| |
Antiretroviral Therapy and the Clinical Evolution of Human Papillomavirus Associated Genital Lesions in HIV-Positive Women
|
| |
| |
"...use of effective antiretroviral therapy appears to be crucial to obtaining regression of high-grade lesions by surgical treatment, as well as to the management of low-grade lesions...there appears to be benefit when combined with surgical..." men and anal HPV-associated lesions, see below
Clinical Infectious Diseases 2004;38:737-742
Annarosa Del Mistro,1 Roberta Bertorelle,1 Marzia Franzetti,2 Annamaria Cattelan,2 Angela Torrisi,3 Maria Teresa Giordani,4 Roberto Sposetti,5 Emanuela Bonoldi,6 Lolita Sasset,2 Laura Bonaldi,1 Daria Minucci,3 and Luigi Chieco-Bianchi1
1Servizio Citologia Diagnostica Molecolare Oncologica, 2Malattie Infettive, and 3Servizio Ginecologia Oncologica, Azienda Ospedaliera di Padova, Padua; and 4Malattie Infettive, 5Ginecologia e Ostetricia, and 6Anatomia Patologica, Ospedale San Bortolo, Vicenza, Italy
ABSTRACT
The effect of antiretroviral therapy on the natural history of human papillomavirus (HPV)associated genital lesions was evaluated in 201 human immunodeficiency virus (HIV)infected women who were followed-up for 16 years.
Gynecologic examinations were performed every 612 months. HPV sequences in cervico-vaginal cells, analyzed by polymerase chain reaction and typed by restriction fragment length polymorphism analysis, were repeatedly detected in 126 women; 29 had transient HPV infection. Genital lesions were found in 137 patients; prevalence was comparable in women who were receiving different antiretroviral regimens. Regression of low-grade lesions was more prevalent among patients receiving highly active antiretroviral therapy than among those receiving other regimens; high-grade lesions regressed in the majority of cases, regardless of antiretroviral therapy. HPV infection persisted in nearly 80% of the cases.
In conclusion, our data show that antiretroviral therapy does not prevent the development of HPV-associated lesions and does not eliminate HPV infection; therefore, early and strict gynecologic follow-up of HIV-infected women is warranted.
Because the continuous increase in the incidence of invasive cervical carcinoma after 1996 indirectly indicates that HAART has no preventive effect, the increased longevity associated with receipt of HAART may increase the incidence of cervical and anal cancer in HIV-infected subjects in the future.
Our data indicate that patients with HIV infection benefit from surgical therapy when performed in conjunction with any antiretroviral therapy regimen.
Clinical outcome of LGSILs was negatively correlated with persistent HPV infection, consistent with the notion that persistent HPV infection has a central role in cervical carcinogenesis. On the other hand, the results of multivariate analysis suggest that the outcome of lesions in HIV-infected patients is the result of a complex interplay among viral and host factors, influenced by the receipt of antiretroviral therapy.
HGSILs were diagnosed in similar proportions in the different antiretroviral therapy groups; also, rates of anal SILs reported in a group of 98 men during the pre-HAART period were similar to those reported during the HAART period. These data seem, therefore, to indicate that HAART, per se, does not prevent the occurrence of high-grade cervical and anal HPV-associated lesions. On the other hand, our findings on the successful clinical outcome associated with surgically excised HGSILs in patients receiving antiretroviral therapy, which are consistent with the findings reported by Robinson et al., suggest that lesion excision in conjunction with an effective antiretroviral therapy achieves a good response in the majority of the cases.
Moreover, early intervention also appears determinant in avoiding more-extensive surgery, such as abdominal hysterectomy, that involves an increased risk of complications in HIV-infected women.
SEE DISCUSSION BY AUTHORS AT END OF THIS REPORT
INTRODUCTION
The natural history of HIV infection has been profoundly modified by the use of protease inhibitors (PIs) in combination with reverse-transcriptase inhibitors (i.e., HAART). Indeed, deaths due to AIDS-related diseases have decreased by 75% since the introduction of HAART, and several opportunistic complications, as well as some opportunistic neoplasms (e.g., Kaposi sarcoma and primary CNS lymphoma), have been greatly reduced.
