RHO : Topics : Cervical Cancer Prevention

Annotated Bibliography

This is page 1 of the Cervical Cancer Prevention Annotated Bibliography. This page contains:

To access more bibliographic entries, visit page 2 or page 3, or return to the complete list of topics covered in the Cervical Cancer Annotated Bibliography. Be sure to use the Glossary if you are unfamiliar with any of the terms on this page.

Please note that PDF files require Adobe Acrobat Reader software, which can be downloaded for free at www.adobe.com/products/acrobat/readstep.html.


Abwao, S. et al., eds. Prevention and Control of Cervical Cancer in the East and Southern Africa Region. Summary of proceedings of regional meeting held in Nairobi, Kenya, March 29-April 1, 1998.
The document provides summaries of the technical papers, the subsequent discussion sessions, and the country profiles. Action plans developed by the 15 country teams that participated, an overview of available funding and technical resources, and the consensus recommendations from the meeting are included.

Bishop, A. et al. Cervical cancer: evolving prevention strategies for developing countries. Reproductive Health Matters 6:60-71 (November 1995).
This article makes a strong case for rational, public health approaches to the prevention and treatment of cervical cancer, focusing on practical strategies that can be used in developing countries. The authors note that scarce resources, limited infrastructure, and competing health priorities have prevented most developing-country health systems from implementing successful programs. Three approaches to cervical cancer screening are suggested for programs with limited resources: (1) screening women aged 30-35 or older; (2) screening women relatively infrequently; and (3) considering alternate approaches to conventional screening techniques. For women identified as needing treatment, the authors discuss outpatient approaches, such as cryotherapy and loop electrosurgical excision procedure (LEEP), that can effectively treat most precancerous lesions and reduce the number of clinic visits. The authors recommend coordinating efforts to achieve broader screening and treatment coverage, and they note that introducing alternate approaches requires policy changes, for which community input is essential.

Fahey, M.T. et al. Meta-analysis of Pap test accuracy. American Journal of Epidemiology 141:680-689 (1995).
This meta-analysis examined 62 studies published by August 1992 that compared Papanicolau (Pap) test results with histology. Data from 59 of the studies were used to assess the accuracy of the Pap test. Estimates of� Pap smear sensitivity ranged from 11 percent to 99 percent and estimates of specificity ranged from 14 percent to 97 percent. The authors found that a specificity in the 90 percent-95 percent range on a Pap test corresponds to a sensitivity in the 20 percent-35 percent range. While many studies had methodological weaknesses, the authors concluded that an increase in sensitivity almost always corresponded to a decrease in specificity or vice versa and that the Pap cannot achieve concurrently high sensitivity and specificity. They recommended that future studies follow methodological standards for diagnostic test evaluation more closely.

Fahs, M.C. et al. Cost-effective policies for cervical cancer screening: an international review. Pharmacoeconomics 3: 211-230 (March 1996).
This international review of the cervical cancer literature focuses on studies that address the cost-effectiveness of cervical cancer screening. The authors conclude that centrally organized screening programs implemented by the public sector are the most cost-effective type of program. Many programs are not effective (and have limited impact) due to over-screening of younger, affluent, lower-risk women and under-screening of older, less affluent, and minority women (for example, with opportunistic screening and screening at the discretion of individual health care practitioners). The authors also conclude that it is more cost-effective to begin screening women between the ages of 25 and 35. The appropriate cost-effective age to end screening is less clear; once an older woman has had several negative Pap smears, screening is less efficient and can be discontinued. The authors found that the interval of screening with the best balance between cost and life years saved was between 3 and 5 years. However, when repeat screening is not feasible, even one test per lifetime can significantly reduce mortality of a population. Efforts to reach unscreened women are particularly important as this population is typically at higher risk of developing cervical cancer. The authors noted three factors that can heavily impact the cost-effectiveness of a screening program: proportion of women screened by the program, quality of the Pap smear, and the cost of the Pap smear. Finally, the review discusses different cost models and analyses for evaluating cervical cancer screening and suggests the need for models that incorporate total costs and benefits of cervical cancer screening to society.

Hernandez-Avila, M. et al. Evaluation of the cervical cancer screening programme in Mexico: a population-based case-control study. International Journal of Epidemiology 27:1-7 (1998).
This case-control study evaluated the preventative effect of the cervical cancer screening program in Mexico City between September 1990 and December 1992. The authors selected 233 cases of cancer in situ and 397 cases of invasive cancer from women attending the gynecological clinics at six hospitals in Mexico City for histological confirmation following cytological diagnosis of cervical neoplasm. The 1,003 controls were an age-stratified random sample of residents of the Mexico City metropolitan area. When interviewed about their Pap smear history, cases were asked about the 12-month period before their diagnosis and controls were asked about the 12-month period before their interview. The authors found that the cervical cancer screening had a protective effect in relation to invasive cervical cancer. Women who had a history of a Pap smear, who did not seek screening due to gynecological symptoms, and who had received their Pap results had a 2.63 times lower risk of developing invasive cervical cancer (OR = 0.38; 95% CI = 0.28-0.52). No protective effect was found for in situ cancer. The authors note that the effect of the screening program is small when examined from a population perspective. They suggest that this results from low Pap-coverage (only 50 percent of those interviewed had ever had the test) and the high proportion of women who seek the test because they have symptoms.

Jones, S.B. Cancer in the developing world: a call to action. British Medical Journal 319:505-508 (August 1999).
This article synthesizes several themes from a 1999 WHO conference on "Cancer Strategies for the New Millennium." The author uses charts to characterize global cancer incidence by country, cancer type, incidence, and mortality rates. He notes that developing countries experience a disproportionate disease impact. For cancers with a known cause, such as cervical cancer, he recommends pursuing affordable approaches to prevention combined with development of new technologies appropriate to low-resource settings. The author also notes that clinical trials are underway for two prophylactic HPV vaccines and a therapeutic vaccine that stimulates cell immunity to viral proteins E6 and E7. Development of vaccines and low-technology approaches to detection offer the best promise for controlling cervical cancer in the developing world.

Parkin, D.M. et al. Estimates of the worldwide incidence of 25 major cancers in 1990.International Journal of Cancer 80:827-841 (1999).
This comprehensive review of the worldwide burden of cancer estimates the annual incidence rates (crude and age-standardized) and numbers of new cases of 25 different cancers in 23 areas of the world as of 1990. In 1990, more than 371,000 new cases of cervical cancer were identified among women worldwide. Nearly 290,000 of these cases are estimated to have occurred in developing countries. The highest age-standardized incidence of cervical cancer in 1990 was reported in southern Africa, Central America, and Melanesia, where the rates were over 40 per 100,000 women. Rates of more than 30 per 100,000 were reported in eastern Africa, the Caribbean, and tropical South America. For a companion study of mortality from 25 cancers in 1990, see Pisani et al. 1999.

Parkin, D.M. and Sankaranarayanan, R. Prevention of cervical cancer in developing countries. Thai Journal of Obstetrics and Gynaecology 11 (Suppl. 1):3-20 (August 1999).
This article provides a review of methods available for preventing cervical cancer in developing countries. It describes the acceptance of the human papillomavirus (HPV) as the most important etiological agent and the difficulty involved in primary prevention of HPV. It describes the methods and difficulties involved with screening and early detection of cervical cancer through health education strategies and downstaging (which the authors conclude is not a cost-effective procedure). It also describes screening for pre-invasive disease through Pap smears, its effectiveness and difficulties in the organization of screening programs in developing countries, and visual inspection using acetic acid, which the authors report to be as sensitive as Pap tests, though generally less specific. Finally, the authors review testing for human papillomavirus, the value of which they conclude will have to await the development of tests that can be applied rapidly and relatively cheaply.

PATH (Program for Appropriate Technology in Health). Preventing cervical cancer in low-resource settings.Outlook 18(1):1-8 (September 2000) (http://www.path.org/files/eol18_1.pdf).
This issue of Outlook describes the problem of cervical cancer in developing countries, the basic principles of cervical cancer control, screening and treatment options, and more.

Pisani, P. et al. Estimates of the worldwide mortality from 25 cancers in 1990. International Journal of Cancer 83:18-29 (1999).
This article presents worldwide estimates of annual mortality from all cancers and for 25 specific cancer sites from 1990. Crude and age-standardized mortality rates and numbers of deaths were computed for 23 geographical areas. The study estimated the global number of deaths from cervical cancer in 1990 at 190,000. Eighty percent of these deaths occurred in developing countries, where cervical cancer remains the largest cause of cancer death among women (followed closely by breast and stomach cancers). Some 40 percent of the estimated 148,500 cervical cancer deaths in developing countries occur in south central Asia, a region that includes India, Pakistan, Bangladesh, Afghanistan, and Iran, among others. For a companion study of worldwide cancer incidence in 1990, see Parkin et al. 1999.

