Use of AffiProbe HPV test kit for detection of human papillomavirus DNA in genital scrapes

0095-1137/90/092076-06$02.00/0

Use of AffiProbe HPV

Test Kit for Detection of

Human

Papillomavirus

DNA in Genital

Scrapes

MARJUT RANKI,l* ANTTI W. LEINONEN,' TARJA JALAVA,' PEKKA NIEMINEN,2 VALÉRIA R. X. SOARES,2 JORMAPAAVONEN,2AND ARJAKALLIO1

OrionPharmaceutica/Biotechnology, OrionCorporation,' andDepartment ofObstetrics andGynecology,

University

Central

Hospital,2

Helsinki, Finland

Received13 February1990/Accepted 25 May 1990

The presence of human papillomavirus (HPV) DNA in cervical and vaginal scrapes was analyzed by the AffiProbe HPV test kit (Orion Corp., Orion Pharmaceutica, Helsinki, Finland), which is a 1-day solution hybridizationtestforHPV type6/11, 16,or18.TheAffiProbetestwascomparedwithacommerciallyavailable

dot blot test (ViraPap and ViraType tests; Life Technologies Inc., Gaithersburg, Md.). The study group consistedof 178patientsseeninagynecological outpatientclinic.Altogether,64specimens(36cervical and28 vaginal scrapes) from 49patientswerepositive bythe AffiProbe test.Concurrently collected cervicalscrapes from 174patientswereavailable for thereferencetest,whichyielded27positiveresultsfor HPVtype6/11or

16/18 and25positive resultsfor HPV type31/33135. Agreementasto thepresenceof HPVtype6/11, 16,or18 bythetwotestswasreached in85 %of thespecimens.Eleven cervicalspecimenswerepositivebytheAffiProbe testonly,and nine cervicalspecimenswerepositivebytheViraTypetestonly.Independentevidenceobtained bythepolymerase chainreaction, repeatexamination, orthe concurrentpresenceof HPVDNAinvaginalor

vulvalepithelium supportedtheAffiProbe and theViraTypetestresults for 6 of the11and 6of the9specimens withdiscrepant results, respectively. Thus, the DNAtests had similar sensitivities for HPVtype 6/11, 16,and 18DNAs,but theresultswereobtained within 1day bythe AffiProbetest,whereas results fortheViraPapand ViraType analyses requiredfrom 4 daysto2 weeks.

Human papillomaviruses (HPVs) cause infections ofthe genitaltract, whichcanbesubclinicalormanifestasvisible

orflat condylomas. HPVinfectionsarecommon;theannual incidence inFinland is 8% in young, sexuallyactive female populations (23a). HPV infection is considered a necessary

butnotasufficientriskfactorfor thedevelopmentof cervical cancer(9, 24, 29). HPV type 16 and 18 DNAs arefound in mostcancerbiopsy specimens (6, 11, 17, 30),whereasmany moretypesarepresentincervical dysplasias (11, 18).

Papillomavirus infections are a diagnostic problem

be-cause HPV cannot be cultivated in vitro and viral antigens

cannotbereliablydetected in clinicalspecimens. The

pres-ence of HPV is suspected by the presence oftypical

mor-phological changes in the epithelial cells detectable by cytology or histopathology. Specific diagnosis is, however,

only possible by nucleic acidhybridization with type-specific probes(20). SeveralDNAhybridization methodsareinuse,

including Southern blot, dot blot, filter in situ, and in situ hybridization ontissue sections (21).

We have developed a solution hybridization method,

basedon sandwich hybridization followed by affinity-based

collection ofthehybrids, for the detection of microbialDNA fromcrudeclinical specimens(14, 16, 26). One suchtest,the AffiProbe HPV test kit, is aimed at the detection ofHPV DNA from genital scrapes. The specimen is divided into

three aliquots for the detection HPV type 6/11, 16, or 18

DNA, respectively. The sensitivity of the test is 5 x 105

target molecules. The results are obtained as numerical values, andthe test standards give a clearcutofffor distin-guishing positive specimens from negativeones (T. Jalava, A. Kallio, A. W. Leinonen, and M. Ranki, Mol. Cell. Probes, in press).

*Corresponding author.

