Abstracts Vol. 1 [International Conference on AIDS (11th: 1996: Vancouver, Canada)]

Tuesday, July 9, 1996

Tu.B.523 -Tu.B.540

predict better clinical outcome. But there is another problem:,irilveti.v'iral trials are
plagued by incompleteness in such visit-driven data. However precise a test mnay be, failure
to collect it limits its utility Laboratory measurements are lost wihen p lients die, miss study
visits or do not have blood drawn. At best, this reduces power for treaitment comparisons;
more likely it introduces bias. Concern about this bias led to the ptesernt study.
Methods: CD4+ lymphocyte counts were used as an instance of a visit driven (as opposed
to event-driven) endpoint. Missing CD4+ counts were analyzed ii ive hir e multicenter USgovernment sponsored antiretroviral trials ACTG 002, 019. I B/I i / and CPCRA 002
and 007. Number of patients alive, under followup, and providirng CD4 counts are summarized at specific time points. Logistic regressions are used to describe pI',ients with missing
data, and regressions (using prior CD4+ counts and AIDS history) ire used to impute missing CD4+ counts.
Results: In all studies, the fraction of missing CD4+ counts incra:ed with time after randomization. Between 20% and 38% were missing about a half ye lr after randomization
(specific time varied according to protocol follow-up schedule), andIc ill fve trials showed significant relationships between presence of CD4+ count and treatment ' oup or baseline
characteristics (prior AIDS, hemoglobin, CD4+ count, and Karnc.,irk.core, as available).
Except for ACTG 019, patients sicker at baseline were more lih kely to miss CD4+ counts.
Except in CPCRA 002, treatment arm was a significant predictor of missing a CD4+ count.
Such relationships can induce biases in analyses. For example, in CPCF RA 002, the averages
of collected CD4+ counts (at baseline and at 2, 6, and 12 mouth's) indicate stable CD4+
levels throughout the trial, but if the missing counts are imputed (i e., replaced by predictions from regression models), the average CD4+ count declines steadily
Conclusions: Many patients cannot or do not provide visit-driven data in antiretroviral trialsThese data are not missing at random, so serious bias can ocor, an ciry interpretation
of these highly selected data is perilous.Visit-driven endpoints niust be chllected and analyzed carefully, especially when they are used as surrogates for clinicil disease progression.
Carlton I-I Hogan, Division of Biostatistics, University of Minnesota, 2221 University Ave SE,
Suite 200 Minneapolis, MN 55414 Ph: [phone]; Fax: (61 ) t626 8892
e- mail: [email protected]
Tu.B.523
TENSIONS BETWEEN THE SCIENCE AND ETHICS OF HIV CLINICAL DRUG TRIALS:
SOCIAL CONSTRUCTIONS OF THE TRIAL
Parsons, Claire D. F., Grubb lan R. National Centre in HIV Socii Rese.irch: People with HIV
and Their Carers, La Trobe University, Bundoora,Victoria, AUSTRAL IA 3083
Issue: There is international concern regarding tensions between the science and ethics of
HIV clinical drug trials. It is important to examine whether the interests of science and
ethics are incommensurable or whether there are avenues for ameliorating the perceived
and actual differences.
Project: Unstructured interviews with key informants (in medicine, nursing, government
and community) who are recognised for their leadership related to HIV clinical trials in
Australia were used to elicit explanatory statements regarding the ethics and science of
clinical drug trials.
Results: Five main findings were: I) doctors overestimate participant iltr!,ismr in trial participation; 2) little consensus exists over appropriate trial designs, or the potential use of surrogate markers (such as the PCR viral load test); 3) tensions predoe'iinntly articulate around
varying levels of confidence in assessing safety and efficacy (requirements of pharmaceutical
companies, medical researchers, regulatory authorities and the community) and inequities in
power to assert particular perspectives; iv) the 'explanatory models revealed multiple realities (narratives) both within and between the expert narratives on clinical trials, not a singular unified disciplinary perspective; v) expert' discourse comprises facts' (professional knowledge) cemented together with discontinuous fragments of knowledge reflecting the speakers' institutional and disciplinary affiliation, the individual's ethical stance, and socio-political
positioning which influences the packaging of objective accounts of how RCT methods and
process function.
Lessons Learned: Analysis suggests alternative trials research designs to amrneliorate scientific
and ethical concerns
Dr CDF Parsons, National Centre in HIV Social Research: PHC, La Tfrobe University
Bundoora,Victoria, Australia 3083.Telephone: [phone]. Fax: 001 I-6 I-3-94 I1869.
Email: [email protected]
Tu.B.530
PROPHYLAXIS OF TUBERCULOSIS INFECTION
Reichman, Lee B. New Jersey Medical School National Tuberc ilosis Center Newark, NJ, USA.
It has long been known that HIV infection was the most potent facilittor of tuberculosis
ever known.This coexistent interaction has led to TB being the most important opportunistic infection associated with HIV, more so because TB even with HIV is almost completely preventable.
With the increasing evidence thatTB enhances the progression utf ft as shown by an
increasingly larger number of studies, the priority in providingTB 5rophylaxis in the dually
infected HIV/TB individual become paramount.The effect ofTB en HIV p-ogression is so
important it should be used as a compliance enhancer in the prescription of preventive therapy in the dually infected to prevent active TB and thus prevent progression of HIV infection.
Preventive therapy (chemoprophylaxis) in HIV-infected individuals consists of prescribing 12
months of INH 300 mg (regardless of age). In individuals who ae unable to mount a reaction to tuberculin (presumably because of anergy), INH may be gen with-out regard to the
tuberculin skin test in HIV-infected populations with a high pi ev ienice if [B infection. Anergy
testing is unnecessary uncomfortable and adds nothing toTB treatum'ent or prevention in
the individual patient.
With proper monitoring, INH side effects, usually hepatits in okr i niciduals, are uncommon. Prophylaxis to contacts of drug-resistant or multidrug-resistant TB should be with
Rifampin, or PZA and a quinolone if sensitive.
Lee B. Reichman. New Jersey Medical School National Tuberacilcasis Ceiteni; 65 Bergen St.,
Scite GB I, Newark, NJ, 07 107-3001, USA.

