The incidence of invasive cervical cancer in HIV-positive women is higher than in the general population [1]. This elevated incidence varies from country to country depending on study site and characteristics of the populations under study [2–4]. The weaker association between HIV-infection and cervical cancer risk in some countries seems to be explained by competing risks of death (e.g.: in Africa), or early detection of pre-invasive lesions (e.g.: in the United States) [5].

Effective screening and early treatment of precancerous cervical lesions are key factors in preventing the progression to invasive cervical cancer in both HIV-positive and -negative women [6]. Recent guidelines recommend that, following two initial normal Pap-smears at a 6-month interval, all HIV-positive women should undergo annual cervical cytologic examination [7]. In addition, it is recommended that all immunosuppressed women with atypical squamous cells undergo colposcopy [7].

In the United States, approximately 80% of HIV-positive women reported a history of Pap-smear in the past year [8, 9]. Low incidence rates of invasive cervical cancer, similar to those among HIV-negative women, were found among adequately screened HIV-positive women [10]. In contrast, there is evidence that HIV-positive women do not participate sufficiently in cervical cancer screening in Southern European countries such as Italy and Spain, where cervical cancer incidence is more than 10-fold higher in women with AIDS than in general population [3, 5, 11, 12].

In Italy, one-half of women with invasive cervical cancer as the AIDS-defining illness (3% of all women with AIDS) had their first HIV-positive test ≥10 years before cancer diagnosis [12]. This long interval suggests a failure to stop the progression of precancerous lesions through screening, despite the knowledge of HIV infection.

A large cross-sectional study showed not only a scanty use of Pap-smear among HIV-positive Italian women, but also that these women did not know that a Pap-test was used to prevent cervical cancer [13]. In addition, a survey conducted among clinicians in 27 Italian HIV centers showed low compliance with the published guidelines on gynecologic follow-up of HIV-positive women in Italy [14].

The aim of the present study was to evaluate the recent history of Pap-smear in HIV-positive women in Italy followed in public HIV clinics. We also examined the sociodemographic, clinical, and organizational factors associated with the lack of adherence to screening recommendations.

A total of 1,002 HIV-positive women completed the questionnaire. The median age was 41 years and the vast majority of women (87%) fell in the 30-to-49-year age range. Eighty percent of our study women were born in Italy and nearly two-thirds reported sexual intercourse as the route of HIV acquisition. One hundred eighty (18%) women had a previous AIDS diagnosis and 56% had their first HIV-positive test 10 or more years prior to the questionnaire.

Nine percent of women reported no history of Pap-smear and 305 (34%) had abnormal Pap-smear results (Figure 1). Women with no history of Pap-smear were younger and more likely to be born in Central-Eastern Europe (15%) or Africa (29%) than women reporting previous Pap-smears (3% and 9%, respectively) (data not shown). HPV test was reported by only 27% of women (Figure 1).

Receiving screening advice from a gynecologist rather than an infectivologist (OR = 0.6; 95%CI: 0.4-0.9) and history of abnormal Pap-smear (OR = 0.5; 95%CI: 0.3-0.6) were associated with better screening participation (Table 1). The multivariate models, including additional terms for area of birth, mode of HIV acquisition, and time since first HIV-positive test, did not materially modify the risk estimates even if associations with age and first HIV-positive test were not statistically significant (Table 1).

Among the 305 women who reported a previous abnormal Pap-smear, treatment was recommended for 178 (58%); of whom, 173 (97%) complied. Among the 127 (42%) women who were referred for further testing but not treatment, 84 (66%) underwent all the recommended follow-up tests.

Among the 178 women to whom treatment was recommended, 145 (81%) reported colposcopy and all follow-up tests. Conization was the surgical procedure most frequently reported by patients (122, 69%), followed by hysterectomy (19, 11%), and cryotherapy (8, 4%).