The purpose of the present study was to evaluate the AffiProbe HPVtest kitin the detection and typing of HPV DNAs fromgenital scrapes and to compare the efficacy of thetestwith that ofareference method, for whichwechose thecommercially availableViraPap and ViraTypetest kits. The latter are dot blot hybridization tests in which the samples are initially screened by the ViraPap testand the positive samples are subsequently typed by the ViraType test. TheViraTypetestallows detection of HPVtypes6/11, 16/18, and 31/33/35. Thetest results arevisualized by

auto-radiography. For interpretation of the ViraPap test, the signal intensity is compared with those of positive controls. Comparison of thetestefficacieswasrestrictedtoanalysis of cervical scrapes forwhichpaired specimens wereavailable and for the presence of HPV types 6/11, 16, and 18. The DNA tests had similar sensitivities, allowing detection of HPV DNAs from 63% ofthe cervical scrapes with benign

atypiaordyskaryosis.

MATERIALSANDMETHODS

Patients. A total of 178 patients attending the outpatient clinic of the Department of Obstetrics and Gynecology, University Central Hospital, Helsinki, Finland, were

in-cluded in thestudy. The patients belonged tothe following categories: 27 patients from the colposcopy clinic, 94

pa-tients from the emergency room, and 57 patients from the

abortion clinic.

Specimens. All patients were subjectedtoagynecological

examination during whichacervicovaginal cytologic smear

(Papanicolaou smear) was obtained. Smears were stained withamodified Papanicolaoustain.Twocervical specimens, both containing endocervical and ectocervical cells, were

collected from each patient for the DNAtests, one with a

cytobrush by using the AffiProbe specimen collection kit (OrionCorp., Orion Pharmaceutica,Helsinki, Finland) and 2076

on April 12, 2020 by guest

http://jcm.asm.org/

the other with a swab by using the ViraPap specimen collection kit (Life TechnologiesInc., Gaithersburg, Md.). The Papanicolaou smear was always taken first, but the orderof collection of the specimens for the DNA tests was randomized. The AffiProbe specimens were processed dur-ing the same day to recover the cells from the collection medium by centrifugation for 5 min at 850 x g. The cells, with100,ul of supernatant, were kept frozen until they were examined. The ViraPap specimens were also frozen until they were studied. A similar scrape from the vaginal epithe-lium of all patients was collected for the AffiProbe test. Additional specimens for the ViraPap and ViraType tests wereavailablefrom the vaginas of 65 patients and the vulvas of54 patients. The specimens were subjected to analysis with theAffiProbeHPV identification kit (Orion Corp.) and the ViraPap and ViraType tests (Life Technologies Inc.), respectively. Comparisonof the test efficacies was restricted toanalysis of174 paired cervical specimens.

AffiProbe HPV identification test. The AffiProbe HPV test kit allowsfor thesimultaneous detection and typing of HPV types 6/11, 16, and 18. For that purpose, the AffiProbe specimen was divided into three 30-pÀl aliquots, each of which was treated separately with the specimen pretreat-ment solution to release and denature the viral DNA. The aliquots were subjected to hybridization with type-specific detector probes and the biotinylated capture DNA. The probes werelabeledwith"Sisotope. After hybridization for 3 h at65°C, the hybridization mixtures were transferred to microdilution wells coated with streptavidin. Binding was for2h at37°C.The plates were then washed, and the labeled DNA that bound to hybrids was eluted with the elution solutionandsubjectedtocountingfor 5 min in a scintillation counter (LKB Wallac, Turku, Finland). Positive and nega-tivecontrolswereincludedin each type-specific test series: three controls for background, two low-positive controls, andtwohigh-positivecontrols. The cutoff of the test was set accordingto themean of the low-positive standards, which, in all cases, was at least 1.5 times the mean of the back-ground,i.e., negative control values. Eight test series were carried out.Themeanof the cutoff-to-background ratio was 2.5 (range, 1.6 to4.1) for HPV type 6/11-, 4.5 (range, 2.4 to 8.3) for HPV type 16-, and 6.0 (range, 1.9 to 11) for HPV type 18-specific tests.

ViraPap and ViraType tests. ViraPap and ViraType tests (LifeTechnologies Inc.) areintendedfordetection of DNAs ofHPV types 6/11, 16/18, and 31/33/35, as groups, by using probemixtures labeledwith32P isotope. The specimen DNA was dotted onto a nylon membrane and hybridized with a pool ofthe probemixtures described above (ViraPap test). ViraPap-positive specimens were subsequently tested as three separate dots by using three group-specific probe mixtures in thehybridizations (ViraTypetest). Resultswere visualized by

autoradiography.

Typing results of positive specimens requirefrom 4days to 2 weeks, depending on the age ofthe probe.