In 1993, human papillomavirus (HPV)associated cervical intraepithelial neoplasia (CIN) was classified as a stage B condition and invasive cervical cancer was classified as an AIDS-defining condition according to the Centers for Disease Control and Prevention (CDC) classification system. With regard to HPV-induced genital lesions, HIV-infected women differ from immunocompetent ones in several respects: (1) HPV DNA is detected more frequently and has a greater tendency to persist in HIV-infected women; (2) HPV-induced dysplastic lesions are both more prevalent and more severe in those women with a greater degree of immunosuppression; and (3) squamous intraepithelial lesions (SILs) managed by conventional surgical treatments show higher persistence and recurrence rates. It has been suggested that HIV may influence the clinical outcome of an HPV infection, mainly by decreasing systemic and local cell-mediated immune response, although virus-virus interactions could also play a role.
The natural history of cervical disease in HIV-positive women is, as yet, poorly understood. Great interest accompanied the large-scale availability of HAART to HIV-infected women (beginning during the period 19961997 in developed countries), and several studies have evaluated the impact of HAART on HPV-associated lesions; preliminary findings of HAART efficacy were followed by data suggestive of limited or no effect on both anogenital and cutaneous HPV infections. Moreover, consistent with a lack of efficacy of HAART in decreasing the development and progression of cervical dysplastic lesions, a meta-analysis involving 47,936 HIV-infected subjects from North America, Europe, and Australia for the period 1992-1999 concluded that there had not been a significant change in the incidence rate of invasive cervical cancer. Here, we report the results of a prospective study, begun in 1994, involving genital HPV infections and associated lesions in HIV-positive women.
METHODS
Study population. Starting in August 1994 (for Vicenza) and March 1996 (for Padova), all HIV-infected women attending the Infectious Disease Units of Vicenza and Padova City Hospitals, located in northeast Italy, were offered gynecologic consultation in the units themselves. The study population consisted of all women who were enrolled between the start date and 30 September 2000 and were followed-up through 30 September 2002. Informed consent was obtained in accordance with institutional guidelines.
Study plan. At entry, sociodemographic data and risk factors for HIV infection were obtained, as well as gynecological, obstetrical, and sexual histories. Clinical status, CD4+ cell counts, HIV viral load (available after 1996), and antiretroviral therapy received were recorded. HAART was defined as any combination of therapy that included >1 PI, 2 nucleoside reverse-transcriptase inhibitors (NRTIs) in combination with a nonnucleoside reverse transcriptase inhibitor (NNRTI), or 3 NRTIs; non-HAART was defined as any regimen that included 1 or 2 drugs but did not include a PI. The HAART, non-HAART, and no therapy groups each consisted of subjects who consistently received the same regimen. Women whose antiretroviral regimen was changed from HAART to non-HAART (or vice versa) during follow-up because of side effects or treatment failure formed a separate group (designated the "miscellaneous antiretroviral therapy" group). After gynecologic and colposcopic examinations, ectocervical and/or vaginal and endocervical samples were obtained using an Ayre spatula, for Papanicolaou (Pap) evaluation, and a cotton swab, for HPV analysis.
Follow-up visits were scheduled for every 612 months. At each visit, data on clinical status, CD4+ cell count, HIV load, and antiretroviral therapies received were updated, and a gynecologic examination was performed. Cervical intraepithelial lesions were managed according to protocols for immunocompetent women, as follows: close follow-up for low-grade SILs (LGSILs), colposcopy-directed biopsy for high-grade SILs (HGSILs), and conization or loop electrosurgical excision for CIN of grade II or greater. Regression was defined as a normal Pap test finding or a diagnosis of a lower degree of SILs at subsequent examination or after excisional therapy. Persistence was defined as the presence of the same grade of lesion at follow-up or after treatment as was present at the initial diagnosis. Progression was determined by the presence of a more severe lesion than that observed at the initial diagnosis.
Laboratory procedures. HIV infection was determined by EIA for HIV antibodies and was confirmed by the results of Western blot testing. CD4+ and CD8+ cell counts were determined by flow cytometry. Plasma HIV RNA levels were assayed by means of an RT-PCR technique (HIV-1 Amplicor Monitor; Roche), or a nucleic acid sequencebased amplification technique (NASBA; Organon Teknika).