Ponten, J. et al. Strategies for global control of cervical cancer. International Journal of Cancer 60:1-26 (1995).
This extensive review summarizes data from around the world on cervical cancer. Topics addressed include the tumor biology and natural history of cervical cancer, etiology of the disease (including impact of HPV), strategies for reducing mortality without screening (including focusing on treatment of early stage disease), and cytological screening. The authors conclude that the natural history and disease patterns of cervical cancer are similar throughout the world. They argue that HPV testing as a strategy for cervical cancer control remains experimental, and that providing treatment for early-stage disease where feasible can help reduce mortality. They discuss the challenges of ensuring maximum coverage with cytological screening without wasting resources on frequent testing of women at lower risk.

Rajkumar, R. et al. Leads to cancer control based on cancer patterns in a rural population in South India. Cancer Causes and Controls 11:433-439 (2000).
This study reviewed cancer patterns from the population-based cancer registry in Palani and Oddanchatram taluk in Dindigul District, Tamil Nadu State, South India. The Ambillikai Cancer Registry (ACR) records invasive and in-situ cancers for a rural population of 359,525 in 384 villages (estimate for 1996-1998). A total of 763 cancer cases were recorded during 1996-1998, 310 male and 473 female. In women, uterine cervical cancer accounted for more than half (53.9%) of all cases, with an age-specific and age-standardized rate of 65.4 cases per 100,000 person years. This age-standardized rate is the second highest in the world, following a rate observed in Harare, Zimbabwe (67.2). The age-specific rate for women ages 40 to 60 years from the ACR is the highest in the world. More than four-fifths of the cervical cancer cases from the ACR were diagnosed in FIGO stages IIB and IIIB. It is possible that the true risk for cervical cancer is even higher, since cases in old age groups may be under-registered. Breast and mouth cancers were the second and third most common cancers among women (age-standardized rate of 14.2 and 10.2 respectively). For men, mouth cancer was the most commonly recorded cancer (age-standardized rate: 11.5), followed by tongue (age-standardized rate: 8.6), hypopharynx (age-standardized rate: 7.8), esophagus (age-standardized rate: 7.8), and larynx (age-standardized rate: 7.8).

Rogo, K.O. et al. Carcinoma of the cervix in the African setting. International Journal of Gynecology and Obstetrics 33:249-255 (1990).
Data on 1,210 patients treated in Kenya between 1974 and 1979 formed the basis for this study. The authors compared Kenyan data to statistics for cervical cancer incidence in developed countries and found Kenyan patients to be both younger (mean age 42 versus 54 years) and more likely to present with late-stage disease (55% Stage 3 compared to 25%). A high loss to follow-up and limited treatment resources contributed to poor survival outcomes. The authors cited some possible additional contributors to high mortality, including tumor bulk, nutrition, and individual immunity. They concluded that further study of the differences in findings is needed and noted that cytologic screening is less likely to succeed in developing country settings.

Sherlaw-Johnson, C. et al. Evaluating cervical cancer screening programmes for developing countries. International Journal of Cancer 72(2):210-216 (July 1997).
This study evaluates infrequent cervical screening and programs in which as many women as possible are screened just once in their lifetime in developing countries. It also compares the effectiveness of cytology and HPV testing for primary screening. Evaluation criteria include: progression of precancer by use of a stochastic model, its relationship to HPV infection, and diagnostic accuracy of alternative screening methods. These criteria are compared in terms of the impact on the incidence of invasive cancer and resource use. The authors found that blanket screening for women aged 30-59 years, with the aim of covering all just once in their lifetime, could reduce the incidence of invasive cancer by up to 30 percent. The effectiveness of cytology and HPV testing as a primary screening method depends on the underlying prevalence of HPV infection, the accuracy of cytology, the cost and the suitability of the testing procedure under field conditions.

Sherris, J.D. Cervical cancer prevention: a strategic opportunity to improve women's reproductive health. International Family Planning Perspectives 25 (Suppl.):556-557 (1999).
This article summarizes the issues associated with preventing cervical cancer in developing countries and the steps needed to strengthen prevention efforts. The article also reports the initial findings of the Cervical Cancer Consultative Group, which in 1997 identified a list of ten questions felt to be of highest priority in guiding strategy development for preventing cervical cancer in low-resource settings. The author concludes by stating that activities focused on answering these questions will contribute to preventing some of the 200,000 deaths from cervical cancer that occur each year in developing countries.

Top of page

Epidemiology and natural history of cervical cancer

Anttila, T. et al. Serotypes of Chlamydia trachomatis and risk for development of cervical squamous cell carcinoma. Journal of the American Medical Association 285(1):47-51 (January 3, 2001).
This study and the accompanying editorial by Zenilman, J.M. provide evidence that Chlamydia trachomatis is associated with an increased risk for developing cervical cancer. This study reports on data from a longitudinal, nested case-control study in a cohort of 530,000 serum samples from women in Finland, Norway, and Sweden. Invasive cervical cancer was identified in 181 women through linked records to national cancer registries. C. trachomatis has 18 different serotypes. Of these the highest risk of cervical cancer was associated with serotype G (adjusted odds ratio 6.6, 95 percent confidence interval (1.6-27.0). Serotype I and D and exposures to multiple serotypes were also found to be associated with an increased risk of subsequent development of cervical cancer.

Biswas, L.N. et al. Sexual risk factors for cervical cancer among rural Indian women: a case-control study. International Journal of Epidemiology 26(3):491-495 (June 1997).
This hospital-based case-control study in India investigates the role of sexual risk factors in cervical cancer among rural women with a low rate of sexual promiscuity. The study included 134 women with invasive cervical cancer and 134 control women. Results from multiple logistic regression analysis showed that cervical cancer is associated with early age at first coitus, extramarital sex partners, and the time interval since first exposure. Independent effects were observed for early age at first coitus. Women who reported their first intercourse at less than 12 years of age were at the greatest risk compared to that of their counterparts at 18 years or older. Increased risk was also observed for women who had extramarital sex partners. The findings confirm the association between early age at first coitus and cervical cancer in women with a low rate of sexual promiscuity.

Bosch, F.X. et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. Journal of the National Cancer Institute 87(11):796-802 (June 7, 1995).
This study confirmed an extensive, global association between human papillomavirus (HPV) infection and cervical cancer. The study had two objectives: to determine whether the association between cervical cancer and HPV was consistent worldwide and to investigate geographic variation in the distribution of over 20 types of cancer-associated HPV. Investigators collected more than 1,000 specimens from cervical cancer patients in 22 countries with high recorded cervical cancer incidence. HPV DNA was detected in 93 percent of the tumors. Although HPV types differed somewhat by geographic region, HPV 16 was present in 50 percent of all specimens. The second most predominant type, HPV 18, was present in 14 percent of all specimens.

Brinton, L.A. Epidemiology of cervical cancer—an overview. In: The Epidemiology of Cervical Cancer and Human Papillomavirus, Mu�oz, N. et al., eds. Lyon: International Agency for Research on Cancer. Scientific Publication Number 119, 3-23 (1992).
Numerous studies of cervical cancer epidemiology are reviewed in this 1992 publication that examines factors other than HPV associated with cervical cancer. In tables summarizing findings from multiple studies, the author demonstrates the strong association between cervical cancer risk and number of sexual partners and age at first intercourse (which may be markers for HPV risk). She then examines epidemiologic evidence for contributory or interactive roles of other suspected risk factors, such as the relationship between HPV and other sexually transmitted diseases. She reviews study findings for more speculative risk factors also, particularly cigarette smoking and use of oral contraceptives. She concludes by noting that disease risk may be affected by changes in recent times including the tendency for women to initiate sexual intercourse at earlier ages, increased exposure by younger women to cigarette smoking and oral contraceptives, and changes in the sexual behavior of their male partners.

Burk, R.D. Pernicious papillomavirus infection [editorial]. New England Journal of Medicine 341(22):1687-1688 (November 25, 1999).
In this editorial, the author reviews some of the evidence that HPV causes cervical cancer. While not all of the criteria proving causation have been established, cervical cancer begins with sexual transmission of HPV to a woman susceptible to persistent infection. A study by Wallin et al. in the same issue found that HPV was present in the Pap smears of most women who later developed cervical cancer. This study also found a relatively high rate of false-negative Pap smear results, which the author notes is cause for concern. The author concludes that it may be time to consider periodic HPV screening along with Pap smears, especially for older women at high risk for high-grade cervical disease.

