HPV Type Distribution in HIV Positive and Negative Women With or Without Cervical Dysplasia or Cancer in East Africa

Abstract

Background: Women living with HIV in sub-Saharan Africa are at increased risk to develop cervical cancer (CC), which is caused by persistent infection with 13 oncogenic human papilloma viruses (HR-HPVs). It is important to accurately identify and target HIV-positive women at highest risk to develop CC for early therapeutic intervention.

Methods: A total of 2,134 HIV+ and HIV- women from South-West Tanzania were prospectively screened for cervical cancer and precancerous lesions. Women with cervical cancer (n=236), high- and low-grade squamous intraepithelial lesions (HSIL: n=68, LSIL: n=74), and without lesion (n=426) underwent high-resolution HPV genotyping.

Results: Eighty percent of women who were diagnosed with HSIL or LSIL were living with HIV. Any lesion, young age, HIV status, and depleted CD4 T cell counts were independent risk factors for HPV infections, which were predominantly caused by HR-HPV types. While multiple HR-HPV type infections were predominant in HIV+ women with HSIL, single-type infections predominated in HIV+ CC cases (p=0.0006). HPV16, 18, and 45 accounted for 85% (68/80) and 75% (82/110) of HIV+ and HIV- CC cases, respectively. Of note, HPV35, the most frequent HPV type in HSIL-positive women living with HIV, was rarely detected as a single-type infection in HSIL and cancer cases.

Conclusion: HPV16, 18, and 45 should receive special attention for molecular diagnostic algorithms during CC prevention programs for HIV+ women from sub-Saharan Africa. HPV35 may have a high potential to induce HSIL in women living with HIV, but less potential to cause cervical cancer in single-type infections.

Keywords: cervical cancer; cervical dysplasia; high-grade intraepithelial lesions; human immunodeficiency virus—HIV; human papilloma virus—HPV; low-grade intraepithelial lesions; molecular diagnosis.

Figure 1

Flow Diagram of subjects included for HPV genotyping analyses. A total of 2,146 women were screened during the 2H study, including patient referrals with cervical cancer and women with a positive visual inspection with acetic acid result during cervical cancer screening program. Valid cytohistologic diagnoses based on Pap smear and/or histology was obtained for 1,964 of these women. HPV genotyping was then performed for 815 women and to include all cervical cancer cases, all women with cervical intraepithelial lesions, as well as the similar number of women living with or without HIV and with no lesions.

Figure 2

Number of different HPV infections in relation to suspected risk factors. Number of HPV infections stratified by HIV status and cervical lesion status (A), by HIV status and age groups (B), and for women living with HIV stratified by CD4 T cell count (below and above 250) and lesion status (C). Boxplots showing the median and interquartile range as well as outliers are indicated. Statistical analyses were performed using the Wilcoxon test. Multivariate linear regression results with number of HPV infections as dependent variable are shown in (D) for all women, adjusting by age, HIV status, and cytohistological diagnosis, and (E) for women living with HIV, adjusting by age in years, cytohistological diagnosis, number of CD4 T cell count (above vs. below 250), and antiretroviral treatment (no vs. yes). The individual risk factors are indicated on the y-axis; beta-coefficients and 95% confidence intervals are shown on the x-axis. In order to better visualize different data points that overlap, we used jitter, a graphical representation strategy in the ggplot R-package.

Figure 3

Frequency of occurrence of HPV types in HIV+ and HIV− women stratified by cytohistological diagnosis. HR-HPV types are indicated in bold. Frequencies are color coded in different shades as follows: 0% (light green), 0.1 to 5.0% (light blue), 5.1 to 10.0% (light yellow), 10.1 to 15.0% (light orange), 15.1 to 20.0% (faint red), 20.1 to 25.0% (orange), 25.1 to 30.0% (red-orange), >30.0 (strong red). The two-sided Fisher’s exact test was for significance testing and was calculated to compare all women with cervical cancer, HSIL, LSIL to women without lesion, regardless of HIV status. nd, not determined.

Figure 4

Decreased HPV type diversity and shifting HPV genotype distribution in HIV+ women with cervical cancer. (A) shows the number of subjects with HSIL and cervical cancer cases that were associated with HPV16, 18, or HPV45 infection (red bars) versus those that were not (gray bar) and (B) those associated with multiple HR-HPV types versus single HR-HPV types. Indicated p-values were calculated using the two-sided Fisher’s exact test. (C) shows the number of HPV types detected in HIV⁺ women with cervical cancer or HSIL in different age groups. Statistical analyses were performed by Wilcoxon test.