**4.1 Cervical cancer screening in low resource countries**

236 Cancer Prevention – From Mechanisms to Translational Benefits

before diagnosis (NIH 1996). However, conventional Pap smears are not perfect: a single Pap smear is associated with false-negative rates of 10–25%, largely because of errors in sampling or interpretation. False-negative Pap smears are associated with 30% of the new

Newer Pap smear screening techniques using liquid-based media appear to decrease inadequate smears and also offer the possibility of direct HPV-DNA testing on collected specimens (ACOG 2009). A recent review of over 400 HIV-infected women who underwent both conventional and liquid-based cytologic screening found a significant decrease in the proportion of smears diagnosed as ASCUS/AGUS as well as the ASCUS/SIL ratio, with liquid-based preparations (Swierczynski 2002). HPV testing for cancer-associated HPV subtypes is currently used as a triage test to stratify risk in women with a cytology diagnosis of atypical squamous cells of undetermined significance (ASC-US), in postmenopausal women with a cytology diagnosis of LSIL and is also often used as an adjunct to cytology for primary screening in women older than 30 years (ACOG 2009). The currently used classification system for Pap smear results is known as the Bethesda classification, which includes an indication of adequacy of sample, whether the result is normal and, if abnormal,

With current US-based guidelines (ACOG 2009), Pap smears are recommended beginning at age 21 and every two years until age 30 if normal. After 30, if the last 3 tests have been normal, screening interval can be increased to 3 years. Providers may consider discontinuation of screening after age 65–70; if there have been no abnormal Pap smears in 10 years and no on-going risk factors. Colposcopy with biopsy of abnormal areas for histologic confirmation is recommended with ASCUS/+HPV or greater abnormality on Pap or with repetitive ASCUS, even if HPV-. The results of both Pap smear and colposcopy/biopsy are used to determine need to treatment, follow-up or further

Women with HIV infection are recommended to have more frequent screening with cervical cytology: twice in the first year after diagnosis of HIV and, if normal, annually thereafter (CDC 2009; ACOG 2010) More frequent Pap smears should be considered with previous abnormal Pap smears, with conservative follow-up of cervical dysplasia without treatment (after colposcopic evaluation to rule out HSIL), with other evidence of HPV infection and

The role of HPV-DNA testing in HIV+ women is unclear. In a WIHS substudy of HIV+ and HIV- women with normal baseline cytology, incidence of squamous intraepithelial lesions (SIL) were examined by baseline HPV DNA results and stratified by CD4 count. Through 3 year follow-up, incidence of any SIL was similar in HIV- and HIV+ with CD4>500 who had negative results for oncogenic HPV or all HPV, suggesting that similar cervical cancer screening practices may be applicable to both groups. On the other hand, after just 2 years follow-up, incidence of any SIL in HIV+ with CD4<500 was increased over HIV-, even among women with negative results for any HPV, suggesting that a closer screening strategy may be needed for women with lower CD4 counts (Harris et al. 1995). In two prospective studies of HIV-infected women with ASC-US, approximately 30% of participants had evidence of oncogenic HPV, a finding that would support the use of HPV testing in this population if HPV testing remained highly sensitive (Massad et al. 2004; Kirby et al. 2004). However, one of these studies reported a sensitivity of HPV testing for the

cases of cervical cancer each year (NIH 1996; Shingleton et al. 1995)

degree of abnormality (Solomon 2002).

after treatment for cervical dysplasia (Anderson 2005).

evaluation.

Cervical cancer screening is often simply not available in developing world settings. Barriers to cytology-based screening programs include poor health infrastructure, lack of trained cytology technicians and cytopathologists and cost. In addition Pap smears are not point-ofcare tests; they require the ability to notify women of abnormal results and to follow-up with further evaluation or treatment. However, in low resource settings, many or most women reside at some distance from health centers, have little access to or cannot afford effective transportation, and there is a lack of effective recall mechanisms for abnormal results (Anorlu 2008). In sub-Saharan Africa there was a 60-80% default rate among those with cytologic abnormalities-(Cronje 2004).

Prevention of cervical cancer by identification and treatment of cervical cancer precursors is key, since treatment resources for invasive disease are scarce. In 2002, the survival rate for invasive cervical cancer was 21% in sub-Saharan Africa vs. 70% in US (Parkin et al. 2005). This is related to the fact that most patients present at late stages, as well as a lack of effective treatment resources, including surgical expertise and radiotherapy (Ashraf 2003). When women present with advanced cervical cancer, palliative care resources are also limited; although morphine is on the WHO list of essential medications (http://whqlibdoc.who.int/hq/2011/a95053\_eng.pdf), one study of 47 African countries found that only 11 used morphine for chronic pain (Harding 2005).

