine
Denmark real-time PCR total nucleic acid Roche L1
[32] (cancer screening) CIN1, CIN2+) after cervical
isolation kit
NILM, ASCUS/LSIL,
Severien Van Keer, colposcopy (HPV ASC-H/HSIL same day, urine
Belgium qPCR Non-commercial — —
2018, [33] surveillance) (normal, CIN1, before cervical
CIN2, CIN3)
primary health
Nicolás Vergara, normal, ASC-US, same day, urine
Chile care centre conventional PCR — Agilent Technologies L1/PGMY 09/11
2018, [3] HSIL, LSIL (—) before cervical
(cancer screening)
normal, ASC-US,
Brenda Y. community urine before cervical
HSIL, LSIL (normal,
Hernandez, 2018, Yap health centre and urine real-time PCR — Roche (Linear Array) L1/PGMY 09/11
CIN2, CIN3,
[34] (cancer screening) after cervical
cervical cancer)
normal,
ASCUS/LSIL,
colposcopy same day, urine
A Leeman, 2017, [35] Spain ASC-H/HSIL conventional PCR — Innogeneticstechnology L1/SPF10
(follow-up of CIN) before cervical
(normal, CIN1, CIN2,
CIN3)
ASCUS, LSIL, HSIL
Jack Cuzick, 2017, colposcopy (normal, CIN1, same day, urine QIAamp DNA
United Kingdom conventional PCR — E1
[36] (follow-up of CIN) CIN2, CIN3, before cervical Mini Kit
cervical cancer)
Pornjarim
normal, LSIL, HSIL PCR based DNA HPV GenoArray
Nilyanimit, 2017, Thailand (cancer screening) urine after cervical HybriBio L1
(—) microarray Diagnostic Kit
[37]
normal, NucliSENS
Alba Lucía Combita, health center same day, urine
Colombia ASCUS/LSIL, multiplex PCR easyMAG Luminex technology E7
2016, [17] (cancer screening) before cervical
ASC-H/HSIL (—) Extraction Kit
normal,
Elena Burroni, 2014, 8 days (median), QIAamp DNA
Italy (cancer screening) ASCUS/LSIL, conventional PCR Innogenetics L1
[16] urine after cervical Mini Kit
ASC-H/HSIL (—)
Int. J. Environ. Res. Public Health 2021, 18, 13314 6 of 15

Table 1. Cont.

Author, Year, Study Context Cytology HPV Detection DNA Extraction DNA Amplification
Country Timing Primers
[Ref] (Purpose) (Histology) Method Method Platform
Vikrant V. NILM, ASCUS/LSIL,
colposcopy (cancer same day, urine QIAamp DNA
Sahasrabuddhe, USA HSIL (normal, CIN1, conventional PCR Roche (Linear Array) —
screening) before cervical Blood Kit
2014, [38] CIN2, CIN3)
ASCUS/LSIL,
Keimari Mendez, gynaecology clinic same day, urine MagNA PureDNA
USA ASC-H/HSIL conventional PCR Roche (Linear Array) —
2014, [39] (cancer screening) before cervical Isolation Kit 1
(CIN1, CIN2)
normal,
A. Ducancelle, 2014, colposcopy (cancer QIAamp viral RNA
France ASCUS/LSIL, — real-time PCR Innogenetics L1
[40] screening) mini kit
HSIL (-)
normal,
Samuel Bernal, 2014, gynaecology clinic ASCUS/LSIL, HSIL same day, urine
Spain real-time PCR Cobas X 480 Roche —
[41] (HPV surveillance) (normal, CIN1, before cervical
CIN2, CIN3)
normal, BioMérieux
Elisabetta Tanzi, genitourinary clinic
Italy ASCUS/LSIL, same day conventional PCR NucliSENS1 Innogenetics L1(MY09/MY11)
2013, [42] (cancer screening)
HSIL (-) MiniMAG1

Table 2. Quantitative characteristics of included studies.

First-Void Storage
Author, Year Sample Recruited Median Age CIN2
Normal ASCUS/LSIL ASC-H/HSIL Normal CIN1 CIN3 (Cancer) Urine (Volume Temperature
[Ref] (Sample Detection) (Range) (CIN2/3)
Analysed in mL) in ◦ C
Hyun-Woong
314 (314) 40 (20–60) – 244/– –/70 – – 21 104 (4) (30) 4
Cho, 2020, [31]
Mette Tranberg,
150 (150) 45 (30–59) – 150/– – 11 10 11 (10–12) 4
2020, [32]
Severien Van
110 (110) 36 (25–64) 58 36/– –/15 7 11 6 9 (median; 19) −80
Keer, 2018, [33]
Nicolás Vergara,
543 (543) (18–64) 483 24/22 –/12 – – – – (10–15) −20
2018, [3]
Brenda Y.
Hernandez, 217 (210) (21–65) 179 31/3 –/4 2 – 2 5 (2) (30) 4
2018, [34]
A Leeman,
113 (91) (18–60) 28 11/28 9/15 50 22 13 6 (16) −80
2017, [35]
Int. J. Environ. Res. Public Health 2021, 18, 13314 7 of 15

Table 2. Cont.

