Int. J.
Cancer: 68,766-769 (1996)
Q 1996 Wiley-Liss, Inc.
PCR-BASED HIGH-RISK HPV TEST IN CERVICAL CANCER SCREENING
GIVES OBJECTIVE RISK ASSESSMENT OF WOMEN WITH
CYTOMORPHOLOGICALLY NORMAL CERVICAL SMEARS
L. ROZENDML~, J.M.M. WALBOOMERSI, J.C. VAN DER LINDEN',
M. VAN BAI.1.EGOOlJEN4, C.J.L.M. MEIJEK~.*
Departments of IPathology, 2Epidemiology and Bwstatistics and -'Obstetrics and Gynaecology, Free University Hospital,
Amsterdam and 4Department of Public Health, Faculty of Medicine, Erasmus University, Rotterdam, The Netherlands.
Cemcalsancer screening programmesusingcytomorphologi-
cal criteria could be more efficient if the screening included
objective individual risk factors for women with normal cytol-
ogy, such as a test for high-risk human papillomavirus (HPV).
The value of a PCR-based test for high-risk HPV types was
studied in a cohort of 1622 women presenting in a routine
triannual population-basedscreening programme. Women were
included in the study when they had no previous history of
cervical dysplasia; and their initial Pap smear was read as
normal (Pap I or 2). The mean age of the women was 42 years
(range 34-54 years) and mean follow-up time was 40 months
(range 5-73 months). Women were referred for colpoxopically
directed biopsies if they had had 2 successive cervical smears
read as Pap 3a (mild to moderate dyskaryosis) or one read as
?Pap 3b (severe dyskaryosis). Women with histologically con-
firmed cervical intraepithelial neoplasiagrade 111 (CIN 111) were
considered positive cases. All women were tested for 14
high-risk HPV genotypes. Of the 86 high-risk HPV-positive
women, 6 developed CIN 111, whereas only I of the 1536
HPV-negative women did. The women with normal Pap smears
containing high-risk HPV genotypes were I 16 times (95% CI,
13-990) more at risk of developing CIN 111, in contrast to
women without high-risk HPV. These results support the view
that the interval between successive smears in cervical-cancer
screening can be increased considerably for women with cyto-
morphologically normal and high-risk HPV-negative cervical
smears as determined by PCR.
c 1996 Wiley-Liss, Inc.
Detection of cytomorphologically abnormal cells in smears
of the uterine cervix is used in cervical-cancer screening
programmes to identify women with cervical cancer and
cervical intraepithelial neoplasia (CIN). However, cytomorpho-
logical examination has major drawbacks (Koss, 1989). The
percentage of false-negative cervical smears varies from 15 to
SO%, while the number of false-positive smears is about 10%
(Coppleson and Brown, 1974). This means that the sensitivity
of cytology for cervical cancer and CIN is quite low ( 5 0 4 5 % )
and the specificity is about 90%. The lack of sensitivity is
compensated by repeating the procedure after a fairly short
time interval, up to 3 years, depending on the result of the
Pap-smear (IARC Working Group, 1986). There is strong
epidemiological and molecular biological evidence that infec-
tion with high-risk human papillomaviruses (HPV) plays an
important role in the development of cervical cancer (IARC
Working Group, 1995). In addition, the relevant high-risk
HPV types can now be detected in one PCR reaction (Wal-
boomers et al., 1994). Thus, the question has been raised
whether cervical-cancer screening may become more efficient
when testing for high-risk HPV DNA is included in thc
screening procedure (Meijer et al., 1992). In this study we
evaluate whether the presence of high-risk HPV types in
cytomorphologically normal (Pap 1 or 2) cervical smears is
predictive for development of CIN 111 among women aged 34
to 54. The women were selected from the routine triannual
population-based cervical-cancer screening programme in The
Netherlands. According to the criteria used in the Dutch
programme, women were referred for colposcopically directed
biopsies if they had had 2 successive cervical smears read as
Pap 3a (mild to moderate dyskaryosis) or one read as 2 Pap 3b
Publicattonof the International Unton Against Cancer
Publication ae I'Union lnlernationale Contre le Cancer
F.J. vOORHORST2,P. KENEMANS3,Th.J.M. HELMERHORST',
(severe dyskaryosis). Women with histologically confirmed
CIN 111were considered positive cases.