Recent work has focused on service delivery models using alternatives to cytology for screening for cervical precancerous lesions in order to improve access to safe and effective treatment, minimize loss to follow-up and prioritize utilization of specialized care. Necessary programmatic components, regardless of the strategy employed, include leadership at national levels making preventing cervical cancer a priority and including development of strategies and guidelines; community awareness building; training of providers and continuing education; data collection systems and outcomes monitoring, including quality assurance measures; administrative management; patient recall and retention plans; and appropriate linkages to assure adequate supplies, timely and high quality laboratory testing, and referrals when needed for higher levels of care. In South Africa, where it is estimated that 1 in 26 women develop cervical cancer in their lifetime, a cervical screening program was initiated in 2001; it called for three free Pap smears, starting

Cervical Cancer Screening and Prevention for HIV-Infected Women in the Developing World 239

increased detection of nonvisible lesions in the endocervical canal (Pretorius et al. 2007;

In a cluster randomized trial in India, 31343 women screened with VIA were compared to 30958 controls, 30–59 years of age. Women who were VIA+ received colposcopy and biopsy with cryotherapy at the same visit for a colposcopic impression of dysplasia. With 7 years of follow-up, VIA was associated with a 24% reduction in cervical cancer incidence, stage 2 or higher, and 35% reduction in cervical cancer mortality (Sankaranarayanan 2007). In a study one year after cryotherapy, 648 women received both VIA by trained nurses and colposcopy and biopsy with VIA+ patients; 42 (6.5%) were referred for colposcopy and three of these had HSIL or cancer. Of those who were VIA-, colposcopically-based diagnosis was HSIL in only two cases (VIA sensitivity 60%, specificity 93.9%, PPV 7.1%, NPV 99.7%, comparable to Pap smear (Chumworathayi et al. 2008). A recent review of published studies of VIA accuracy with histology as the standard and CIN 2 as the outcome measure found

A pooled analysis of approximately 30,000 women from 17 population-based studies in China assessed the diagnostic accuracy of HPV testing for the detection of CIN 3 or greater; all positive tests were referred for colposcopy and biopsy. HPV-DNA testing had higher sensitivity of 97.5% and lower specificity of 85.1% , as compared to cytology (sensitivity 87.9%, specificity 94.7%) and VIA (sensitivity 54.6%, specificity 89.9%) (Zhao et al. 2010). HPV testing was evaluated and compared to both VIA and cytology in a cluster randomized trial in India. The trial had four arms, with >31,000 women aged 30–59 in each arm and 8 years of follow-up. In this study HPV by hybrid capture (detects 13 high risk subtypes) was compared to cytology, VIA and standard of care, which was no screening. With a positive result with any of the three screening tests, colposcopy and biopsy were performed and treatment with cryotherapy or LEEP was offered. When compared to the no screening group, HPV testing was associated with an approximately 50% reduction in detection of advanced cervical cancer and deaths from cervical cancer; neither VIA nor cytology was

In an analysis of the accuracy of five cervical cancer screening tests assessed in 11 studies in Africa and India, using colposcopically-directed biopsy as the standard and high grade CIN as the outcome, pooled sensitivity for VIA was 79.2%; for VILI 91.2%; for cytology 57% and for HPV testing (using Hybrid-Capture 2 assay) 62%(and pooled specificity for VIA was 84.7% for VILI 84.5% for cytology 93% and for HPV testing 94%). In this study pooled prevalence of CIN2+ was 2.3% and PPV was 11.6% and 12.9%, respectively for VIA and VILI and NPV >99% for both techniques. Accuracy of visual methods and cytology increased

In the setting of HIV infection, there are more limited data. In one study of 205 women correlating VIA with cytology with biopsy as the standard, VIA was more sensitive than Pap smear (76% vs. 57%, respectively) but less specific (83% vs. 95%, respectively); PPV for VIA was only 34% but was also low for cytology at 55%, but NPV for both techniques was high (97% for VIA, 95% for cytology). The prevalence for CIN in this patient population was 10.2% (Akinwuntan 2008). More recently VIA, HPV testing and cytology were compared to colposcopically-directed biopsy in 498 women in Kenya. Both HPV testing and Pap smear had higher sensitivity than VIA, with HPV showing greatest sensitivity (94%, 89% and 79%, respectively), while VIA was superior to HPV testing in terms of specificity (51% for HPV

sensitivity 79-82%, specificity 91-92% with PPV 9-10% (Sauvaget et al. 2011).

associated with statistically significant benefit (Sankaranarayanan et al. 2009).

over time, while performance of the HPV test was constant (Arbyn et al. 2008).

