Comparison of HPV detection in HIV-infected and HIV-uninfected Kenyan women with or without cervical dysplasia.

Document Type



Obstetrics and Gynaecology (East Africa)


Background: Cervical cancer, a malignancy caused by human papillomavirus (HPV) infection, is the most common malignancy in women living in sub-Saharan African countries including Kenya. HIV co-infection accelerates the natural course of cervical cancer. To determine the specific HPV type distribution in HIV-infected women compared to HIV-uninfected women, with and without evidence of cervical dysplasia.

Methods: Demographic information, behavioral data, and a cervical swab were collected from women 18 and 45 years of age, HIV-infected or HIV-uninfected, who presented for cervical cancer screening at Moi Referral and Teaching Hospital in Eldoret, Kenya. Women were triaged based on the presence or absence of cervical dysplasia. HPV testing was performed using the Roche Linear Array Assay. Results were compared between women with or without HIV co-infection and between those with or without cervical dysplasia, using Chi-square tests or Fisher’s exact tests.

Results: 223 women had normal VIAs. All had HPV testing, 221 had valid results: 115 HIV-infected women (mean age 37 years) and 106 HIV-uninfected (mean age 33 years). 175 women had abnormal VIAs. 143 women had HPV testing performed, 140 had valid results: 70 HIV-infected women (mean age 38.5 years) and 70 HIV-uninfected (mean age 31.3 years). Greater than 90% of all HIV-infected women in both projects were receiving anti-retroviral therapy at enrollment. HPV of any type was detected in 48% of all women with normal VIA vs. 61% of women with abnormal VIA (P = 0.018). High risk (HR)-HPV was detected in 38% of all women with normal VIA vs. 51% of all women with abnormal VIA (P = 0.012). HIV-uninfected women with normal VIA had significantly lower detection of all HPV (P = 0.026), high risk-HPV (P = 0.018), IARC high risk-HPV (P = 0.047), A9 types (P = 0.050), and individual types HPV 16 (P = 0.0274), HPV 18 (P = 0.007), and HPV 51 (P = 0.009) than HIV-uninfected women with abnormal VIA. Among HIV-infected women, there was no difference in detection of any group of HPV types or individual types with respect to VIA results.

Conclusions: HIV-uninfected women without cervical dysplasia had lower detection of oncogenic HPV than HIV-uninfected women with dysplasia. In contrast, HPV detection did not differ among HIV-infected women between those with or without cervical dysplasia. In addition, VIA appears to lack specificity for HPV-associated cervical dysplasia, as 39% of women with abnormal VIA examinations did not have any HPV detected, and 49% of women with abnormal VIA examinations did not have any HR-HPV detected.


This work was published before the author joined Aga Khan University.

Publication (Name of Journal)

Journal of Clinical Oncology