MATERIAL AND METHODS
Study group
Of the women living in Amstelveen and Ouder-Amstel, 2
suburbs of Amsterdam, who were invited by the municipal
health service to attend the general practitioner for routine
cervical-cancer screening from September 1988 till April 1991,
4079 gave oral consent and volunteered to participate in a
study investigating the relation between cervical HPV and CIN
lesions. In 1995 the Pathology database of the Free University
Hospital and the Dutch nationwide pathology registration
service (PALGA) was searched for past and present data on
cervical pathology of this cohort. Of the 4079 women, 9 (0.3%)
women who had a previous history of cervical dysplasia, and
120 (2.9%) women whose initial Papanicolaou smear had been
read as >Pap 3a were excluded from the study. Follow-up
data on cervical cytology was available for 2454 out of 3950
(62%) women with initially normal cervical cytology (Pap 1 or
Pap 2). HPV scrapes collected in the period from September
1988 till April 1991 were still available for 1622 out of 2454
(66%) women. These 1622 women are included in the analysis.
The mean age of the women was 42 years (range 34-54 years)
and the mean follow-up time was 40 months (range 5-73
months).
Cervical cytology
For cytological analysis and HPV detection, 2 cervical
scrapes were taken with Cervex brushes (International Medi-
cal Products, Zutphen, The Netherlands) as described (Jacobs
et al., 1995). The first brush was used for routine cytomorpho-
logical examination. Morphological abnormalities were catego-
rized according to KOPAC, the standard classification in thc
Netherlands, adapted from that described by Vooijs et al.
(1987). By means of KOPAC, cervical smears are cytomorpho-
logically classified as Pap 0 (inadequate smear), Pap 1 (normal
smear), Pap 2 (very mild dyskaryosis, including atypical squa-
mous cells of unknown significance; ASCUS), Pap 3a (mild or
moderate dyskaryosis), Pap 3b (severe dyskaryosis), Pap 4
(suspected carcinoma in situ), or Pap 5 (suspected at lcast
micro-invasive carcinoma). According to the criteria used in
the Dutch cervical-cancer scrcening programme, women were
referred for colposcopically directed biopsies if they had had 2
successive cervical smears read as Pap 3a (mild to moderate
dyskaryosis) within an interval of at least 3 months, or one read
as 2 Pap 3b (severe dyskaryosis). The initial Pap smears of the
women who developed abnormal cytology ( 2 Pap 3a) and of
the women with initially high-risk HPV positive smears and thc
*To whom corres ondence and reprint requests should be sent, at
Department of PatEology, Free University Hospital, De Boelelaan
1117, 1081 HV Amsterdam, The Netherlands. Fax: (31)-20-4442964.
Email: pathol@,azvu.nl
Received: July 1,1996 and in revised form September 19, 1996.
 IIIGH-RISK HPV TEST IN CERVICAL CANCER SCREENING
follcw-up histology were reviewed. To keep review of the Pap
reading as blind as possible, the smears to be reviewed were
mixed with a 10-fold number of randomly selected smears.