First-Void Storage
Author, Year Sample Recruited Median Age CIN2
Normal ASCUS/LSIL ASC-H/HSIL Normal CIN1 CIN3 (Cancer) Urine (Volume Temperature
[Ref] (Sample Detection) (Range) (CIN2/3)
Analysed in mL) in ◦ C
Jack Cuzick,
652 (501) 30 (18–69) – 160/292 –/49 185 99 64 79 (0.5) –
2017, [36]
Pornjarim
Nilyanimit, 164 (164) (19–69) 95 –/50 –/19 – – – – (15) 4
2017, [37]
Alba Lucía
Combita, 540 (530) (18–25) 462 45/17 2/1 – – – – (9) −20
2016, [17]
Elena Burroni,
271 (215) 25 205 3/4 1//1 – – – – (60) −20
2014, [16]
Vikrant V. Sa-
hasrabuddhe, 72 (72) 28 (20–61) 18 23/11 –/16 17 28 16 10 (0.53) 20
2014, [38]
Keimari Mendez,
52 (50) (21–60) – 27/13 2/5 – 42 7 – (6) −20
2014, [39]
A. Ducancelle,
245 (230) (18–55) 34 70/59 –/25 – – – – (1) −80
2014, [40]
Samuel Bernal,
125 (125) 36 (21–65) 65 21/22 –/14 43 17 4 16 (20) –
2014, [41]
Elisabetta Tanzi,
107 (107) 42 (22–70) 79 3/21 –/4 – – – – (15) −20
2013, [42]
 3.2. 18,
Int. J. Environ. Res. Public Health 2021, Quality
13314 of Studies 8 of 15

A quality evaluation of the studies is shown in Figure 2. Due to narrow patient spec
trums
3.2. for 6 ofofthe
Quality studies, the high-risk of bias for patient selection was recorded: 3 studie
Studies
focusedAonly on patients with CIN of high grade [31,32,39], 2 studies recorded only youn
quality evaluation of the studies is shown in Figure 2. Due to narrow patient
women (18–25
spectrums forage) [16,17],
6 of the andthe
studies, 1 study included
high-risk human
of bias for immunodeficiency
patient virus (HIV
selection was recorded:
patients
3 studies focused only on patients with CIN of high grade [31,32,39], 2 studies recordedbias; 8/1
[42]. In most studies, the patient flow and timing reduced the risk of
analysed all recruited
only young participants,
women (18–25 andand
age) [16,17], 7 studies analysedhuman
1 study included (1.9–23.2%) of recruited partic
immunodeficiency
ipants. In 8 of 15 studies, both tests completed during the same day, and in the
virus (HIV) patients [42]. In most studies, the patient flow and timing reduced risk
8 studies, urin
of bias; 8/15 analysed all recruited participants, and 7 studies analysed (1.9–23.2%) of
samples were collected prior to taking cervical samples. In all low-risk-of-bias studies, th
recruited participants. In 8 of 15 studies, both tests completed during the same day, and in
reference standard was applied. Out of 15 studies, 1 used an index test with in-hous
8 studies, urine samples were collected prior to taking cervical samples. In all low-risk-
methods
of-bias that did the
studies, notreference
specify astandard
threshold,wasi.e., the bias
applied. Outof ofthis study was
15 studies, considered
1 used an index unclea
risktest
[33].
with in-house methods that did not specify a threshold, i.e., the bias of this studytest
In other studies (14/15), a predetermined threshold of the index waswith low
riskconsidered
of bias was unclear
used.risk [33].
The In other studies
publication (14/15),
bias did a predetermined
not appear threshold of the
in this study.
index test with low risk of bias was used. The publication bias did not appear in this study.

Figure 2. Qualitative assessment of 15 studies included in the meta-analysis using QADAS-2.