EnzymaticamplificationofHPV DNA by usingtype-specific primers (PCR). For the polymerase chain reaction (PCR) (13), DNA was isolated from the remaining aliquot of the AffiProbe specimen by phenol extraction and ethanol pre-cipitation. The resulting DNA was dissolved in H20. PCR analysis forHPV types 6, 11, 16, and 18 usedtype-specific biotinylated primers followed by quantification in a hybrid-ization test as described previously (8a, 25). All primers weredesigned accordingtopreviously published sequences (3, 4, 22, 23). As controls, weused linearized recombinant plasmidDNAscontainingtheappropriateHPV genome and

U.

b-

é--I

m y>

w

cc

<n

b-< w

cc 10

un 0 w

oùw

Ln à

-w

zz

A Positive specimens

;

e

e

.

e s

se

r

r

Negativespecimens

O-HPV6/11 HPV1M HPV18

'o~~~~~~~~~~~~~~~~~~~f

50s

LL 0.0 1. 11 UO .v .L LIO m

Off

e

L

e0

RATIO (TESTRESULT/CUT OFF)

FIG. 1. ThepresenceofHPVtype6/11, 16,and 18DNAsin352

specimenswasassessedbyusing the AffiProbeHPVidentification kit. Theratio ofeachtestresulttoitsrespectivetype-specific cutoff valuewascalculated.(A)Positiveresults (ratio, 21)for HPVtypes

6/11, 16, and 18 aredepicted above the cutoSline. The means standard deviations(verticalbars)of the negativeresults (ratio,<1)

arebelow thecutoffline. Note that the scalebelowthecutoffline is linear. Thenumber of negative resultswas 328for HPVtype6/11, 324for HPVtype16,and 333 for HPVtype18. (B)Distributionsof theratios of thenegativetestresults.

pools of positive and negative patient samples from which the DNAwas purified. A sensitivity similar to that of the AffiProbe test was obtained with

104

molecules of control

plasmid DNA inthe reaction. A sample was interpretedas

positiveifthecountsperminuteobtainedafterhybridization

were the same as or higher than those obtained with the

controlplasmid.

RESULTS

PerformanceoftheAffiProbeHPVtestkit. Altogether,178

cervical and 174 vaginal scrapes were analyzed for the

a

e

a

OLZ 0.6 1.0

Q2

0.6 1.0 0.2 (.6 Lu

on April 12, 2020 by guest

http://jcm.asm.org/

TABLE 1. Detection of HPV DNA from the cervical and vaginal scrapes from 178 patients by theAffiProbeHPV test Cervical No. of vaginal specimenresults

specimen result + - ND"

+ 15 20 1

- 13 126 3

a ND, Not done.

presence of HPV type 6/11, 16, and 18 DNAs by the AffiProbeHPV test kit.The cutoffvalue in each test series was defined according to the respective low-positive stan-dard. Theratio of eachtype-specifictestresult to thecutoff value was calculated. The results are shown in Fig. 1. Altogether,24,28,and 19positiveresults wereobtained for HPV types 6/11, 16, and 18, respectively. In 39% of the positive tests thesignal-to-cutoifratio was 1 to2, in44% it was between 2.1 and 10, andin17% itwasover 10(Fig. 1A). Of 985 negative tests, 867 (88%) gave counts-per-minute values beloworequaltohalf ofthecutoffvalue (Fig. 1B).

Thepositivetest results (seeabove)werederivedfrom 36 cervical and28 vaginal scrapescollected from atotal of 49 patients. In 15 (31%) patients both the cervical and the vaginalscrapeswere positive,in20patientsonly thecervical scrape was positive, and in 13 patients only the vaginal scrape waspositivefor HPV DNA (Table 1). Double infec-tions were found in seven specimens (Table 2). In one patient (patient3), thedoubleinfectionwasdetectedinboth cervical andvaginalscrapes,andinfourpatients,oneof the HPV types was also found in a second site (vagina or cervix).

Correlation between theAffiProbetestand thecontroltest in the detection of HPV DNA. The ViraPap and ViraType tests served as a reference method. For the comparison study, two simultaneously obtained cervical scrapes were available from 174 patients. A total of 18 specimens were positive and 129 specimens were negative for HPV types 6/11,16, and 18by bothtests(Table3).Thus, the agreement between the two tests was 85%. Eleven AffiProbe-positive specimens were negative by the control test. Seven addi-tional AffiProbe-positive specimens (five positive for HPV type 16 or 18 andtwo positive for HPV type 6/11; see Table 5) were suggested to contain HPV type 31/33/35 by the controltest. Eighteenadditional specimens that were

nega-TABLE 2. Double infections detected from cervical and vaginal scrapesby the AffiProbeHPV test

Test result Patient Spcmn

no. Specimen HPV HPV HPV

type 6/11 type 16 type 18

1 Cervix - + +

Vagina - - +

2 Cervix - +

-Vagina + +

-3 Cervix + +

-Vagina + +

-4 Cervix + -

-Vagina + +

5 Cervix - + +

Vagina - - +

6 Cervix

Vagina + - +

aSpecimenswithdouble infections are in boldface.

