ARTICLES

Prevalence of Human Papillomavirus in

Cervical Cancer: a Worldwide Perspective
F. Xavier Bosch, M. Michele Manos, Nubia Munoz, Mark Sherman,
Angela M. Jansen, Julian Peto, Mark H. Schiffman, Victor Moreno,
Robert Kurman, Keerti V. Shah, International Biological Study on
Cervical Cancer (IBSCC) Study Group*

Downloaded from http://jnci.oxfordjournals.org/ at Univ of Southern California on April 6, 2014

Background: Epidemiologic studies have shown that the
association of genital human papillomavirus (HPV) with
cervical cancer is strong, independent of other risk factors,
and consistent in several countries. There are more than 20
different cancer-associated HPV types, but little is known
about their geographic variation. Purpose: Our aim was to
determine whether the association between HPV infection
and cervical cancer is consistent worldwide and to inves-
tigate geographic variation in the distribution of HPV types.
Methods: More than 1000 specimens from sequential
patients with invasive cervical cancer were collected and
stored frozen at 32 hospitals in 22 countries. Slides from all
patients were submitted for central histologic review to con-
firm the diagnosis and to assess histologic characteristics.
We used polymerase chain reaction-based assays capable of
detecting more than 25 different HPV types. A generalized
linear Poisson model was fitted to the data on viral type and
geographic region to assess geographic heterogeneity.
Results: HPV DNA was detected in 93% of the tumors, with
no significant variation in HPV positivity among countries.
HPV 16 was present in 50% of the specimens, HPV 18 in
14%, HPV 45 in 8%, and HPV 31 in 5%. HPV 16 was the
predominant type in all countries except Indonesia, where
HPV 18 was more common. There was significant geo-
graphic variation in the prevalence of some less common
virus types. A clustering of HPV 45 was apparent in western
Africa, while HPV 39 and HPV 59 were almost entirely con-
fined to Central and South America. In squamous cell
tumors, HPV 16 predominated (51 % of such specimens), but
HPV 18 predominated in adenocarcinomas (56% of such
tumors) and adenosquamous tumors (39% of such tumors).
Conclusions: Our results confirm the role of genital HPVs,
which are transmitted sexually, as the central etiologic factor in
cervical cancer worldwide. They also suggest that most geni-
tal HPVs are associated with cancer, at least occasionally.
Implication: The demonstration that more than 20 different
genital HPV types are associated with cervical cancer has
important implications for cervical cancer-prevention
strategies that include the development of vaccines targeted
to genital HPVs. [J Natl Cancer Inst 87:796-802,1995]
Cervical cancer represents the second most common cancer in
women worldwide. Age-adjusted incidence rates vary from
about 10 per 100 000 per year in many industrialized nations to
more than 40 per 100 000 in some developing countries. Four of
five new cases are currently diagnosed in the developing parts
of the world (7).
Epidemiologic studies (2-5) have shown that the association
of human papillomavirus (HPV) with cervical neoplasia is
strong, independent of other risk factors, and consistent in
several countries.
More than 35 distinct HPV types are known to infect the
genital tract (6"), complicating the detection and distinction of
these agents. Twenty or more of these HPV types are cancer-as-
sociated. HPVs appear to represent the most common sexually
transmitted agent studied to date. In some populations, cross-
sectional studies of cytologically normal women (7-10) suggest
•F. X. Bosch and M. M. Manos contributed equally to the project and are to
be considered co-first authors.
Affiliations of authors: F. X. Bosch, V. Moreno, Servei d'Epidemiologia i
Registre del Cancer, Institut Catala d'Oncologia, Hospital Duran i Reynals,
CSUB, Autovia Castelldefels, km. 2.7, E-08907 L'Hospitalet del Llobregat, Bar-
celona, Spain.
M. M. Manos, A. M. Jansen, K. V. Shah, Department of Molecular Microbiol-
ogy and Immunology, School of Public Health, Johns Hopkins University, Bal-
timore, Md.
N. Mufloz, Unit of Field and Intervention Studies, International Agency for
Research on Cancer, Lyon, France.
M. Sherman, R. Kurman, Department of Pathology, Johns Hopkins Medical
Institutions, Baltimore.
J. Peto, Institute of Cancer Research, Belmont, Surrey, England.
M. H. Schiffman, Environmental Epidemiology Branch, National Cancer In-
stitute, Bethesda, Md.
Correspondence to: Nubia Mufioz, M.D., International Agency for Research
on Cancer, 150, Cours Albert Thomas, F-69372 Lyon Cedex 08, France.
See "Notes" section following "References."