Cagle et al. 2010).

at age 30, at ten year intervals (Cronje 2003; Moodley 2006). By 2005–2006, 100% of primary health care clinics in South Africa had health professionals trained to conduct Pap smears, yet the screening rate was only 1.3% (van Schalkwyk 2008). Several studies have found a lack of awareness of cervical cancer as a disease among women, as well as stigma and cultural beliefs or perceptions related to the reproductive organs and symptoms that may delay care-seeking (Anorlu et al. 2008; Wellensiek et al. 2002; Anorlu et al. 2000); however, studies have also documented that health care workers also often have poor knowledge about cervical cancer (Tarwireyi et al. 2003; Ayinde and Omigbodun et al. 2003).

#### **4.2 Alternatives to cervical cytology**

Two primary strategies have been developed as alternatives to cytologic screening. The first technique utilizes visual inspection of the cervix without magnification after application of a dilute solution (3–5%) of acetic acid (VIA) or, less commonly, Lugol's iodine solution (VILI). Most studies report results with VIA. With visual inspection techniques, there are three possible results: negative, positive, or suspicious for cancer requiring referral and further evaluation and management. The accuracy of VIA/VILI depends on the ability to visualize the cervical transformation zone, the area where the original columnar epithelium covering the ectocervix has been replaced by squamous epithelium, and the area where oncogenesis begins. As women approach menopause and afterwards, the transformation zone recedes into the cervical canal and may no longer be visible, reducing accuracy of VIA (Cremer 2011). However, in younger women and women in whom the transformation zone is visible, the high negative predictive value of VIA (see below) suggests that significant lesions can generally reliably be excluded if VIA is negative. A major advantage of these techniques is the ability to offer treatment the same day, known as the single-visit approach (SVA).

The other major alternative to cytology is HPV testing. HPV-DNA testing, with detection of high-risk HPV subtypes, is reproducible and objective. HPV testing has been suggested as primary screening in place of cytology in the US and Europe (Kitchener et al. 2009; Cusick et al. 2006) and a negative HPV test predicts a less than 2% risk of developing cervical dysplasia (Naucler et al. 2009; Lonky et al. 2010; Mesher et al. 2010; Kitchener et al. 2009; Cusick et al. 2006). An advantage of HPV testing is that a pelvic exam is not required, but simply insertion of a swab into the vagina to obtain the sample; furthermore, studies have shown that accurate results can be obtained with self-testing, where the woman inserts the swab into her own vagina (Ogilvie et al. 2005; Balasubramanian et al. 2010), that this compares favorably to collection by clinicians (Bhatla et al. 2009; Petignat et al. 2007) and is acceptable to women (Mitchell et al. 2011; Lack et al. 2005). However, treatment of positive results clearly requires access to and good visualization of the cervix.

Sensitivity and specificity values for these screening strategies vary depending on the comparison technique used as "gold standard", as well as the detection goal, i.e., any cervical intraepithelial neoplasia (CIN) or only high grade CIN (CIN2 or CIN3) which are the immediate precursors to invasive cancer. Low-grade lesions (CIN1) may regress spontaneously up to 60% of the time (Cox et al. 2003) and are not routinely treated. Therefore, it is thought that the most appropriate detection goal is CIN 2 or higher (CIN2+), since these lesions are the ones likely to progress to cancer. Some studies have suggested that use of colposcopically-directed biopsy may overestimate sensitivity of VIA when compared to expanded biopsies, including endocervical curettage (ECC), likely due to

at age 30, at ten year intervals (Cronje 2003; Moodley 2006). By 2005–2006, 100% of primary health care clinics in South Africa had health professionals trained to conduct Pap smears, yet the screening rate was only 1.3% (van Schalkwyk 2008). Several studies have found a lack of awareness of cervical cancer as a disease among women, as well as stigma and cultural beliefs or perceptions related to the reproductive organs and symptoms that may delay care-seeking (Anorlu et al. 2008; Wellensiek et al. 2002; Anorlu et al. 2000); however, studies have also documented that health care workers also often have poor knowledge

Two primary strategies have been developed as alternatives to cytologic screening. The first technique utilizes visual inspection of the cervix without magnification after application of a dilute solution (3–5%) of acetic acid (VIA) or, less commonly, Lugol's iodine solution (VILI). Most studies report results with VIA. With visual inspection techniques, there are three possible results: negative, positive, or suspicious for cancer requiring referral and further evaluation and management. The accuracy of VIA/VILI depends on the ability to visualize the cervical transformation zone, the area where the original columnar epithelium covering the ectocervix has been replaced by squamous epithelium, and the area where oncogenesis begins. As women approach menopause and afterwards, the transformation zone recedes into the cervical canal and may no longer be visible, reducing accuracy of VIA (Cremer 2011). However, in younger women and women in whom the transformation zone is visible, the high negative predictive value of VIA (see below) suggests that significant lesions can generally reliably be excluded if VIA is negative. A major advantage of these techniques is the ability to offer treatment the same day, known as the single-visit approach (SVA).