Critchlow, C.W. and Kiviat, N.B. Old and new issues in cervical cancer control (editorial). Journal of the National Cancer Institute 91(3):200-201 (February 3, 1999).
This editorial comments on the Holowaty et al. article on natural history and supports the practice of following women with cytologic diagnosis of mild dysplasia rather than immediate referral for colposcopy and biopsy. The authors note that the Holowaty et al. results confirm the results of previous studies that indicate that most mild and moderate dysplasia regresses spontaneously. Given the methodological rigor of the study, Holowaty et al.'s results add strength to earlier findings. The editorial also comments on the Palefsky et al. study of HPV infection in HIV-positive and HIV-negative women, which showed that HIV-positive women are at markedly increased risk of HPV infection (probably reactivated latent infection), particularly with high-risk HPV types.

Gustafsson, L. et al. International incidence rates of invasive cervical cancer after introduction of cytological screening. Cancer Causes Control 8(5):755-763 (September 1997).
This study compares the changes in cervical cancer incidence at different ages after the introduction of cytologic screening in different populations, and addresses the impact of organized and opportunistic smear taking. The relative reduction in age-specific incidence rates and in incidence rates age-standardized to the world population after the introduction of screening were calculated for each of the 17 cancer registries identified. In 11 of the 17 populations, age-standardized incidence rates declined markedly, from 27 percent in Norway to 77 percent in Finland. Age-specific declines were confined to women aged 30 to 70 years old; incidence rates were lowest among women aged 40 to 55. The authors state that because age-specific declines in cervical cancer incidence rates were strikingly similar in populations with widely different screening practices, organized screening may not be markedly superior to opportunistic screening.

Ho, G.Y. et al. Persistent genital human papillomavirus infection as a risk factor for persistent cervical dysplasia. Journal of the National Cancer Institute 87(18):1365-1371 (September 20, 1995).
This study analyzed the factors that determine cervical intraepithelial neoplasia (CIN) persistence or regression and found that persistent CIN is linked to chronic HPV infection, particularly HPV infection with a high viral load. The study enrolled 100 U.S. women diagnosed with CIN II. About one third of the women experienced regression; the remaining 70 were evaluated at three-month intervals for 15 months. Women who had chronic HPV infection had a fourfold higher risk for persistent CIN than those without HPV. The authors noted that repeated testing for HPV infection may help clinicians to differentiate between women who are likely to experience spontaneous regression and women whose lesions will persist or progress.

Holowaty, P. et al. Natural history of dysplasia of the uterine cervix. Journal of the National Cancer Institute 91(3):252-258 (February 3, 1999).
This article reviews a historical cohort of women in Toronto, Canada, whose Pap smear histories were recorded at a major cytopathology laboratory. This authors studied progression and regression of cervical dysplasia in this cohort during the period 1962 to 1980. The cohort size and time period covered by the analysis distinguishes this study from others, which tend to be much smaller and of shorter duration (the article includes a useful summary table illustrating the results of previous studies). The results of the study confirm those of smaller studies: the risk of mild dysplasia progressing to cervical cancer in situ (CIS) is small�about 1 percent per year. The study results support the recommendation of following patients with mild dysplasia with periodic cytology rather than immediate referral to colposcopy. The study also found that the risk of moderate dysplasia progressing to CIS is intermediate between the risks associated with mild and severe dysplasia�16 percent in 2 years.

Kjaer, S.K. et al. Type specific persistence of high risk human papillomavirus (HPV) as indicator of high grade cervical squamous intraepithelial lesions in younger women: population based prospective follow up study. British Medical Journal 325(7364):572 (September 14, 2002). Available at: http://bmj.bmjjournals.com/cgi/content/full/325/7364/572?maxtoshow=&HITS=10&hits=10&RESULTFORMAT=1&searchid= 1063662947341_14425&stored_search=&FIRSTINDEX=0&sortspec=relevance&volume=325&firstpage=572&resourcetype=1,2,3,4,10.)
Much of what is known about HPV infection and the development of cervical neoplasia is drawn from cross-sectional studies and a few prospective cohort studies. The authors of this study conducted a population-based prospective study that followed 10,758 women between 20 and 29 years of age. Cervical smears and cervical swabs for HPV DNA testing were taken at enrollment and at a two-year follow-up visit. In the final study population, consisting of 653 women with normal cytology, and 252 incident cases of cervical dysplasia (25 atypical cells, 115 low-grade lesions, and 112 high-grade lesions), HPV 16 was the most common HPV type detected. For women who had normal cytology at enrollment, but who subsequently developed high-grade lesions, 80 percent were also HPV-positive at enrollment. Among women who had normal cytology but were HPV-positive at their first exam, 71 percent (62/87) cleared the infection by the second exam (45 became HPV-negative, 17 cleared that specific type of HPV but acquired a different HPV type). The study found an increased risk of developing cervical lesions for women who were HPV-positive at enrollment and at the follow-up visit. The most pronounced risk was for women with high-grade lesions who were HPV-positive at both exams as compared to women who were HPV-negative at both exams (odds ratio 413.9, 96.3 to 1779.5). The odds ratio increased most substantially for women with high-grade lesions who were positive with the same HPV type at both exams as compared to HPV negative women (odds ratio 813, 168.2 to 3229.2), suggesting that persistent infection with a high risk HPV type is a good predictor of subsequent high-grade lesions.

Kjellberg, L. et al. Smoking, diet, pregnancy and oral contraceptive use as risk factors for cervical intra-epithelial neoplasia in relation to human papillomavirus infection. British Journal of Cancer 82:1332-1338 (2000).
This study sought to distinguish if smoking, nutrition, parity, and oral contraceptive use are independent risk factors for cervical cancer, or whether they may act as cofactors with human papillomavirus (HPV) infection. This population-based case-control study was undertaken in the V�sterbotten county of Northern Sweden of 137 women with high-grade cervical cancer (CIN II or III) and 253 healthy age-matched women. The women completed a 94-item questionnaire, and donated specimens for HPV testing. Of the environmental risk factors being studied, smoking appeared to be a significant risk factor for cervical cancer. Smoking was associated with CIN II-III (odds ratio 2.6, 95 percent confidence interval of 1.7-4.0), and was dose-dependent. The excess risk was not affected by adjusting for HPV infection, suggesting that smoking is a causal risk factor for CIN II-III. The study did not find diet to be protective (as other studies had suggested). Pregnancy appeared to be a risk factor for CIN II-III, and further study is warranted. Prolonged oral contraceptive use and sexual history were not associated with CIN II-III when HPV infection was taken into account.

Lancaster, E.J. et al. Carcinoma of the uterine cervix: results of Ka-Ngwane screening programme and comparison between the results obtained from urban and other unscreened rural communities. East African Medical Journal 76(2):101-104 (February 1999).
This study revived data from 10,000 consecutive Pap smears from women attending Ob/Gyn clinics in Ka-Ngwane, Pretoria (urban area), and Transkei (rural area) in South Africa. The study found positive Pap smears (mild dysplasia through cancer) in 3 percent of patents in Ka-Nguane, 5 percent in Greater Pretoria, and more than 6 percent in Transkei. Of the positive cases, cervical cancer accounted for 12 percent of cases in Ka-Ngwane, 5 percent in Pretoria, and 26 percent in Transkei. The majority of positive cases were younger than 41 years (the article did not indicate the age distribution of all 10,000 patients). The authors concluded that there is a high incidence of dysplasia and cervical cancer in previously unscreened populations in South Africa, and that there is an urgent need to develop education and screening programs in the region. The authors also noted that a significant proportion of cervical cancer cases occurred in women below 40 years of age.

Li, H.Q. et al. The decline in the mortality rates of cervical cancer and a plausible explanation in Shandong, China.  International Journal of Epidemiology 29:398-404 (2000).
This study analyzed the decline in cervical cancer deaths in Shandong Province, China, from 1970 to 1992. China has had one of the most rapid declines in the incidence of cervical cancer of any Asian country. This is largely due to social changes and health policies put in place by the Chinese government after the 1949 founding of the People's Republic of China. These changes reduced exposure to sexually transmitted diseases; they also resulted in improved screening and treatment. Lifetime exposure to risk factors and co-factors (such as smoking) declined with successive birth cohorts. The authors conclude that if the recent increase in rates of sexually transmitted diseases continues in China, an increase in rates of cervical cancer is likely to follow.

Moreno, V. et al. Effect of oral contraceptives on risk of cervical cancer in women with human papillomavirus infections: the IARC multicentric case-control study. Lancet 359:1085-1092 (March 30, 2002).
This study analyzed data from a large multi-center cancer registry. Cases were matched with hospital-based controls in five countries and population-based controls in two countries. Women were questioned about their life-long use of oral contraceptive pills (OCPs); however, no information was gathered on the type or dosage. HPV infection was determined using PCR. Analysis by logistic regression adjusted for confounders and results showed that HPV positive women who had ever used OCPs were 1.5 times more likely to develop cervical cancer than controls. Women who used OCPs for 5-9 years had a significant association with cervical cancer (odds ratio, 2.82; 95% CI = 1.46-5.42). Comment by David Skegg (Lancet 359:1080-1081) discusses study weakness that make interpreting the data complex. These shortcomings include possible recall bias, lack of distinction between OCPs and other hormonal contraceptives, small number of HPV-positive controls, and the use of logistic regression analysis when controlling for confounding factors.