Figure 2. Qualitative assessment of 15 studies included in the meta-analysis using QADAS-2.
3.3. Meta-Analysis
3.3. Meta-Analysis
The heterogeneity of sensitivity and specificity between individual urine detection
studies of any HPV (10 studies),
The heterogeneity high-risk
of sensitivity HPV
and (12 studies),
specificity and HPVindividual
between 16 and 18 (7 urine
studies)detectio
studies of any HPV (10 studies), high-risk HPV (12 studies), and HPV 16 and 18 in
is shown in Figure 3. The individual sensitivities and specificities of any HPV detection (7 studies
urine varied from 54% [37] to 99% [38,42] and from 67% [29] to 99% [38,39], respectively.
is shown in Figure 3. The individual sensitivities and specificities of any HPV detection i
Individual sensitivities (51% [30] to 92% [33]) and specificities (59% [36] to 98% [35]) for
urine variedHPV
high-risk from 54% [37]
detection to 99%
studies [38,42]
in urine wereand from 67%
observed. [29] to
According to 99% [38,39],
analysis respectively
conducted
Individual
on HPV 16 and 18, sensitivities ranged from 27% [32] to 96% [33] and specificities ranged [35]) fo
sensitivities (51% [30] to 92% [33]) and specificities (59% [36] to 98%
high-risk
from 92%HPV [37]detection studies
to 99% [32,35,42] inin urine were observed.
urine-detection studies. According to analysis conducte
A SROC plot for pooled sensitivity and specificity
on HPV 16 and 18, sensitivities ranged from 27% [32] to 96% for the three[33]
groups,
and (a) any HPV, range
specificities
(b) high-risk HPV and (c) HPV 16 and 18 is shown in Figure 4. Pooled sensitivity for
from 92% [37] to 99% [32,35,42] in urine-detection studies.
high-risk HPV detection in urine of 78% (70% to 84%) and specificity of 89% (81% to 94%)
were calculated. For any HPV detection in urine of 87% (74% to 94%) and 89% (81% to
93%) sensitivity and specificity, respectively, were pooled. HPV 16 and 18 had a pooled
sensitivity of 77% (76% to 77%) and specificity of 98% (98% to 98%). The whole upper-left
quadrant in Figure 4 represents the 95% prediction region for the SROC plots, i.e., between
studies was heterogeneity. For any HPV detection, the 95% prediction region covers the
largest portion of the plot, i.e., it had the most heterogeneity between studies (Figure 4a).
For any HPV detection, the LR+ was 15.62 (95% CI 4.60 to 53.05) and the LR− was 0.14
(95% CI 0.08 to 0.24). For high-risk HPV detection, the LR+ was 6.81 (4.07 to 11.41) and
the LR− was 0.25 (0.18 to 0.34). For HPV 16 and 18 detection, the LR+ was 39.73 (39.33 to
40.14) and the LR− was 0.24 (0.24 to 0.24).
Int. J. Environ. Res. Public Health 2021, 18, 13314 9 of 15
Int. J. Environ. Res. Public Health 2021, 18, x FOR PEER REVIEW 9 of 15

Figure3.3.Forest
Figure Forestplots
plotsofof(a)(a)any
any HPV,
HPV, (b)(b) high-risk
high-risk HPVHPV
andand (c) HPV
(c) HPV 16 18
16 and andsensitivity
18 sensitivity and specificity
and specificity for studies
for studies
evaluating accuracy of first-void urine human papillomavirus (HPV) detection compared to cervical
evaluating accuracy of first-void urine human papillomavirus (HPV) detection compared to cervical HPV. HPV.

3.4. Meta-Regression
A SROC plotAnalyses
for pooled sensitivity and specificity for the three groups, (a) any HPV,
(b) A
high-risk HPV andwith
meta-regression (c) HPV 16 and 18covariates
the following is shown (bias
in Figure 4. Pooled
in patient sensitivity
selection, for high-
purpose,
risk HPV detection in urine of 78% (70% to 84%) and specificity of 89% (81% to 94%) were
sample timing, storage temperature and HPV detection method) was conducted to identify
the possible sources
calculated. For anyofHPVheterogeneity.
detection Using theofCochran’s
in urine 87% (74% Q to
test, likelihood
94%) and 89% ratios
(81%and
to 93%)
diagnostic
sensitivityodds
andratios were tested
specificity, for homogeneity
respectively, between
were pooled. HPVstudies.
16 andHeterogeneity and sensi-
18 had a pooled
variation
tivity ofbetween
77% (76% studies
to 77%)were
andnot confirmedofusing
specificity the covariates
98% (98% listed
to 98%). The aboveupper-left
whole (Table 3). quad-
rant in Figure 4 represents the 95% prediction region for the SROC plots, i.e., between
Table 3. Multivariate meta-regression results for characteristics with backward regression analysis.
studies was heterogeneity. For any HPV detection, the 95% prediction region covers the
largest portion of theMeta-Regression
plot, i.e., it had(Inverse
the most heterogeneity
Variance Weights 1 )between studies (Figure 4a).
For any HPVVar. detection, the LR+ wasStd.
Coeff. 15.62
Err. (95%p-Value 53.05) 2and the
CI 4.60 to RDOR LR−
(95% CI) was 0.14
(95% CI 0.08
Cte. 3 to 0.24). For high-risk
4.202 HPV detection,
0.900 the
0.006LR+ was 6.81 (4.07 to 11.41) and the
LR− was 0.25S 4 (0.18 to 0.34).−For
0.628HPV 160.307
and 18 detection,
0.096 the LR+ was 39.73 (39.33 to 40.14)
Bias the
and in patient
LR− was selection
0.24 (0.240.170
to 0.24). 0.866 0.852 1.19 (0.13; 10.98)
Purposes −2.708 1.759 0.184 0.07 (0.00; 6.13)
Sample timing −0.318 1.056 0.776 0.73 (0.05; 11.00)
Storage temperature 0.269 1.068 0.811 1.31 (0.08; 20.39)
HPV detection method 1.556 1.451 0.333 4.74 (0.11; 197.51)
1 Variables were retained in the regression model if p < 0.05. 2 Relative diagnostic odds ratio. 3 Constant coefficient.
4 Statistic S.
Int. J. Environ. Res. Public Health 2021, 18, x FOR PEER REVIEW 10 of 15
Int. J. Environ. Res. Public Health 2021, 18, 13314 10 of 15

Figure
Figure 4. 4. SROC
SROC plotfor
plot forstudies
studiesevaluating
evaluating accuracy
accuracy of
ofdetecting
detecting(a)
(a)any
anyHPV,
HPV,(b)(b)
high-risk HPV
high-risk and and
HPV (c) HPV 16 and
(c) HPV 1618 in 18
and
first-void urine
in first-void urinecompared
comparedwith withinincervix.
cervix.

