Effect of Patient Adherence to Antiretroviral Therapy on the Progression of Cervical Dysplasia Among HIV-Infected Women

Human papillomavirus (HPV) infection is a common infection of the human genital tract and is believed to be the cause of most cervical cancers worldwide.^1-4 Estimates of the prevalence of HPV infection among women with cervical cancer range from 93% to 99%.^1,5 Factors in the host immune system appear to strongly influence the course of HPV infection. As such, HPV infection tends to be persistent in women with HIV/AIDS. In these women, it is more aggressive in terms of causing warts and dysplasia and is more difficult to treat. In addition, studies have documented a high prevalence of HIV/HPV coinfection; some investigators have described HIV itself as an independent risk factor, while others have attributed the increased risk of cervical cancer merely to HIV immunosuppression.^6-8

Current standard highly active antiretroviral regimens contain drugs that have played an important role in the reduction of most opportunistic infections and HIV-related malignancies.^9,10 However, studies have reported conflicting results on the effect of antiretroviral therapy on the development and disease course of cervical dysplasia.^11,12

Our study evaluated the effects of patient adherence to antiretroviral treatment on the relative risk for progression of cervical dysplasia. We further examined the potential impact of other factors on progression of cervical neoplasia, such as age, baseline CD4⁺ cell count, HIV RNA level, smoking status, drug use, ethnicity, and length of time on an antiretroviral regimen.

METHODS
We employed a case-control design to evaluate the effect of patient adherence to antiretroviral therapy on the relative risk of disease progression of cervical dysplasia among HIV-infected women. Data were obtained from chart reviews from HIV outpatient clinics in a large urban hospital system. This study was approved by the Institutional Review Boards of the University of Texas Southwestern Medical Center at Dallas and the University of Texas Health Science Center at Houston.

Information was drawn from chart entries between January 1, 2001, and December 31, 2004, which had been coded by the same investigator (A.B.). Among 308 patients who met the inclusion criteria for having HIV infection and cervical dysplasia, 94 were excluded because they had no prescription refill information, 31 were excluded because they had only 1 assessment during the study time frame, and 2 of Asian ethnicity were excluded because of the small sample size of that ethnic group. The remaining 181 women were designated as cases or controls based on their cervical dysplasia status at the time of enrollment in the study. Case patients were those who had evidence of progression before the start of the study, and control patients were those who had been stable for the 2 years before the study and did not have progression or regression during the course of the study.

HPV infection was not an inclusion criterion; however, reflex HPV DNA testing was automatically performed on the residual fluid using liquid-based cytology, described in the 2001 guidelines developed by the American Society for Colposcopy and Cervical Pathology (ASCCP). Cervical dysplasia status (progressed or nonprogressed) was determined on the basis of cytological/histological findings before the start of the study. Following hospital procedures, patients were first screened by liquid-based cytology with HPV testing; those with a positive result were referred for colposcopy for histological evaluation.

Description of Study Population
Case patients and control patients were in their late 30s, and most were African American, which is consistent with the larger patient demographic of the urban hospital. Case patients were not more likely to be alcohol users, be smokers, use drugs, adhere to treatment, be of a particular ethnic group, have a sex

ually transmitted disease, have an opportunistic infection, or have AIDS, than control patients (P = ns for all variables). Case and control patients were not statistically different for CD4 count and viral load. Case patients were not significantly older or younger, had not received antiretroviral therapy for significantly more or less days, and were not in the study for a significantly shorter time than control patients. Table 1 displays sample characteristics by group (chi-square [χ²] tests of association and independent samples t tests; P = ns).

Table 2 shows the patient histology and cytology data at the beginning and end of the study. Treatment in the study population was limited to modalities in common use. Low-grade neoplasia was usually observed over time, while high-grade dysplasia and invasive cancers were treated. At both the

start and end of the study, most of the patients had a Papanicolaou test and/or colposcopy. At the start of the study, 2 women had cold knife conization biopsies. Also, 1 woman had had a hysterectomy and another had undergone a loop electrosurgical excision procedure (LEEP). By the end of the study, 8 women had undergone LEEP, 1 had had a hysterectomy, and 1 had had radiation therapy.

The 154 case patients were placed in 1 of 3 groups based on their histological/cytological test results during the study: regressors, persisters, or progressors. We used the 27 control patients as a comparison group in the multinomial logistic regression, which predicts disease progression. Diagnosis of cervical dysplasia and clinical surveillance of disease were carried out by colposcopically directed biopsy with histological examination, liquid-based (ThinPrep) cytology, or a combination of the 2 methods. The 2001 Bethesda system nomenclature was used to classify cytological results. Treatment in the population was limited to modalities in accordance with ASCCP protocols.¹³

Disease progression was measured from January 2001 through December 2004. Case patients were considered regressed if the cytological/histological interpretation at follow-up had an acuity rank less than the rank at the initial diagnosis; persistent if the cytological/histological interpretation at follow-up had the same acuity rank as the initial visit diagnosis; and progressed if the cytological/histological interpretation at follow-up had an acuity rank greater than the rank at the initial diagnosis. At completion, a random 15% of the patient files were recoded by the same reader (A.B.) to determine consistency of the ratings. Consistency was approximately 94%. Based on an estimated effect size of 0.20, an alpha of .05, and power equal to 0.80, the sample size was adequate for the various statistical tests.¹⁴

Hospital prescription refill files were used as a proxy measure of patient adherence to treatment. Patients were coded as adherent if they obtained 70% to 100% of the expected refills during the study period and coded as nonadherent if they obtained less than 70%.