Sample preparation and detection of HPV by PCR
For HPV detection, the remaining material of the first brush
and the material of a second brush were placed in 5 ml of PBS
containing 0.05% merthiolate. After vortcxing vigorously, the
suspensions were centrifuged for 10 min. at 3,000g. The cells
pelleted were re-suspended in 1.0 ml lOTE (10 mM Tris, 1 mM
EDTA, pH 7.8) and frozen at -80C. HPV detection was
performed by PCR on crude cell suspensions as described
(Jacobs et al., 1995). Briefly, the quality of target DNA was
determined by PCR using P-globin-specific primer. Scrapes
showing successful amplification of p-globin sequences were
subjected to HPV-general-primer-mediated PCR (GP-PCR)
with the general primers GP5+ and GP6' (De Roda Husman
et a/., 1995). High-risk and low-risk HPV genotypes were
determined as a group by a cocktail of HPV-type-specific
oligoprobcs (Jacobs et al., 1995), and individually after South-
ern-blot hybridization of GP5+/6+-generated products with
HPV-type-specific oligoprobes. Analysis was performed for 14
high-risk HPV types (16, 18,31,33,35,39,45,51,52,56,58,59,
66 and 68) and 6 low-risk HPV typcs (6,11,40,42,43 and 44).
Statistical methods
The logistic-regression model was used to calculate odds
ratio(s) (OR) and 95% confidence interval(s) (CI). Age was
omitted from the analysis, since the number of the women who
developed a CIN-111 lesion (n = 7) was too small to draw
age-adjusted conclusions. The time until development of CIN
111 was not included in the model, e.g., by using Cox's
regression, since in all but one case the CIN-I11 lesion was
detected at the next screening round, and the interval of 3
years until the next screening is too uniform for a meaningful
estimate of the time from the first test till development of CIN
111.
767
RESULTS
Baseline HPVprevalence and development of CIN lesions
Table I shows the HPV types detected at the beginning of
the study. Of all the women studied, 6.0% (n = 98) were HPV
(any type)-positivc. W e identified 17 different H P V genotypes.
In 0.7% (n = l l ) , we detected 10 double infections and 1 triple
infection. HPV 16, 18 and 31 were the most prevalent HPV
types, accounting for 66 out of 98 (67%) of the HPV-positive
cases. Based on the HPV phylogenetie tree (Van Ranst et al.,
1992), 0.7% (n = 12) of the women had a single low-risk HPV
type in their cervical smears, whereas 4.6% (n = 75) had a
single high-risk HPV infection. All the double infections
(n = 10) and the one triple infection contained at least one
high-risk HPV type. Thus, at baseline 5.3% (n = 86) of the
1622 wornen included in this study were high-risk HPV-DNA-
positive. As summarized in Table 11, of these 86 high-risk
HPV-positive women, 7% (n = 6) developed CIN 111, 2%
(n = 2) developed CIN 11, 2% (n = 2) developed CIN I, 8 %
(n = 7) had a single, transient Pap smear read as Pap 3a, and
80% (n = 69) had normal follow-up Pap smears (Pap 1 or 2).
Of the 1536 high-risk HPV-DNA-negative women, 0.1%
(n = 1) developed CIN 111, 1% (n = 1) CIN 11, 0.3% (n = 5)
CIN I, 2% (n = 31) had a single, transient Pap smear read as
Pap 3a, and 98% (n = 1498) had normal follow-up Pap smears
(Pap 1 or 2).
In the setting of cervical-cancer screening by cytology, the
unadjusted O R for development of CIN I11 among high-risk
HPV-positive women was 115 (95% CI, 14970; p < 0.0001)
when compared with high-risk HPV-negative women. When
the baseline high-risk HPV status was adjusted for the corre-
sponding Pap classification (Table 111), the O R hardly changed
(116; 95% CI, 13-99O;p < 0.0001).