The other major alternative to cytology is HPV testing. HPV-DNA testing, with detection of high-risk HPV subtypes, is reproducible and objective. HPV testing has been suggested as primary screening in place of cytology in the US and Europe (Kitchener et al. 2009; Cusick et al. 2006) and a negative HPV test predicts a less than 2% risk of developing cervical dysplasia (Naucler et al. 2009; Lonky et al. 2010; Mesher et al. 2010; Kitchener et al. 2009; Cusick et al. 2006). An advantage of HPV testing is that a pelvic exam is not required, but simply insertion of a swab into the vagina to obtain the sample; furthermore, studies have shown that accurate results can be obtained with self-testing, where the woman inserts the swab into her own vagina (Ogilvie et al. 2005; Balasubramanian et al. 2010), that this compares favorably to collection by clinicians (Bhatla et al. 2009; Petignat et al. 2007) and is acceptable to women (Mitchell et al. 2011; Lack et al. 2005). However, treatment of positive

Sensitivity and specificity values for these screening strategies vary depending on the comparison technique used as "gold standard", as well as the detection goal, i.e., any cervical intraepithelial neoplasia (CIN) or only high grade CIN (CIN2 or CIN3) which are the immediate precursors to invasive cancer. Low-grade lesions (CIN1) may regress spontaneously up to 60% of the time (Cox et al. 2003) and are not routinely treated. Therefore, it is thought that the most appropriate detection goal is CIN 2 or higher (CIN2+), since these lesions are the ones likely to progress to cancer. Some studies have suggested that use of colposcopically-directed biopsy may overestimate sensitivity of VIA when compared to expanded biopsies, including endocervical curettage (ECC), likely due to

results clearly requires access to and good visualization of the cervix.

about cervical cancer (Tarwireyi et al. 2003; Ayinde and Omigbodun et al. 2003).

**4.2 Alternatives to cervical cytology** 

increased detection of nonvisible lesions in the endocervical canal (Pretorius et al. 2007; Cagle et al. 2010).

In a cluster randomized trial in India, 31343 women screened with VIA were compared to 30958 controls, 30–59 years of age. Women who were VIA+ received colposcopy and biopsy with cryotherapy at the same visit for a colposcopic impression of dysplasia. With 7 years of follow-up, VIA was associated with a 24% reduction in cervical cancer incidence, stage 2 or higher, and 35% reduction in cervical cancer mortality (Sankaranarayanan 2007). In a study one year after cryotherapy, 648 women received both VIA by trained nurses and colposcopy and biopsy with VIA+ patients; 42 (6.5%) were referred for colposcopy and three of these had HSIL or cancer. Of those who were VIA-, colposcopically-based diagnosis was HSIL in only two cases (VIA sensitivity 60%, specificity 93.9%, PPV 7.1%, NPV 99.7%, comparable to Pap smear (Chumworathayi et al. 2008). A recent review of published studies of VIA accuracy with histology as the standard and CIN 2 as the outcome measure found sensitivity 79-82%, specificity 91-92% with PPV 9-10% (Sauvaget et al. 2011).

A pooled analysis of approximately 30,000 women from 17 population-based studies in China assessed the diagnostic accuracy of HPV testing for the detection of CIN 3 or greater; all positive tests were referred for colposcopy and biopsy. HPV-DNA testing had higher sensitivity of 97.5% and lower specificity of 85.1% , as compared to cytology (sensitivity 87.9%, specificity 94.7%) and VIA (sensitivity 54.6%, specificity 89.9%) (Zhao et al. 2010). HPV testing was evaluated and compared to both VIA and cytology in a cluster randomized trial in India. The trial had four arms, with >31,000 women aged 30–59 in each arm and 8 years of follow-up. In this study HPV by hybrid capture (detects 13 high risk subtypes) was compared to cytology, VIA and standard of care, which was no screening. With a positive result with any of the three screening tests, colposcopy and biopsy were performed and treatment with cryotherapy or LEEP was offered. When compared to the no screening group, HPV testing was associated with an approximately 50% reduction in detection of advanced cervical cancer and deaths from cervical cancer; neither VIA nor cytology was associated with statistically significant benefit (Sankaranarayanan et al. 2009).

