Determinants of Antiretroviral Therapy Initiation and Treatment Outcomes for People Living With HIV in Vietnam Dam Anh Tran,1,2 Anthony Shakeshaft,1,2 Anh Duc Ngo,3 Kylie-Ann Mallitt,2 David Wilson,2 Christopher Doran,4 and Lei Zhang2 1
National Drug Alcohol Research Centre, 2Kirby Institute, The University of New South Wales, Sydney, Australia; 3 School of Population Health (formerly School of Health Sciences) and Samson Institute for Health Research, Adelaide, University of South Australia; 4University of Newcastle, Hunter Medical Research Institute, Australia
Objectives: This study explores patient characteristics that are significantly associated with very late combination antiretroviral therapy (cART) initiation (CD4 count ≤100 cells/mm3) and examines the association between patient characteristics and treatment outcomes, CD4 recovery, and mortality. Design: Data were obtained from the clinical records of 2,198 HIV/AIDS patients in 13 outpatient clinics across 6 provinces in Vietnam. Methods: Multivariate logistic regression and Cox proportional hazards regression were used to identify patient characteristics that are significantly associated with very late cART initiation and to measure relationships between patient characteristics and treatment outcomes. Results: Very late cART initiation was significantly associated with being male compared with female (odds ratio [OR], 0.36; 95% CI, 0.23–0.58), becoming HIV infected through injecting drugs (OR, 2.13; 95% CI, 1.09–4.14), and having opportunistic infections at cART initiation (OR, 1.69; 95% CI, 1.02–2.86). Being male (female vs male: hazard ratio [HR], 0.45; 95% CI, 0.20–0.98), very late cART initiation (timely vs late: HR, 0.18; 95% CI, 0.04–0.72), low baseline body mass index (BMI) (HR, 0.95; 95% CI, 0.92–0.98), and later baseline WHO clinical stage (WHO clinical stage IV vs combined group of stage I and II: HR, 5.70; 95% CI, 3.90–7.80) were significantly associated with death, whereas being female compared with male (HR, 1.51; 95% CI, 1.14–1.99) and timely cART initiation (HR, 35.45; 95% CI, 13.67–91.91) were significant predictors of CD4 recovery. Conclusions: Timely testing of patients for HIV, increasing use of CD4 count testing services, and starting cART earlier are essential to reduce mortality and improve treatment outcomes. Key words: baseline CD4 count, cART, clinic, combination antiretroviral therapy, HIV/AIDS, Vietnam
lthough the optimum CD4 cell count at which to start combination antiretroviral therapy (cART) among HIV-infected patients is yet to be definitively determined,1 initiation of cART when the CD4 cell count is 350 cells/mm3 has been accepted as the benchmark for timely intervention in many countries, including Vietnam,2 and is associated with improved virological outcome and reduced incidence of morbidity and mortality.3–8 Late cART initiation (CD4 cell count ≤ 200 cells/mm3) is more prevalent in low and middle income countries9 and has been associated with increased risk of HIV comorbidity and mortality.6 Patient demographics, time interval between HIV diagnosis and treatment, and HIV-related risk behaviors and transmission routes have all been identified as
A
significant predictors of late cART initiation. In China, for example, late cART initiation is associated with being male, single, or infected with HIV through heterosexual contact, whereas being infected through shared injections and homosexual contact were protective factors.10 Similarly, European studies indicate that injecting drug
