A single round of screening for human papillomavirus (HPV) among women in India's Osmanabad district substantially reduced the risk of death from cervical cancer during eight years of follow-up, according to a randomized trial that examined three screening methods.1 Compared with women who lived in villages where cervical cancer screening was rarely performed, those who lived in areas where HPV screening was offered were half as likely to develop advanced cervical lesions or to die from cervical cancer (hazard ratios, 0.5 for each). Two alternative detection strategies—cytologic testing (Pap smear) and visual inspection of the cervix with acetic acid—did not reduce the incidence of advanced disease or cervical cancer mortality.
The study was conducted in 52 clusters of villages in Osmanabad. During visits by female health workers, all healthy, nonpregnant women aged 30–59 who had ever been married and had an intact uterus were invited to participate. Women completed a questionnaire and received information about the causes, prevention, detection and treatment of cervical cancer. In addition, women in 39 of the clusters received an appointment card for a screening (HPV testing, cytologic testing or visual inspection with acetic acid); to simplify the process, all women in a particular cluster were assigned to receive the same type of screening. Women in the remaining 13 clusters served as a control group: They were given information about how to obtain screening at local hospitals, but were not given appointments.
All screenings were conducted by trained nurse-midwives under the supervision of physicians. The screening process differed according to the nature of the test. For HPV or cytologic screening, the nurse-midwives collected cervical cells for testing; women received results within two weeks, and those who tested positive were given appointments for colposcopy, biopsy and treatment. Women in the visual inspection group who screened positive underwent immediate colposcopy and biopsies, and received an appointment for treatment. In general, after colposcopy, women with low- or high-grade precancerous lesions were offered immediate treatment; those with probable invasive cancers were referred to the hospital.
The study was launched in January 2000, and women were followed through December 2007. To identify cases of cervical cancer and deaths from the disease during the follow-up period, the researchers used a cancer registry, hospital records, death certificates, house visits and interviews with relatives and friends.
The HPV-testing, cytologic-testing, visual inspection and control groups each included 31,000–34,000 women. About 80% of women in the first three groups were screened; however, because the trial used a cluster design, all eligible women in each of these clusters were included in the analyses, regardless of whether they had participated in the interviews or screenings. Overall, the analyses included 131,746 women, only eight of whom had ever undergone previous cervical screening. (Because so few women in the region ever receive screening, the study's inclusion of an unscreened control group was ethically justified, the researchers note.) Adverse effects from screening and treatment were rare: Among the nearly 80,000 women who participated, 123 had mild events and one developed uncontrolled bleeding.
The proportion of women with positive screening results ranged from 7% in the cytologic-testing group to 14% in the visual inspection group. The vast majority of women with positive results underwent colposcopy (more than 88% in all three intervention groups); the proportion of women with low-grade lesions was highest in the visual inspection group, but the proportion with high-grade lesions or invasive cancers did not differ among the three screening groups. The proportion of women with low-grade lesions who underwent treatment ranged from 33% in the HPV group to 45% in the cytologic-testing group; among those with high-grade lesions, about 90% of women in each intervention group underwent treatment. Six percent of women in the control group requested screening, which was done by colposcopy.
During the initial screening and follow-up, 127 cancers were detected in the HPV group, 152 each in the cytologic-testing and visual inspection groups, and 118 in the control group. The number of advanced cancers—i.e., those that had spread beyond the cervix and uterus—ranged from 39 in the HPV-testing group to 86 in the visual inspection group. Deaths from cervical cancer were lowest in the HPV-testing group (34), intermediate in the other two screening groups (54–56) and highest in the control group (64). Overall, HPV testing was the only screening procedure that reduced the risk of advanced cervical cancer (hazard ratio, 0.5) and deaths from cervical cancer (0.5) relative to not screening. Because the results encompass all women in each study area, regardless of whether they chose to participate, many of the advanced cancers and deaths occurred among women who were not screened. For example, nearly two-thirds of the cervical cancer deaths in the HPV-testing group and half of those in the cytologic-testing group occurred among women who had not been screened.
The findings indicate that "a single round of HPV testing" can produce "a significant decline in the rate of advanced cervical cancers and associated deaths," the researchers note. Moreover, compared with the other two methods, HPV testing is "less demanding in terms of training and quality assurance." The cost of HPV testing (US$20–30 per test) and the need for laboratory facilities to analyze samples remain obstacles to utilizing this approach in developing countries, but "a simple, affordable and accurate HPV test that provides results within three hours" is expected to become commercially available "in the near future." This development, together with the findings from Osmanabad, suggests that HPV testing is "appropriate as a primary screening approach in low-resource settings for women who are at least 30 years of age."—P. Doskoch
1. Sankaranarayanan R et al., HPV screening for cervical cancer in rural India, New England Journal of Medicine, 2009, 360(14):1385–1394.