796 ARTICLES Journal of the National Cancer Institute, Vol. 87, No. 11, June 7, 1995
that 20%-40% of sexually active young women have detectable
HPV infection and that prevalence decreases with age. In most
studies, HPV 16 has been found to be the most prevalent HPV
type in cytologically normal women, women with cervical in-
traepithelial neoplasia (CIN), and women with cervical cancer.
There is, at least within the United States, some geographic
and ethnic variation in the HPV types detected {10-13), but less
is known about the international variation of HPV types in cer-
vical cancer. Moreover, it is debatable whether HPV-negative
cervical cancer is a biologic entity that is etiologically unrelated
to HPV, an artifact attributable to limitations of current detec-
tion methods, or the result of loss of HPV DNA during the
evolution of the tumor.
Through an extensive multinational collaboration, we or-
ganized the International Biological Study on Cervical Cancer
(IBSCC) to further our understanding of the relationship be-
tween HPV and cervical cancer. We sought to determine
whether the high prevalence of HPV in cervical tumors is con-
sistent worldwide and whether HPV types associated with can-
cer vary geographically. For that purpose, we used polymerase
chain reaction (PCR)-based HPV assays that afforded the
specific detection of more than 25 different genital HPV types.
This information is essential to the development of vaccina-
tion strategies to curb cervical cancer. Furthermore, the recogni-
tion of a sexually transmitted etiologic agent has extensive
implications for prevention. We report here on the prevalence of
HPV in our survey of specimens from patients with invasive
cervical cancer, mostly from high-incidence countries.
Patients and Methods
The field study took place in 32 hospitals from 22 countries between June
1989 and June 1992. Patients with invasive cervical cancer (sequential cases)
were identified by the following criteria: 1) clinical and histologic confirmation
of the diagnosis, 2) one to four frozen biopsy specimens or tissue specimens, 3) a
10-mL serum sample, and 4) the completion of an epidemiologic questionnaire.
Local pathologists were asked to provide a stained histologic slide for diagnostic
confirmation and histologic typing. Biopsy specimens were obtained from 1050
patients and were kept frozen at -20 "C to -70 'C without any additives. Instruc-
tions were given to avoid contamination during the collection of specimens.
Stage of the disease was originally coded according to International Federa-
tion of Gynecology and Obstetrics (FIGO) (14) and/or International Union
Against Cancer (UICC) (15) staging schemes. Whenever formal staging had not
been performed in the field, the gynecologist was asked to grade the disease as
local, regional, or disseminated. Stage I was grouped with local, stages II and III
with regional, and stage FV with disseminated.
Informed, written consent was obtained from all participants, and the study
protocol was cleared by the International Agency for Research on Cancer
(IARC) and by local ethic committees.
Histology Review
All submitted histologic slides or slides prepared from the submitted frozen
tissue were reviewed by one of us (M. Sherman) to establish the histologic type
and grade of differentiation, with uniform criteria to avoid interobserver varia-
tion. Equivocal and unusual specimens were examined jointly with a second
pathologist (R. Kurman). The following specimens were included and recorded
by original histologic type: 49 from patients for whom slides were not submitted
or were of poor quality and 85 from patients with disease not confirmed as in-
vasive carcinoma when reviewed (76 were classified as CIN III and nine slides
did not contain tumor tissue). It is likely that, for a substantial proportion of the
76 cases classified as CIN III, the slides that were sent for review were different
from those on which the original diagnosis was made or that the biopsy
specimens were taken from in situ cancer adjacent to the invasive tumor.
Journal of the National Cancer Institute, Vol. 87, No. 11, June 7, 1995
Patients for whom histologic confirmation of invasive cancer was not available
in the medical records and for whom the review of the slides did not confirm the
presence of cancer were ultimately excluded from the study.
HPV Detection and Typing
Crude DNA samples were prepared from the tumor specimens for subsequent
HPV DNA detection. Each tumor was thawed slightly, placed in a sterile dish,
and liberally rinsed with sterile saline to remove blood and debris. When pos-
sible, presumptive tumor tissue was identified macroscopically and used for
processing. A 2- to 3-mm3 segment was excised from the tumor, minced to a
pulp, and then proteinase treated (16). Of each prepared sample, 2.5% was used
in the HPV DNA analysis. All utensils were disposable and were discarded im-
mediately after each use (single use for each specimen). Standard methods to
avoid and monitor for contamination were used throughout the laboratory
analysis (17).
A widely used PCR DNA amplification-based method was employed for the
detection and typing of HPV, and 26 type-specific probes were utilized
(16,18,19). Primers for a fragment of p^globin gene served as an internal control
to assess the sufficiency of each specimen for amplification. All specimens were
prepared and analyzed in one laboratory (M. M. Manos and K. V. Shah), follow-
Downloaded from http://jnci.oxfordjournals.org/ at Univ of Southern California on April 6, 2014
ing standard procedures. Positive controls showed that the assay detected fewer
than 25 copies of HPV 16 per reaction. None of the negative controls (one
human kidney tissue fragment per 25 tumors analyzed) revealed HPV DNA,
suggesting that laboratory contamination was not common. Quality control (in-
ternal reproducibility) specimens (one masked, repeated tumor specimen per 25
tumors analyzed) showed perfect agreement, suggesting both reliability and lack
of contamination.
Specimens in which HPV was not detected (126 cases) were subjected to
more extensive evaluation. Another portion of each frozen specimen was ex-
cised and bisected. One half was fixed, embedded, and used for histologic con-
firmation of cancer, and the adjacent material was processed as above for use in
HPV detection analysis. Negative (kidney) and positive (HPV detected in initial
analysis) specimens were included in the process.
Statistical Methods
A generalized linear Poisson model was fitted to the data on viral type and
geographic region. The independent model showed significant lack of fit, and
the standardized residuals were used to explore the region and HPV-type com-
binations that were responsible for the heterogeneity. Fisher's exact test was per-
formed, collapsing the data to 2 x 2 tables on the cells where residuals were
greater than 1.96. Two-sided P values were calculated, and P<.005 was con-
sidered significant to take into account the multiple comparisons.
The chi-squared test was used to assess statistical significance of differences
in prevalence of HPV types by histologic type of cervical cancer.
Results
Collaborators from 22 countries participated in the study.
Frozen biopsy material for virology testing was received for
1050 patients. Fifteen samples were excluded because the diag-
nosis could not be confirmed, and all 27 from one hospital were
excluded because sample contamination during collection was
suspected. The remaining 1008 samples were tested by PCR.
Twenty-seven specimens were not sufficient for PCR. The over-
all prevalence of HPV in the remainder was 87.2% (855 of 981).
Frozen material from the 126 women with HPV-negative dis-
ease was recut for histologic review and further PCR analysis.
Eleven were HPV positive on retesting. Forty-nine samples
were excluded because no tumor tissue was seen (46 samples)
or were not PCR sufficient (three samples). All analyses were
based on specimens from the remaining 932 women, of whom
866 (92.9%) were HPV positive either initially or on retesting.
The mean age of the patients (±SD) was 47.8 years (±12.9
years), ranging from 33.9 years (±11.4 years) in the African
ARTICLES 797
 region to 56.5 years (±14.3 years) in the southern European
region. The mean number of cases included per country was
47.0 (range, 12-123) (Table 1).
The diagnosis of invasive cervical carcinoma was confirmed
in 798 (85.6%) of 932 of the included patients, based on review
of the slides provided by the participant institutions. For the
remaining 134 patients, the histologic data were extracted from
the medical records as indicated in the "Patients and Methods"
section. The distribution of disease was as follows: squamous
cell carcinoma—881 (94.5%), adenocarcinoma—25 (2.7%),
adenosquamous—18 (1.9%), and other or unknown—eight
(0.9%). Based on the degree of nuclear and architectural abnor-
malities and the proportion of cells showing cytoplasmic dif-
ferentiation (squamous and glandular), disease was graded as
follows: well differentiated—82 (10.8%), moderately differen-
tiated—405 (53.1%), and poorly differentiated—275 (36.1%);
this information was missing for the remaining 170 (18.2%)
HPV 18-related group (HPV 18, 39, 45, 59, and 68) represented
27.3%.
Twenty different viruses were detected by hybridization with
type-specific probes at least once each in the tumor collection.
HPVs 40, 42, 53, 54, and 66 and papillomavirus novel PAP 155
(20) were not identified in any of the specimens. HPV types 6
and 11 were each found once in the collection.
Two different viruses were detected in 36 (4%) of the
specimens, and these are counted twice in Tables 2 and 3.
Double infections did not cluster geographically, and there was
no relationship between double infections and histologic type,
grade of differentiation, or stage of the disease (data not shown).
Twelve specimens harbored HPVs that could not be typed
with our battery of 26 type-specific probes. By restriction frag-
ment length polymorphism analysis and DNA sequencing of the
amplification products, 10 specimens appeared to be either pre-
viously identified types for which we did not probe (e.g., HPV