Initial Pap smear and development of CIN lesions
In 21% (n = 346) of the 1622 women, the initial Pap smear
was read as Pap 2. Of these 346 women, 0.6% (n = 2)
developed CIN 111, 0.3% (n = 1) CIN 11, 1.2% (n = 4) CIN I
and 4% (n = 14) had a single, transient Pap smear read as Pap
3a. Of the remaining 1276 women with initially a Pap 1

TABLE I - HPV GENOTYPES ATTHE START OF THE STUDY

- HPV risk type' --

Women HR n = 75 (4.64) H R n = 11
(number) ( HPV-
~ ~ ~- LRn r= 12,
~(0.7%) _(0.7%)
-
6 I1 42 44 16 18 31 33 35 39 45 51 52 56 58 66 M2
--
1622 98(6.0%) 2 1 7 2 29 14 13 3 4 1 2 1 1 3 2 1 11
ILR, low-risk; HR, high-risk.-*Detected multiple infections: 1 x 11/16; 6 x 16/ 18; 1 x 16/31; 1 x
31/40; 1 x 35/58; 1 x 6/11/31.

TABLE I1 - HPV RELATED DE.VEI.OPMENT OF CIN LESIONS AMONG WOMEN (n = 1622)

WIIO METTHE CRITERIA'

Status after follow-up?

Initial ~

I x ~Pap3bor2xPap3a Total
hi h risk - 1 x Pap3a' Pap 1.2 % (4
H P 6 &,US CIN 111 CIN I1 CINl ~

96 (n)
(5 (n\ Vo (nl 5% In\ 4 (n)

Positive
Negative
Total
7.0% 6
0.1% 111
0.4% 7
2.3% 2
0.1% 11)
0.2% 3
2.3% 2
0.3%
0.4% 7
13 8.1% 7)
2.0% (311
2.3% 38
80% 69)
98% 11498
97% (15671
5.3% 86)
94.7% 11536
100% (16221
'(a) Screening population (34-54 years of age); (b) no past history of cervical dys lasia; (c)
initially normal cervical smear; (d) known follow-up data on cervical cytology; (e) initiay high-risk
HPV status assessed by PCR. Crude OR associated with development of CIN 1-111, CIN 11-111 and
CIN 111 are 29 (95% CI, 10-78), 79 (95% CI, 16-380) and 115 (95% C1, 14-970) respectively.-
2According to the criteria used in the Dutch cervical-cancer screening programme, women were
referred for colposcopically directed biopsies if they had had 2 successive cervical smears read as
Pap 3a (mild to moderate dyskaryosis) or one read as 2 Pap 3b (severe dyskaryosis).-'Women with
a single, transient Pap smear read as Pap 3a have no biopsy.
768 ROZENIMALETAL

reading, 0.4% (n = 5) developed CIN 111, 0.2% (n = 2) CIN was shown that women with cytomorphologieally normal
11, 0.3% (n = 3) CIN I and 1.9% (n = 24) had a single, cervical smears, who were positive in a HPV-PCR tcst that can
transient Pap smear read as Pap 3a. identify 14 high-risk HPV types, were 116 (95% CI 13-990)
The crude O R for dcvclopment of CIN I11 among women times more at risk of developing CIN I11 than womcn with
with Pap 2 vs. those with Pap 1 was calculated for the women high-risk HPV-negative scrapes. The womcn with an initial
with a known baseline HPV status (n = 1622; OR, 1.5; 95% Pap smear read as Pap 2 (very mild dyskaryosis. including
CI, 0.3-7.6;~ = 0.6) and for all women with follow-up data on atypical squamous cells of unknown signifieancc; ASCUS)
ccrvical cytology (n = 2454; OR, 1.8; 95% CI, 0.5-7;p = 0.4). showed an increased risk of 1.5 (95% CI, 0.3-7.7) for develop-
The high-risk HPV-adjusted O R for dcvcloping CIN III was mcnt of CIN I11 when compared with women with initially Pap
1.0 (95% CI, 0.2-5.9; Table HI). This mcans that the O R of 1.5 1. The O R of 1.5 is cntirely explained by the higher rate of
of developing CIN 111 among women with Pap 2 vs. those with high-risk HPV typcs in women with a smear read as Pap 2.