Moscicki, A., Hills, N., Shiboski, S., et al. Risks for incident human papillomavirus infection and low-grade squamous intraepithelial lesion development in young females. Journal of the American Medical Association 285(23):2995-3002 (June 20, 2001).
Results from the longest longitudinal study of incident human papillomavirus infection (HPV) reveal that, over a period of 36 months, 55 percent of women without prior HPV infection acquired an incident HPV infection. Associated risks for incident HPV infections, as indicated by multivariate analysis, include sexual behavior, smoking, prior herpes simplex virus infection, and history of vulvar warts. A woman�s risk of HPV infection increased 10-fold with each new sexual partner per month. Current oral contraceptive use provided a significantly protective effect, though other literature has suggested a long-term increased risk of developing cervical lesions among oral contraceptive users. Consistent with previously reported findings by the authors, the results indicate that approximately 90 percent of young women clear the HPV virus within 36 months. Findings from this study shed light on the time between HPV infection to the development of low-grade SILs. Seventy percent of women with HPV infection, followed for 60 to 80 months, did not develop LSIL and the majority of women who did develop LSIL had the lesions spontaneous clear later.

Muñoz, N. et al. Role of parity and human papillomavirus in cervical cancer: the IARC multicentric case-control study. Lancet 359:1093-1101 (March 30, 2002).
This multi-center study pooled data on HPV-positive women to evaluate the association between parity and increased risk of cervical cancer. HPV infection was determined using PCR. Using logistic regression analysis to calculate pooled odds ratios, the authors reported a significant association between high parity and increased risk of squamous-cell cervical cancer. The association was strongest for women with seven or more full-term pregnancies (odds ratio, 3.8; 95% CI 2.7 to 5.5). For women with one or two full term pregnancies, the odds ratio was 2.3 (95% CI = 1.6 to 3.2). Comment by David Skegg (Lancet 359:1080-1081) outlines several weaknesses that should be kept in mind when interpreting the data. These shortcomings include the use of hospital-based controls, small number of HPV-positive controls, and the use of logistic regression analysis when controlling for confounding factors.

Nasiell, K. et al. Behavior of mild cervical dysplasia during long-term follow-up. Obstetrics and Gynecology 67(5):665-669 (May 1986).
This Swedish study evaluated follow-up data from 555 women diagnosed with mild cervical dysplasia between 1962 and 1983. Of these mild dysplasias, 62 percent regressed, 22 percent persisted, and 16 percent progressed. Patients with regression were followed for an average of 39 months; patients with persistent dysplasia were followed for an average of 52 months. Where mild dysplasia progressed to more severe disease, the average time to progression was 48 months. Two cases of invasive cancer occurred in women lost to follow-up for several years during the study. The invasive cancers were diagnosed at 79 and 125 months after initial diagnosis of mild dysplasia. This study is often cited as evidence that a significant proportion of mild dysplasia does not progress to more severe disease. At the same time, the study results highlight the importance of regular follow-up of women diagnosed with mild dysplasia, given the potential of progression in more than a sixth of cases.

Pisani, P. et al. Cancer and infection: estimates of the attributable fraction in 1990. Cancer Epidemiology, Biomarkers, and Prevention 1:387-400 (June 1997).
This article, written by researchers from the International Agency for Research on Cancer (IARC), reviews the evidence linking various cancers to infectious agents, including cervical cancer and HPV infection. The article includes data from around the world, and concludes that over 90 percent of cervical cancer cases in the developing world can be directly attributed to HPV infection.

Prokopczyk, B. et al. Identification of tobacco-specific carcinogen in the cervical mucus of smokers and nonsmokers. Journal of National Cancer Institute 89(12):868-873 (June 18, 1997).
The goal of this study was to determine for the first time whether tobacco-specific carcinogen N-nitrosamines (NNK) are present in the cervical mucus of cigarette smokers and of nonsmokers. Cervical mucus specimens from 15 smokers and 10 nonsmokers were tested for NNK. NNK was found in all cervical mucus specimens of smokers, and only one nonsmoker specimen did not contain detectable NNK. NNK concentrations in specimens from cigarettes smokers (11.9 to 115.0 ng/g of mucus) were significantly higher than those from nonsmokers (4.1 to 30.8 ng/g of mucus). The authors suggested that the presence of NNK in the cervical mucus of nonsmokers is likely due to environmental exposure or to the fact that some of the women in the study may not have revealed that they occasionally smoked cigarettes. The finding of this study further strengthens the association between cervical cancer and tobacco smoking.

Robles, S.C. et al. Trends in cervical cancer mortality in the Americas. Bulletin of PAHO 30(4):290-301 (1996).
This article provides an assessment of cervical cancer mortality trends in the Americas using data from the Pan American Health Organization. While cervical cancer in Canada and the United States has declined steadily over the past 30 years (to about 1.4 and 1.7 deaths per 100,000 women, respectively, in 1990), most Latin American and Caribbean countries with available data have experienced constant or increasing levels of cervical cancer mortality (generally in the range of 5 to 6 deaths per 100,000 women). The authors suggest that while not all changes in cervical cancer mortality can be directly attributed to screening, a correlation clearly can be drawn. They suggest that screening services in Latin America have been linked to family planning and prenatal care services, and have not appropriately targeted older women with the highest risk of cervical cancer.

Roteli-Martins, C.M. et al. Cigarette smoking and high-risk HPV DNA as predisposing factors for high-grade cervical intraepithelial neoplasia (CIN) in young Brazilian women. Acta Obstetrica et Gynecologica Scandinavica 77:678-682 (1998).
This cross-sectional study of young Brazilian women aged 20-35 investigated the role of cigarette smoking and high risk HPV types as risk factors for CIN 2 and 3. The study included 77 women with biopsy confirmed CIN 1, 2, and 3. Women with CIN 1, 2, and 3 did not differ significantly from one another with regard to age, race, schooling, marital status, lifetime number of sexual partners, age of first intercourse, use of oral contraceptives, or parity. Results from multivariate logistic regression analysis, however, showed that smoking and HPV type were significantly associated with CIN 2 and 3. Smoking increased the risk of developing CIN 2 and 3 by 6.6 fold and the detection rate of high-risk HPV types was significantly higher among cigarette smokers than among nonsmokers. These findings indicate the relationship between severe CIN lesions and both high-risk HPV types and current cigarette smoking. They also suggest that there may be a synergistic action between these two factors in the development of cervical cancer.

Schiffman, M.H. New epidemiology of human papillomavirus infection and cervical neoplasia. Journal of the National Cancer Institute 87(18):1345-1347 (September 20, 1995).
This article discusses the link between persistent detection of HPV DNA (especially high levels of DNA) and persistent diagnosis of CIN. A companion study by Ho et al. in the same journal issue tracked HPV and CIN transience versus persistence, and Schiffman notes that the findings of Ho et al. have implications for the development of screening strategies that include HPV-DNA testing. The author notes several findings and factors that complicate epidemiologic analysis: (1) HPV-negative CIN does exist, although it occurs in 10 percent or fewer cases; (2) up to 10 percent of women may develop CIN2 or CIN3 lesions initially instead of progressing from lower- to higher-grade lesions; (3) diagnostic uncertainty and the lack of a reference standard complicate interpretations of data; (4) the cervix may contain discrete lesions with separate natural histories (for example, CIN1 lesions may progress to CIN3 over time or simply emerge adjacent to CIN3 lesions).

Sellors, J.W. et al. Prevalence of infection with carcinogenic human papillomavirus among older women. Canadian Medical Association Journal 167(8):871-873 (2002).
Researchers obtained results of HPV DNA testing from 154 women over the age of 50 living in Ontario, Canada. This study investigated the prevalence of carcinogenic HPV in older women and the associated risk factors. In this study, HPV prevalence peaked for women older than 60 years. These findings also are consistent with another study of the natural history of HPV among Costa Rican women that found a high prevalence peaking in women older than 59 years. Risk factors that have a significant association to HPV infection in the Canadian study include the absence of a stable sexual partnership and early age at first intercourse.

