Document Type

Dissertation

Date of Degree

2012

Degree Name

PhD (Doctor of Philosophy)

Degree In

Epidemiology

First Advisor

James C. Torner

Abstract

Infection with human immunodeficiency virus (HIV) and progression to acquired immune deficiency syndrome (AIDS) often result in neurologic and neuropsychiatric changes, although the prognostic information available for patients affected by HIV/AIDS-related neurologic diagnoses has been limited. The objective of the present study was to characterize the patterns and predictors of survival, including the impacts of antiretroviral therapy (ART) use and potential factors in healthcare access and disparity, among patients with one or more of the following conditions: cryptococcosis, toxoplasmosis, primary central nervous system lymphoma, progressive multifocal leukoencephalopathy, and HIV-associated dementia. To accomplish this, a cohort was drawn from the Iowa HIV/AIDS reporting system, and a non-independent, university-based cohort was then used to validate the analyses conducted for the statewide sample. Patterns of ART use were identified in each cohort using logistic regression, and survival analyses were conducted using Kaplan-Meier analysis, Cox regression, and accelerated failure time modeling.

Survival was poor in both cohorts, although the university-based setting (University of Iowa Hospitals and Clinics) was associated with better overall survival. Of 230 persons in the statewide cohort, 77.0% were deceased by the end of the study period (1982-2008), and the median survival was 1.13 years (95% CI: 0.90 to 1.86 years, n=225). By contrast, 56.4% of the university-based cohort was deceased by the end of the study period (1984-2009), and the median survival in this group was 3.04 years (95% CI: 1.79 to 11.62 years, n=172). Both cohorts were predominantly male, non-Hispanic white, and residents of a small metropolitan area at the time of the AIDS diagnosis.

ART use had a strong protective effect on survival in both cohorts. Use of ART among patients diagnosed during the era of highly active antiretroviral therapies (HAART) was associated with an 80% reduction in the rate of death (HR=0.20, 95% CI: 0.08 to 0.46) compared to the non-users diagnosed during the pre-HAART era (that is, prior to 1996), after adjustment for age, race, birth sex, healthcare facility type, opportunistic infection count, HIV transmission risk category, neurologic condition, years since AIDS diagnosis, and timing of neuro-AIDS in a Cox regression model. In the UIHC cohort, the adjusted expected survival time among ART/HAART users was 37.71 (95% CI: 14.44 to 99.48) times that among non-users.

Women had significantly poorer outcomes than men in the statewide cohort (adjusted HR=2.31, 95% CI: 1.22 to 4.35), and a similar, non-significant trend was observed among university-based cases. Secondary analyses suggested that this difference persisted over the course of the epidemic and was not attributable to differential ART response among men and women. Evidence for a role of disease severity, psychosocial support, and/or psychiatric comorbidity in the differential survival of men and women was identified.

This study provides useful prognostic data for patients and providers and may guide future research efforts aimed toward improved survival for neuro-AIDS patients. The survival disadvantage of women compared to men should be confirmed and the mechanisms underlying this disparity elucidated. Meanwhile, clinical and public health efforts might be directed towards screening, treatment, and support for women affected by neuro-AIDS, including potential assessment of comorbid psychiatric disorders.

Keywords

AIDS, Cox regression, HIV, Neurological, Survival

Pages

xv, 170 pages

Bibliography

Includes bibliographical references (pages 168-170).

Copyright

Copyright 2012 Martha L. Carvour

Share

COinS