Pap 1 is entirely explaincd by thc higher rate of high-risk HPV These findings underline the potential value of HPV testing in
typcs in women with a smcar read as Pap 2 (7.5% high-risk cervical-cancer screening programmes.
HPV) vs. Pap 1 (4.7% high-risk HPV). We found that 5.3% of the women had high-risk HPV-
Characterization of women developing high-grade CIN lesions, positive cervical smears. From earlicr data (Walboomers et al.,
and review of cytology 1994) and currcnt findings (data not shown) of cross-sectional
studies we estimatc that about 4% of women with normal
Data on the 10 women who dcvcloped high-grade CIN cytology at 34 to 54 ycars of age have high-risk HPV-positive
lesions, i.e., 7 cases of CIN 111 and 3 cases of CIN 11, are given cervical smcars. The higher prcvalcnce of high-risk HPV in our
in Table IV. In 1 case, the high-grade CIN lesion was detected cohort may be explained by the requirement that follow-up
as a result of the first follow-up smear aftcr 5 ycars, in 5 cases data on cervical cytology is available for the women included in
after 3 years and in 4 cases within 2 ycars aftcr the initial Pap the analysis, which is also related to the slight over-
smear. In 3 of the latter 4 cases, it had been advised to rcpeat rcprcsentation of women with a Pap-2 smear (21%), who have
the Pap smear within one year, bccause the initial Pap smcar a stricter follow-up scheme, with an additional smear after 12
had been classified as Pap 2 (n = 2) or endoccrvical cells had months. However, thc rclatively high initial high-risk HPV
bccn missing (n = 1). Blind review of the initial Pap smcars prevalcncc does not influence the estimated OR or the
revealed that 2 of them could have been classified as Pap 3a, cstimatcd proportion of high-risk HPV-positive womcn dcvcl-
moderate dyskaryosis. Review of the (follow-up) histology oping CIN Ill.
yiclded identical rcsults for all cases.
The results of our study are in line with those of Koutsky et
al. (1992): in a prospectivc cohort of womcn selected in a clinic
DISCUSSION for scxually transmitted diseases they found a relative risk of 11
This study of women with cytomorphologically normal (95% CI 4 . 6 2 6 ) of developing CIN I1 and CIN Ill lcsions
cervical smcars participating in a triannual population-based among HPV-positive women whcn compared with HPV-
cervical-cancer screening programme and a follow-up time of negative womcn. Our strongcr correlation bctwccn initial
up to 73 months cvaluatcs the significancc of high-risk HPV HPV status and development of CIN I1 and CIN Ill (crude
types as a risk factor for development of CIN lesions. Women O R , 79; 95% CI, 1 6 3 8 0 ) might be attributed to the population
with histologically confirmed CIN Ill were considcrcd positive studicd and the detection method used. The dot-filtcr-
cases, since CIN I11 lesions are most closely related to cervical hybridization proccdure used by Koutsky et al. (1992) is
cancer and should be treatcd (Cannistra and Niloff, 1996). It considered to bc lcss sensitive than the PCR-based mcthod we
applied (Walboomers et al., 1994), leaving possiblc HPV-
positive cases with progression unidentified. In addition, Kout-
TABLE 111 - LOGISTIC-REGKESSIOS AVALYSIS FOR DEVELOPMENT OF sky el al. (1992) tested for only 5 high-risk HPV genotypes. The
CIS 111 WITH TlIE INITIAL HPV STATUS A N D INITIAL PAP READING AS
INDEPEKDEST RISK FACIOKS PCR method used in our study results in almost equal
amplification of the different HPV gcnotypes, including all 14
risk ;F
: py,r;,l; Unagwd Mutually ddiuslcd
OR 1 9 5 9 C1)
high-risk HPV types detected in a worldwide study on ccrvical
~ __ -
cancer (Bosch eta/., 1995).