Smith, J.S. et al. Herpes simples virus-2 as a human papillomavirus cofactor in the etiology of invasive cervical cancer. Journal of the National Cancer Institute 94(21):1604-1612 (November 6, 2002).
This article examined the association between herpes simplex virus-2 (HSV-2), HPV, and the development of invasive cervical cancer. Authors conducted a pooled analysis of seven case-control studies conducted in Morocco, Brazil, Peru, Columbia, Thailand, Philippines, and Spain. The pooled analysis examined a total of 1,263 patients with invasive cervical cancer and 1,117 controls to determine whether HSV-2 plays a role as a cofactor, in conjunction with HPV infection, to increase the risk of developing cervical cancer. Results showed that women with invasive cervical cancer had a higher rate of HSV-2 infection (44%) compared to women in the control group (26%). Women who were HPV-positive, infection with HSV-2 was associated with increased the risk of invasive cervical cancer. The authors conclude that while persistent HPV infection is still the strongest predictor of subsequent development of invasive cervical cancer, infection with HSV-2 may increase the risk.

Sukvirach, S. et al. Population-based human papillomavirus prevalence in Lampang and Songkla, Thailand. Journal of Infectious Diseases 187(8):1246-1256 (April 15, 2003).
There is a need for population-based, age-specific and type-specific regional data on HPV prevalence to further help define strategies for cervical cancer prevention. This study reports the results of 2 population-based prevalence surveys conducted as part of the International Agency for Research on Cancer’s multicenter study on global prevalence of HPV infection. Overall, the prevalence of HPV DNA was 6.3 percent, although Lampang had a significantly higher HPV prevalence (8.0%; 95% CI = 6.4-9.8%) than Songkla (3.8%; 95% CI = 2.5-5.5%). The most frequently reported HPV types were 16, 52, and 72. Similar to other population-based surveys, women less than 35 years of age had the highest HPV prevalence (9.8%). HPV prevalence then remained stable, between 4.5 and 6.0 percent for women ages 35 years and older. HPV infection was statistically significantly associated with having previously had a sexually transmitted infection (odds ratio [OR], 3.2) or having Herpes simplex virus 2 (HSV-2) (OR, 2.1). HPV infection also appeared related to a woman’s husband’s sexual behaviors. For example, the risk of HPV DNA detection was higher for women who reported their husband had other sexual partners (OR, 1.9) or had sexual relations with prostitutes (OR, 1.6). The authors note that the lower than expected HPV prevalence may be due to a large number of younger women refusing to participate in the study because of the need for a pelvic exam and the fact that 85 percent of women reported only having had one sexual partner in their lifetime.

Thomas, D.B. et al. Prostitution, condom use, and invasive squamous cell cervical cancer in Thailand. American Journal of Epidemiology 143(8):779-86 (1996).
This article reports on a case-control study conducted in three hospitals to investigate the role of male sexual behavior in the development of cervical cancer in their wives. Data were obtained from interviews with 225 married women with invasive cervical cancer, 791 hospitalized controls, and from interviews with their husbands. The study found that the rise of cervical cancer was strongly related to the women's husbands having visited sex workers without using a condom when the husbands were less than 30 years old. The authors conclude that regular condom use by sex worker clients could reduce the number of invasive cervical cancer cases in Thailand's general population by at least 25 percent.

Wallin, K.L. et al.  Type-specific persistence of human papillomavirus DNA before the development of invasive cervical cancer. New England Journal of Medicine 341(22):1633-1638 (November 25, 1999).
This study compared Pap smear results from routine screening of Swedish women during the years 1969 to 1995. The authors compared normal smears in 118 women who later developed cervical cancer to the smears of 118 age-matched women who did not develop cancer. HPV DNA was detected in 30 percent of the smears of women with cancer, and in only 3 percent of those who remained healthy. At the time of diagnosis with cancer, 77 percent of the tissue samples tested positive for HPV DNA, while 4 percent of the matched controls were HPV positive. The HPV-DNA type was the same in the baseline smear and in the tissue sample for all of the women with cancer, but not for any of the healthy women (controls). While there is some concern that stored samples may underestimate the proportion of women who were HPV positive before developing cancer, the authors conclude that HPV DNA in normal Pap smears is associated with an increased risk for cervical cancer, and persistence of HPV infection is related to the development of cancer.

Zenilman, J.M. Chlamydia and cervical cancer: a real association? Journal of the American Medical Association 285(1): 81-83 (January 3, 2001).
This editorial accompanies Anttila et al. (2001) with a discussion of the evidence for an association between Chlamydia trachomatis and cervical carcinoma. The author states studies prior to the large multi-center study conducted by Anttila et al. have been weak methodologically and have been unable to adjust for confounders. The author interprets the Anttila et al. findings as providing strong evidence for a causal relationship between C. trachomatis and subsequent cervical cancer development. The author poses several questions for future research on the biological mechanisms and suggests that cervical cancer be considered a potential adverse outcome of Chlamydia infections.

Top of page

Role of HPV tests in cervical cancer prevention

Bollen, L.J. et al. Human papillomavirus deoxyribonucleic acid detection in mildly or moderately dysplastic smears: a possible method for selecting patients for colposcopy. American Journal of Obstetrics and Gynecology 177(3):548-553 (September 1997).
The goal of this study was to determine whether human papillomavirus testing is capable of discriminating between high- and low-grade squamous intraepithelial lesions so as to be useful in reducing the number of colposcopic examinations. One hundred and ninety women with smears of mild or moderate dysplasia were tested for presence of HPV DNA using the CPI/IIG and MY09/11 tests. The HPV test results were compared with histologic diagnosis, the reference standard. The study found that 56 of the 190 women had high-grade squamous intraepithelial lesions. The sensitivity was 96 percent for the CPI/IIG test and 95 percent for the MY09/11 test; specificity was 33 percent and 40 percent, respectively. The authors concluded that use of HPV tests as a secondary triage in women with mild or moderate dysplasia could prevent those women from undergoing unnecessary colposcopy.

Clavel, C. et al. Hybrid Capture II-based human papillomavirus detection, a sensitive test to detect in routine high-grade cervical lesions: a preliminary study on 1518 women. British Journal of Cancer 80(9):1306-1311 (1999).
Hybrid Capture II (HC-II) is an HPV-detection test designed to detect 18 high-risk and low-risk HPV types. This study evaluated how effective the HC-II test is for assessing high-grade cervical lesions as compared with cytological screening. Twenty-four percent of the samples taken from women attending routine cytological screening tested positive for HPV: 18 percent for high-risk types and 6.3 percent for low-risk HPV types. High-risk HPV was found in all of the cases presenting a high-grade lesion confirmed by biopsy. The overall sensitivity of the HC-II in detecting high-grade SILs and cervical cancers is 98.1 percent, and the specificity is 85.2 percent in cervical samples. Cytological testing was more specific (94.9%), but less sensitive (85.3%) in detecting high-grade SILs. The use of HC-II along with cytological screening would accurately diagnose most all high-grade lesions.

Cuzick, J. Human papillomavirus testing for primary cervical cancer screening [editorial]. JAMA 283(1);108-109 (January 5, 2000).
This editorial reviewed two articles appearing in the January 5, 2000 JAMA issue, Schiffman et al. and Wright et al. The editorial noted that the two articles made important contributions to assessing use of HPV for primary screening for cervical cancer. The Schiffman article focused on performance of HPV-DNA testing according to different thresholds of test positivity (a test using a 1 pg/mL cutoff for HPV positivity performed best) and according to patient age (specificity was highest for test results among older women). The Wright study evaluated results of HPV-DNA testing in an unscreened population using self-collected and clinician-collected samples. While the self-collected samples resulted in lower sensitivity than the clinician-collected samples, sensitivity was higher than Pap smear testing in that setting. Specificity remained a problem, however. The author concluded that HPV-DNA testing has significant potential for making cervical cancer screening available in settings where traditional approaches may not be feasible. Further study is necessary to clarify optimal approaches that will ultimately lead to reduced morbidity and mortality.

Denny, L. et al. Evaluation of alternative methods of cervical cancer screening for resource-poor settings. Cancer 89(4):826-833 (August 15, 2000).
This study compared the performance of four methods of screening for cervical cancer and precancerous lesions. Conventional cytology (Pap smear), direct visual inspection after an acetic acid wash (DVI), HPV-DNA testing, and cervicography were applied in sequence to 2,944 women in a resource-poor settlement outside of Cape Town, South Africa. After adjusting for loss to follow up, calculations were done similar to standard measures of sensitivity and specificity. Of the 2,944 women, 842 (29%) tested positive by one or more of the tests. Colposcopy and cervical biopsies were performed on 790 (90%) of the women testing positive. Overall test performance was best for cytology, which correctly identified approximately 78 percent of women with SIL or invasive cervical carcinoma (confirmed by biopsy). For women whose Pap smear was classified as normal or ASCUS, 97 percent were classified as negative by DVI, HPV-DNA testing, cervicography, and/or colposcopy. DVI and HPV testing using a standard cutoff level (RLU>1x positive control) had similar sensitivities to detect high-grade SIL and invasive carcinoma to cytology. The specificity of DVI, HPV testing, and cervicography, however, was much lower than cytology, and resulted in many more false-positives. The authors conclude that DVI and HPV-DNA testing have adequate sensitivities to make them effective tests for identifying women with cervical disease in low-resource settings (that is, in the absence of colposcopic evaluation). The authors caution that because these two tests identify considerably more women as having disease who are really disease free, the impact of providing immediate treatment of women based on the results of either DVI or HPV-DNA testing warrants further investigation.