Downloaded from http://jnci.oxfordjournals.org/ at Univ of Southern California on April 6, 2014

patients.
The overall HPV prevalence was 92.9% (95% confidence in-
terval 91.1-94.5) and ranged from 75%-100% by country
(P value for heterogeneity = .57) (Table 1). Of the 902 viral
types identified, HPV 16 accounted for 51.5%. HPVs 16, 18,31,
and 45 comprised 79.8% of the viruses found (Table 2). The
HPV 16 phylogenetic group (6) (including HPV 16, 31, 33, 35,
52, and 58) represented 66.5% of the viruses found, and the
64) or variants that would not have been detected by hybridiza-
tion. Two specimens, one from Argentina and one from Cuba,
contained a novel HPV sequence designated IS39 (21).
Table 3 shows the relationship between HPV type and histol-
ogy. With the exception of 58 cases (those with no slides or
poor-quality slides), the histologic assessment made on review
was used in these analyses. HPV 16 or related viruses accounted
for 68% of the viral types found in squamous cell rumors; HPV

Table 1. HPV prevalence of major HPV types by country

Specimens adequate for HPV testing, No. (%)*

total
Total HPV Any HPV
No. of
Geographic region patients No. negative positive HPV 16 HPV 18 HPV 31 HPV 45

Africa
Algeria 41 12 3 (25.0) 9 (75.0) 4 (33.3) 5(41.7) 0 (0.0) 0 (0.0)
Benin 12 6 1 (16.7) 5(83.3) 3 (50.0) 1 (16.7) 0 (0.0) 0 (0.0)
Guinea 21 18 0 (0.0) 18(100.0) 7 (38.9) 1 (5.6) 0 (0.0) 6 (33.3)
Mali 59 58 6(10.3) 52 (89.7) 20 (34.5) 7(12.1) 4 (6.9) 10(17.2)
Uganda 48 43 3 (7.0) 40 (93.0) 23 (53.5) 7(16.3) 1 (2.3) 2 (4.7)
Tanzania 51 49 6(12.2) 43 (87.8) 22 (44.9) 12 (24.5) 0 (0.0) 5(10.2)
Central and
South America
Argentina 61 57 3 (5.3) 54 (94.7) 34 (59.6) 8(14.0) 3 (5.3) 3 (5.3)
Bolivia 56 49 4 (8.2) 45(91.8) 17 (34.7) 2(4.1) 13 (26.5) 4 (8.2)
Brazil 49 46 6(13.0) 40 (87.0) 24 (52.2) 4(8.7) 2(4.3) 2 (4.3)
Chile 85 80 6(7.5) 74 (92.5) 36 (45.0) 4 (5.0) 7 (8.8) 7 (8.8)
Columbia 42 38 2(5.3) 36(94.7) 20 (52.6) 3 (7.9) 3 (7.9) 2 (5.3)
Cuba 51 45 3(6.7) 42 (93.3) 26 (57.8) 3 (6.7) 2 (4.4) 3 (6.7)
Panama 80 73 5 (6.8) 68 (93.2) 34 (46.6) 11(15.1) 2(2.7) 7 (9.6)
Paraguay 123 117 7 (6.0) 110 (94.0) 64 (54.7) 13(11.1) 3 (2.6) 9 (7.7)
Southeast Asia
Indonesia 52 47 2(4.3) 45 (95.7) 15(31.9) 23 (48.9) 0 (0.0) 4 (8.5)
Philippines 26 24 1 (4.2) 23 (95.8) 11 (45.8) 2 (8.3) 0 (0.0) 2(8.3)
Thailand 27 27 0 (0.0) 27(100.0) 16(59.3) 6(22.2) 1 (3.7) 2(7-4)
North America
Canada 49 46 3 (6.5) 43 (93.5) 27 (58.7) 8(17.4) 2 (4.3) 5 (10.9)
United States 13 11 1(9.1) 10 (90.9) 6 (54.5) 1 (9.1) 1 (9.1) 3 (27.3)
Europe
Germany 17 17 1 (5.9) 16(94.1) 13 (765) 0 (0.0) 0 (0.0) 0 (0.0)
Poland 25 23 0 (0.0) 23 (100.0) 18 (78.3) 4(17.4) 2(8.7) 0 (0.0)
Spain 47 46 3 (6.5) 43 (93.5) 25 (54.3) 3(6.5) 3 (6.5) 2 (4.3)

Total 1035 932 66 a n 866 (92.9) 465 (49.9) 128(13.7) 49 (5.3) 78 (8.4)

*Histologically confirmed and PCR sufficient.

798 ARTICLES Journal of the National Cancer Institute, Vol. 87, No. 11, June 7, 1995
 Table 2. Prevalence of individual HPV types by geographic region

Region, No. (% of total specimens)

Central and Southeast North

Type* Africa South America Asia Europe America Total

HPV 16 and
16 related
16 79 (42.5) 255 (50.5) 42 (42.9) 56(65.1)t 33 (57.9) 465 (49.9)
31 5(27) 35 (6.9) 1 (1.0) 5 (5.8) 3 (5.3) 49 (5.3)
33 5(27) 18(3.6) 2 (2.0) 1 (1.2) 0 (0.0) 26 (2.8)
52 4 (2.2) 16(3.2) 2 (2.0) 3 (3.5) 0 (0.0) 25 (2.7)
58 5(2.7) 11 (2.2) 2 (2.0) 1 (1.2) 0 (0.0) 19(2.0)
35 4 (2.2) 10(2.0) 1 (1.0) 1 (1.2) 0 (0.0) 16(1.7)
HPV 18 and
18 related
18 33(17.7) 48 (9.5)§ 31 (31.6)t 7(8.1) 9(15.8) 128(13.7)
45 23 (12.4)4: 37(7.3) 8 (8.2) 2 (2.3) 8(14.0) 78 (8.4)
59 0 (0.0) 14(2.8)t 1 (1.0) 0 (0.0) 0 (0.0) 15(1.6)
39 0 (0.0) 13(2.6)t 1 (1.0) 0 (0.0) 0 (0.0) 14(1.5)
68 4 (2.2) 2 (0.4) 1 (1.0) 3 (3.5) 1 (1.8) 11(1.2)

Downloaded from http://jnci.oxfordjournals.org/ at Univ of Southern California on April 6, 2014

Other
56 6 (3.2) 3 (0.6)1 3(3.1) 2 (2.3) 2(3.5) 16(1.7)
51 2(1.1) 5(1.0) 0 (0.0) 0 (0.0) 0 (0.0) 7 (0.8)
PAP 238a 1 (0.5) 4 (0.8) 0 (0.0) 1 (1.2) 0 (0.0) 6 (0.6)
W13b 1 (0.5) 1 (0.2) 4(4.1)t 0 (0.0) 0 (0.0) 6 (0.6)
26 0 (0.0) 4(0.8) 0 (0.0) 0 (0.0) 0 (0.0) 4 (0.4)
55 0 (0.0) 2 (0.4) 0 (0.0) 0 (0.0) 0 (0.0) 2 (0.2)
PAP 291 1 (0.5) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 1(0.1)
6 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 1 (0.1)
11 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0) 0 (0.0) 1(0.1)
Undetermined 2(1.1) 8(1.6) 0 (0.0) 1 (1.2) 1 (1-8) 12(1.3)
HPV negative 19(10.2) 36(7.1) 3(3.1) 4(4.7) 4(7.0) 66(7.D
Total HPV positives 175 488 99 83 57 902
Total No. of 186(100) 505 (100) 98 (100) 86(100) 57(100) 932 (100)
specimens

•HPV types 40, 42, 53, 54, 66 and PAP 155 were not identified. Double infections were counted twice. Samples that were HPV negative were excluded from
statistical analysis.
tSignificant increase (/><.0O5).
^Significant increase (/><.05).
§Significant decrease (V<.005).

