PREDICTORS FOR LONGITUDINAL CD4 CELL COUNT AND WEIGHT WITH TIME TO DEATH OF HIV/TB CO-INFECTED PATIENTS UNDER TREATMENT AT UNIVERSITY OF GONDAR COMPREHENSIVE SPECIALIZED HOSPITAL, NORTH WEST ETHIOPIA

Abstract

Background: Co-infections with Human Immunodeficiency Virus and Tuberculosis are extremely taxing on health care systems and present unique diagnostic and treatment difficulties. TB is the most frequent cause of mortality from HIV patient. Thus, Mycobacterium tuberculosis and HIV work together to accelerate the reduction of immune system. The objective of this study was joint modeling of bivariate longitudinal outcomes and time to death of HIV/TB co-infected patients at University of Gondar Comprehensive Specialized Hospital, North West Ethiopia. Methods: A retrospective cohort study was conducted with data from 192 HIV/TB co-infected patients, covering the period from September 1, 2014, to September 1, 2024. The analysis utilized Cox proportional hazards models for survival outcomes and linear mixed-effects models for longitudinal data, focusing on CD4 cell count and weight changes. A joint model was constructed to assess the correlation between these longitudinal markers and survival outcomes. Result: The cox proportional hazards model indicated that functional status significantly impacted mortality among HIV/TB co-infected patients, with ambulatory and bedridden individuals having hazard ratios of 1.721 (p = 0.005) and 1.694 (p = 0.008), increased risk of death compared to worker patients, respectively. Patients in WHO stage IV had a 1.239-fold increased risk of death compared to those in stage I (p = 0.003). In the linear mixed effects models, changes in CD4 cell count were significantly influenced by age (-0.002, p = 0.036) and treatment adherence (0.122, p = 0.029). The joint modeling analysis revealed a significant negative association between the natural log of CD4 cell count and mortality risk (association parameter = -0.613, p < 0.0001), indicating that higher CD4 counts are linked to lower mortality risk. Conclusion: In the joint model analysis patients with functional status, WHO stage, religion, and marital status had significant factors for time to death of HIV/TB co-infected patients. And age, WHO stage, functional status, opportunity infected, adherence, smoking status, occupation and visiting time were significant factors of CD4 cell count of patients. As a recommendation, to reduce the risk of death, the government health agencies and healthcare centers should provide nutritional and psychological support for functional status of ambulatory and bedridden individuals. Keywords. Joint model, HIV/ TB co-infected, CD4 cell count, Weight, LMM, Survival