The Pap smears were classified according to the Bethesda System. The cervico-vaginal cell suspensions were processed and analyzed for HPV sequences, as described elsewhere. In brief, DNA was amplified with the consensus primers MY09/MY11, followed by RFLP analysis of the amplified fragments for HPV type definition. Classification of HPV types as either low or high risk was done according to Munoz et al.
In reference to HPV and SIL status, subjects were classified on the basis of data recorded at baseline as well as on the basis of their first abnormal results, and these classifications were correlated with baseline CD4+ cell counts and HIV RNA levels. Clinical outcome was then analyzed in relation to CD4+ cell count, HPV type, and age at SIL detection, as well as in relation to evolution of HPV infection (i.e., whether the infection was transient or persistent) and receipt of antiretroviral therapy during follow-up; for patients with HGSILs, the gynecological treatment was also taken into account.
RESULTS
Between 1 August 1994 and 30 September 2002, 201 women with HIV infection attended >1 follow-up visit. The median age at study entry was 33 years (range, 21-76 years), and all but 17 of the subjects (14 from Africa, 1 from Brazil, 1 from Argentina, and 1 from China) were white. In all cases, HIV infection had been acquired by injection drug use or through heterosexual contact (in nearly equal proportions). Length of follow-up was <1 year for 35 (18%) of the subjects, 13 years for 67 (33%), and >3 years for 99 (49%). During follow-up, 19 subjects died; cause of death was linked to HIV infection in all but 2 of these patients (who died of posthysterectomy surgical complications and breast cancer, respectively).
During the observation period, 22 patients (11%) did not receive any antiretroviral therapy, 49 (24%) received treatment with a non-HAART combination, 74 (37%) received HAART, and 56 (28%) received >1 type of drug regimen (i.e., switched between non-HAART and HAART). At study entry, the CD4+ cell count was <200 cells/mm3 in 68 (34%) of the patients, between 200-500 cells/mm3 in 97 (48%), and >500 cells/mm3 in 36 (18%); the median CD4+ cell count was 292 cells/mm3 (range, 2-1445 cells/mm3). The HIV load (available for 124 subjects) was <4000 HIV RNA copies/mL in 73 (59%) of the patients, 4000-100,000 in 34 (27%), and >100,000 in 17 (14%).
HPV infection was detected at least once in 155 (77%) of the subjects, and only 46 (23%) had test results that were repeatedly negative for HPV. Among the many different HPV types detected, type 16 was the most frequent and was identified at least once (at entry or during follow-up) in 31 (20%) of 155 subjects. It was followed in frequency by types 61 and 53 (each in 13% of subjects); types 58, 31, 6, 11, 18, and 56 (each in 5%-10% of subjects); and types 33, 52, 54, 62, 66, and 84 (each in 4% of subjects). A mixed infection was detected at least once in 26% of the HPV-positive patients. HPV sequences were detected in >2 samples obtained from each of 126 patients (62.5%). Persistent infection with the same virus type(s) was present in 83 of the HPV-positive patients (41.5%) (high-risk types were present in 71% of 83 patients, and low-risk types were present in 29%), and 43 patients (21%) had different types, suggesting regression of one infection and reinfection with others. In 13 subjects, HPV type changed after surgical excision of a CIN lesion of grade II or greater (at 36 months of follow-up, 4 of these patients had negative results of tests for HPV). Only 29 (14.5%) of the patients experienced a transient infection (of these 29 patients, 59% had high-risk infections and 41% had low-risk infections). No statistically significant differences in CD4+ cell count or HIV load were observed in relation to HPV infection status (data not shown). Prevalence of LGSILs was comparable among subjects with no HPV infection (17 [37%] of 46 had LGSILs), transient HPV infection (16 [55%] of 29), and persistent HPV infection (49 [36%] of 126). The prevalence of HGSILs, in contrast, was 2% (1 of 46 patients) among HPV-negative subjects, 14% (4 of 29) among patients with transient HPV infection, and 39% (48 of 126) among patients who were consistently HPV positive (considering same and different HPV types together for statistical analysis, P < .0001).