Table 3. Prevalence of HPV type by stage of the disease and histologic tumor characteristics4

Histologic type, No. Histologic differentiation degree, No. Clinical stage, No.
(% of total specimens) (% of total specimens) (% of total specimens)

Adeno- Adeno- Local Regional Disseminated

Type Squamous carcinoma squamous Well Moderate Poor orTl orT2-T3 orT4-M

HPV 16 451(51.2) 7 (28.0) 3(16.7) 46(56.1) 199(49.1) 126(45.8) 105 (52.2) 301 (48.0) 16(57.1)
HPV 16 131 (14.9) 0 (0.0) 2(11.1) 9(11.0) 66(16.3) 34(124) 28(13.9) 92(14.7) 3 (10.7)
related
HPV 18 107(12.1) 14(56.0) 7 (38.9) 6(7.3) 59(14.6) 49(17.8) 32(15.9) 86(13.7) 2(7.1)
HPV 18 109(12.4) 3(12.0) 5 (27.8) 8 (9.8) 46(11.4) 47(17.1) 20(10.0) 79(12.6) 5 (17.9)
related
Other 55 (6.2) 0 (0.0) 0 (0.0) 10(122) 24(5.9) 12(4.4) 11(5.5) 39 (6.2) 3(10.7)
HPV negative 64(7.3) 1 (4.0) 1 (5.6) 7(8.5) 26 (6.4) 18(6.5) 13(6.5) 47 (7.5) 2(7.1)
Total HPV 853 24 17 79 394 268 196 597 29
positives
Total No. of 881 (100) 25(100) 18(100) 82(100) 405 (100) 275 (100) 201 (100) 627(100) 28(100)
specimens
x2= 61.2
2 2
/> = .001 X = 203 P = .03 Z = 5.3 P = .87

•Double infections were counted twice. Samples that were HPV negative and for which clinical data were unknown were excluded from statistical tests.