This study was the first attempt to provide estimates of self-reported Pap-smear history in HIV-positive women attending infectivology units of a large Italian region. We found that 91% of women had had at least one Pap-smear in their lifetime, and 61% reported a Pap-smear in the last year. These proportions are similar to the ones reported by the general population of the same age and geographical area (95% and 47%, respectively) (Carrozzi and Bertozzi, personal communication), and substantially higher than those reported (43%) in 2001 by a physician-based Italian survey [14]. Increase in the number of Pap-smears performed, however, is difficult to evaluate, given the different study designs of previous studies conducted in Italy [13, 14]. The proportion of women who had a Pap-smear in the last year, however, remains suboptimal with respect to current guidelines regarding HIV-positive women (annual Pap-smear recommended to all patients), and is lower than the estimates (approximately 80%) reported in the USA [8, 9].

Factors associated with lack of recent Pap-smear were young age (<35 years), recent HIV diagnosis (≤2 years), and more advanced diseases (CD4 count at last visit <200 cells/μL). Conversely, receiving screening advice from a gynecologist significantly improved the adherence to Pap-smear. We found less influence of education level or country of birth than a previous cross-sectional study conducted in Rome [13]. As expected, study women reported a high proportion of previous abnormal Pap-smears (34%), in agreement with previous reports among HIV-positive women in Italy [13] and the United States [8, 9].

Overall, our study highlights the difficulties in following the recommended protocol of combining routine follow-up of HIV infection with gynecological examination, which is generally not performed in infectivology units [5, 10]. In addition, patient compliance needs to be improved [14] and health professionals (e.g., infectious disease specialists and gynecologists) should thoroughly inform HIV-positive women (particularly the disadvantaged ones) of the importance of Pap-smear to their health.

The present study has strengths and weaknesses. Strengths included the high participation among unselected women attending infectivology units. Approximately 1,500 HIV-positive women are estimated to live in the Emilia-Romagna Region [15, 16] and all women attending HIV follow-up (recommended every 6 months) were contacted in a 1-year period. A previous study in Northern Italy [19] showed that patients infected with HIV through injecting drug use, patients without AIDS diagnosis, or patients with higher CD4 counts are more likely to miss medical appointments and discontinue their follow-up. It is unlikely that HIV-positive women in the region received their follow-up visits elsewhere, given the high standard of medical care provided (free of charge) in Emilia-Romagna in comparison with other Italian areas. However, the possibility of some socio-demographic bias in women who did not attend regularly follow-up cannot be totally ruled out.

The number of self-reported Pap-smears among HIV-positive women in our study may be an overestimate as women tend to over-report their participation in cervical cancer screening in a given timeframe [20]. Confidentiality prevented us from linking women's reports with gynecological and cytological records, thus leaving substantial uncertainty about the actual additional tests and treatments performed. To validate self-reported Pap-smear use and to follow-up the treatment of HIV-positive women diagnosed with gynaecological lesions, a new study is being planned, which will adopt broader confidentiality rules and obtain a written consent to follow-up. The most important limitation of our survey, however, is the restriction to HIV-positive women followed by public clinics in one of the best organized regions of Italy, in terms of participation and quality of cervical cancer screening [21, 22].

Our study showed that progresses have been made to screen adequately HIV-positive women in Italy and that such efforts can also reach the most vulnerable populations (e.g.: immigrants). The opportunity to prevent cervical cancer in women living with HIV infection should not be missed; successful implementation of such screening programs will also teach valuable lessons that could then be applied to all women [5]. There is, however, still some scope for improvement, especially in the completeness of treatment and follow-up of HIV-positive women with an abnormal Pap-smear. Most importantly, much remains to be done to make sure that the experience of the Emilia-Romagna Region can be extended to other parts of Italy.

LDM is a Senior Scientist at Aviano Cancer Institute and coordinated a surveillance study on cancer in people with HIV/AIDS in Italy. SF is Head of Infections and Cancer Epidemiology Group at International Agency for Research on Cancer (Lyon, France) and has been carrying out numerous international studies on the association between infection and cancer. ACF is Head of Public Health department of Emilia-Romagna Region. ML and JP collaborated with LDM, PSdB collaborated with ACF. FG is president of AIDS Commission, Emilia-Romagna Region, FF is Chief of Romagna Cancer Registry.