LGSILs were detected in 73 subjects at study entry and in 15 during follow-up. HGSILs (confirmed as CIN grade II or greater by histology in 87.5% of cases) were diagnosed in 42 patients at study entry and in 10 patients during follow-up. Thirty-five patients underwent excisional therapy (28 patients had conizations and 11 had loop electrosurgical excisions; 4 patients had 2 subsequent excisions). Four patients underwent hysterectomy: in 1 of these patients (who died due to surgical complications), hysterectomy was the primary therapy for CIN grade III; in 2, it was performed for recurrent disease after conization; and in 1 (whose test results had been persistently negative for both HPV infection and genital lesions), it was performed for conspicuous uterine bleeding caused by leiomyomatosis.
At the time of LGSIL diagnosis, mean CD4+ cell counts were comparable among subjects not receiving antiretrovirals (mean CD4+ cell count, 376 cells/mm3) and those consistently receiving HAART (mean CD4+ cell count after a mean of 14 months of therapy, 346 cells/mm3) or non-HAART (mean CD4+ cell count after a mean of 10 months of therapy, 337 cells/mm3); however, mean CD4+ cell count was significantly lower (mean, 204 cells/mm3) in subjects who had changed therapy (HAART group vs. miscellaneous antiretroviral therapy group, P = .0093). Overall, 31 (35%) of 88 subjects with LGSILs showed regression, 51 (58%) had persistence, and 6 (7%) had progression of the lesions; the mean values of CD4+cell counts at LGSIL diagnosis for patients who experienced lesion regression and for those who did not were similar (323 and 320, respectively. HPV infection, in contrast, persisted in >70% of the subjects, irrespective of the HPV type involved. Interestingly, in 10 patients, regression was observed only at the second or third follow-up visitin other words, >1 year after diagnosis. Lesion regression occurred in one-half of untreated subjects and one-half of subjects treated with HAART, but it occurred in only 17% of those in the other 2 antiretroviral therapy groups (X2 = 10.11; P = .0176). By HPV status, regression occurred in 12 (60%) of 20 HPV-negative subjects, in 8 (45%) of 18 with transient HPV infection, and in 11 (22%) of 50 with test results consistently positive for HPV (X2 = 9.9; P = .0071); no significant differences emerged in relation to HPV types (comparing patients with types classified as low risk vs. high risk and those with a single type vs. those with mixed types; data not shown). Progression to HGSILs occurred in 6 subjects; 5 had persistent infection, and 1 was HPV negative. One subject received no antiretroviral therapy, 3 received non-HAART, 1 received HAART, and 1 received miscellaneous antiretroviral therapy. On multivariate analysis, younger age at time of SIL detection, classification in the non-HAART group, and classification in the miscellaneous antiretroviral therapy group were independently associated with persistence or progression of the lesions.
HGSILs were diagnosed in 3 (14%) of 22 subjects not receiving antiretrovirals, in 14 (29%) of 49 receiving non-HAART, in 22 (30%) of 74 receiving HAART, and in 13 (23%) of 56 receiving miscellaneous antiretroviral therapy regimens. Of the 52 patients with HGSILs, surgical excision and biopsy were performed for 35, biopsy alone was performed for 14, and neither procedure was performed for 3. Overall, 36 (67%) of the patients with HGSILs showed regression and 16 (33%) showed persistence or progression of the lesions. HPV infection persisted in 83% of patients with HGSILs, irrespective of the type or types of HPV present; nonetheless, in all of the cases in which HPV clearance was documented, lesion regression occurred. Regression rates were comparable among patients in each of the 3 antiretroviral therapy groups (only 3 women were not receiving antiretroviral therapy, and the number is too small to draw any conclusion regarding regression rates). Subjects who experienced regression had a higher CD4+ cell count at the time of HGSIL diagnosis (mean CD4+ cell count, 320 cells/mm3) than did those who had persistence or progression (mean CD4+ cell count, 223 cells/mm3), but the difference is not statistically significant (P = .1259). Lesions associated with mixed infections were less likely to regress (4 of 12 cases) than were those linked to single infections (30 of 38 cases) (P = .0100). In relation to the gynecological management, lesion regression occurred in 27 (77%) of 35 patients who underwent excisional therapy and in 7 (50%) of 14 who only underwent biopsy of lesions (P = .0890). Multivariate analysis could not be performed because the validity of the model fit was questionable.