TABLE 3. Detection of HPV DNA from cervicalscrapesby AffiProbe HPVcompared with that by the controltest AffiProbe HPV No. of scrapes by the ViraType test

testresult for Toa

HPV type

6/11,

no.of HPVtype HPVtype HPVDNA 16,or18 DNA scrapes 6/11,16/18 31333 negative

DNApositive DNApositive

+ 36 18 7 il

- 138 9 18 111a

aIncludeseight scrapesthat werepositive bytheViraPaptestonly.

tive by the AffiProbe test contained HPV type 31/33/35 by the controltest(probesfor HPVtypes31, 33,and 35 are not included in the AffiProbe HPVtest).

HPV DNA in relation tocytologie findings. APapanicolaou smear wastaken fromeachpatient duringthesamevisit that scrapes for the DNA tests were obtained. HPV DNA was detected by either test in 23% of women with normal cytologyand in63% ofpatientswith abnormalcytology,i.e., Papanicolaou smear findings consistent withbenign atypia (class 2) or cervicaldysplasia (class 3) (Table 4).

Analysis of discrepant results. Independent control tests were performed on the 18 AffiProbe-positive and the 9 ViraType-positive clinical specimenswithdiscordantresults (Tables5and6). These control tests included a PCR for the detection ofHPV DNA from the remaining original speci-men, when available, detection of HPV DNA from the vagina or vulva, and repeat cervical scrapes collected on reexamination of patients who were not yet treated. The cervical scrapewas divided intotwoidentical aliquots, one fortheAffiProbeHPVtestand theother for theViraPap and ViraTypetests.

ForPCR, materialwasavailablefrom17of27specimens. Of these, 6 were

positive

and 11 were

negative

for HPV DNA (Tables5and6). In 2specimens, the amountof DNA was at least 500 ng, andin15 specimens itwasless than 250 ng. In the groupofspecimenswith alowyield of DNA, four specimens were positive, whereas both specimens with a substantialquantity ofDNA were positive.

Of11AffiProbe-positive but ViraType-negative specimens (patients 1 to 11, Table 5), only 5 were available forrepeat sampling.Threeofthese werepositive. HPV DNA was also detectedineitherthevaginaorvulvasamplefromtwoof the patients.ThePapanicolaousmear wasabnormal,suggesting HPVinfection in four patients (Table 5).

Of nine ViraType-positive but AffiProbe-negative speci-mens(patients 19 to 27; Table 6) only one wasavailable for repeatsampling (patient 21), but the specimen wasnegative. HPV DNA was present in vaginal specimens from five patients,and thePapanicolaousmear suggested HPV infec-tion infour patients.

Ofsevenpatientswithdiscrepanttypes(patients 12 to 18; Table 5), patient 12 waspositive for HPV type 6/11 DNA, as

TABLE 4. Correlation of presence of HPV DNA to cytologyincervicalscrapes

Cytology Totalno. HPV DNAa

Cyof patients No. (%) positive

No.

(%)

negative

Normal 122 28 (23) 94 (77)

Benignatypia 43 25 (58) 18 (42)

Dyskaryosis 13 10 (77) 3 (23)

aResults obtained by theAffiProbe and ViraType assays are combined.

on April 12, 2020 by guest

http://jcm.asm.org/

TABLE 5. Specificity ofAffiProbeHPV test

Original scrape HPVtype HPV typeon reexamination by:

Patient concurrently

no.' AffiProbe PCR Papanicolaou detected at other AffiProbe ViraType

(HPV type) (HPVtype) smearresult genital site

1 6/11 N NDd ND

2 6/liC N ND ND

3 6/11e BA 6/11, 16

4 16 16 N

5 16 16 BA 16/18

6 16 ND N ND ND

7 16C D 16/18 16

8 16C ND N 16 ND ND

9 18 ND BA ND ND

10 18 18 N ND ND

il 18C N

12 6/11 6/11 N ND ND

13 16C N ND ND

14 16C + 18C 16 BA 18 16e 16/18e

15 6/11 ND BA 31/33/35 31/33/35

16 16 BA 31/33/35 ND ND

17 16 ND D 31/33/35 ND ND

18 18 D 31/33/35 ND ND

a Patients1through 11 had a cervicalspecimenpositive by the AffiProbe HPV test butnegative by theViraType assay. Specimens from patients 12 through 18 were positive by theAffiProbeHPVtestandweretyped to containHPVDNAofHPVtypes31,33,and 35bythecontroltest.

bN,Normal; BA,benign atypia; D, dyskaryosis.

