Objective
We sought to identify patient characteristics associated with cervical intraepithelial neoplasia (CIN) 3 in adolescents.
Study Design
A cross-sectional cohort study was done using a colposcopic database. Data collected included demographic and clinical information as well as cytologic and histologic results. Variables associated with CIN 3 in the univariable analysis ( P ≤ .10) were included in a multivariable logistic regression model.
Results
Of 614 adolescents evaluated, 48 (8%; 95% confidence interval, 6–10%) had CIN 3. No cancers were detected. In the univariable analysis, parity, ≥4 lifetime male sexual partners, and a history of sexually transmitted infections were associated with an increased risk of CIN 3. In the final model, only ≥4 sexual partners (compared to <4; adjusted odds ratio, 3.66; 95% confidence interval, 1.26–10.61) was significantly associated with the finding of CIN 3.
Conclusion
CIN 3 was infrequently encountered in adolescents, and those with multiple sexual partners were at increased risk for CIN 3.
Over the last decade, significant differences have emerged between the adolescent and adult populations in terms of human papillomavirus (HPV) infection, cervical intraepithelial neoplasia (CIN), and cancer risk. Guidelines for screening initiation, as a result, have undergone substantial modification. In 2002, the American Cancer Society published revised guidelines recommending that cervical cancer screening begin approximately 3 years after the initiation of sexual intercourse, but no later than age 21 years. In 2009, the American College of Obstetricians and Gynecologists endorsed an even less aggressive approach by recommending that screening not be initiated until age 21 years, regardless of when sexual initiation occurred.
These changes reflect findings that indicate both the higher prevalence and clearance of HPV infection in adolescent and young women compared to older women, as well as the increased likelihood among younger women of regression of mildly abnormal cytologic and histologic changes. Unlike their adult counterparts for whom colposcopy is indicated, adolescents with atypical squamous cells of undetermined significance (ASC-US), low-grade squamous intraepithelial lesion (LSIL), and CIN 1 are currently managed conservatively. Evidence also supports observation of CIN 2 in young women with satisfactory colposcopy, again based on significant regression rates for those followed expectantly. For adolescents and young women, then, the focus is on the detection and treatment of the cervical cancer precursor lesion, CIN 3.
Whether or not the natural history of cytologic abnormalities in adolescents is similar to adults remains unclear. For example, reported rates of CIN 3 among teens and young women with ASC-US and/or LSIL range from 6-16%. Compared to adult women, teens and young women with mildly abnormal cytology appear to be less likely to harbor CIN 3. Identification of potential risk factors for the development of CIN 3 among teens with cytologic abnormalities, and the degree to which they are similar or dissimilar to those historically seen in older populations, remains relevant as many adolescents will likely continue to be screened despite recent recommendations. The primary objectives of this study, therefore, are to estimate the prevalence of and risks associated with the findings of CIN 3 among a cohort of adolescents (defined as young women aged ≤20 years) referred to colposcopy for the evaluation of abnormal cervical cytology.
Materials and Methods
Beginning in 1999, women seen in the Colposcopy Clinic at Women and Infants Hospital in Providence, RI, have been entered into a colposcopic database. This database of 3137 patients was searched for adolescent girls referred for colposcopy from August 1999 through June 2007. Patients were included in the study if they met the age criteria defined above, had available histologic results, and had recorded demographic information on the number of lifetime male sexual partners, as this information served as the exposure variable for the purposes of the study. Pregnant adolescents were not excluded. The institutional review board approved the use of the database, and the medical record and computerized record review, for this study in October 2007.
The database includes information on a number of demographic variables including age, race/ethnicity, gravidity/parity, smoking history, human immunodeficiency virus status, use and type of birth control, history of sexually transmitted infections, age at first coitus, years since first coitus, and number of lifetime sexual partners. Cervical cytologic smears were obtained using thin-layer cytology and were generated from PreservCyt (Cytyc Corp, Marlborough, MA) samples by use of a semiautomated processor (ThinPrep 3000; Cytyc Corp). Colposcopic examinations were performed by resident physicians and nurse colposcopists under the guidance of attending physicians, and all histologic specimens (including cervical biopsies and/or endocervical samples) were placed in formalin prior to transport to pathology. Hospital cytotechnicians and pathologists evaluated all specimens and classified them according to 2001 Bethesda terminology or the CIN histologic grading system as appropriate. If a patient had >1 biopsy, the most severe diagnosis was selected for the purposes of analysis.
The detection of histologic high-grade disease (defined as ≥CIN 3) was determined, and descriptive statistics were then calculated and compared for young women by histologic grade. Differences in patient characteristics were compared by the Student t test (means) or Wilcoxon rank sum test (medians) for continuous variables, and χ 2 or Fisher’s exact test for categorical variables. All reported P values were 2-sided, with P < .05 considered statistically significant.
Multivariable logistic regression was used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for ≥CIN 3 vs ≥CIN 1. Continuous variables were dichotomized at their mean or median values and levels of categorical variables were combined to provide sufficient numbers for regression analysis. Characteristics associated with CIN 3 with P values of ≤ .1 in the univariable analysis were included in the final model along with number of lifetime male sexual partners (≥4 vs <4). The remaining variables were entered and retained in the model if the OR estimates changed by >20%. Data analysis was performed with software (SAS, version 9.1; SAS Institute, Cary, NC).
An initial sample size calculation assumed a probability of ≥CIN 3 of 20% among adolescents with ≥4 lifetime male sexual partners and 10% among those with <4 partners. With alpha set at 0.05 and beta at 0.20, we calculated a required sample size of 398. Given the likelihood of incomplete data (estimated at 20% of charts), the final sample size estimate increased to 478.
Results
In all, 841 adolescents were referred to colposcopy from August 1999 through June 2007. Of the 841, 644 (77%) had available histologic results (cervical biopsy and/or endocervical sampling). Of the 644, data were missing on the number of lifetime sexual partners for an additional 30 women (5%), leaving a final study size of 614 eligible young women ( Figure ). Adolescents who were not biopsied were more likely to have been pregnant and less likely to be current smokers compared to those who were biopsied (full data not presented).
The median age of the study cohort was 19 (range, 13–20) years ( Table 1 ). In general, the population was racially/ethnically diverse, unmarried, and on government assistance or uninsured. The mean age of first coitus was 15.3 years, the mean number of years since first coitus was 3.2, the median number of lifetime sexual partners was 4, and 51% of the population reported a history of ≥1 sexually transmitted infections, most commonly Chlamydia trachomatis. Approximately one third of the population was not using any form of contraception (188 or 31%, 49 of whom were pregnant at the time of colposcopy), followed by other hormonal contraception (159 or 26%), oral contraceptives only (142 or 23%), and barrier methods only (87 or 14%). The most common cytologic diagnoses were LSIL (330 or 56%) and ASC-US (184 or 30%). In all, 64 (10%) young women presented with high-grade squamous intraepithelial lesion (HSIL); 28 (5%) with atypical squamous cells, cannot exclude HSIL; and 3 (0.5%) with atypical glandular cells.
