Visual inspection of the cervix following application of acetic acid is an effective method of reducing the incidence of and mortality from cervical cancer in developing countries, according to findings from a randomized trial conducted in Tamil Nadu, India.1 During up to seven years of follow-up, the incidence of cervical cancer was reduced by 25% among women who lived in areas where visual inspection was offered, and mortality from the disease was reduced by 35%, compared with the incidence and mortality in areas without screening. The benefits of screening were especially great for women aged 30–39.
Although the Pap smear and, to an increasing extent, testing for the human papillomavirus (HPV) are the gold standards for screening for cervical cancer and its precursor lesions, these approaches are difficult to implement in developing countries. Visual inspection—even when performed only once in a woman's lifetime—may be an acceptable alternative, but its effectiveness in real-world settings had not been adequately assessed.
In the new trial, researchers randomly chose 57 municipal units in Dindigul, Tamil Nadu, as sites for a visual inspection intervention. All healthy women aged 30–59 in these areas who had an intact uterus and no history of cervical cancer were interviewed, educated about cervical cancer and given a printed invitation to attend an upcoming screening. At the screening clinics, specially trained nurses examined each woman's cervix after applying 4% acetic acid; women who had well-defined lesions or cervical growths upon inspection were offered immediate colposcopy and, if appropriate, biopsy, followed by cryotherapy to remove any precancerous lesions. Women with suspected invasive cancer were referred for treatment. The vast majority of screenings were performed in 2000–2003, although a small fraction were performed later.
To serve as a comparison group, women in 56 other randomly selected municipal units (authorities in a 57th area did not wish to participate) were interviewed and educated about cervical cancer but were not screened; instead, they were given information about relevant government and private sector services and encouraged to use those facilities. The researchers note that the use of an unscreened control group was ethically appropriate because organized screening programs do not currently exist in India; in fact, only two women in the study had ever undergone cervical cancer screening.
To determine cervical cancer incidence and mortality in the intervention and control groups, staff at the Dindigul cancer registry, who were blinded to each woman's intervention status, examined records from hospitals, clinics and pathology laboratories in Dindigul and nine surrounding districts (where cancer patients from Dindigul would likely have been treated); they and other project workers also visited municipal death registration offices, death registries at churches and mosques, and households to collect information about deaths and migrations. Women who had died from cervical cancer were categorized as having lived in an intervention area, a control area or elsewhere (i.e, an area not involved in the study).
In total, 49,311 women in the intervention group and 30,958 women in the control group were interviewed. Nearly two-thirds (64%) of women in the intervention group attended a screening, of whom 10% had suspected lesions. Low-grade lesions were diagnosed in 5% of those screened, and high-grade lesions were diagnosed in 1%; sixty-seven women (0.2%) had cervical cancer. Most women (71–81%) with precancerous lesions received cryotherapy or other treatment to remove the growths; there were no instances of severe bleeding or other serious adverse effects from treatment.
Between January 2000, when screening started, and December 2006, when follow-up ended, a total of 158 cases of cervical cancer were diagnosed in the control group and 167 cases (including those detected during screening) in the intervention group; those numbers translate to incidence rates of 89 cases per 100,000 person-years in the control group and 61 per 100,000 person-years in the intervention group, or a 25% lower incidence of cancer in the intervention group (hazard ratio, 0.8) after adjustment for age, education and other characteristics. The reduction in incidence was greatest in women aged 30–39 (0.6). Cancers in the intervention group were more likely than those in the control group to be stage I (20% vs. 10%).
Similarly, the incidence of cervical cancer mortality was substantially lower in the intervention group than in the control group (52 vs. 30 per 100,000 person-years); cervical cancer mortality was reduced for the entire sample (hazard ratio, 0.7) as well as for women aged 30–39 (0.3) and those aged 40–49 (0.6). Women in the intervention group also had reduced rates of all-cause mortality (0.9).
Overall, the findings show that visual inspection is "a simple, feasible and effective method to prevent cervical cancer and death" in developing countries, where Pap screening and "multiple visits and diagnosis and treatment are impractical" and unaffordable. The fact that benefits of visual screening were greatest for women aged 30–39 "makes biological sense," the authors note, because the cervical transformation zone, where cervical cancer occurs, is fully visible in young women but less so in older ones. Visual inspection is not likely to be the ultimate solution to the challenge of reducing cervical cancer in low-resource countries: HPV vaccination is likely to partially address the problem, and an affordable, rapid version of the HPV test (which is more sensitive and specific than visual inspection) is on the horizon and may prove useful in developing countries. In the meantime, the researchers believe that visual inspection screening "should be established in routine health services in India" and other low-resource countries, as the implementation of such screening programs will not only reduce the disease burden but will create the infrastructure that will eventually be needed to administer HPV screening efforts.—P. Doskoch
REFERENCE
1. Sankaranarayanan R et al., Effect of visual screening on cervical cancer incidence and mortality in Tamil Nadu, India: a cluster-randomised trial, Lancet, 2007, 370(9585):398–406.