C Ahighcounts-per-minute valuewasobtained withasignal-to-cutoffratio of

.2.

dND,Not done.

eResults for thevaginal specimenonreexaminationareshown. Theconcurrentcervical scrapewasnegative bytheAffiProbe andViraTypetests.

confirmedbyPCR. Inpatient 13, the signal in the HPV type 16 DNA test was high. Patient 14 had a double infection with HPVtypes 16 and18 bytheAffiProbe test, and the presence of HPV type 16 DNA wasconfirmed byPCR. HPV type 18 was also detected in the vagina. On reexamination, HPV type 16DNA was detected by both tests. In the other four patients (patients 15 to 18), the AffiProbe test showed low counts per minute. In three patients, DNA of HPV type 31/33/35 was also found in the vulval specimen, and in the fourth patient it was found in the vaginal specimen. In one patient (patient 15), the presence of HPV type31/33/35 was confirmed by repeat sampling.

Inconclusion, independentevidence, including cytology, supported the specificity of the AffiProbe HPV test in specimensfrom 7of 11 patients (64%) that were negative by the ViraType assay. The specificity of the ViraType assay was suggested by similar analysis of specimens from seven

TABLE 6. Specificity ofthe ViraType test

Original scrape HPVtype

con-Patient currently

de-no.a ViraType PCR Papanicolaou tected atother (HPV type) (HPVtype) smearresult genitalsite

19 6/11 N 6/11

20 16/18 18 BA

21 16/18 N 16/18

22 16/18 N

23 16/18 NDC BA

24 16/18 ND N 16/18

25 16/18 ND N

26 16/18 ND BA 16/18

27 16/18 ND D 18

aThecervical scrapesfrom patients 19 through27werepositive by the ViraType assaybutnegative bytheAffiProbeHPVtest.

b N,Normal; BA,benignatypia; D,dyskaryosis.

cND,Notdone.

of nine

patients

(78%)

that were

positive

by

the

ViraType

testbut

negative

by

theAffiProbetest.

Of

theseven

patients

with

discrepant specimen typing results,

independent

evi-dence

supported

the AffiProberesult intwo and HPV type 31/33/35 DNA waspresentinfour.

DISCUSSION

The AffiProbe HPV test kit was

developed

to achieve a

standardizable, easy-to-perform

DNA

hybridization

test suitable for use in any

laboratory.

The

specimens

require

minimalpretreatment, and becausethereaction takes

place

in

solution,

hybridization

is short and crude

specimens

can

be tested without

background problems.

The test

yields

numerical

counts-per-minute values,

allowing

unambiguous

interpretation

ofthe results.

Type-specific

internal

positive

standards define the

cutoff

values. To

study

the clinical

applicability

ofthe new AffiProbe HPVtest

kit,

352

genital

scrapes wereexamined.

Overall,

288

specimens

were

nega-tiveforHPV

DNA,

57

specimens

were

positive

foroneHPV

type, and 7

specimens

were

positive

fortwoHPV

types.

Of

the

negative

specimens,

88% had a

signal

that was

half

or

less than

half

ofthe

cutoff

value. In

only

15tests

(1.5%),

the

signal

was between 0.8 and 0.99 times the

cutoff

value

(Fig.

1).

Erythrocytes

or mucus that were

frequently

present

in

clinical

specimens

did not affectthe

signal

levels.

Thus,

the

negative

results were

clearly

distinguishable

from the

posi-tive ones.

Ithas

previously

beendocumented that unrelated bacteria

or viruses do not cause

background problems

in the Af-fiProbe HPVtest

(Jalava

et

al.,

in

press).

Nocross-reactions between HPV type

6/11,

16,

or 18 DNA occurred in the

type-specific

tests. A

quantity

of about 109 unrelated HPV

DNA molecules was

required

for a

positive

signal.