High-risk HPV 6 80 119 116?(13-990) Our observations are consistent with the current epidemio-
positive logical model of cervical carcinogencsis (for review, sec Schiff-
Pap 2 2 344 1.5 1.0 (0.2-5.9)
man and Brinton, 1995). According to this model, H P V is the
OR were mutually adjusted for the baseline hi h-risk HPV primary risk factor for cervical cancer. Most HPV infections
status and Pap classification.-2Statisticallysignificant < 0.WOl). &, are transient (Meijer et al., 1992; Walboomers et al., 1994) and
do not lead to CIN. Mild to moderate dysplastic cervical
TABLE W - CHARACTERIZATION OF WOMEN W t 1 0 DEVELOPED lesions are associated with productive, and still transient, HPV
CIS 11-111 LESIONS
. ~.
infections. A minority of HPV-infected women will progress to
Status at intake (t,,) Status after follow-up ( t i )
CIN Ill and an even smallcr numbcr to ccrvical cancer. On the
- __ ~
basis of data from thc British Columbia cytological screening
programme (Boyes et al., 1982). Van Oortmarssen and
Habbcma (1991) estimated a mean time of 12 ycars for the
detectable phase of pre-malignant ccrvical lesions. Gustafsson
53 2 3a1 16 9 4 111
and Adami (1989) estimated a similar time (13 ycars) on the
42 1 1 18 61 4 I11
38 I 2 35 18 4 111 basis of Swedish ccrvical-cancer screening data. A detectable
38 1 3a 35 37 3b 111 HPV infection that precedes development of prc-malignant
35 1 1 18 38 3b 1I1 lesions dctcctablc by cytology would lengthen the total detcct-
35 2 2 16 31 3b Ill able phase. More cxtcnsive studies should be performed to
34 1 I? negative 25 3b 111 estimate the mean time between acquisition of high-risk HPV
47 2 2 negative 13 3al I1 and development of pre-malignant ccrvical lesions detectable
44 1 1 16, 18 31 3a2 II by cytology.
36 1 1 ~-
16 31 3b II
~

In agreement with thc model of ccrvical carcinogenesis,

Pap 3a, moderate dyskaryosis.-2No endocervical cells. Tablc I1 shows that of the majority of thc women with initially
 HIGH-RISK HPV TEST IN CERVICAL CANCER SCREENING
high-risk HPV-positive and cytomorphologically normal cervi-
cal smears, the smear remained cytomorphologically normal
(69 out of 86 cases; 80%), 8% (n = 7) had a single, transient
smear read as Pap 3a, 2% (n = 2) developed CIN I, 2%
(n = 2) developcd CIN 11, and 7% (n = 6) developed CIN 111.
The high-risk HPV-associated crude OR for development of
CIN 1-111, CIN 11-111 and CIN I11 are 29 (95% CI, 10-78), 79
(95% CI, 16380), and 115 (95% CI, 14-970) respectively. The
OR increase for women who develop more severe CIN lesions.
The very high OR of high-risk HPV for development of CIN
I11 is consistent with the hypothesis that these lesions develop
among women with persistent high-risk HPV infection (Rem-
mink et al., 1995).
In this study, the PCR-based HPV test is used in primary
screening to differentiate women at low risk for development
of CIN I11 from those at high risk. In such a setting, women
with a cytomorphologically normal, high-risk HPV-negative
cervical smear may need less frequent screening, whereas
women with a high-risk HPV-positive smear could be offered a
more intensive screening scheme or clinical management.
More intensive clinical managcmcnt could be immediate
colposcopy, as suggested by Cuzick et al. (1995), who used a
high-risk HPV tcst in primary screcning, and thereby detected
more prcvalent cases of CIN I11 than did screening using
cytology. Since some cases of CIN I11 were detected by
cytology and not by HPV, both tcsts might be combined to
detect more cases of CIN 111. In our study, after a blind review
of the smears containing high-risk HPV, 2 cases of initially
REFERENCES
BOSCH,F.X., MANOS, M.M., Mufioz, N., SHERMAN, M., JANSEN, A.M.,
PETO, J., SCHIFFMAN, M.H., MOKENO, V., KL'RMAN,R. and
SHAH,K.V., Prevalence of human papillomavirus in cervical cancer: a
worldwide perspective.J. nut. Cancer Insr., 87,796402 (1995).