Elfgren, K. et al. A population-based five-year follow-up study of cervical human papillomavirus infection. American Journal of Obstetrics and Gynecology 183(3):561-567 (2000).
This study reviewed results from a population-based cervical screening program in Sweden in 1991 to determine the persistence of cervical HPV DNA. Women with normal Pap smears but who tested positive for HPV DNA were age-matched with controls (women with normal cytology, absence of HPV DNA) and both groups were reexamined five years later. Of the women infected with HPV DNA in 1991, only four showed infection with the same type of HPV five years later. All of these women had HPV 16. This represents a five-year HPV clearance rate of 92 percent. Eleven women acquired HPV infection during the five-year follow-up period. Although this study did not attempt to compare HPV-DNA tests, it noted differences between the MY09 and MY11 and the GP5+ and GP6+ primer systems for detection of different types of HPV. Because of the high likelihood that HPV infection will clear spontaneously, the authors conclude that a single HPV screening test is of limited value. Strategies that rely on repeated HPV testing to identify persistent infections will be more cost-effective.

Hernandez-Avila, M. et al. Human papilloma virus 16-18 infection and cervical cancer in Mexico: a case-control study. Archives of Medical Research 28(2):265-271 (1997).
The goal of this case-control study was to evaluate the association between HPV types 16 and 18 and cervical cancer in women living in Mexico City. In the study, 148 cases and 204 controls were tested for HPV types 16 and 18 using polymerase chain reaction (PCR) technique. HPV 16 was detected in 48.3 percent of 60 in situ cases, 48.8 percent of 88 invasive cases, and 13.2 percent in controls. HPV 18 was detected only in 6.7 percent of invasive cases. The odds ratio for HPV 16 infection in in situ cases was 5.17 and the odds ratio for invasive cases was 3.84. The results showed that all women with a strong positive PCR reaction had the greatest risk; the odds ratio was 38 compared to 23 and 11 of intermediate and weak PCR reactions respectively. The finding provided further evidence for the association between HPV types 16 and 18 infection and cervical cancer. The authors considered that the Pap test is still the single most effective measure in cervical cancer control.

Hillemanns, P. et al. Screening for cervical neoplasia by self-assessment for human papillomavirus DNA [research letter]. Lancet 354:1970 (December 4, 1999).
This study evaluated the sensitivity of HPV tests on self-collected specimens and compared it with the detection rate of specimens obtained from the cervix by gynecologists. A total of 247 patients attending a German colposcopy clinic self-introduced a cytobrush into the vagina for specimen collection and were examined by gynecologists who took Pap smears and an additional cytobrush specimen. Ninety-four percent of patients reported preferring the self-sampling to sampling by a physician. HPV positive results were found by 53 percent of the patients' samples compared to 42 percent by physician-collected specimens. Both test methods had a sensitivity of 93 percent for high-grade dysplasia and invasive cervical cancer. The authors concluded that self-sampling is a reliable method of testing for HPV and may be useful in settings in which cytology is not readily available, especially if these study findings can be reproduced in larger studies.

Jacobson, D. et al. Concordance of human papillomavirus in the cervix and urine among inner city adolescents. Pediatric Infectious Disease Journal 19(8):722-728 (August 2000).
A study of HPV in cervical samples and in urine found a high prevalence of HPV in both types of samples. Specimens from 80 sexually active adolescent women in Baltimore, Maryland, were tested for HPV DNA using both PCR and Probe B of Hybrid Capture. The presence of HPV was extremely high in the cervix (90%), and lower in the urine (75%). One adolescent tested positive for HPV in urine alone. Between the two sample sites, there was 82 percent agreement for presence of any type of HPV, and 40 percent agreement for specific HPV types. Despite the differences in sensitivity between the samples, the results indicate that HPV testing of urine samples could be useful outside of clinical situations. Urine testing can be useful in epidemiologic studies and in monitoring infected women, situations where obtaining genital specimens is difficult.

Johnston, C. Quantitative tests for human papillomavirus. Lancet 355:2179-2180 (June 24, 2000).
This commentary focuses on the relationship between infection with HPV DNA type 16, viral load, and risk of developing cervical cancer. Studies in this issue of the Lancet indicate that infection with a high viral load of HPV 16 many years prior to diagnosis of cervical cancer increases the risk of developing cancer, and those with persistently high viral loads are most at risk. These studies also suggest that cervical cancer is a slowly developing disease. Since HPV 16 has not been shown to be a useful predictor of cervical cancer risk by itself, except among those with a high viral load, is it a useful screening test? If viral load with HPV 16 is used to identify women at risk for high-grade disease, what should be done for these women now in the absence of any vaccine or treatment for HPV infection? The author concludes that thought needs to be given to whether these new viral tests offer substantial improvement over the Pap smear.

Josefsson, A. et al. Viral load of human papilloma virus 16 as a determinant for development of cervical carcinoma in situ: a nested case-control study. Lancet 355:2189-2193 (June 24, 2000).
This study investigated whether the amount of HPV DNA detected in cervical smears is a useful predictor of progression to cervical cancer in situ. Analysis of multiple smears from 478 women with cervical cancer (cases) and 608 matched controls showed that the risk of cervical cancer increased as the amount of HPV 16 DNA in the sample increased. The presence of high amounts of HPV 16 DNA is a major risk factor for developing cervical cancer. The positive predictive value of these results was too low to enable the test to be used as a single test for cancer risk except in women with the highest amounts of HPV DNA. However, the authors indicate that a quantitative test for HPV DNA has the potential to predict cancer risk at an earlier stage than current screening tools are able to do.

Kaufman, R.H. et al. Relevance of human papillomavirus screening in management of cervical intraepithelial neoplasia. American Journal of Obstetrics and Gynecology 176(1):87-92 (January 1997).
The goals of this study were to evaluate the utility of HPV-DNA testing in facilitating management decisions for women with abnormal Pap smears and to determine the prevalence of a positive HPV test result in these women. A total of 1,128 women who had abnormal Pap smears and were referred to a colposcopy clinic were evaluated in the clinic with a repeat Pap smear, HPV-DNA testing, and colposcopy. Of these, 172 (35.4%) of 486 women with low-grade lesions had high-risk HPV DNA detected; 263 (44.4%) of 592 women with high-grade lesions had high-risk DNA detected. High-risk DNA was detected in 38.7 percent of 527 and 56.2 percent of 267 women with biopsies showing CIN I and CIN II or III, respectively. The DNA test for detecting biopsy-confirmed CIN II or III had a sensitivity value of 55.7 percent; its positive predictive value was 34.9 percent. The results showed that HPV is causally associated with cervical cancer and that HPV-DNA testing does not appear to be a cost-effective measure in detecting high-grade CIN.

Koss, L. Human papillomavirus testing as a screening tool for cervical cancer. JAMA 283(19):2525-2526 (2000).
In this letter, the author questions the conclusions of Schiffman et al. 2000, that HPV testing of women older than age 35 is as effective or better than the Papanicolaou test in screening for cervical cancer. Koss argues that waiting until age 35 or 40 to test for high-risk HPV infection puts the lives of many younger women at risk. On the other hand, HPV testing of women ages 18 to 35, the ages at which most precancerous infection occur, would be misleading because of the high false positive rate of the test. Although the Pap test has a significant false-negative rate, the technique of testing for HPV is not mature enough to be its replacement, according to Koss. Schiffman and colleagues reply that HPV testing and Pap testing can be used to improve cervical cancer screening. At minimum, HPV-DNA testing can be recommended to standardize cytologic diagnoses in different laboratories.