Journal of the National Cancer Institute, Vol. 87, No. 11, June 7,1995 ARTICLES 799
 18 and related viruses accounted for 71% of the viral types
found in adenocarcinomas and 71% in adenosquamous car-
cinomas. HPV 18 and related types were also associated with
degree of differentiation. Their prevalence showed an increasing
trend with decreasing degree of differentiation, and the associa-
tion persisted when adjusted for histology. No relationship was
observed between viral type and clinical stage of disease. When
the above analyses were repeated including only the cases con-
firmed on review, the direction and strength of the associations
remained unchanged.
The distribution of HPV types varied geographically. Table 2
and Fig. 1 show the HPV-specific prevalences by geographic
region. HPV 16 was the predominant type in all countries except
Indonesia, where HPV 18 accounted for more than 50%. This
predominance of HPV 18 was not explained by tumor type be-
cause it persisted when the analysis was restricted to squamous
cell tumors. In Mali (western Africa), HPV 45 was in significant
excess and this clustering of HPV 45 was also suggested in the
collection from the adjacent country of Guinea. Most of the
HPV 39 (13 of 14 specimens) and HPV 59 (14 of 15 specimens)
were found in Central and South America. In three (12.5%) of
the 24 specimens from the Philippines, novel type W13B (20)
was present.
Characteristics of the 66 (7.1%) patients in whom no HPV
was identified were compared with those that were conclusively
HPV positive. The only variable that appeared to be associated
with no detection of HPV was the presence of inflammatory
signs in the histologic specimen (not shown). Neither the clini-
cal appearance of the lesion (ulcerated, necrotic, bulky, or
bloody) nor the presence of concurrent local infection was re-
lated to the detection of HPV.
Discussion
This international survey of invasive cervical cancer revealed
that 92.9% of the analyzable tumors contained HPV DNA. The
HPV prevalence was remarkably homogeneous among the
800 ARTICLES
countries, ranging from 75.0% to 100%. These results are com-
patible with the hypothesis that only a small proportion of in-
vasive cervical cancer cases do not contain HPV DNA. As the
techniques to detect viral DNA have improved, the reported
prevalence among cervical cancer patients has increased
dramatically (22). Whether our results are contingent on the
HPV detection methods remains to be established. Consistent
with this idea, analysis of the 66 HPV-negative specimens using
additional HPV detection methods (e.g., other primers) suggests
that fewer than 5% of cervical cancers are truly HPV negative.
The prevalence of HPV in the general population has been es-
timated in case-control studies conducted in six of the countries
included in our study. The prevalence in control subjects with
the same age distribution as the cervical cancer patients was 5%
in Spain and between 13% and 20% in Thailand, the Philip-
pines, Paraguay, Brazil, and Colombia. Prevalence in western
Europe and the United States is generally less than 10% at 40
years of age or older (2324).
Downloaded from http://jnci.oxfordjournals.org/ at Univ of Southern California on April 6, 2014
Together with data from case-control studies showing that
HPV represents the central risk factor for cervical neoplasia (2-
5) and that HPV infection precedes the development of CIN
lesions (25), the IBSCC results demonstrate that HPV confers
the major attributable risk for the estimated 500 000 cases of
cervical cancer developing per year worldwide.
Although the IBSCC collection did not include some regions
of the world, such as central Asia and India, this study provides
the most extensive global view of HPV in cervical cancer ob-
tained to date. The countries that participated in the study
belong to 10 of the 18 regions of the world for which estimates
of the incidence of cancer have been recently provided (26). Al-
though, in each country, the size of the study is limited and the
cases can hardly be claimed to be representative, regions with
extreme differences in the incidence of cervical cancer are in-
cluded and no obvious selection bias within centers has been
identified.
This large, multicenter study was further strengthened by
centralized HPV detection performed in one expert research
Fig. 1. HPV type-specific distribu-
tion among HPV-positive individuals
are given as percentages of total
HPVs identified for the following
regions: Europe, North (N) America,
Central and South (C-S) America,
Africa, and Southeast (SE) Asia.
Journal of the National Cancer Institute, Vol. 87, No. 11, June 7,1995
laboratory and by histologic review conducted by one expert
pathologist. In our final analyses, we chose to accept HPV-nega-
tive results only from specimens with adjacent, confirmed tumor
tissue in order to avoid false-negative results. Our prevalence
estimate (93%) may therefore be slightly inflated because the
only restrictions for HPV-positive specimens were adequate
diagnostic confirmation and PCR sufficiency. The overall prev-
alence was 87.2% when all 126 HPV-negative specimens were