3.5. Publication Bias

3.4. Meta-Regression Analyses
We investigated the potential publication bias by using Deek´s funnel plot asymmetry
A meta-regression with the following covariates (bias in patient selection, purpose,
test, as shown in Figure 5. The regression test showed no significant publication bias
sample timing, storage temperature and HPV detection method) was conducted to iden-
(p = 0.19).
tify the possible sources of heterogeneity. Using the Cochran’s Q test, likelihood ratios
and diagnostic odds ratios were tested for homogeneity between studies. Heterogeneity
and variation between studies were not confirmed using the covariates listed above (Table
3).
 stant coefficient. 4 Statistic S.

3.5. Publication Bias

We investigated the potential publication bias by using Deek´s funnel plot as
Int. J. Environ. Res. Public Health 2021, 18, 13314 11 of 15
metry test, as shown in Figure 5. The regression test showed no significant publica
bias (p = 0.19).

Figure5.5.Deek’s
Figure Deek’sfunnel plot.plot.
funnel The The
regression test showed
regression no significant
test showed publicationpublication
no significant bias (p = 0.19).
bias (p = 0.1
4. Discussion
4. Discussion
The purpose of diagnostic tests in healthcare settings is to confirm or exclude diag-
noses.The purpose
Assessment of of diagnostic
accuracy tests inby
is determined healthcare
comparing settings is to test
the diagnostic confirm
resultsor exclude d
with
the “gold
noses. standard” according
Assessment of accuracy to which individuals’by
is determined true diagnosis can
comparing thebe determined.
diagnostic In results
test
our study, the HPV DNA in cervix samples represented the gold standard
the “gold standard” according to which individuals’ true diagnosis can be determine test, to compare
with the HPV DNA in first-void urine samples.
our study, the HPV DNA in cervix samples represented the gold standard test, to com
In Pathak’s review, accuracy of urinary HPV testing for cervical human papillomavirus
withinvestigated
was the HPV DNA throughin first-void
meta-analysis.urineThere
samples.
was only one source of heterogeneity
In Pathak’s
identified, which was review, accuracyi.e.,
urine sampling, of the
urinary HPV
accuracy testingcollected
of samples for cervical human
as random or papillo
virus was investigated through meta-analysis. There was only one source of heterogen
midstream, as opposed to first-void samples, decreased by more than 22 times [23]. The first-
void urine contains
identified, whichhigher
was urinelevels sampling,
of high-risk i.e.,
HPVthe
as expected,
accuracyi.e.,
of 4.8–160
samples times higher in
collected as rando
comparison to the other fraction [24]. The first-void urine can produce
midstream, as opposed to first-void samples, decreased by more than 22 times [23]. more HPV DNA-
positive results than paired cervical samples when using sensitive HPV DNA assays [43–45].
first-void urine contains higher levels of high-risk HPV as expected, i.e., 4.8–160 t
Therefore, in our meta-analysis we used studies with first-void urine samples.
higher in comparison
To evaluate to the other
the performance fraction [24].
of a diagnostic Thesynthesized
test, we first-void sensitivity
urine canand produce m
specificity from a meta-analysis of diagnostic test accuracy studies. In our meta-analysis, a
heterogeneity between the pooled sensitivities and specificities was detected, i.e., pooled
sensitivity for high-risk HPV detection in urine was 78% (70% to 84%) and specificity was
89% (81% to 94%). For any HPV detection in urine of 87% (74% to 94%) and 91% (83%
to 96%), we pooled sensitivity and specificity, respectively. HPV 16 and 18 had a pooled
sensitivity of 77% (76% to 77%) and a specificity of 98% (98% to 98%).
The bivariate model has been shown to be mathematically identical to the HSROC
model when covariates are not included. The HSROC parameters were estimated using
parameters of the bivariate model and the equivalence equations of Harbord et al. The
SROC plot was drawn using the resulting HSROC parameters [29], and it shows the
relationship between sensitivity (y-axis) and 1-specificity (x-axis), illustrating variations in
sensitivity and specificity for different thresholds of a test. The whole upper-left quadrant
in Figure 4 represents the 95% prediction region for the SROC plots, i.e., between studies
there was heterogeneity. For any HPV detection, the 95% prediction region covers the
Int. J. Environ. Res. Public Health 2021, 18, 13314 12 of 15