Kuhn, L. et al. Human papillomavirus DNA testing for cervical cancer screening in low-resource settings. Journal of the National Cancer Institute 92(10):818-825 (May 17, 2000).
Given the many barriers to effective cytology screening in low-resource settings, this study evaluated HPV-DNA testing as an alternative. Cervical samples from 2,944 South African women ages 35 to 65 were tested for high-risk types of HPV using the Hybrid Capture I assay (HC-I). The screening included Pap smear, direct visual inspection of the cervix, and Cervicography (35 mm photograph of cervix). Women with any positive result were referred for colposcopy. Samples from a group of women with disease and from a random sample of disease-free women were retested for HPV DNA using the Hybrid Capture II assay (HC-II). Of the 86 women with high-grade squamous intraepithelial lesions (SILs) or invasive cancer, 73.3 percent were positive for a high risk type of HPV DNA using HC-I assay, and 88.4 percent were positive using the HC-II assay. HPV-DNA testing using the HC-II assay was more sensitive than cytology for detecting high-grade SILs and invasive cancer, and testing with the HC-I assay was equally sensitive. Cytology, on the other hand, had better specificity (96.8%) than either HC-I (87.8%) or HC-II (81.9%). By adjusting the test cut-off level, HPV-DNA testing can identify 57 percent of women with high-grade SIL or cancer, and limit the number of women with no cervical disease as HPV-DNA positive (false positives) to less than 5 percent. Given the equivalent or better sensitivity of HPV-DNA testing as compared to cytology, and because HPV-DNA testing may be easier to implement than cytology screening, HPV testing should be considered for primary cervical cancer screening in low-resource settings. Cost-effectiveness studies of HPV-DNA testing are needed to further evaluate its application.

Lytwyn, A. et al. Comparison of human papillomavirus DNA testing and repeat Papanicolaou test in women with low-grade cervical cytologic abnormalities: a randomized trial. Canadian Medical Association Journal 163(6):701-707 (September 19, 2000).
This Canadian study compared the accuracy of HPV-DNA testing with 6-month repeat Pap tests in detecting cervical intraepithelial neoplasia (CIN) II or III among women with atypical squamous cells of undetermined significance (ASCUS) or low-grade squamous intraepithelial lesions (LSIL). The authors found that HPV-DNA testing detected significantly more cases of CIN II or III than delayed repeat Pap smears, but at greater cost. A cost-effectiveness analysis determined that the cost of HPV-DNA testing was Canadian $3,003 per additional case of CIN II or III detected. The authors conclude that HPV-DNA testing resulted in better detection and in fewer women lost to follow-up.

Manos, M.M. et al. Identifying women with cervical neoplasia: using human papillomavirus DNA testing for equivocal Papanicolaou results. Journal of the American Medical Association 281(17):1605-1610 (May 5, 1999).
This study, conducted in California, USA, compared a newer screening method with a more traditional method to determine whether women with atypical squamous cells of underdetermined significance (ASCUS) could be better served by the newer screening method. A total of 995 women with ASCUS Pap test results participated in the study. Under the new method, HPV-DNA testing was conducted on 995 ASCUS specimens, and those cases found to be HPV positive were referred directly to colposcopy. The authors found that the sensitivity of this method for detecting high-grade squamous intraepithelial lesions (HSILs) was equivalent to or greater than the traditional method of taking repeat Pap smears. In addition, because this screening method uses the same specimen for both cytology and HPV testing, it reduces the number of colposcopic examinations and follow-up visits (as well as patient anxiety and loss to follow-up). The authors believed the savings achieved by fewer visits and procedures would offset the cost of implementing the HPV-DNA testing.

Schiffman, M. et al. Human papillomavirus DNA remains detectable longer than related cervical cytologic abnormalities. Journal of Infectious Diseases 186:1169-1172 (2002).
HPV persistence is typically measured by using repeat HPV DNA tests or cervical cytological tests over time. To study the timing of HPV DNA clearance as it relates to the regression of cervical abnormalities, researchers assessed a sample of 840 women at six-month intervals for two years. Results showed that HPV DNA persists longer than cervical cytologic abnormalities. These results support the conclusion that women with negative HPV tests are at low risk of having undetected cervical lesions. They also indicate that HPV DNA testing may have a role in triage of women with ambiguous cytology results.

Schiffman, M. et al. HPV DNA testing in cervical cancer screening: results from women in a high-risk province of Costa Rica. JAMA 283(1):87-93 (January 5, 2000).
This study evaluated use of an HPV screening test to detect high-grade cervical lesions and cancer among more than 9,000 sexually active women age 18 and older in Guanacaste province, Costa Rica. In addition to HPV testing, women underwent conventional Pap smears and visual screening using cervicography. Samples also were prepared for ThinPrep cytologic analysis. The study found that HPV screening (using the Digene Hybrid Capture II test with a detection threshold of 1.0 pg/mL for 13 oncogenic HPV types) resulted in detection of 88.4 percent of high-grade cervical lesions and cancers, with a specificity of 89 percent. When results were calculated by age tertile (18 to 30, 31 to 40, and 41 and older), specificity was highest (94%) for older women. The percentage of lesions and cancers detected was lower (75%) for the original Hybrid Capture test (which has a detection threshold of 10pg/mL for 11 oncogenic types). Overall, HPV-DNA testing using the Hybrid Capture II test was more sensitive than conventional Pap testing (88.4 versus 77.7%) for detection of high-grade lesions and cancers, but less specific (89% versus 94%). The authors concluded that, while more data are needed on various factors that can affect the results of HPV testing (such as analytic cut offs and geographic variation in HPV types), the method clearly has "come of age" technically and should be increasingly useful in cervical factors screening efforts.

Walboomers, J.M.M. et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. Journal of Pathology 189(1):12-19 (September 1999).
In this study, researchers reexamined data from an earlier study (Bosch et al. 1995) that had confirmed a global association between HPV infection and cervical cancer by detecting HPV DNA in 93 percent of more than 1,000 cervical cancer specimens collected worldwide. The reexamination of data found many HPV-positive samples that had been false-negative in the previous study due to integration events affecting L1 sequences, a possibility that had been suggested by Bosch et al. By reanalyzing samples and excluding inadequate specimens, the current study found the worldwide HPV prevalence in cervical carcinoma to be 99.7 percent. This represents the highest percentage reported to date. The authors argue that the extreme rarity of HPV-negative cervical cancer reinforces the rationale for HPV testing in addition to, or perhaps even instead of, conventional screening methods.

Womack, S.D. et al. HPV-based cervical cancer screening in a population at high risk for HIV infection. International Journal of Cancer 85:206-210 (2000).
This study, involving 466 Zimbabwean women, evaluated the potential for using HPV testing to screen HIV-infected women for cervical dysplasia. Pap smears and samples for HIV and HPV testing (using Digene's HC II test) were obtained from all women. The study found that over half of the women tested were HIV-positive. HIV positivity was associated with a twofold increase in HPV prevalence and a threefold increase in HGSIL prevalence. The sensitivity and specificity of HPV-DNA testing for HGSIL were 91 percent (95% CI = 78-97%) and 41 percent (95% CI = 35-48%) for HIV-positive women; 62 (95% CI = 32-86%) and 75 percent (95% CI = 68-80%) for HIV-negative women. The negative predictive value for the test for each group was over 95 percent. The authors conclude that HPV-DNA testing is not an ideal test for HGSIL in HIV-positive women because of its relatively low specificity. At the same time, they noted that since HIV-positive women infected with HPV are at high risk of cervical disease, they should be provided with regular follow-up care. The authors also note that the ultimate usefulness of HPV-DNA testing as a screening test for cervical cancer must be decided based on local health priorities, policies, and capabilities.

Wright, T.C. et al. HPV DNA testing of self-collected vaginal samples compared with cytologic screening to detect cervical cancer. Journal of the American Medical Association 283(1):81-86 (January 5, 2000).
This study evaluated use of HPV screening tests (using self-collected and clinician-obtained samples) to detect high-grade cervical lesions and cancer among more than 1,400 previously unscreened black South African women aged 35 to 65. In addition to HPV testing, women underwent conventional Pap smears, STI testing, visual inspection of the cervix with acetic acid wash (magnified and unmagnified), and cervicography. The sensitivity of HPV-DNA testing (see Schiffman 2000 for more information about the Digene Hybrid Capture II test) of self-collected vaginal samples was 66.1 percent for detection of high-grade lesions and cancer; the false-positive rate was 17.1 percent. The sensitivity of HPV-DNA testing of clinician-collected samples was 83.9 percent; the false-positive rate was 15.5 percent. In comparison, the sensitivity of conventional Pap smear (with low-grade SIL and higher cytologic abnormality classified as positive) was 60.7 percent, with a false-positive rate of 3.2 percent. In summary, HPV-DNA testing using self-collected vaginal samples was less specific, but as sensitive as conventional Pap testing for detecting high-grade cervical disease in women age 35 or older (it is important to note that self-collection occurred at a clinic after specific instructions on how to obtain the sample). The authors concluded that HPV-DNA testing using self-collected vaginal samples holds promise for simplifying cervical cancer screening in many settings, although low specificity remains a problem. They discussed the need to assess two-stage screening protocols (for example, HPV testing followed by a Pap smear or visual inspection) to improve specificity.

