Abstract Book Vol. 1 [International Conference on AIDS (16th: 2006: Toronto, Canada)]

Results: 1186/3863 patients were hospitalised at least once, resulting in
1854 hospitalisations for medical concerns. The yearly rate of hospitalisation
/ hospitalised patients decreased over time between 2000 and 2004 from
172 to 92 and from 132 to 78 per 1000 patients, respectively. 21% of events
were bacterial infections, 20% AIDS events (24% in 2000 to 11% in 2004),
10% psychiatric disorders, 9% cardiovascular events (5 to 15%), 7% hepatic/
gastro-intestinal disorders (cirrhosis 60%), 6% other viral infections, 5% nonAIDS cancers, and 4% iatrogenic events (5 to 3%). CD4 count at the time of
event was below 200/mm3 in 37% of patients and above 500/mm3 in 19%.
As compared to other patients, patients with events were older (41.5 vs 39.9
years; P<0.001), with lower CD4 count (250/mm3 vs 489/mm3; P<0.001).
They were also more likely to be contaminated through intra venous drug use
(30 vs 19%; P<0.001), positive for hepatitis C (37 vs 27%; P<0.001) and at
AIDS stage (43 vs 14%).
Conclusions: Severe morbidity has shifted from AIDS-related to non AIDSrelated causes. Ageing, co-morbidities (hepatitis, addictive behaviours) might
explain this evolving distribution. Immunodepression, even moderate may also
have a role in the development of non-AIDS morbidity. Limiting endpoints to
AIDS events and death could be misleading in the interpretation of prognostic
and therapeutic studies.
MOPDBO3
Attributes to death of HIV/AIDS patients in the era
of HAART
C.F. Ho, K.L.S. Cheng, K.H. Wong, K.C.W. Chan. Special Preventive
Programme, Department of Health, Hong Kong, China
Background: In HAART era, there is a substantial reduction in HIV-associated
mortality but notable shifted causes of death. Understanding attributes to death
can help in formulating strategies and establishing priorities to improve care.
Although there is no consensus on the context to be examined, it is essential
to include all possible attributes, i.e. biomedical, behavioural and psychosocial
factors. We set out to study attributes to death of HIV/AIDS patients in Hong
Kong.
Methods: The clinical records of HIV/AIDS patients who died between 1st Jan
2003 and 31st December 2004 were reviewed. Information was collected on:
(a) demographics, (b) last psychosocial assessment result, (c) medical and
treatment history, (d) adherence to medical follow-up and HAART, and (e)
medical cause of death.
Results: Of 698 patients, 24 died. All were Chinese and two were female. The
median age at death was 48.5 years (range: 24 to 76 years). The median time
from HIV diagnosis to death was 3 years (range: 3 months to 19 years). Sixteen
deaths were causally related to HIV (including 6 AIDS); four were not; and four
were unknown. No patient was known to die of antiretroviral toxicity. Six deaths
(25%) occurred in those who presented with CD4 <50 cells/uL and died within
1 year of diagnosis. Unfavourable behavioural and psychosocial factors were
common. Thirteen (54%) had missed medical appointment for more than 6
months, had refused HAART, or had a self-reported adherence rate <95%.
Conclusions: Only one-fourth of deaths in the HAART era resulted from
AIDS. Beyond biomedical factors, one-fourth of deceased patients were late
presenter of HIV diagnosis and half had poor adherence to medical followup and HAART. Both groups deserve special attention as their death may be
avoidable. Interventions should be formulated to encourage early diagnosis,
modify unfavourable behaviour and address psychosocial needs of HIV-infected
persons.
MOPDBO4
Predictors of mortality in HIV-infected adult African
patients receiving highly active antiretroviral
therapy
A.M. Siika1, K. Wools-Kaloustian2, S.N. Kimaiyo', A. Mwangi3, L.O. Diero1,
P.O. Ayuo1, W.D. Owino-Ong'or', J.E. Sidle4, R.M. Einterz4, C. Yiannoutsoss,
B. Musicks, W.M. Tierney4. IMoi University School of Medicine, Medicine,
Eldoret, Kenya, 'Indiana University School of Medicine, Infectious Diseases,
Indianapolis, United States, 3Academic Model for Prevention and Treatment
of HIV (AMPATH), Biostatistics, Eldoret, Kenya, 'Indiana University School of
Medicine, Medicine, Indianapolis, United States, sIndiana University School of
Medicine, Biostatistics, Indianapolis, United States
Background: There is little data available on mortality in African patients on
Highly Active anti-retroviral Therapy (HAART). This study was undertaken to
determine predictors of mortality in such patients.
Methods: This was a retrospective study of prospectively collected data
from consecutively enrolled adult HIV-infected patients in nine HIV clinics in
western Kenya. Data from records of deceased patients started on HAART
between November 2001 and December 2005 were analyzed and compared
with those from records of living patients started on HAART during the same
period. Analyses of time to death were undertaken using Kaplan-Meyer and
Multivariate Cox proportional hazard regression models.
Results: Data from 527 deceased patients were compared with 1054 patients

