Risk factors for cervical intraepithelial neoplasia recurrence after

Risk factors for cervical intraepithelial neoplasia recurrence after loop electrosurgical excision procedure in HIV-1 infected and non-infected women Maria Inês Miranda LIMA (1) Victor Hugo MELO (2) Celso Pedro TAFURI (1) Luiza Miranda LIMA (3) Angela Cristina Labanca ARAÚJO (1) Mark Drew Crosland GUIMARÃES (1) School of Medicine, Federal University of Minas Gerais, Belo Horizonte, Brazil (2) Belo Horizonte City Health Department (3) Medical Science School of Medicine

Background n CERVICAL CANCER: n n High incidence of Cervical Cancer in Brazil (22/100, 000 women in 2005) Cervical Intraepithelial Neoplasia (CIN) is a precursor of cervical cancer and is highly associated with HPV infection Early Dx of CIN can prevent new cases of CC (Cytology, Colposcopy, Biopsy) LEEP (Electrosurgical Excision Procedure) has been extensively used in Brazil

Background n AIDS epidemic in Brazil (up to 2005: 370, 000 cases) n n Women Heterosexual transmission Lower income and education HIV – Cervical cancer: n n n Higher incidence of CIN and CC among HIV positive women Evidence of higher incidence of recurrence of CIN among HIV positive women There is no published data in Brazil regarding recurrence of CIN comparing HIV positive and negative women

Objective Ø To assess factors associated with recurrence of cervical intraepithelial neoplasia after conization by LEEP in HIV infected and non-infected women

Methods Ø DYNAMIC COHORT STUDY Ø Population: Public Cervical Pathology Referral Service, Belo Horizonte, Brazil Referred from other public primary care units or HIV services Ø Screening criteria: Abnormal cytology (Bethesda, 1999) OR Normal cytology with Positive Schiller test n Colposcopy n Biopsy

Methods Ø Elegibility criteria for LEEP: q Age > 18 years old q Informed consent q Dx of Cervical Intraepithelial Neoplasia (CIN): HSIL lesions or persistent LSIL lesions for HIV negative women HSIL or LSIL for HIV positive women Ø Exclusions: q Pregnant women q Other Dx: Cervicitis, Invasive Cancer, Micro-invasive Cancer

Methods Ø Recurrence of CIN: q. First recurrence q. Cytology and Colposcopy If abnormal or normal with positive Schiller test Biopsy: HSIL lesions or persistent LSIL lesions

Methods Ø Statistical analysis: q. Cox Proportional Hazard Model (Univariate and Multivariate) q. Kaplan-Meier Survival Curves q. Variables of interest: HIV status, Histology, Glandular involvement, Margins, Number of sex partners, Age, Smoking Hx, HPV, Viral Load and CD 4+ Cell count.

Results n ( % ) or Mean SCREENED AND BIOPSIED EXCLUDED PARTICIPANTS 206 (100) 5( 2) 201 ( 98) BIOPSIED DURING FOLLOW-UP 73 ( 36) Cumulative incidence 40 ( 20) Incidence / 1, 000 women-months Mean number of visits Mean Time of follow-up (Median) 10. 2 4. 6 19. 4 (18. 6)

Results Variable Age (> 35 y. o. ) Number of lifetime partners (< 6) Smoking Hx HIV Positive Cytology (HSIL) Biopsy (HSIL) Histopathology (HSIL) Positive margins Glandular involvement n (%) 126 142 68 94 62 114 129 45 21 (63) (71) (34) (47) (31) (57) (64) (24) (10)

Results: Multivariate Analysis Variable HIV Status (Pos) Glandular involvement Positive Margin RH=Relative Hazard RH (95% CI) p-Value 3. 00 (1. 38 – 6. 48) 0. 005 3. 46 (1. 71 – 7. 01) 0. 000 2. 04 (1. 05 – 3. 98) 0. 035

Positive margins and glandular involvement Normal Duct Positive margin Glandular involvement

Conclusions n n n The incidence of CIN recurrence in this population is high. HIV infection, positive margins and glandular involvement are independent co-factors. More careful follow-up of these women is necessary, specially among HIV positive ones. Histopathology indicates HPV infection in over 95%. However, PCR is recommended for sub-typing and is currently under way. Adherence to ARVT and Viral load will be further explored among HIV positive women.