included, but this percentage is certainly too low because 36.5%
of these specimens contained no tumor material.
Despite suboptimal collection, storage, and transport condi-
tions in many facilities, a surprising majority (more than 95%)
of the specimens yielded material that was sufficiently intact for
PCR-based HPV detection. During preparation of the specimens
for HPV detection, great care was taken to remove blood,
debris, or obviously necrotic tissue as well as to macroscopical-
ly identify tumor material. We believe that these procedures
functioned very well, as reflected by the high proportion of
specimens that were found to be adequate for analysis.
The distribution of HPV types found in tumors differed some-
what by geographic region. These observed geographic varia-
tions require confirmation in future studies because, given the
number of HPV types detected and countries examined, it was
not possible to determine if variation in the detection frequency
was significant for less common types. Of particular interest is
whether these HPV type distributions in cancer reflect prevalen-
ces of these viruses in cytologically normal women. Alternative-
ly, specific HPV types may pose different relative risks for
cancer in different geographic regions. Recent evidence of a
linkage between specific immune response haplotypes (HLA
class II) and HPV type-specific risk of cervical cancer (27) sug-
gests that host genetics may need to be studied in concert with
HPV typing to fully understand geographic variation in type-
specific HPV prevalence.
The overall HPV prevalence did not vary between histologic
tumor types, suggesting an etiologic consistency among squa-
mous cell tumors, adenocarcinomas, and adenosquamous car-
cinomas.
As previously reported by' other investigators (28J29), HPV
18 was the predominant type found in adenocarcinoma speci-
mens. We also found HPV 18 to predominate significantly in
adenosquamous tumors; this predominance has not been pre-
viously described for this rare tumor type.
Our results suggest that most genital HPVs are associated
with cancer, at least occasionally. The identification of only one
novel virus, to date, from this large and diverse collection (21)
suggests that few cancer-associated genital HPVs remain to be
identified. Further studies are necessary to confirm the presence
of types that were detected very rarely (HPVs 6 and 11 and PAP
291) and to clarify which types are present in tumor cells in
doubly infected specimens. For example, in situ hybridization
studies are in progress to determine whether the virus types
detected by our methods are localized to tumor cells.
Diagnostic, therapeutic, and vaccination strategies must con-
sider the complexity posed by the existence of at least 20 dif-
ferent cancer-associated HPVs. Further efforts are necessary to
determine whether less prevalent types [often termed "inter-
mediate risk" (30)] confer cancer risks to an individual that are
Journal of the National Cancer Institute, Vol. 87, No. 11, June 7, 1995
equivalent to those posed by infection with the most common
types (e.g., HPVs 16 and 18, which are termed "high-risk"). Al-
though our understanding of cervical cancer has been en-
lightened by defining the role of HPV, it may be complicated by
the intricacies of this diverse group of viruses (6).
Cervical cancer is the second most common cancer in women
worldwide. It takes its greatest toll in the developing world,
where it is the most common cancer among women and where
health care resources are limited. The knowledge that a
heterogeneous family of sexually transmitted agents confers the
large majority of risk for this disease has far-reaching implica-
tions for prevention strategies, including the development of ef-
fective vaccines.
References
(/) Parkin DM, Muir CS, Whelan SL, et al: Cancer incidence in five continents.
Downloaded from http://jnci.oxfordjournals.org/ at Univ of Southern California on April 6, 2014
Comparablity and quality of data. IARC Sci Publ 120:45-173,1992
(2) Mufloz N, Bosch FX, de Sanjos^ S, et al: The causal link between human
papillomavirus and invasive cervical cancer a population-based case-con-
trol study in Colombia and Spain. Int J Cancer 52:743-749, 1992
(3) Bosch FX, Mufioz N, de Sanjose' S, et al: Human papillomavirus and cervi-
cal intraepithelial neoplasia grade m/carcinoma in situ: a case-control
study in Spain and Colombia. Caocer Epidemiol Biomarkers Prev 2:415-
422,1993
(4) Eluf-Neto J, Booth M, Mufioz N, et al: Human papillomavirus and in-
vasive cervical cancer in Brazil. BrJ Cancer 69:114-119, 1994
(5) Schiffman MH, Bauer HM, Hoover RN, et al: Epidemiologic evidence
showing that human papillomavirus infection causes most cervical in-
traepithelial neoplasia [see comment citations in Medline]. J Natl Cancer
Inst 85:958-964,1993
(6) Bernard HU, Chan SY, Manos MM, et al: Identification and assessment of
known and novel human papillomaviruses by polymerase chain reaction
amplification, restriction fragment length polymorphisms, nucleotide se-
quence, and phylogenetic algorithms [see comment citation in Medline].
J Infect Dis 170:1077-1085,1994