DISCUSSION
HIV-positive women, compared with immunocompetent women, show a higher prevalence of persistent HPV infections, a higher prevalence of high-grade lesions, a higher prevalence of invasive carcinoma of the cervix and the vulva, and a poorer response to excisional therapy. Antiretroviral regimens that include PIs have shown great efficacy in reducing the rate of several opportunistic infections, Kaposi sarcoma, and primary brain lymphomas, and studies addressing the effect of HAART on the clinical history of the anogenital lesions have reported both efficacy and little or no effect. These discrepant results could be related to differences in the number of women evaluated, in the length of follow-up, and in the geographical, ethnic and clinical characteristics of the women. In this regard, African American women have been shown to be at higher risk for HPV infection than are white women, and rates of invasive cervical cancer in HIV-infected women are elevated in Italy, France, and, to a lesser extent, in the United States, but not in Africa and Australia. Indeed, because the continuous increase in the incidence of invasive cervical carcinoma after 1996 indirectly indicates that HAART has no preventive effect, the increased longevity associated with receipt of HAART may increase the incidence of cervical and anal cancer in HIV-infected subjects in the future.
We prospectively evaluated 201 HIV-infected women for HPV genital infections and associated lesions in relation to immunological and virological parameters, as well as in relation to receipt of antiretroviral and excisional therapies, for periods of up to 6 years. Women with a diagnosis of LGSILs were followed-up without intervention (with very few exceptions); the majority of those with CIN of grade II or greater underwent surgical excision. Overall, our results indicate that: (1) neither the prevalence of HPV infection nor the rate of clearance or persistence of HPV infection are modified by receipt of any antiretroviral therapy combination; (2) the rate of LGSIL regression is higher among those receiving HAART than among those receiving other antiretroviral therapy regimens; and (3) HGSIL regression occurs in the majority of women whose lesions are surgically excised and who are using antiretroviral therapy. Robinson et al. had shown that the rate of persistence of HGSILs was significantly lower if surgical excision was associated with receipt of HAART; our data indicate that patients with HIV infection benefit from surgical therapy when performed in conjunction with any antiretroviral therapy regimen. Regression rates were higher for patients with HGSILs than for those with LGSILs, as has already been reported. A slightly higher probability of LGSIL regression was observed among older women (OR, 1.163; 95% CI, 1.0231.323).
Women with LGSILs who received HAART showed results similar to those of women whose HIV infection was managed without the use of antiretroviral therapy. No correlation between CD4+ cell counts at SIL detection and at outcome was observed. Indeed, although lesion regression is regulated by CD4+ T cell dependent mechanisms, the number of circulating CD4+ cells might not reflect the immune status at the cervical mucosa. Clinical outcome of LGSILs was negatively correlated with persistent HPV infection, consistent with the notion that persistent HPV infection has a central role in cervical carcinogenesis. On the other hand, the results of multivariate analysis suggest that the outcome of lesions in HIV-infected patients is the result of a complex interplay among viral and host factors, influenced by the receipt of antiretroviral therapy.
HGSILs were diagnosed in similar proportions in the different antiretroviral therapy groups; also, rates of anal SILs reported in a group of 98 men during the pre-HAART period were similar to those reported during the HAART period. These data seem, therefore, to indicate that HAART, per se, does not prevent the occurrence of high-grade cervical and anal HPV-associated lesions. On the other hand, our findings on the successful clinical outcome associated with surgically excised HGSILs in patients receiving antiretroviral therapy, which are consistent with the findings reported by Robinson et al., suggest that lesion excision in conjunction with an effective antiretroviral therapy achieves a good response in the majority of the cases.
Moreover, early intervention also appears determinant in avoiding more-extensive surgery, such as abdominal hysterectomy, that involves an increased risk of complications in HIV-infected women. It is of interest that, because progression of LGSILs occurred within 1 year after diagnosis and because regression rates increased with a longer follow-up period, strict follow-up of patients with LGSILs without intervention also appears to be a safe approach in treating HIV-positive women. This also suggests that a lag time between the improvement in systemic immune function and the local cervical effect of that improvement might exist.
In conclusion, our data provide evidence that, in HIV-positive women, a close and continuous gynecological follow-up is warranted to monitor HPV infection and SIL development, and prompt treatment of all high-grade lesions is recommended. The use of effective antiretroviral therapy appears to be crucial to obtaining regression of high-grade lesions by surgical treatment, as well as to the management of low-grade lesions.
|
|
| |
| |
|
|
|