Address for correspondence: Dam Anh Tran, National Drug and Alcohol Research Centre, Kirby Institute, The University of New South Wales, Sydney, Australia; phone: 61 2 9385 0239; e-mail:
[email protected] HIV Clin Trials 2013;14(1):21–33 © 2013 Thomas Land Publishers, Inc. www.thomasland.com doi: 10.1310/hct1401-21
21
22
HIV CLINICAL TRIALS •
users, males, and young people are associated with late cART initiation.4,11,12 In Vietnam, cART was first introduced in 3 clinics in Hanoi and Ho Chi Minh City in 2004. The number of cART sites increased from 74 in 2005 to 287 in 2009. A total of 63 provinces and cities now have cART clinics. Accordingly, the number of people receiving cART increased from 14,969 people in 2007 to 37,995 in 2009 (14% to 45% cART coverage, respectively).13 In 2011, Vietnam changed its benchmark for timely cART initiation from a CD4 cell count ≤250 cells/mm3 to ≤350 cells/mm3, to align with the World Health Organization (WHO) guidelines. HIV patients in Vietnam are eligible for cART when (a) their baseline CD4 count is ≤350 cells/mm3 or (b) their WHO clinical stage is III or IV.14 Despite these criteria, most people living with HIV/AIDS (PLHIV) in Vietnam initiate cART at a very late stage. One report, for example, indicates that the median baseline CD4 cell count among HIV patients who initiated cART before June 2010 was only 75 cells/mm3.15 Identifying the determinants of late cART initiation may help target interventions aimed at promoting timely access to cART more effectively, which in turn would improve treatment outcomes. To date, no studies in Vietnam have investigated the determinants of and treatment outcomes associated with late cART initiation. This study aims to (a) identify risk factors for very late cART initiation, (b) identify predictors that are significantly associated with key outcomes after 6 months of treatment (opportunistic infection, body mass index [BMI], WHO clinical stage), and (c) identify the demographic and HIV-related risk factors associated with mortality (from cART initiation time to November 30, 2010) and CD4 recovery (CD4 count > 350 cells/mm3 after 6 months of cART initiation).
METHOD Study Design This is a retrospective cohort study. It received ethics approval from Research Ethics Committee at Hanoi School of Public Health. Patients were categorized, depending on their CD4 cell count at cART initiation, as very late (CD4 count ≤100 cells/ mm3), late (CD4 count >100 cells/mm3 and ≤200 cells/mm3), and timely (CD4 count >200 cells/mm3 and ≤350 cells/mm3).
14/1
•
JAN-FEB 2013
Sampling and Sample Size Study participants were HIV-infected patients who were at least 18 years of age and who initiated cART between January 1, 2005 and December 31, 2009 in one of 13 cART outpatient clinics in 6 provinces of Vietnam: Dien Bien, Bac Giang, Hai Phong, Quang Ninh, Ho Chi Minh City, and Binh Duong. These provinces represent diverse geographical regions in Vietnam (eg, urban, rural, mountainous), including 4 provinces with a relatively high HIV prevalence (Dien Bien [0.8%], Hai Phong [0.4%], Quang Ninh [0.3%], and Ho Chi Minh City [0.6%]) and 2 provinces with a relatively low HIV prevalence (Bac Giang [0.1%] and Binh Duong [0.2%]).16 CD4 count at cART initiation (baseline CD4) was defined as the latest CD4 count measured within 3 months prior to cART initiation. Those who did not have a CD4 test in the 3 months prior to their cART initiation were excluded, as were patients with a baseline CD4 count >350 cells/mm3.