largest portion of the plot, i.e., it had the most heterogeneity between studies (Figure 4a).
Regarding the method used in the present meta-analysis, we acknowledge as a limitation
that hierarchical models (such as the bivariate model) used in this meta-analysis are likely
to be vulnerable when the number of studies is small and also when sample sizes are highly
variable, which is partly the case of the present data [46].
The estimates of logit sensitivity and logit specificity were used to calculate LR+ and
-
LR . In our study, higher values of the positive likelihood ratio were detected, i.e., for any
HPV detection, the LR+ was 15.62 (95% CI 4.60 to 53.05) and the LR- was 0.14 (95% CI 0.08
to 0.24). For high-risk HPV detection, the LR+ was 6.81 (4.07 to 11.41) and the LR- was 0.25
(0.18 to 0.34). For HPV 16 and 18 detection, the LR+ was 39.73 (39.33 to 40.14) and the LR-
was 0.24 (0.24 to 0.24).
QUADAS-2 was used as a revised tool for the quality assessment of diagnostic accu-
racy studies [26]. The patient selection, index test, standard test, and patient flow were
the factors involved in quality evaluation. Generally, these studies had a high quality, i.e.,
an appropriate patient spectrum and a consecutive or random recruitment of participants
were used, the majority of recruited participants were included in analyses and all of them
used the same reference standard. However, the main weakness in some studies was that
they included only patients with CIN2+ [31,32,39], young women (18–25) [16,17] and HIV
patients [42]. In addition to resulting in a high prevalence, these factors could also lead to a
biased evaluation of test accuracy [47,48].
To determine whether these differences in testing methods influenced results, a meta-
regression was used. In the meta-regression analysis, the variation in accuracy was not seen
by using covariates (bias in patient selection, purpose, sample timing, storage temperature,
and HPV detection method). However, a heterogeneity between the pooled sensitivities
and specificities, and higher values of the positive likelihood ratio were detected. These
factors could have a significant impact on the probability of infection in HPV-positive
women. Therefore, the false positive results could lead to unnecessary invasive examina-
tion and costs, which is the advantage of the urine-testing method. However, the high
specificity of our test suggests that this scenario is less likely to occur. For these reasons,
our results should be interpreted cautiously because there is always the risk of over- or
underestimating data. Testing methods need to be more consistent and reproducible if
the test is to be successfully implemented in current practice. Therefore, we recommend
standardizing urine testing methods, i.e., before incorporating urine testing for HPV into
cervical cancer screening guidelines, it is important to minimise variation.
Based on the above-mentioned facts, it is necessary to optimise the HPV DNA detec-
tion in first-void urine in order to minimise variation of the first-void urine test (sensitivity
and specificity) for the presence cervical HPV in women. Optimised HPV DNA detection in
urine should include the following: (1) use of the first-void urine (morning or later during
the day) captured with a urine collection device [49]; (2) immediately mix first-void urine
with a conservation medium to prevent HPV DNA degradation during extraction and
storage; (3) provide sufficient first-void urine volume for subsequent sample concentration;
(4) recover cell-associated HPV DNA as well as cell-free DNA [43]; (5) use of HPV tests
meeting the criteria for primary cervical cancer screening [50]; (6) not cleaning the genital
area before collecting the sample [21]; and (7) collect the first-void urine samples before
cervical samples since this may reduce mucus and debris [51].

5. Conclusions
Our meta-analysis demonstrates the accuracy of detection of HPV in urine for the pres-
ence of cervical HPV. Although progress is continuously made in urinary HPV detection,
further studies are needed to evaluate and to improve the accuracy of the first-void urine
test in order to be comparable with other screening methods. Different testing platforms
and conditions were used in these studies. Therefore, all results should be interpreted
carefully, as they may have been over- or underestimated.
Int. J. Environ. Res. Public Health 2021, 18, 13314 13 of 15