In an analysis of the accuracy of five cervical cancer screening tests assessed in 11 studies in Africa and India, using colposcopically-directed biopsy as the standard and high grade CIN as the outcome, pooled sensitivity for VIA was 79.2%; for VILI 91.2%; for cytology 57% and for HPV testing (using Hybrid-Capture 2 assay) 62%(and pooled specificity for VIA was 84.7% for VILI 84.5% for cytology 93% and for HPV testing 94%). In this study pooled prevalence of CIN2+ was 2.3% and PPV was 11.6% and 12.9%, respectively for VIA and VILI and NPV >99% for both techniques. Accuracy of visual methods and cytology increased over time, while performance of the HPV test was constant (Arbyn et al. 2008).

In the setting of HIV infection, there are more limited data. In one study of 205 women correlating VIA with cytology with biopsy as the standard, VIA was more sensitive than Pap smear (76% vs. 57%, respectively) but less specific (83% vs. 95%, respectively); PPV for VIA was only 34% but was also low for cytology at 55%, but NPV for both techniques was high (97% for VIA, 95% for cytology). The prevalence for CIN in this patient population was 10.2% (Akinwuntan 2008). More recently VIA, HPV testing and cytology were compared to colposcopically-directed biopsy in 498 women in Kenya. Both HPV testing and Pap smear had higher sensitivity than VIA, with HPV showing greatest sensitivity (94%, 89% and 79%, respectively), while VIA was superior to HPV testing in terms of specificity (51% for HPV

Cervical Cancer Screening and Prevention for HIV-Infected Women in the Developing World 241

collection by clinicians (Bhatla et al. 2009; Petignat et al. 2007) and is acceptable to women (Mitchell et al. 2011; Lack et al. 2005). However, treatment of positive results clearly requires access to and good visualization of the cervix. On the other hand, VIA is inherently subjective and less reproducible; however a number of studies have shown its sensitivity to be similar to or higher than that of cytology, it is inexpensive, can be task-shifted to lower level health workers and, most notably allow single visit treatment. It remains unclear how the greater prevalence of HPV infection in HIV-infected women will affect performance

**Outcome Sensitivity (%)** 

**VIA** CIN 2 + 65-91-79 74-95-85 **VILI** CIN 2 + 74-98-91 73-92-85 **Cytology** CIN 2 + 33-82-57 87-99-93 **HPV Hybrid Capture 2** CIN 2 + 48-68-62 92-95-95

Table 1. Sensitivity and specificity of 4 screening tests for CIN 2+ cervical lesion, minimum,

Minimum, maximum and pooled measure of sensitivity and specificity adapted from Table 3. Marc Arbyn1, Rengaswamy Sankaranarayanan, Richard Muwonge, Namory Keita, Amadou Dolo, Charles Gombe Mbalawa, Hassan Nouhou, Boblewende Sakande, Ramani Wesley, Thara Somanathan, Anjali Sharma, Surendra Shastri and Parthasarathy Basu. Pooled analysis of the accuracy of five cervical cancer screening tests assessed in eleven studies in

As yet, there has been fairly limited programmatic experience with these screening techniques in the setting of HIV. Jhpiego has introduced a VIA/SVA program in Guyana, Cote d'Ivoire and Tanzania, screening over 16,000 women from 2009–2010. Services were provided by trained nurses and midwives at HIV care and treatment sites and general health facilities. In all 3 countries HIV+ women were more likely to be VIA+ than HIV-/unknown women. In all 3 countries HIV+ women who were VIA+ were more likely to have large lesions (occupying >75% cervix) and therefore ineligible for cryotherapy. Eighty-two% of eligible women had same-day treatment with cryotherapy; of those who postponed, 44% did not return for treatment (Anderson 2011). These findings confirm cytology-based studies that HIV+ women are at greater risk for cervical dysplasia and is consistent with other studies suggesting that a larger volume of the cervix involved. It also supports the feasibility of VIA/SVA from a programmatic standpoint and suggests that this approach results in reduction of loss to follow-up as compared to screening requiring a

As HPV testing becomes more affordable and accessible, and particularly if it can be done as a genuinely rapid point-of-care test, it is possible that a hybrid approach to cervical screening may maximize accuracy and minimize unnecessary treatment. In this scenario, HPV testing could be the primary screening method, with VIA performed on those who are

**(min-max-pooled)** 

**Specificity (%) (min-max-pooled)** 

characteristics of HPV testing as a primary screen in this population.

maximum and pooled measures

subsequent visit.