RESULTS
From the study population of 181 patients, 572 entries were included in the analyses. A nonparametric χ² test of association revealed a significant pattern in the percentages of adherence and disease progression (χ² = 23.56; P < .05). The patient group considered nonadherent to antiretroviral therapy had the highest percentage of patients who had progression of dysplasia (47.5%), followed by persistence (40.0%) and regression (7.5%); 5% were nonprogressors (control patients). In contrast, the patient group adherent to therapy had the highest percentage of patients who had regression of dysplasia (42.9%), followed by persistence (27.0%) and progression (15.9%); 14.3% were nonprogressors (χ² = 23.56; P < .05).

To control for confounding variables and to combine the effects of multiple variables, we used a multivariate polytomous regression. Covariates and predictors included viral loads, CD4 counts, age, smoking status, drug use, alcohol use, type of treatment, and treatment adherence. Drinking alcohol and smoking increased the likelihood of being placed in any of the 3 case patient groups.

Controlling for the effects of viral load, CD4 count, age, smoking status, drug use, alcohol use, and treatment, a patient was less likely to have regression of cervical dysplasia if she either moderately adhered (odds ratio [OR] = 0.357; P < .05) or did not adhere to antiretroviral therapy (OR = 0.430; P < .05). A patient was less likely to be in the persistence group if she moderately adhered to antiretroviral therapy (OR = 0.358; P < .05) and more likely to be in the persistence group if she did not adhere to therapy (OR = 2.667; P < .05). Similarly, a patient was less likely to be in the progression group if she moderately adhered to antiretroviral therapy (OR = 0.480; P < .05) and more likely to be in the progression group if she did not adhere to therapy (OR = 5.542; P < .01).

DISCUSSION
The present study found that after controlling for stage of HIV disease, adherence to antiretroviral therapy was associated with regression of cervical dysplasia among HIV-infected women. Being nonadherent to treatment significantly decreased the likelihood of regression and increased the likelihood of progression of cervical dysplasia. We also found associations between progression and other variables, such as AIDS diagnosis, alcohol use, and smoking status. Although the present study is enriched by the use of histology, cytology, HPV DNA testing, and a medication refill profile as a measurement of patient adherence to treatment, the study is limited by a nonexperimental design and our inability to control for external factors or to examine other measures of medication adherence. The information collected was also limited to the medical record and clinical database of a relatively small group of patients from a single hospital. Because this particular study was conducted at an urban hospital where 96% of patients are below the federal poverty level, the results may not be applicable to a more economically advantaged population.

This study adds to the body of knowledge on the potential benefits of antiretroviral therapy with regard to the progression of cervical dysplasia among HIV-infected women. The results showed that there were significant differences in disease progression, as predicted by antiretroviral exposure based on patient adherence to a drug regimen. The results also showed that antiretroviral therapy is associated with decreased progression and increased regression of cervical dysplasia. This study suggests that practitioners may want to communicate to patients the potential benefits of antiretroviral therapy for reducing the risk of progression of cervical dysplasia. These findings may also have implications for the development of educational interventions that promote patient adherence to antiretroviral therapy.

No potential conflict of interest relevant to this article was reported by the authors.

References:

References1. Walboomers JM, Jacobs MV, Manos MM, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol. 1999;189:12-19.
2. Sun XW, Ellerbrock TV, Lungu O, et al. Human papillomavirus infection in human immunodeficiency virus-seropositive women. Obstet Gynecol. 1995;85(5, pt 1):680-686.
3. Garland SM. Human papillomavirus update with a particular focus on cervical disease. Pathology. 2002;34:213-224.
4. Bosch XF, de Sanjose S. Human papillomavirus and cervical cancer––burden and assessment of causality. J Natl Cancer Inst Monogr. 2003;31:3-13.
5. Herrington CS. Do HPV-negative cervical carcinomas exist?––revisited. J Pathol. 1999;189:1-3.
6. Williamson AL, Marais D, Passmore JA, Rybicki E. Human papillomavirus (HPV) infection in Southern Africa: prevalence, immunity, and vaccine prospects. IUBMB Life. 2002;53:253-258.
7. Marais DJ, Vardas E, Ramjee G, et al. The impact of human immunodeficiency virus type 1 status on human papillomavirus (HPV) antibodies in serum and cervical secretions. J Infect Dis. 2000;182:1239-1242.
8. Sun XW, Ellerbrock R, Lungu O, et al. Human papillomavirus infection in human immunodeficiency virus-seropositive women. Obstet Gynecol. 1995;85:680-686.
9. International Collaboration on HIV and Cancer. Highly active antiretroviral therapy and incidence of cancer in human immunodeficiency virus-infected adults. J Natl Cancer Inst. 2000;92:1823-1830.
10. Murphy EL, Collier AC, Kalish LA, et al; Viral Activation Transfusion Study Investigators. Highly active antiretroviral therapy decreases mortality and morbidity in patients with advanced HIV disease. Ann Intern Med. 2001;135:17-26.
11. Minkoff H, Ahdieh L, Massad LS, et al. The effect of highly active antiretroviral therapy on cervical cytologic changes associated with oncogenic HPV among HIV-infected women. AIDS. 2001;15:2157-2164.
12. Schuman P, Ohmit SE, Klein RS, et al; HIV Epidemiology Research Study (HERS) Group. Longitudinal study of cervical squamous intraepithelial lesions in human immunodeficiency virus (HIV)-seropositive and at-risk HIV-seronegative women. J Infect Dis. 2003;188:128-136.
13. Solomon D, Davey D, Kurman R, et al; Bethesda 2001 Workshop. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA. 2002;287:2114-2119.
14. Cohen J, Cohen P. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. Oxford, UK: Lawrence Erlbaum; 1975.