Author Contributions: Conceptualisation, P.B. and J.S.; methodology, P.B.; software, P.B.; validation,
J.S.; writing—original draft preparation, P.B.; writing—review and editing, J.S. and P.F.; visualisation,
P.B.; supervision, J.S. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by the Slovak Research and Development Agency, under project:
APVV-19-0476.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are available on request from the
corresponding author.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Bosch, F.X.; Lorincz, A.; Munoz, N.; Meijer, C.J.L.M.; Shah, K.V. The causal relation between human papillomavirus and cervical
cancer. J. Clin. Pathol. 2002, 55, 244–265. [CrossRef]
2. Ferlay, J.; Soerjomataram, I.; Dikshit, R.; Eser, S.; Mathers, C.; Rebelo, M.; Parkin, D.M.; Forman, D.; Bray, F. Cancer incidence and
mortality worldwide: Sources, methods and major patterns in GLOBOCAN 2012. Int. J. Cancer 2014, 136, E359–E386. [CrossRef]
3. Vergara, N.; Balanda, M.; Hidalgo, W.; San Martín, H.; Aceituno, A.; Roldán, F.; Villalón, T.; Hott, M.; Espinoza, G.; Quiero, A.;
et al. Detection and genotyping of HPV in urine samples from Chilean women attending primary health care centers. Med.
Microbiol. Immunol. 2017, 207, 95–103. [CrossRef] [PubMed]
4. Crosbie, E.J.; Einstein, M.H.; Franceschi, S.; Kitchener, H.C. Human papillomavirus and cervical cancer. Lancet 2013, 382, 889–899.
[CrossRef]
5. Clifford, G.M.; Smith, J.S.; Aguado, T.; Franceschi, S. Comparison of HPV type distribution in high-grade cervical lesions and
cervical cancer: A meta-analysis. Br. J. Cancer 2003, 89, 101–105. [CrossRef]
6. Sanjose, S. de Human papillomavirus genotype attribution in invasive cervical cancer: A retrospective cross-sectional worldwide
study. Lancet Oncol. 2010, 11, 1048–1056. [CrossRef]
7. Greig, J.M.; Ellis, C.J. Occupational Hygiene of Chemical and Biological Agents, 3rd ed.; Elsevier Science: Amsterdam, The Netherlands,
2008; Volume 100, pp. 344–359. [CrossRef]
8. Bruni, L.; Albero, G.; Serrano, B.; Mena, M.; Gómez, D.; Muñoz, J.; Bosch, F.; de Sanjosé, S. Human Papillomavirus and Related
Diseases in the World-Summary Rep; ICO/IARC Inf Cent HPV Cancer (HPV Inf Centre): Geneva, Switzerland, 2019; p. 307.
9. Peto, J.; Gilham, C.; Fletcher, O.; Matthews, F.E. The cervical cancer epidemic that screening has prevented in the UK. Lancet 2004,
364, 249–256. [CrossRef]
10. Safaeian, M.; Solomon, D.; Castle, P.E. Cervical Cancer Prevention—Cervical Screening: Science in Evolution. Obstet. Gynecol.
Clin. N. Am. 2007, 34, 739–760. [CrossRef] [PubMed]
11. Sasieni, P.; Adams, J. Effect of screening on cervical cancer mortality in England and Wales: Analysis of trends with an age period
cohort model. BMJ 1999, 318, 1244–1245. [CrossRef]
12. Louie, K.S.; Silvia de Sanjose, P.M. Epidemiology and prevention of human papillomavirus and cervical cancer in sub-Saharan
Africa: A comprehensive review. Trop. Med. Int. Heal. 2009, 14, 1287–1302. [CrossRef]
13. Cronjé, H.S. Cervical screening strategies in resourced and resource-constrained countries. Best Pract. Res. Clin. Obstet. Gynaecol.
2011, 25, 575–584. [CrossRef] [PubMed]
14. Monsonego, J.; Bosch, F.X.; Coursaget, P.; Cox, J.T.; Franco, E.; Frazer, I.; Sankaranarayanan, R.; Schiller, J.; Singer, A.; Wright, T.;
et al. Cervical cancer control, priorities and new directions. Int. J. Cancer 2003, 108, 329–333. [CrossRef] [PubMed]
15. Jin, X.W.; Sikon, A.; Yen-Lieberman, B. Cervical cancer screening: Less testing, smarter testing. Cleve. Clin. J. Med. 2011, 78,
737–747. [CrossRef] [PubMed]
16. Burroni, E.; Bonanni, P.; Sani, C.; Lastrucci, V.; Carozzi, F. Human papillomavirus prevalence in paired urine and cervical samples
in women invited for cervical cancer screening. J. Med. Virol. 2014, 87, 508–515. [CrossRef]
17. Cómbita, A.L.; Gheit, T.; González, P.; Puerto, D.; Murillo, R.H.; Montoya, L.; Vorsters, A.; Van Keer, S.; Van Damme, P.;
Tommasino, M.; et al. Comparison between Urine and Cervical Samples for HPV DNA Detection and Typing in Young Women in
Colombia. Cancer Prev. Res. 2016, 9, 766–771. [CrossRef] [PubMed]
18. Franceschi, S.; Umulisa, M.C.; Tshomo, U.; Gheit, T.; Baussano, I.; Tenet, V.; Tshokey, T.; Gatera, M.; Ngabo, F.; Van Damme, P.;
et al. Urine testing to monitor the impact of HPV vaccination in Bhutan and Rwanda. Int. J. Cancer 2016, 139, 518–526. [CrossRef]
[PubMed]
19. Vorsters, A.; Micalessi, I.; Bilcke, J.; Ieven, M.; Bogers, J.; Damme, P. Detection of human papillomavirus DNA in urine. A review
of the literature. Eur. J. Clin. Microbiol. Infect. Dis. 2011, 31, 627–640. [CrossRef]
20. Vorsters, A.; Van Keer, S.; Biesmans, S.; Hens, A.; De Coster, I.; Goossens, H.; Ieven, M.; Damme, P. Van Long-Term Follow-up of
{HPV} Infection Using Urine and Cervical Quantitative HPV DNA Testing. Int. J. Mol. Sci. 2016, 17, 750. [CrossRef]
Int. J. Environ. Res. Public Health 2021, 18, 13314 14 of 15