known to be alive. At initiation of therapy, median age was 38 years (range
16-77) for deceased and 36 years (range 15-73) for alive patients. Median
duration of time on HAART was 7.7 weeks (range 0-110) and 42 weeks (range
0-208) (p<0.0001) for deceased and alive patients respectively. Patients with
CD4 count < 100 were more likely to die than those with CD4 count 100 - 200
(HR=1.94, CI (1.63,2.53), p<0.0001) who in turn had a higher chance of dying
than those with CD4 cell counts >200 (HR=1.63, CI (1.13,2.42), p<0.0093).
The hazard for death of perfectly adherent patient was 0.6 times that of non
adherent patients (HR=0.61, CI (0.44, 0.86), p=0.0025) while that for patients
attending urban clinic was a third of that among patients attending rural satellite
clinics (HR=0.35, CI (0.24, 0.51), p <0.0001). Male gender, higher WHO Stage
and hemoglobin level <10 grams % were significantly associated with death.
Age, marital status, educational level, employment status and weight were not
associated with mortality.
Conclusions: Patients on HAART who died were in poor health based on CD4
cell count and WHO Stage and often did so soon after initiation of therapy.
MOPDB05
Predictors of mortality in patients initiating
antiretroviral therapy in Durban, South Africa
B. Ojikutu', H. Zheng2, R. Walensky2, Z. Lu2, E. Losina2, J. Giddy3,
K. Freedberg2. 1Harvard Medical School, Division of AIDS, Boston, United
States, 2Massachusetts General Hospital, Boston, United States, 'McCord
Hospital, Durban, South Africa
Background: As an increasing number of patients start antiretroviral therapy
(ART) in resource-constrained settings understanding predictors of mortality
will help to maximize future clinical outcomes. Our objective was to characterize
factors present prior to ART initiation that are associated with mortality in HIVinfected patients who start ART in Durban, South Africa.
Methods: We conducted a retrospective cohort study of patients who initiated
3-drug ART (AZT, 3TC and efavirenz or nevirapine) at McCord Hospital's
Sinikithemba HIV Clinic from January 1, 1999 to February 29, 2004. The impact
of demographic, clinical and laboratory parameters on mortality were evaluated
using univariate and multivariate Cox proportional hazards models. The KaplanMeier method was used to assess survival.
Results: Three hundred and nine patients were included in the study. Mean
age was 38 (+/-8.6). 56% of patients were female. Overall one-year survival
was 0.78 [95%CI (0.71, 0.85)]. In univariate analysis, pre-ART history of oral
candidiasis [HR 3.17(1.70, 5.87)], history of cryptococcal meningitis [HR 2.76
(1.80, 19.2)], CD4 cell count <50/mm3 [HR 3.70 (1.96, 7.14)], and hemoglobin
< 8 g/dl [HR 1.23 (1.08, 1.40)] were the strongest predictors of mortality
once ART was initiated. Age, gender, history of pulmonary or disseminated
tuberculosis [HR 0.93 (0.51, 1.71)] did not predict mortality. History of oral
candidiasis and CD4 cell count <50/mm3 remained significant in multivariate
analysis [HR 2.58 (1.37, 4.88) and HR 3.12 (1.56, 5.88)]. No difference in
mortality was noted between patients with CD4 cell count <20/mm3 and 20 -50/mm3 [HR 1.06 (0.51, 2.18)].
Conclusions: History of oral candidiasis and CD4 count <50/mm3 are
independent predictors of mortality in patients initiating ART in Durban, South
Africa. Further interventions to reduce mortality associated with these clinical
markers should be developed and urgently implemented.
MOPDB06
Very low CD4 T cell counts and low total lymphocyte
counts at initiation of HAART are associated with a
poor outcome in the first 6 months of antiretroviral
treatment
H. Mayanja-Kizzal, F. Lutwama2, M. Kamya3, C. Kikawa2, L. Spacek4,
T. Quinn4. 1Makerere University, Faculty of Medicine, Academic Alliance for
AIDS Care and Prevention, Kampala, Uganda, 2Makerere University Infectious
Diseases Institute, Kampala, Uganda, 3Makerere University Infectious
Diseases Institute/ Academic Alliance for AIDS Care and Prevention, Kampala,
Uganda, 4Johns Hopkins University, Public Health, Baltimore, United States
Background: Global HAART scale up efforts have increased access to AIDS
treatment in developing countries, but mortality remains high despite adequate
management.
Methods: A cohort of 550 patients starting HAART were prospectively followed
up over thirty months at an AIDS treatment center in Kampala, Uganda.
Regular clinical review; and CD4 T cell and viral load counts were done every
3 and 6 months respectively. Early home visits to locate those who missed
clinical appointments were done. Data was compared to patients alive after
over twelve months of HAART.
Results: Seventy two (13%), patients (45 females, mean age 36.7 (SD
19) and 27 men, 39.3 (SD 6) years), died while on ARVs despite over 90%
adherence. Twenty nine died within 6 months of commencement of ARVs,
24 between 6-12 months and 19 died after one year. HAART was started at
very low CD4s, median 24 cells/mm3 among those who died, (deaths within
6, 6-12, over 12 months of HAART, median CD4T cells 14 cells/mm3, 31 cells/
mm3, 73 cells/mm3 respectively) compared to 110 cells/mm3 among over 12
months survivors (p=0.002). Initial total lymphocyte counts were significantly
lower among deaths within 6 months, mean 1200/mm3, (SD 720) compared
to deaths over one year 1800 cells/mm3, (SD 1038) and survivors 1780 (SD
1120) (p=0.028). Viral load at onset of HAART was similar in all groups, mean
5.5 log among survivors, and 5.7 log among all deaths. There was better viral
load suppression at 12 months 2.9 log among survivors, compared to 0.4 log
among deaths over one year (p=0.005).

Conclusions: High mortality on HAART may be associated with late
commencement of treatment at very low CD4 counts, where despite adequate
treatment immune recovery lags behind virological suppression. Commencing
HAART at earlier CD4 counts should be considered, even in resource limited
countries to improve outcome.

XVI INTERNATIONAL AIDS CONFERENCE * 13-18 AUGUST 2006 * TORONTO CANADA * ABSTRACT BOOK VOLUME 1