(7) Ley C, Bauer HM, Reingold A, et al: Determinants of genital human papil-
lomavirus infection in young women [see comment citations in Medline]. J
Natl Cancer Inst 83:997-1003,1991
(8) Melkert PW, Hopman E, van den Brule AJ, et al: Prevalence of HPV in
cytomorphologically normal cervical smears, as determined by the poly-
merase chain reaction, is age-dependent. Int J Cancer 53:919-923, 1993
(9) Bauer HM, Ting Y, Greer CE, et al: Genital human papillomavirus infec-
tion in female university students as determined by a PCR-based method
[see comment citation in Medline]. JAMA 265:472^77,1991
(10) Bauer HM, Hildesheim A, Schiffman MH, et al: Determinants of genital
human papillomavirus infection in low-risk women in Portland, Oregon.
Sex Transm Dis 20:274-278; 1993
(11) Becker TM, Wheeler CM, McGough NS, et al: Cervical papillomavirus in-
fection and cervical dysplasia in Hispanic, Native American, and non-
Hispanic white women in New Mexico. Am J Public Health 81:582-586,
1991
(12) Hildesheim A, Gravitt P, Schiffman MH, et al: Determinants of genital
human papillomavirus infection in low-income women in Washington,
D.C. Sex Transm Dis 20:279-285,1993
(13) Wheeler CM, Parmenter CA, Hunt WC, et al: Determinants of genital
human papillomavirus infection in cytologically normal women attending
the University of New Mexico student health center. Sex Transm Dis
20:286-289,1993
(14) Creasman W: New gynecologic cancer staging. Obstet Gynecol 75:287-
288,1990
(15) Hermanek P, Sobin LH, eds: TNM Classification of Malignant Tumours,
4th ed. Geneva: UICC, 1987
(16) Bauer H, Greer C, Manos M: Detection of genital human papillomavirus
infection using PCR. In Diagnostic Molecular Pathology: a Practical Ap-
proach (Herrington CS, McGee JO'D, eds). Oxford: Oxford Univ Press,
1992, pp 131-152
(17) Gravitt PE, Manos MM: Polymerase chain reaction-based methods for the
detection of human papillomavirus DNA. IARC Sci Publ 119:121-133,
1992
(IS) Manos MM, Ting Y, Wright DK, et al: The use of polymerase chain reac-
tion amplification for the detection of genital human papillomaviruses. In
ARTICLES 801
 Molecular Diagnostics of Human Cancer (Furth M, Greaves MF, eds).
Cold Spring Harbor, NY: Cold Spring Harbor Press, 1989, pp 209-214
(19) Hildesheim A, Schiffman MH, Gravitt PE, et al: Persistence of type-
specific human papillomavirus infection among cytologically normal
women [see comment citation in Medline]. J Infect Dis 169:235-240, 1994
(20) Manos MM, Waldman J, Zhang TY, et al: Epidemiology and partial
nucleotide sequence of four novel genital human papillomaviruses. J Infect
Dis 170:1096-1099,1994
(21) Peyton CL, Jansen AM, Wheeler CM, et al: A novel human papillomavirus
sequence from an international cervical cancer study. J Infect Dis
170:1093-1095,1994
(22) Mufioz N, Bosch FX: HPV and cervical neoplasia: a review of case-control
and cohort studies. IARC Sci Publ 119:255-266,1992
(25) Schiffman MH: Recent progress in defining the epidemiology of human
papillomavirus infection and cervical neoplasia. J Natl Cancer Inst 84:394-
398, 1992
(24) Mufioz N, Bosch FX, Deacon J, et al: HPV and cervical cancen a multi-
centre case-control study. Abstract presented at the 13th International
Papillomavirus Conference, Amsterdam, October 8-12,1994
(25) Koutsky LA, Holmes KK, Critchlow CW, et al: A cohort study of the risk
of cervical intraepithelial neoplasia grade 2 or 3 in relation to papil-
lomavirus infection. N Engl J Med 327:1272-1278, 1992
(26) Parkin DM, Pisani P, Ferlay J: Estimates of the worldwide incidence of
Notes
LBSCC Study Group: E. Alihonou, University Nationale du Benin, Cotonou,
Benin; S. Bayo, Institut National de Recherche en Sant6 Publique, Bamako,
Mali; H. Cherif Mokhtar, Centre Hospitalo-Universitaire, Setif, Algeria; S.
Chichareon, Prince of Songkla University, Hat-Yai, Thailand; A. Daudt, Hospi-
tal of the Clinics of Porto Alegre, Brazil; E. de los Rios, National Oncology In-
stitute, Panama; P. Ghadirian, Hfltel-Dieu de Montreal, Montreal, Canada; J. N.
Kitinya, Muhimbili Medical Centre, Dar es Salaam, Tanzania; M. Koulibaly,
Centre National d'Anatomie Pathologique, Conakry, Guinea; C. Ngelangel,
University of the Philippines, Manila, Philippines; L. M. Puig Tintore, Hospital
Clfnic I Provincial, Barcelona, Spain; J. L. Rios-Dalenz, Cancer Registry of La
Paz, La Paz, Bolivia; Sarjadi, Diponegoro University Medical Faculty,
Semarang, Indonesia; A. Schneider, Friedrich Schiller University, Jena, Federal
Republic of Germany; L. Tafur, University of Valle, Cali, Colombia; A. R.
Teyssie, National Institute of Microbiology, Buenos Aires, Argentina; P. A.
Rol6n, Registro Nacional de Patologfa Tumoral, Asuncion, Paraguay; M. Tor-
roella, National Cancer Institute, Havana, Cuba; A. Vila Tapia, Regional Clini-
cal Hospital, Ministry of Health, Concepci6n, Chile; H. R. Wabinga, Makerere
Medical School, Kampala, Uganda; W. Zatonski, M. Sklodowska-Curie
Memorial Cancer Centre, Warszawa, Poland; B. Sylla, P. Vizcaino, D. Magnin,
International Agency for Research on Cancer, Lyon, France; J. Kaldor, National
Centre in HIV Epidemiology and Clinical Research, Darlinghurst, Australia; C.