21. Pattyn, J.; Van Keer, S.; Téblick, L.; Van Damme, P.; Vorsters, A. HPV DNA detection in urine samples of women: ‘An efficacious
and accurate alternative to cervical samples? Expert Rev. Anti. Infect. Ther. 2019, 17, 755–757. [CrossRef]
22. Lefeuvre, C.; Pivert, A.; Guillou-Guillemette, H.L.; Lunel-Fabiani, F.; Veillon, P.; Le Duc-Banaszuk, A.-S.; Ducancelle, A. Urinary
HPV DNA testing as a tool for cervical cancer screening in women who are reluctant to have a Pap smear in France. J. Infect.
2020, 81, 248–254. [CrossRef]
23. Pathak, N.; Dodds, J.; Zamora, J.; Khan, K. Accuracy of urinary human papillomavirus testing for presence of cervical HPV:
Systematic review and meta-analysis. BMJ 2014, 349, g5264. [CrossRef] [PubMed]
24. Vorsters, A.; Van Damme, P.; Clifford, G. Urine testing for HPV: Rationale for using first void. BMJ 2014, 349, g6252. [CrossRef]
25. Liberati, A.; Altman, D.G.; Tetzlaff, J.; Mulrow, C.; Gotzsche, P.C.; Ioannidis, J.P.A.; Clarke, M.; Devereaux, P.J.; Kleijnen, J.;
Moher, D. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare
interventions: Explanation and elaboration. BMJ 2009, 339, b2700. [CrossRef]
26. Whiting, P.F. QUADAS-2: A Revised Tool for the Quality Assessment of Diagnostic Accuracy Studies. Ann. Intern. Med. 2011,
155, 529. [CrossRef]
27. Šimundič, A.M. Measures of Diagnostic Accuracy: Basic Definitions. EJIFCC 2009, 19, 203–211.
28. Chu, H.; Cole, S.R. Bivariate meta-analysis of sensitivity and specificity with sparse data: A generalized linear mixed model
approach. J. Clin. Epidemiol. 2006, 59, 1331–1332. [CrossRef]
29. Harbord, R.M.; Deeks, J.J.; Egger, M.; Whiting, P.; Sterne, J.A.C. A unification of models for meta-analysis of diagnostic accuracy
studies. Biostatistics 2006, 8, 239–251. [CrossRef]
30. Lijmer, J.G.; Patrick, M.M.; Bossuyt, H.; Siem, H. Heisterkamp2, Exploring sources of heterogeneity in systematic reviews of
diagnostic tests. Statist. Med. 2002, 21, 1525–1537. [CrossRef] [PubMed]
31. Cho, H.-W.; Hong, J.H.; Min, K.J.; Ouh, Y.-T.; Seong, S.J.; Moon, J.H.; Cho, S.H.; Lee, J.K. Performance and Diagnostic Accuracy of
Human Papillomavirus Testing on Self-Collected Urine and Vaginal Samples in a Referral Population. Cancer Res. Treat. 2021, 53,
829–836. [CrossRef] [PubMed]
32. Tranberg, M.; Jensen, J.S.; Bech, B.H.; Andersen, B. Urine collection in cervical cancer screening–analytical comparison of two
HPV DNA assays. BMC Infect. Dis. 2020, 20, 926. [CrossRef]
33. Van Keer, S.; Tjalma, W.A.A.; Pattyn, J.; Biesmans, S.; Pieters, Z.; Van Ostade, X.; Ieven, M.; Van Damme, P.; Vorsters, A. Human
papillomavirus genotype and viral load agreement between paired first-void urine and clinician-collected cervical samples. Eur.
J. Clin. Microbiol. Infect. Dis. 2018, 37, 859–869. [CrossRef]
34. Hernandez, B.Y.; Tareg, A.C.; Reichhardt, M.; Agapito, A.; Zhu, X.; Sy, A.; Yuji, A.; Killeen, J.; Chan, O.; Buenconsejo-Lum, L.E.
Randomized controlled trial evaluating the utility of urine HPV DNA for cervical cancer screening in a Pacific Island population.
J. Glob. Heal. Rep. 2018, 2. [CrossRef]
35. Leeman, A.; del Pino, M.; Molijn, A.; Rodriguez, A.; Torné, A.; de Koning, M.; Ordi, J.; van Kemenade, F.; Jenkins, D.; Quint, W.
HPV testing in first-void urine provides sensitivity for CIN2+ detection comparable with a smear taken by a clinician or a
brush-based self-sample: Cross-sectional data from a triage population. BJOG Int. J. Obstet. Gynaecol. 2017, 124, 1356–1363.
[CrossRef]
36. Cuzick, J.; Cadman, L.; Ahmad, A.S.; Ho, L.; Terry, G.; Kleeman, M.; Lyons, D.; Austin, J.; Stoler, M.H.; Vibat, C.R.T.; et al.
Performance and Diagnostic Accuracy of a Urine-Based Human Papillomavirus Assay in a Referral Population. Cancer Epidemiol.
Biomark. Prev. 2017, 26, 1053–1059. [CrossRef] [PubMed]
37. Nilyanimit, P.; Chansaenroj, J.; Karalak, A.; Laowahutanont, P.; Junyangdikul, P.; Poovorawan, Y. Comparison of human
papillomavirus (HPV) detection in urine and cervical swab samples using the HPV GenoArray Diagnostic assay. PeerJ 2017,
5, e3910. [CrossRef] [PubMed]
38. Sahasrabuddhe, V.V.; Gravitt, P.E.; Dunn, S.T.; Brown, D.; Allen, R.A.; Eby, Y.J.; Smith, K.; Zuna, R.E.; Zhang, R.R.; Gold, M.A.;
et al. Comparison of Human Papillomavirus Detections in Urine, Vulvar, and Cervical Samples from Women Attending a
Colposcopy Clinic. J. Clin. Microbiol. 2013, 52, 187–192. [CrossRef]
39. Mendez, K.; Romaguera, J.; Ortiz, A.P.; López, M.; Steinau, M.; Unger, E.R. Urine-based human papillomavirus DNA testing as a
screening tool for cervical cancer in high-risk women. Int. J. Gynecol. Obstet. 2013, 124, 151–155. [CrossRef]
40. Ducancelle, A.; Legrand, M.C.; Pivert, A.; Veillon, P.; Le Guillou-Guillemette, H.; De Brux, M.A.; Beby-Defaux, A.; Agius, G.;
Hantz, S.; Alain, S.; et al. Interest of Human Papillomavirus DNA quantification and genotyping in paired cervical and urine
samples to detect cervical lesions. Arch. Gynecol. Obstet. 2014, 290, 299–308. [CrossRef]
41. Bernal, S.; Palomares, J.C.; Artura, A.; Parra, M.; Cabezas, J.L.; Robles, A.; Mazuelos, E.M. Comparison of urine and cervical
samples for detecting human papillomavirus (HPV) with the Cobas 4800 HPV test. J. Clin. Virol. 2014, 61, 548–552. [CrossRef]
[PubMed]
42. Tanzi, E.; Bianchi, S.; Fasolo, M.M.; Frati, E.R.; Mazza, F.; Martinelli, M.; Colzani, D.; Beretta, R.; Zappa, A.; Orlando, G. High
performance of a new PCR-based urine assay for HPV-DNA detection and genotyping. J. Med. Virol. 2012, 85, 91–98. [CrossRef]
43. Vorsters, A.; Van den Bergh, J.; Micalessi, I.; Biesmans, S.; Bogers, J.; Hens, A.; De Coster, I.; Ieven, M.; Van Damme, P. Optimization
of HPV DNA detection in urine by improving collection, storage, and extraction. Eur. J. Clin. Microbiol. Infect. Dis. 2014, 33,
2005–2014. [CrossRef]
Int. J. Environ. Res. Public Health 2021, 18, 13314 15 of 15