Africa and India. Int. J. Cancer: 123, 153–160 (2008).

testing, 60% for cytology, 63% for VIA). PPV was low for all three methods (18% for HPV testing, 20% for Pap, 17% for VIA) and NPV was high (99% for HPV testing, 98% for PAP, 95% for VIA) (Chung CROI 2011). In a "see and treat" program in Uganda HIV-infected women had higher likelihood of inflammation, resulting in an increase in false-positive results (Mutyaba et al. 2010).A randomized clinical trial of VIA and HPV testing , with cryotherapy treatment for positive results, was performed among over 6500 women in South Africa, of whom 956 were HIV+. Women were followed for up to 36 months after randomization with colposcopy and biopsy to determine the study endpoint of CIN2+. Screen-and-treat using HPV testing significantly reduced CIN2+ in both HIV+ and HIVwomen at follow-up (relative risk 0.20 [95% CI 0.06–0.69] and 0.31 [95% CI 0.20–0.50], respectively), compared to controls with sensitivity of 94% and PPV of 29.9% in HIV+ women; VIA also reduced the likelihood of CIN2+ at follow-up, but to a lesser degree and only reached statistical significance in HIV+ women (RR 0.51 [95% CI 0.29–0.89]), where sensitivity was 63.9% and PPV was (Kuhn et al. 2010). Because HIV+ women had higher rates of CIN2+, both screen-and-treat strategies had a stronger impact at the population level in HIV+ women than in HIV- women. It was estimated that for every 100 women screened HPV screen and treat could prevent 11.9 CIN2+ cases in HIV+ women and VIA screen and treat could prevent 7.4 cases of CIN2+ in HIV+ women.

A comparison of VIA and HPV testing is found in Table 1. In computer-based models, both VIA and HPV testing are cost-effective alternatives to conventional cytology-based programs, which usually require three visits, in low resource settings (Goldie et al. 2005). While HPV testing is more objective and reproducible and has higher sensitivity than VIA, it has some significant disadvantages for lower resource countries. A rapid HPV test has now been developed (Care-HPV, Qiagen Inc.) but is not yet commercially available), but costs (estimated at \$5–10 US) are still largely prohibitive in areas where annual health expenditures are often under <\$5 per person. Furthermore, although results can be available the same day, they are not available instantly, but are run in batches and require up to 3 hours for actual testing. VIA is inherently subjective and less reproducible, as reflected by low inter-rater agreement was seen among both midwives and gynecologists compared to lead reference physicians regarding cryotherapy treatability (Gage et al. 2009), but performance seems to improve with experience. A number of studies have shown the sensitivity of VIA to be similar to or higher than that of cytology. It is inexpensive, with low costs, including cost of supplies and cost to the patient because of the ability to treat abnormalities at the same visit. VIA can be task-shifted to lower level health workers and, most notably, allows single visit screening and treatment. Furthermore, VIA has been shown to be safe, feasible and acceptable in multiple studies (Phongsavan et al. 2011; Sankaranarayanan et al. 2007; Palanuwong 2007; Sanghvi et al. 2008). The positive predictive value of VIA is low and fairly common conditions such as cervicitis may cause false positive VIA results (Davis-Dao 2008). It remains unclear how the greater prevalence of HPV infection in HIV-infected women will affect performance characteristics of HPV testing as a primary screen in this population.

An advantage of HPV testing is that pelvic exam is not required, but simply insertion of a swab into the vagina to obtain the sample; furthermore, studies have shown that accurate results can be obtained with self-testing, where the woman insert the swab into her own vagina (Ogilvie et al. 2005; Balasubramanian et al. 2010), that this compares favorably to

