Servei d’Epidemiologia i Registre del Cancer, Institut Catal#{224} d’Oncologia,
Hospital Duran i Reynals. Ciutat Sanit#{224}ria i Universit#{224}ria de Bellvitge. Autovia
de Castelldefels km 2.7 L’Hospitalet de LI. 08907 Barcelona, Spain [V. M.,
F. X. B.. S. d-S., P. V.1; Laboratori de BioestadIstica i Epidemiologia, Facultat
de Medicina, Universitat AutOnoma de Barcelona, Barcelona, Spain [V. M.J;
International Agency for Research on Cancer, Lyon, France [N. M.];
Delegaci#{243}n Territorial de Bienestar Social. Salamanca, Spain [L. C. G.j;
Universidad del Valle, Cali, Colombia [L. TI; Hospital Universitario del
RocIo, Sevilla, Spain [M. G.l: Registro de Cancer de Euskadi, Departamento
de Sanidad y Consumo. Gobierno Vasco, Vitoria-Ga.steiz, Spain [I. I.J;
ConsejerIa de Sanidad y Ausuntos Sociales, Murcia, Spain [C. NI; Registro de
Cancer, DiputaciOn General de Arag#{243}n,Zaragoza, Spain [A. V.1; Hospital
Santa Caterina. Girona. Spain [P. V.1; Departamento de Salud, Gobiemo de
Navarra, Pamplona, Spain IN. Al; and School of Hygiene and Public Health,
The Johns Hopkins University. Baltimore, Maryland 21218 [K. V. S.
Received 10/17/94: revised 4/17/95; accepted 4/20/95
Financial support was received from the ARC, the European Community
[CI 1-0371-F (CD)l. the Fondo de Investigaciones Sanitarias of the Spanish
Government (86/753. 87/1513, 88/2t)49, and 90/0901) and by the Spanish Mi-
nisterio de Educacion y Ciencia (PR94-053). We acknowledge the support re-
ceived in Spain from the Health Departments of the Autonomous Communities
and from the local administrators in the areas that participated in the study.
2 To whom all requests for reprints should be addressed.
Vol. 4, 459- 467, July/August 1995 Cancer Epidemiology, Biomarkers & Prevention 459
Risk Factors for Progression of Cervical Intraepithelial Neoplasm
Grade III to Invasive Cervical Cancer1
V.Moreno,2 N. Muiioz, F. X. Bosch, S. de Sanjos#{233},
L. C. Gonzalez, L. Tafur, M. Gili, I. Izarzugaza,
C. Navarro, A. Vergara, P. Viladiu, N. Ascunce, and
K. V. Shah
Abstract
Data from four case-control studies on invasive cervical
cancer and on cervical intraepithelial neoplasia grade III
(CIN III) that were concurrently conducted in Spain and
Colombia were used to look for factors that might favor
the progression from CIN III to the invasive stage. These
studies were compared in two ways: a case-case
comparison and an estimation of the ratio of odds ratios
with the use of a special logistic regression model that
took into account the different design of each study and
possible confounding factors. Variables studied were
human papillomavirus status, viral load, viral types,
sexual behavior, sexually transmitted diseases,
reproductive patterns, oral contraceptives, and smoking.
Both CIN III and invasive cervical cancer have a very
similar profile of risk factors and none of them was
different in a consistent way to suggest a role in the
progression from CIN III to invasive cervical cancer.
Some methodological problems such as cohort-specific
differences and some selection biases could be adjusted
for with a careful statistical analysis. Other problems
derived from the cross-sectional nature of the design are
unavoidable and should be considered in the
interpretation of the results.
Introduction
Several studies have been carried out to assess the natural
history of CIN3 in relation to its progression to ICC (1-3), but
little is known about the factors that determine such progres-
sion. The ideal design to address the problem would be a cohort
study with ICC as an end point, but ethical reasons make such
studies impossible. Any attempt at a longitudinal design must
end the follow-up when CIN III lesions (i.e., severe dysplasia
on carcinoma in situ) are found and treated. As an alternative,
data from cross-sectional or case-control studies may prove
useful to look for factors that may determine the progression of
CIN lesions by comparing ICC cases with CIN III cases.
We had the opportunity to explore this possibility using
data from four case-control studies, two on ICC and another
two on CIN III that were conducted in Spain and in Colombia
concurrently. These two countries have an 8-fold difference in
the incidence of invasive cervical cancer. However, the esti-
mated prevalence of CIN III among screened women appears to
be similar in both countries (4), which suggests that the differ-
ence in incidence of invasive cancer could be due to a different
rate of progression of CIN III or to very different screening
strategies. In this report, we compare the ICC and the CIN III
studies for each country, to look for differences in risk factors
that could be interpreted as progression factors. We also ad-
dress the difficulties in the interpretation of this comparison due
to the nonlongitudinal design of the studies.
Materials and Methods
Study Population
The main results of these studies, concerning the role of HPV
and other risk factors, have been published elsewhere (5-8).
The design and the methods used have been described in detail
in the main papers. In brief, the field work was conducted in
nine provinces in Spain (Alava, Girona, Guipuzcoa, Murcia,
Navarra, Salamanca, Sevilla, Vizcaya, and Zaragoza) from
June 1985 to December 1987 and in the city of Cali in Colom-
bia from June 1985 to December 1988.
Cases
Eligible cases were, for the ICC studies, incident histologically
confirmed invasive squamous cell carcinomas of the cervix
identified among residents in the predefined study areas. All
cases of CIN III, (including severe dysplasia and in situ cancer)
diagnosed during the study period were included in the CIN III
studies. Cases were identified by periodic visits to all hospitals,
pathology laboratories, and screening clinics of each area.
3 The abbreviations used are: CIN, cervical intraepithelial neoplasia; ICC, inva-
sive cervical cancer; HSV, herpes virus; HPV, human papillomavirus; 5Th.
sexually transmitted disease; OC, oral contraceptive; OR. odds ratio.
on December 8, 2021. © 1995 American Association for Cancer Research. cebp.aacrjournals.org Downloaded from