44. Tshomo, U.; Franceschi, S.; Tshokey, T.; Tobgay, T.; Baussano, I.; Tenet, V.; Snijders, P.J.F.; Gheit, T.; Tommasino, M.; Vorsters, A.;
et al. Evaluation of the performance of Human Papillomavirus testing in paired urine and clinician-collected cervical samples
among women aged over 30~years in Bhutan. Virol. J. 2017, 14, 74. [CrossRef] [PubMed]
45. Clifford, G.M.; Vaccarella, S.; Franceschi, S.; Tenet, V.; Umulisa, M.C.; Tshomo, U.; Dondog, B.; Vorsters, A.; Tommasino, M.;
Heideman, D.A.M.; et al. Comparison of Two Widely Used Human Papillomavirus Detection and Genotyping Methods,
GP5+/6+-Based PCR Followed by Reverse Line Blot Hybridization and Multiplex Type-Specific E7-Based PCR. J. Clin. Microbiol.
2016, 54, 2031–2038. [CrossRef] [PubMed]
46. Begg, C.B. Meta-analysis methods for diagnostic accuracy. J. Clin. Epidemiol. 2008, 61, 1081–1082. [CrossRef] [PubMed]
47. Lijmer, J.G. Empirical Evidence of Design-Related Bias in Studies of Diagnostic Tests. JAMA 1999, 282, 1061. [CrossRef] [PubMed]
48. Leeflang, M.M.G.; Bossuyt, P.M.M.; Irwig, L. Diagnostic test accuracy may vary with prevalence: Implications for evidence-based
diagnosis. J. Clin. Epidemiol. 2009, 62, 5–12. [CrossRef] [PubMed]
49. Pattyn, J.; Van Keer, S.; Biesmans, S.; Ieven, M.; Vanderborght, C.; Beyers, K.; Vankerckhoven, V.; Bruyndonckx, R.; Van Damme, P.;
Vorsters, A. Human papillomavirus detection in urine: Effect of a first-void urine collection device and timing of collection. J.
Virol. Methods 2019, 264, 23–30. [CrossRef] [PubMed]
50. Arbyn, M.; Snijders, P.J.; Meijer, C.J.; Berkhof, J.; Cuschieri, K.; Kocjan, B.J.; Poljak, M. Which high-risk HPV assays fulfil criteria
for use in primary cervical cancer screening? Clin. Microbiol. Infect. 2015, 21, 817–826. [CrossRef]
51. Sehgal, A.; Gupta, S.; Parashari, A.; Sodhani, P.; Singh, V. Urine HPV-DNA detection for cervical cancer screening: Prospects and
prejudices. J. Obstet. Gynaecol. 2009, 29, 583–589. [CrossRef]