Downloaded from http://jnci.oxfordjournals.org/ at Univ of Southern California on April 6, 2014

eighteen major cancers in 1985. Int J Cancer 54:594-606, 1993
(27) Apple RJ, Erlich HA, Klitz W, et a l HLA DR-DQ associations with cervical
carcinoma show papiUomavirus-type specificity. Nat Genet 6:157-162, 1994
(28) Tase T, Okagaki T, Clark BA, et al: Human papillomavirus types and
localization in adenocarcinoma and adenosquamous carcinoma of the
uterine cervix: a study by in situ DNA hybridization. Cancer Res 48:993-
998,1988
(29) Wilczynski SP, Bergen S, WalkeT J, et al: Human papillomaviruses and
cervical cancer: analysis of histopathologic features associated with dif-
ferent viral types [see comment citation in Medline]. Hum Pathol 19:697-
704,1988
(30) Lorincz AT, Reid R, Jenson AB, et al: Human papillomavirus infection of
the cervix: relative risk association of 15 common anogenital types. Obstet
Gynecol 79:328-337,1992
Greer, Chiron Corporation, Emeryville, Calif.; C. Wheeler, University of New
Mexico, Albuquerque, N.M.
Supported in part by grants from the International Agency for Research on
Cancer (IARC), the Cancer Research Campaign of the U.K., the European Com-
munity [CI 1-0371-F (CD)], and by Public Health Service grant MA5623-41 (M.
M. Manos and K. V. Shah), National Cancer Institute, National Institutes of
Health, Department of Health and Human Services.
We thank Roche Molecular Systems (RMS) for the generous donation of
reagents for PCR. We acknowledge the many contributions of P. Gravitt and T.
Zhang (RMS) and thank A. C. Stewart and C. Peyton (University of New
Mexico) for essential and invaluable help in investigating potential novel
HPVs.
Manuscript received December 21, 1994; revised March 3, 1995; accepted
March 17, 1995.

The repository of the Biological Response Modifiers Program (BRMP), Division of

BIOLOGICALS Cancer Treatment (DCT), NCI, NIH, announces the availability of recombinant
human lymphokines IL-la, IL-lfJ, and IL-2; the monoclonal antibody 11B.11 against
mouse IL-4; and the monoclonal antibody 3ZD against human IL-1(5.
AVAILABLE Use of these materials is limited solely to in irivo and in vitro basic research studies
and is not intended for administration to humans.
FROM The lymphoklne materials are aliquoted in 100 ug amounts (>106 units) and are
available to investigators with peer-reviewed support However, manufacturers'
restrictions prohibit distribution of these materials to for-profit institutions or
THE commercial establishments.
The monoclonal antibodies are available to peer-reviewed investigators, for-profit
NATIONAL institutions or commercial establishments. The 1 IB. 11 antibody is available in either
3 or 20 mg vials. The 3ZD antibody is available in 5 or 20 mg amounts.
CANCE1 Investigators wishing to obtain any of these materials should send requests to:
Dr. Craig W. Reynolds
Biological Response Modifiers Program
INSTITU' NCI-FCRDC
Building 1052, Room 253
Frederick, MD 21702-1201
PAX: 301-846-5429
All requests should be accompanied by.
(1) A brief paragraph outlining the purpose far which materials are to be used, (2) the amount desired, (3) description of Investigator's
peer-reviewed support Recipients will be required to sign a Materials Transfer Agreement and to pay shipping and handling costs. Please
allow 4 to 6 weeks for dettvny.
NATIONAL CANCER INSnTUTE-
FREDERICK CANCER RESEARCH ft DEVELOPMENT CENTER

802 ARTICLES Journal of the National Cancer Institute, Vol. 87, No. 11, June 7,1995