testing, 60% for cytology, 63% for VIA). PPV was low for all three methods (18% for HPV testing, 20% for Pap, 17% for VIA) and NPV was high (99% for HPV testing, 98% for PAP, 95% for VIA) (Chung CROI 2011). In a "see and treat" program in Uganda HIV-infected women had higher likelihood of inflammation, resulting in an increase in false-positive results (Mutyaba et al. 2010).A randomized clinical trial of VIA and HPV testing , with cryotherapy treatment for positive results, was performed among over 6500 women in South Africa, of whom 956 were HIV+. Women were followed for up to 36 months after randomization with colposcopy and biopsy to determine the study endpoint of CIN2+. Screen-and-treat using HPV testing significantly reduced CIN2+ in both HIV+ and HIVwomen at follow-up (relative risk 0.20 [95% CI 0.06–0.69] and 0.31 [95% CI 0.20–0.50], respectively), compared to controls with sensitivity of 94% and PPV of 29.9% in HIV+ women; VIA also reduced the likelihood of CIN2+ at follow-up, but to a lesser degree and only reached statistical significance in HIV+ women (RR 0.51 [95% CI 0.29–0.89]), where sensitivity was 63.9% and PPV was (Kuhn et al. 2010). Because HIV+ women had higher rates of CIN2+, both screen-and-treat strategies had a stronger impact at the population level in HIV+ women than in HIV- women. It was estimated that for every 100 women screened HPV screen and treat could prevent 11.9 CIN2+ cases in HIV+ women and VIA

A comparison of VIA and HPV testing is found in Table 1. In computer-based models, both VIA and HPV testing are cost-effective alternatives to conventional cytology-based programs, which usually require three visits, in low resource settings (Goldie et al. 2005). While HPV testing is more objective and reproducible and has higher sensitivity than VIA, it has some significant disadvantages for lower resource countries. A rapid HPV test has now been developed (Care-HPV, Qiagen Inc.) but is not yet commercially available), but costs (estimated at \$5–10 US) are still largely prohibitive in areas where annual health expenditures are often under <\$5 per person. Furthermore, although results can be available the same day, they are not available instantly, but are run in batches and require up to 3 hours for actual testing. VIA is inherently subjective and less reproducible, as reflected by low inter-rater agreement was seen among both midwives and gynecologists compared to lead reference physicians regarding cryotherapy treatability (Gage et al. 2009), but performance seems to improve with experience. A number of studies have shown the sensitivity of VIA to be similar to or higher than that of cytology. It is inexpensive, with low costs, including cost of supplies and cost to the patient because of the ability to treat abnormalities at the same visit. VIA can be task-shifted to lower level health workers and, most notably, allows single visit screening and treatment. Furthermore, VIA has been shown to be safe, feasible and acceptable in multiple studies (Phongsavan et al. 2011; Sankaranarayanan et al. 2007; Palanuwong 2007; Sanghvi et al. 2008). The positive predictive value of VIA is low and fairly common conditions such as cervicitis may cause false positive VIA results (Davis-Dao 2008). It remains unclear how the greater prevalence of HPV infection in HIV-infected women will affect performance characteristics of HPV testing as a

An advantage of HPV testing is that pelvic exam is not required, but simply insertion of a swab into the vagina to obtain the sample; furthermore, studies have shown that accurate results can be obtained with self-testing, where the woman insert the swab into her own vagina (Ogilvie et al. 2005; Balasubramanian et al. 2010), that this compares favorably to

screen and treat could prevent 7.4 cases of CIN2+ in HIV+ women.

primary screen in this population.

collection by clinicians (Bhatla et al. 2009; Petignat et al. 2007) and is acceptable to women (Mitchell et al. 2011; Lack et al. 2005). However, treatment of positive results clearly requires access to and good visualization of the cervix. On the other hand, VIA is inherently subjective and less reproducible; however a number of studies have shown its sensitivity to be similar to or higher than that of cytology, it is inexpensive, can be task-shifted to lower level health workers and, most notably allow single visit treatment. It remains unclear how the greater prevalence of HPV infection in HIV-infected women will affect performance characteristics of HPV testing as a primary screen in this population.


Table 1. Sensitivity and specificity of 4 screening tests for CIN 2+ cervical lesion, minimum, maximum and pooled measures

Minimum, maximum and pooled measure of sensitivity and specificity adapted from Table 3.

Marc Arbyn1, Rengaswamy Sankaranarayanan, Richard Muwonge, Namory Keita, Amadou Dolo, Charles Gombe Mbalawa, Hassan Nouhou, Boblewende Sakande, Ramani Wesley, Thara Somanathan, Anjali Sharma, Surendra Shastri and Parthasarathy Basu. Pooled analysis of the accuracy of five cervical cancer screening tests assessed in eleven studies in Africa and India. Int. J. Cancer: 123, 153–160 (2008).