In this

study,

the

quantity

ofHPV DNA in the cervical and

vaginal

scrapeswas

mostly

below

107

molecules

butwasnever

i109

molecules. The

specificity

of the test was studied

by

using

on April 12, 2020 by guest

http://jcm.asm.org/

crude clinicalspecimens. Specimens(11%) that gaveasignal fortwo HPV typesprobably represented truedouble

infec-tions,

and the rate was within the same order as that observed in previous studies. The HPV type 16 reagent of the AffiProbetest kitdetected 1.5 x 107 molecules of cloned HPV type 31 DNA (data not shown) because of the 70% sequencehomologybetween HPV type 16 and 31 DNAs(7). This cross-reaction apparently accounted for the four false

typing

results obtained bythe AffiProbe test. Nevertheless,

a standardized test with quantifiable numerical results is

likely

to

yield

typing results equally reliable as those ob-tainedbySouthernblotting, whichis hampered by technical

as well as interobserver interpretation problems between laboratories (1).

We chose another commercially available test (the Vira-Pap and ViraTypetests) foruseas a reference method. The AffiProbetestandthedot blot test were assumed to detect a similar

quantity

ofHPV DNA, with the sensitivities of the tests

being

unaffected by the numberof viral DNA copies percell.Agreementas tothe presenceof HPV type 6/11, 16,

or 18 was obtained in 85% of the specimens, and the sensitivities of both tests for HPV types 6/11, 16, and 18 were indeed comparable. The ViraType test has probes for HPVtypes31, 33,and 35; but theseprobes are not included in the AffiProbe test. Therefore, its overall sensitivity in

detecting

HPV infections was higher than that of the Af-fiProbe test. Control tests or other independent evidence

applied

to analysis of the discrepant results suggested that both DNAtests are specific.

Previous studies have shown variable agreement as to HPV DNA positivity in concurrently collected separate

specimens

(2, 8). Furthermore, the HPV DNA detection rate iscritically affected by thenumberof cells collected (12, 15,

28).

Wetherefore feel that the mostplausible explanation for the

discrepant

results in this study was an uneven distribu-tion of exfoliated cells intothe two separate specimens. The

specimens

were not pooled and divided into two aliquots,

because,

according to the instructions of the test manufac-turer, sampling is considered partof the identification kit.

In this study, HPV type 6/11, 16, 18, 31, 33, or 35 DNA

was present in 63% of the cervical scrapes collected from women with cytologic atypia. The HPV DNA positivity as correlated with cytologic atypia was in the same range as that obtained by using Southern blotting for studying exfo-liatedcells (19, 28). By studying vaginalscrapes, inaddition to cervical scrapes, for the presence of HPV DNA, addi-tional information is obtained. Genital HPV infections seem

to be multifocal (10, 18). It was recently reported (28) that theaddition of cells fromtheposterior vagina to the cervical

specimen

increasedthe rate of HPV DNA detection

signifi-cantly.

The presence of HPV DNA detected by such com-bined

specimens

couldregularly be confirmed by concurrent cervicalpunchbiopsy (19).

For our study, defined groups of mostly young females

were selected. In these groups the incidence of HPV infec-tions was high and the observed high number of

false-negative

Papanicolaou smears wasthereforenotunexpected

(5,

6, 27). By collecting multiple genital scrapes, the total number of HPV DNA-positive patients increased from 63 to 76 (Table 1). The rate of HPV DNA-positive patients was 33% (31 patients) in the group of women visiting the emer-gency room, 60% (34 patients) in the patients visiting the abortionclinic,and 40% (11 patients) in the patients from the

colposcopy

clinic. Unexpectedly, the number of patients withcytologicalsigns of HPV infection in the last group was small

(37%).

All HPV DNA-positive women with cervical

dysplasia (10 patients) had

positive

cervical scrapes, but specimens from other

genital

sites were also

frequently

positive (Tables 5 and 6).

The sensitivitiesof the conventional HPV DNA detection methods areapparently sufficient for

detecting

HPV-associ-ated lesions. When combined with a

Papanicolaou

smear, HPV DNA detection increases the

probability

of

predicting

histologically proven HPV lesions (19). An

important

ques-tion in the

diagnosis

ofHPVinfection isthestandardization of the assays to

improve

the interobserver

reproducibility

betweendifferent studies

(1).

Weconcludethat theAffiProbe HPV test is a standardized hybridizationtest which is easy toperform, and theresultsare availablein 1

working

day.

It has asensitivitycomparable to orslightly higherthan thatof the ViraType test in

detecting

HPV type

6/11, 16,

and 18 DNAs.