Data Collection and Measurement Clinical records for eligible patients were obtained from participating clinics. Information recorded at cART initiation was manually extracted by the first author and clinic staff (nurses and doctors) and included the following variables: demographics (age, gender, residential location, and height and weight to calculate body mass index [BMI]); HIV confirmation test date; HIV transmission route (injecting drug use [IDU], unsafe sex); cART start date; referral source; initial cART regimen; presence of tuberculosis and other opportunistic infections (eg, thrush, toxoplasmosis, Penicillium marneffei infection, cytomegalovirus infection, Pneumocystis jirovecii pneumonia); WHO clinical stage17; and cotrimoxazol prophylaxis use before cART initiation. The term opportunistic infections instead of AIDS-defining diseases was used as it was the defined terminology on the survey and in the patient records. WHO clinical stage was consistently used across all 13 outpatient clinics in the study. Because physical and disease progression indicators were measured every 6 months after cART initiation, the following variables were also obtained: weight, height, presence of opportunistic infection(s), WHO clinical stage, CD4 count, treatment adherence,14 periods of treatment interruption, changes in cART regimen during treatment,
ART TREATMENT AND OUTCOMES IN VIETNAM •
and treatment status (death, loss to follow-up). Data were transferred into Microsoft Excel by 3 research assistants and were double-checked (10% of the records) by the first and the third authors. Where data were unclear or missing, we crosschecked with the original patient records. Data collection occurred between May and November 2010. Statistical Analyses Data analyses consisted of 3 sequential steps, implemented using SAS (version 9.2; SAS, Inc, Cary, North Carolina, USA). First, descriptive statistics were generated to examine demographic and risk factor characteristics separately for patients whose cART initiation was classified as very late, late, and timely. Chi-square and Kruskal-Wallis tests were performed to identify baseline differences between these 3 groups. Adjusted risk factors for very late cART initiation were identified using a multivariate ordinal logistic regression model. All baseline demographics and HIV risk characteristics (summarized in Table 1) were included as explanatory variables. Univariate analysis was performed to identify the baseline characteristics that were likely to be associated with the outcome (P < .25), and these variables were subsequently fitted into the initial multivariate logistic regression model. Of the variables that were not significantly associated with the outcome (P > .05), the one that was least significant (highest P value) was removed and the model was refitted with the remaining variables. This procedure was repeated in a backward stepwise manner until the final model was obtained, containing only those variables where P ≤ .05. Second, predictors associated with key outcomes after 6 months of treatment (presence of opportunistic infections, BMI, and WHO clinical stage IV) were identified using multivariate linear regressions for BMI (a continuous outcome variable) and multivariate logistic regressions for opportunistic infections (yes or no) and WHO clinical stage IV (yes or no). These 3 outcome variables were chosen because of their significance in improving morbidity among cART patients.8,18,19 All baseline demographics and HIV risk characteristics (summarized in Table 1) were again included as explanatory variables, although cART initiation stage (very late, late, or timely) was retained in all models regardless of its significance in the univariate analysis.
TRAN ET AL
23
Third, the demographic and HIV-related risk factors associated with mortality (from cART initiation time to November 30, 2010) and CD4 recovery (CD4 count >350 cells/mm3 after 6 months of cART initiation) were identified using multivariable Cox proportional hazards regression. Again, all baseline demographics and HIV risk characteristics (summarized in Table 1) were included as explanatory variables, including cART initiation stage (very late, late, or timely).
RESULTS Sample Characteristics A total of 3,449 patients attended one of the 13 participating cART clinics during the data collection phase, 1,191 of whom (34.5%) were excluded due to the absence of a baseline CD4 count and 60 (1.8%) were excluded because they had a baseline CD4 count >350 cells/mm3. As a result, 2,198 patients (63.7%) were included in analysis. Demographics As summarized in Table 1, the mean age of patients was 30 years and 75% were male. Most patients (63%) initiated cART very late, 26% initiated late, and 11% had a timely initiation. Compared to other sources, voluntary counselling and testing (VCT) services referred the highest proportion of patients for all cART initiation groups (35%), although a substantial proportion of patients self-presented for treatment (23%). The highest proportion of patients commenced cART at WHO clinical stage III (43%). The median CD4 count at HIV diagnosis was 59 cells/mm3. The 3 cART initiation groups were statistically significantly different in terms of gender, transmission routes, WHO clinical stage at baseline, cotrimoxazol use before cART initiation, year of cART initiation, median time interval, and median CD4 count at HIV diagnosis. Risk Factors for Very Late Treatment Seeking Patients Table 2 shows that females were 0.36 times less likely to have very late cART initiation relative to males (odds ratio [OR], 0.36; 95% CI, 0.23–0.58). Patients with a history of IDU were twice as likely
HIV CLINICAL TRIALS •
24
14/1
•
JAN-FEB 2013
Table 1. Demographic and characteristics of HIV-infected patients in 13 HIV outpatient clinics in Vietnam by cART initiation stage (N = 2,198) cART initiation stage Characteristics
Timely Very late Late (n = 243; (n = 1,393; 63.4%) (n = 562; 25.6%) 11.1%)
Total (N=2,198)
1.150 (82.7) 30.3 (±6.5)
2,565 (74.5)