As yet, there has been fairly limited programmatic experience with these screening techniques in the setting of HIV. Jhpiego has introduced a VIA/SVA program in Guyana, Cote d'Ivoire and Tanzania, screening over 16,000 women from 2009–2010. Services were provided by trained nurses and midwives at HIV care and treatment sites and general health facilities. In all 3 countries HIV+ women were more likely to be VIA+ than HIV-/unknown women. In all 3 countries HIV+ women who were VIA+ were more likely to have large lesions (occupying >75% cervix) and therefore ineligible for cryotherapy. Eighty-two% of eligible women had same-day treatment with cryotherapy; of those who postponed, 44% did not return for treatment (Anderson 2011). These findings confirm cytology-based studies that HIV+ women are at greater risk for cervical dysplasia and is consistent with other studies suggesting that a larger volume of the cervix involved. It also supports the feasibility of VIA/SVA from a programmatic standpoint and suggests that this approach results in reduction of loss to follow-up as compared to screening requiring a subsequent visit.

As HPV testing becomes more affordable and accessible, and particularly if it can be done as a genuinely rapid point-of-care test, it is possible that a hybrid approach to cervical screening may maximize accuracy and minimize unnecessary treatment. In this scenario, HPV testing could be the primary screening method, with VIA performed on those who are

Cervical Cancer Screening and Prevention for HIV-Infected Women in the Developing World 243

a loop and electrosurgical generators; the loops are available in a variety of sizes and can also be passed through tissue with several passes if needed to remove the entire area of abnormality, therefore tailoring the procedure to the size of the lesion on the extocervix. The depth of excision can be fairly precisely determined and controlled. LEEP can be performed under local anesthesia and removes less tissue than conization. Complications are less common than after conization (postoperative bleeding 0–8%, infection 0–2%, cervical stenosis 4.3–7.7%; preterm delivery-no increased risk) (Hoffman and Mann 2010; Arbyn et

Cryotherapy uses nitrous oxide or carbon dioxide to freeze the cervix and destroy precancerous tissues and can be performed without anesthesia or electricity. Freezing is done in two cycles of three minutes with five minutes of thawing in between (Sellors 2003). Mid-level providers have been trained successfully to perform cryotherapy safely and with a high degree of acceptability (Bradley 2006; Nene et al. 2008). Carbon dioxide (CO2) provides a less expensive alternative to nitrous oxide and produces equally low temperatures (Sirivongrangson 2007). However, some studies report that CO2 may cause cryotherapy blockage up to 50% (Sirivongrangson 2007; Winkler 2010). The quality of cryotherapy devices rather than gas could be responsible for variant temperature (Winkler 2010). The freeze-clear-freeze technique, often employed to reduce gas blockage, may also produce temperatures not sufficiently low enough for treatment to be effective (Winkler 2010). It is important for providers to give adequate treatment by paying attention to visual cues that confirm freezing. Cryotherapy does not provide a tissue specimen. For cryotherapy to be optimally effective, the entire lesion and ideally the entire transformation zone must be visible and the lesion should occupy less than three-fourths of the transformation zone (Sankaranarayanan 2008). Adverse effects after cryotherapy are relatively uncommon and generally minor, reported in 1–2% of women (Cirisano 1999; Sankaranarayanan 2007; Nene 2008). Discomfort usually resolves within a week after treatment (Chamot 2010; Bradley 2006). Significant bleeding after cryotherapy is uncommon. After cryotherapy, watery discharge generally continues for several weeks. Post-treatment infection and pelvic inflammatory disease for LEEP and cryotherapy are both rare (Chamot 2010). The risk of cervical stenosis after cryotherapy is low (<1%) (Loobuyck and Duncan 1993) and the risk of obstetric complications, particularly preterm delivery, is lower after cryotherapy than after excisional procedures. Complications after

cryotherapy do not differ among developed and developing countries.

In lower resource areas, the advantages of cryotherapy are significant: the ability to perform the procedure without anesthesia, lack of need for electricity, the lower level of technical expertise required to perform the procedure and lower costs. A significant advantage of LEEP is the ability to examine a tissue sample histologically; however, in areas with few resources, pathological evaluation is usually not available. Studies have generally found LEEP to be associated with somewhat higher cure rates (absence of persistent or recurrent disease) than cryotherapy (Melnikow et al. 2009). A randomized clinical trial of cryotherapy and LEEP for treatment of histologically confirmed high-grade cervical dysplasia found that LEEP had higher overall cure rate of 96.4% as compared to 88.3% for cryotherapy (p=0.026) (Chirenje et al. 2001). Treatment methods are generally consistent among high-income and low-income countries (Luciani 2008). Cryotherapy is less effective in older women, where the transformation zone and any cervical lesions are more likely to recede into the cervical

al. 2008).

HPV+ to assess the presence of disease and the feasibility of treatment. Those who are HPVnegative would receive no further screening. Alternatively, VIA could be the initial screen, with those who are VIA+ having HPV testing to improve PPV.