ACKNOWLEDGMENTS

We thank Seija Lehto for excellent technical assistance,

Marja-LiisaOgàrd for helpincollecting the specimens,andHeiniJarvifor processing themanuscript.

LITERATURE CITED

1. Brandsma, J., R. D. Burk, W. D. Lancaster, H. Pfister, and M. H. Schiffman. 1989. Inter-laboratory variation as an

expla-nation forvarying prevalence estimates of human

papillomavi-rusinfection.Int. J.Cancer 43:260-262.

2. Caussy, D., W. Orr, A. D. Daya, P. Roth, W. Reeves, and W. Rawls. 1988. Evaluation ofmethods fordetectinghuman

papil-lomavirus deoxyribonucleotidesequencesinclinicalspecimens.

J. Clin. Microbiol. 26:236-243.

3. Cole, S. T., and O. Danos. 1987. Nucleotide sequence and comparative analysis of the human papillomavirus type 18 genome. Phylogenyof papillomaviruses and repeated structure of the E6 and E7 gene products. J. Mol. Biol. 193:599-608. 4. Dartmann, K., E. Schwarz, L. Gissmann, and H. zur Hausen.

1986. The nucleotide sequence and genome organization of human papillomavirus type 11. Virology 151:124-130.

5. Ferenczy, A., M. Mitao, N. Nagai, S. J. Silverstein, and C. P. Crum. 1985. Latentpapillomavirus andrecurrentgenitalwarts. N. Engl. J. Med. 313:784-788.

6. Fuchs, P. G., F. Girardi, and H. Pfister. 1988. Human papillo-mavirus DNA in normal, metaplastic, preneoplastic and neo-plasticepithelia ofthe cervix uteri. Int. J. Cancer 41:41-45. 7. Goldsborough, M. D., D. DiSilvestre, G. F. Temple, and A. T.

Lorincz. 1989. Nucleotide sequence of human papillomavirus type 31: a cervical neoplasia-associated virus. Virology 171: 306-311.

8. Hallam, N., P. Gibson, J. Green, and M. Charnock. 1989. Detection and typingof human papillomavirus infection of the uterine cervix by dot blot hybridisation: comparison of scrapes

andbiopsies. J. Med. Virol. 27:317-321.

8a.Harju, L., P. Janne, A. Kallio, M.-L. Laukkanen, I. Lauten-schlager, S. Mattinen, A. Ranki, M. Ranki, V. R. X. Soares, H. Soderlund,and A.-C.Syvanen.1990. Affinity-based collection of amplified viral DNA: application to the detection of human immunodeficiency virus type 1, human cytomegalovirus and human papillomavirus type 16. Mol. Cell. Probes 4:223-235. 9. Koss, L. G. 1987. Carcinogenesis in the uterine cervix and

human papillomavirus infection, p. 257-262. In K. J.

Syjanen,

L. Gissmann, and L. G. Koss (ed.), Papillomaviruses and human disease. Springer-Verlag, Berlin.

10. Kulski, J. K., T. Demeter, P.Rakoczy,G. F. Sterrett, and E. C. Pixley. 1989. Human papillomavirus coinfections of the vulva anduterinecervix.J. Med. Virol. 27:244-251.

11. Lorincz,A. T., G. F. Temple, R. J. Kurman, A. B. Jenson, and W. D. Lancaster. 1987. Oncogenic association of specific human papillomavirus types with cervical neoplasia. JNCI79:671-677. 12. Metchers, W., A. van denBrule, J. Walboomers, M. de Bruin, M. Burger, P. Herbrink, C. Meijer, J. Lindeman, and W. Quint.

on April 12, 2020 by guest

http://jcm.asm.org/

1989. Increased detection rate of human papillomavirus in cervicalscrapesbythepolymerasechainreaction ascompared to modified FISH and Southern-blot analysis. J. Med. Virol. 27:329-335.

13. Mullis, K., and F. Faloona. 1987. Specific synthesis of DNAin vitro viaapolymerase-catalyzed chain reaction. Methods En-zymol. 155:335-350.

14. Parkkinen, S., R.Mantyjarvi, K.Syrjanen,A.-C.Syvanen,and M. Ranki. 1989. Sandwich hybridization in solution: a rapid methodto screen HPV 16 DNA incervicalscrapes. Mol. Cell. Probes3:1-11.

15. Parkkinen, S., S.Syrjanen, K. Syrjanen,M. Yliskoski, J. Ten-hunen, R.Mantyjarvi,andM. Ranki. 1988.Screening of prema-lignant cervical lesions for HPV 16DNA bysandwich and in situ hybridization techniques.Gynecol. Oncol. 30:251-264. 16. Ranki, M., A. Palva, M. Virtanen, M. Laaksonen, and H.