BOYES, D.A., MORRISON,B., KNOX, E.G., DRAPER,G.J. and
MILLER,A.B., A cohort study of cervical-cancer screening in British
Columbia. Clin. invest. Med., 5, 1-29 (1982).
CANNISTRA, S.A. and NILOFF,J.M.. Cancer of the uterine cervix. New
Engl. J. Med., 334,103G1038 (1996).
COPPLESON, L.W. and BROWN, B., Estimation of the screening error
rate from the observed detection rates in repeated cervical cytology.
Amer. J. Obstet. Gynecol., 119,953-958 (1974).
CUZICK, J., SZAREWSKI, A., TERRY,G., HO, L., HANBY, A, MADDOX,
P., AKDERSON, M., KOCIAN,G., STEELE.S.T. and GUILLEBAUD. J.,
Human a illomavirus testing in primary cervical screening. Lancer,
345,153f-f536 (1995).
DE RODAHUSMAN, A.M., WALLWOMERS, J.M.M., VAYDEN BKULE, A.J.C.,
MEIJER,C.J.L.M. and SNIJDERS, P.J.F., The use of general primers
GP5 and GP6 elongated at their 3' ends with adjacent highly
conserved sequences improves human papillomavirus detection by
PCR. J. gen. Vir., 76,1057-1062 (1995).
GUSTAFSSON, L. and ADAMI, KO., Natural history of cervical neopla-
sia: consistent results obtained by an identification technique. Brit. J.
Cancer, 60,132-141 (1989).
IARC WORKING GKOUPON EVALUATION 01 CERVICAL-CANCER
SCREF.NING PROGKAMMES, Screening for s uamous cervical cancer:
01
duration of low risk after negative results cervical cytolo
implication for screening policies. Brit. med. J., 293,659-664 r9$. its
IARC WORKINGGROUP,Human papillomaviruses, IARC Scientific
Publication, 64, IARC, Lyon (1995).
JACOBS,M.V., DE RODA HUSMAN,A.M., VAN DEN BRULE,
A.J.C., SNIJDERS, P.J.F., MEIJER, C.J.L.M. and WALBOOMERS, J.M.M.,
Group-specific differentiation between high- and low-risk human-
papillomavirus genotypes by general primer-mediated PCR and two
cocktails of oligonucleotide probes. J. clin. Microbiol., 33, 901-905
(1995).
JENKINS, D., SHERIAW-JOHNSON, C. and GALLIVAN, S., Can pa illoma-
0
virus testin be used to improve cervical-cancer screening. Inf. J.
Cancer, 65, $68-773 (1996).
Koss, L.G., The Pa anicolaou test for cervical-cancer defection: a
triumph and a wage&. J. Arner. med. Ass., 261,737-743 (1989).
769
moderate dyskaryosis were detected among the 7 women who
developed CIN I11 in the follow-up. This indicates that HPV
might function as a quality control in cytomorphological
screening on abnormal cells, as proposed by Sherman et al.
(1994).
Finally, from our data we conclude that HPV testing may be
of potential value in cervical-cancer screening programmes.
Jenkins et al. (1996) havc cvaluated 3 screening policies based
on cytology and/or HPV by a stochastic model; from this, they
estimate that a screening strategy on the basis of routine HPV
testing could perform better than the current screening pro-
grammes based on cytology. In agreement with Jenkins et al.
(1996), we show that the PCR-based test for high-risk HPV
genotypes could identify, among women with normal cytology,
those at risk of developing CIN 111. The very low risk of
developing CIN I11 among women with cytomorphologically
normal and high-risk HPV-negative cervical smears indicates
that testing for high-risk HPV should be considered, in order
to increase the efficiency of cervical-cancer screening pro-
grammcs.