Ylitalo, N. et al. Consistent high viral load of human papillomavirus 16 and risk of cervical carcinoma in situ: a nested case-control study. Lancet 355: 2194-2198 (June 24, 2000).
The authors sought to clarify the association between HPV 16 viral load and cervical cancer by analyzing multiple smears from 478 women with cervical cancer (cases) and 608 controls up to 26 years prior to diagnosis. The results showed that smears from the cases showed an increased viral load of HPV 16 thirteen years or more before diagnosis, often when the smear was cytologically normal. Women with high HPV 16 viral loads had a relative risk of cancer at least 30 times greater than women negative for HPV 16 more than a decade before diagnosis. Among women infected with a high viral load before age 25, about one-quarter developed cervical cancer within 15 years. Based on these findings, the authors conclude that quantitative HPV-DNA testing, in addition to cytological screening, could identify women at high risk of cervical cancer.

Top of page

The promise of HPV vaccines

Coursaget, P. and Muñoz, N. Vaccination against infectious agents associated with human cancer. Cancer Surveys 33:355-381 (1999).
Infectious agents account for about one-quarter of all cancer cases in the developing world, which creates enormous potential for vaccines. This article reviews the rationale for and current state of research on prophylactic and therapeutic vaccines against hepatitis B virus, human papillomavirus, and Helicobacter pylori. The authors conclude that many technical and practical problems remain before a safe, effective, inexpensive HPV vaccine can be produced for mass use, but that lessons learned from the introduction of the hepatitis B vaccine should expedite the process.

De Vuyst, H. et al. Distribution of human papillomavirus in a family planning population in Nairobi, Kenya. Sexually Transmitted Diseases 30(2):137-142 (February 2003).
HPV type prevalence has been reported to vary by region. This article reports results from a cross-sectional study of 429 women to determine the prevalence and distribution of HPV types among women attending a family planning clinic in Nairobi, Kenya. The mean age of the women was 35.2 years (standard deviation 6.5 years). A positive finding on biopsy or curettage was referred to as the reference test. Results showed 30 women (7.0%) with LSIL (CIN 1), 29 cases (6.8%) of HSIL (9 CIN 2 and 20 CIN 3), and one case (0.23%) of invasive cancer. Results of the HPV tests using PCR revealed 190 women (44.3%) were HPV positive. Among HPV positive samples, the most common HPV types were HPV 52 (17.9%), HPV 16 (14.7%), HPV 35 (11.6%), and HPV 66 (9%). Among women with HSIL, the most common HPV types were HPV 16 (35.7%), HPV 52 (25%), and HPV 35 (17.9%). As expected, strong associations were reported between having HSIL and being infected with a high-risk HPV type. The relative risk (RR) for HSIL was greatest for HPV 16 (RR=88.5). HPV 35 had a RR of HSIL of 54.3 and HPV 52 was associated with a RR of 49.2. In contrast to other studies, HPV 16 was not the most prevalent type and the authors conclude that the high prevalence of HPV 52 and 66 in this East African population warrants further research to tailor vaccines to take into account regional variations.

Duggan-Keen, M.F. et al. Papillomavirus vaccines. Frontiers in Bioscience 3:1192-1208 (1998). Available at: www.bioscience.org/1998/v3/d/duggan/d1208.htm).
Both prophylactic and therapeutic vaccines are a promising approach to fighting HPV, given the epidemiology of HPV-related disease, the life cycle of the virus, and natural immunity against HPV. This review article compares the mechanisms, advantages, and disadvantages of various candidate vaccines, including virus-like particles, proteins, peptides, viral vectors, DNA vaccines, bacterial vectors, and dendritic cells. Assessing the efficacy of prophylactic vaccines raises the problem of choosing an appropriate study size and trial endpoint, and deciding how to measure significant antibodies and cytotoxic T-cells. Cost, ease of implementation, and stability are important considerations in developing a vaccine for widespread prophylactic use. Generating mucosal immunity poses a challenge, as does the genetic instability of tumors.

Galloway, D.A. Is vaccination against human papillomavirus a possibility? Lancet 351:22-24 (1998).
HPV potentially is amenable to vaccination because it is not prone to mutation and has just eight genes. Animal models suggest that both prophylactic and therapeutic vaccination is feasible. To make a substantial impact on the prevalence of cervical cancer, a prophylactic vaccine must be multivalent and must generate high titers of neutralizing antibodies at the mucosal surface. Goals for a therapeutic vaccine include eliminating residual cancer, the regression of existing disease, and preventing the progression of low-grade disease. The scientific basis for a therapeutic vaccine is weakened by two unresolved questions: Which antigen should be targeted? And what type of immune response mediates regression? Despite these and other challenges, the author concludes that the prospects for HPV vaccination are strong.

Goldie, S.J. et al. A comprehensive natural history model of HPV infection and cervical cancer to estimate the clinical impact of a prophylactic HPV-16/18 vaccine. International Journal of Cancer 106:896-904 (2003).
This study utilizes a decision-analytic model to estimate the impact of a prophylactic HPV-16/18 vaccine on LSIL, HSIL, and HPV prevalence and the reduction of cervical cancer. The model simulated the age-specific prevalence of LSIL, HSIL, and cervical cancer for a cohort of 100,000 women immunized at age 13. The model also explored other assumptions about vaccine efficacy, coverage, and waning immunity. According to model projections, with 100 percent coverage, a vaccine that is 98 percent effective will produce an approximate 98 percent decrease in cancers related to HPV type 16/18 and a 51 percent reduction in total cancers. The lower percent reduction in total cancers (51%) takes into account competing risks from non-16/18 HPV associated cancers; women vaccinated against HPV type 16/18 could acquire an alternative high-risk type HPV that develops into cervical cancer. A vaccine with 75 percent effectiveness would result in a reduction of 70 to 83 percent of HPV-16/18 associated cancers. The model demonstrated how the impact of the HPV-16/18 vaccine varied with the proportion of the population vaccinated. When a vaccine with 98 percent efficacy covered only 50 percent of the population, the reduction in cervical cancer dropped to less than 30 percent. However, a partially effective vaccine (50 to 75 percent effective) that covers 75 percent of the population would be more effective than a highly effective vaccine with low coverage.

Kahn, J.A. et al. Attitudes about human papillomavirus vaccine in young women. International Journal of STD & AIDS 14(5):300-306 (May 2003).
This study explored young women’s attitudes toward and intention to receive a human papillomavirus (HPV) vaccine if available. Fifty-two women between the ages of 18 and 30 years answered survey questions on vaccine acceptability. The majority of respondents knew that HPV was sexually transmitted and could be asymptomatic, that women with HPV may need more frequent Pap smears, and that an abnormal Pap smear may indicate HPV infection. Other knowledge questions received widely varied percentages of correct answers. Attitudes toward receiving the vaccine were positive for the majority, with 89 percent responding that it would be a good idea for themselves and for their daughters (81%). Receiving the vaccine did not appear to change women’s risk behaviors, with the majority of respondents stating that it would not influence the number of sexual partners, use of condoms, or anxiety about smoking cigarettes. Several variables were statistically significantly associated with young women’s intention to receive the vaccine, including knowledge level, personal beliefs about vaccination, beliefs that others would approve, and having a higher number of sexual partners.

Koutsky, L.A. et al. A controlled trial of a human papillomavirus type 16 vaccine. New England Journal of Medicine 347(21):1645-1651 (2002).
This article presents promising evidence of the effectiveness of a vaccine against HPV type 16, the most common HPV type linked to cancer. The authors conducted a randomized double-blinded trial of the effectiveness of the vaccine versus a placebo in 2,392 women with no history of prior abnormal cytology findings. Primary analysis was conducted on 1,533 of the women after a median of 17.4 months of follow-up. At follow-up, the vaccine was 100 percent effective: none of the women who received the vaccine developed persistent HPV 16 infection, precancerous cervical lesions, or cancerous growth. In contrast, 41 women who received the placebo developed persistent HPV 16 infection, with 9 of the 41 developing precancerous lesions. HPV 16 is responsible for about half of all cervical cancers; however, the authors caution that a vaccine that prevents more than one type of HPV may be "more advantageous." Researchers also caution that the long-term efficacy of the vaccine is currently unknown.

Lowy, D.R. and Schiller, J.T.  Papillomaviruses: prophylactic vaccine prospects. Biochimica et Biophysica Acta 1423: M1-M8 (1998).
This article concentrates on one type of potential prophylactic HPV vaccine: virus-like particles. The molecular structure and immunogenicity of viral capsids make virus-like particles a strong candidate vaccine. Studies in dogs, rabbits, and cows confirm that these particles protect against challenge with papillomavirus, probably by inducing neutralizing antibodies. However, it is not yet certain whether systemic immunization with virus-like particles will confer protection against genital mucosal HPV infections in humans. This may require a different vaccine protocol. Alternatively, nonstructural proteins might be added to increase their effectiveness as vaccines; these chimeric virus-like particles might be able to eliminate early cervical lesions that escaped neutralization.

Top of page