Sôderlund. 1983. Sandwich hybridization as a convenient method for the detection ofnucleic acids in crude samples. Gene21:77-85.

17. Reeves, W. C., L. A. Brinton, M. Garcia, M. M. Brenes, R. Herrero, E. Gaitan, F. Tenorio, R. C. de Britton, and W. E. Rawls.1989. Humanpapillomavirus infection and cervical can-cerin LatinAmerica. N. Engl. J. Med.320:1437-1441. 18. Reid, R., M. Greenberg, A. B. Jenson, M. Husain, J. Willett, Y.

Daoud, G. Temple, C. R.Stanhope,A.I.Sherman,G. D.Phibbs, and A. T. Lorincz. 1987. Sexually transmitted papillomaviral infections. The anatomic distribution and pathologic grade of neoplastic lesions associated with different viral types. Am.J. Obstet. Gynecol. 156:212-222.

19. Ritter, D. B., A. S. Kadis, S. H. Vermund, S. L. Romney, D. Villari,andR. D. Burk. 1988.Detection of human papillomavi-rusdeoxyribonucleicacid inexfoliatedcervicovaginal cellsas a predictor of cervical neoplasia inahigh-risk population. Am.J. Obstet. Gynecol. 159:1517-1525.

20. Roman,A., and K. H. Fife.1989. Humanpapillomaviruses:are we readytotype?Clin. Microbiol. Rev. 2:166-190.

21. Schneider,A.1987. Methods ofidentificationofhuman papillo-maviruses,p. 19-39.In K.J.Syrjanen,L.Gissmann, andL.G. Koss (ed.), Papillomaviruses and human disease.

Springer-Verlag, Berlin.

22. Schwarz, E., M. Durst, C. Demankowski, O. Latterman, R. Zech,E.Wolfsperger, S. Suhai, and H. zur Hausen. 1983. DNA sequenceandgenomeorganization ofgenital human papilloma-virustype6b. EMBO J.2:2341-2348.

23. Seedorf, K., G. Krfmmer, M. Durst, S. Suhai, and W. G. Rowekamp. 1985. Human papillomavirus type 16 DNA

se-quence. Virology 145:181-185.

23a.Syrjanen, K., M. Hakama, S. Saarikoski, M. Vayrynen, M. Yliskoski,S.Syrjanen,V.Kataja,and O.Castrén. 1990. Preva-lence, incidenceandestimatedlife-time risk of cervical human papillomavirus(HPV) infections in nonselected Finnishfemale population. Sex. Transm. Dis. 17:15-19.

24. Syrjanen, K.J. 1987. Papillomavirusinfections andcancer, p. 467-503. In K. J.Syrjanen,L.Gissmann,and L. G. Koss(ed.), Papillomaviruses andhuman disease. Springer-Verlag, Berlin. 25. Syvanen, A.-C., M. Bengtstrom, J. Tenhunen, and H.Sôderlund.

1988. Quantificationofpolymerasechain reaction productsby affinity-based hybrid collection. Nucleic Acids Res. 16:11327-11338.

26. Syvanen, A.-C., M. Laaksonen, and H. Sôderlund. 1986. Fast quantification of nucleic acid hybrids by affinity-based hybrid collection. Nucleic Acids Res. 14:5037-5048.

27. Toon, P. G., J. R. Arrand, L. P. Wilsom, and D. S. Sharp. 1986. Humanpapillomavirus infection of the uterine cervix ofwomen without cytological signsofneoplasia. Br. Med. J. 293:1261-1265.

28. Vermund, S. T., M. H. Schiffman, G. L. Goldberg, D. B.Ritter, A. Weltman, and R. D. Burk. 1989. Molecular diagnosis of genital human papillomavirus infection: comparison of two methods usedtocollectexfoliated cervical cells.Am. J. Obstet. Gynecol. 160:304-308.

29. zur Hausen, H.1989.Papillomaviruses inanogenitalcanceras a model to understand the role of viruses in human cancers. Cancer Res. 49:4677-4681.

30. zurHausen, H., and A. Schneider. 1987. The role of papilloma-viruses inanogenitalcancer, p.245-263.In N. P.Salzman and M. Howley (ed.), The papovaviridae 2. Plenum Publishing Corp.,New York.

on April 12, 2020 by guest

http://jcm.asm.org/