ACKNOWLEDGEMENTS
We thank Mrs. M.M. Kreike, Mr. R. Pol and Mr. B. Bakker
for technical assistance. The general practitioners of Am-
stelveen and Ouder-Amstel are thanked for their co-opera-
tion. This work was supported by grants from the Ziekenfonds-
raad and the Preventiefonds (project 28-1502,2).
KOU~SKY, L.A., HOLMES,K.K., CKITCHLOW, C.W., STEVENS,C.E.,
PAAVONBN, J., BECKMANN. A.M., DEROUEN, T.A., GALLOWAY, D.A.,
VERNON, D. and KIVIAT, N.B., A cohort study of the risk of cervical
intraepithelial neo lasia grade 2 or 3 in relation to papillomavirus
infection. New EngfJ. Med., 327,1272-1278 (1992).
MEIJER,C.J.L.M., VAN DEN BRL'IX, A.J.C., SNIJDERS,P.J.F.,
HELMERHORST, TH.J.M., KENEMANS,P.and WALBOOMERS, J.M.M.,
Detection of human papillomavirus in cervical scrapes by the polymer-
ase chain reaction in relation to cytology: possible implications for
cervical-cancer screening. In: N. Mu~ioz,F.X. Bosch, K.V. Shah and A.
Meheus (eds.), The epidemiology of human papillomavirus and cervical
cancer, IARC Scientific Publication 119, pp. 271-281, IARC, Lyon
(1992).
REMMINK, A.J., WALBOOMERS, J.M.M., HFLMERHORST,TH.J.M.,
VOORHORST, F.J., ROZENDAAL, L.. RISE, E.K.J., MEIJER,C.J.L.M.
and KENEMANS, P., The presence of persistent high-risk HPV geno-
types in dysplastic cervical lesions is associated with ro ressive
disease: natural history up to 36 months. Inf. J. Cancer,
(1995).
A, fO6--311
SCHIFFMAN, M.H. and BRINI'ON,L.A., The epidemiology of cervical
carcinogenesis. Cancer, 76,1888-1901 (1995).
SHERMAN, M.E., SCHIFFMAN, M.H., LORINCZ,A.T., MANOS,M.M.,
SCOTT,D.R., KURMAN, R.J., KIVIAT,N.B., STOLER,M., GLASS,A.G.
and RUSH, B.B., Toward objective quality assurance in cervical
athology: correlation of cytopathologic diagnoses with detection of
E.igh-risk
' human papillomavirus types. Amer. J. clin. Parhol., 102,
182-187 (1994).
VAN OoH'rMARssEN, G.J. and HABBEMA,J.D.F., Epidemiological
evidence for age-dependent regression of pre-invasive cervical cancer.
Brit. J. Cancer, 64,559-565 (1991).
VANRANST,M., KAPLAN,J.B. and BURK,R.D., Phylogenetic classifica-
tion of human papillomaviruses: correlation with clinical manifesta-
tions. J. gen. Krol., 73,2653-2660 (1992).
VOOIJS, G.P., De advisering bij afwijkende bevindingen van cytologisch
onderzoek van de cervix uteri. Ned. Tijdschr. Geneesk., 131,1662-1663
(1987).
WALBOOMERS, J.M.M., DE RODAHUMAN,A.M., SNIJDERS, P.J.F.,
STEL,H.V., RISSE,E.K.J., HELMERHORST, TH.J.M., VOORHORST, F.J.
and MEIJER,C.J.L.M., Detection of genital human papillomavirus
infections. Critical review of methods and revalence studies in
relation to cervical cancer. In: P.L. Stern an8M.A. Stanley (eds.),
Human apillomaviruses and cervical cancer. Biology and immunology,
pp. 4 1 - 4 , Oxford University Press, Oxford (1994).