Tu.B.532
RECENT ADVANCES IN OPPORTUNISTIC INFECTION PROPHYLAXIS
Powderly,Willam G.Washington University School of Medicine, St. Louis, MO, USA
Prophylaxis for Pneumocystis carinii pneumonia (PCP) has been standard care in HIV therapeutics for almost a decade. It is clear from recent clinical trials that any form of PCP prophylaxis is very effective in patients with higher CD4 counts. As immunodeficiency progresses, systemic prophylaxis is superior to local, with TMP/SMX providing the most effective protection. Most breakthroughs of PCP on prophylaxis occur in patients with CD4
counts <50 who cannot take TMP/SMX. Prospective trials and observational data indicates
that TMP/SMX is also the most effective prophylaxis of toxoplasmosis. Several recent trials
have demonstrated that clarithromycin and azithromycin provide effective protection (60
70%) against MAC bacteremia in patients with advanced HIV disease, supplementing the
prior experience with rifabutin. Clarithromycin also demonstrated a survival advantage
compared with placebo. Trials comparing the macrolides, clarithromycin and azithrorrmycin,
with rifabutin showed significant superiority to the macrolides, although resistance in breakthrough infections is a serious concern. Nevertheless, the data fr-om the MAC trials is sufficiently compelling to recommend routine prophylaxis against MAC in advanced disease.
The data is less clear for CMV and fungal infections. Conflicting data cones from the two
trials of ganciclovir as prophylaxis for CMV, although differences in design make conpparisons
difficult. Fluconazole is highly effective in preventing cryptococcosis; however; in spit of
effectiveness, survival is not prolonged. Multiple prophylaxis is complicated by increased
toxicity, drug--drug interactions, microbial resistance, and cost. Thus the main effort in fu ture
strategies of prophylaxis ought to be directed to methods of more effectively targeting prophylaxis to those at greatest risk.
WG Powderly, 45 II Forestpark, Suite 304, St. Louis, MO, 63108 USA T-el: 314-454-0058
Fax: [phone] E-mail: wpowderl~imgate.wustl.edu
Tu.B.533
PROPHYLAXIS ISSUES IN DEVELOPING COUNTRIES
Perriens, los. UNAIDS, Geneva, Switzerland.
Reportedly the pattern of opportunistic infections (OI) among patients with AIDS in developing countries differs from that in industrialized countries. In most of sub-Saharan Africa
PCP is rarely reported. In Thailand PCP is also rarer than in early AIDS cases in the USA
and Europe. Other Ols are reported much more frequentlyThe best-known example is
tuberculosis, but other Ols such as penicilliosis (Thailand) and Chagas disease (Brazil) have
also been reported with increased frequencyThese differences in clinical presentation and
the potential vulnerability of the Ols should drive the orientation of OI prophylaxis in
developing countries.The problems of implementing prophylaxis policies are however many.
First, resource allocation to prophylaxis services should be perceived as equitable at a societal level, i.e., not divert resources from essential health services for the general population.
Second, it is not always known which diseases (toxoplasmosis, community-acquired pneumonia, septicemrnia, fungal infection) in addition to TB could be targeted because, due to limitations in the diagnostic infrastructure and reporting mechanisms, not enough is known of
their incidence and impact on mortalityThird, the underdevelopment and inefficiency of
Western-type medical services in developing countries, combined with the difficulties to
earn a living and illiteracy, leaves significant proportions of people in those countries uninformed on and uninterested in the potential benefits they could derive from such services if
they were available or improved. Lastly, in most settings the diagnosis of HIV infection is only
made when patients present with life-threatening Ols, which limits the window of opportunity for prophylaxis. In view of those difficulties it is encouraging that several teams are now
investigating the cost-effectiveness of different prophylaxis options in Africa and Asia.
Jos Perrilns, UNAIDS, 20 avenue Appia, CH- 121 1 Geneva, Switzerland.
Tel.: (4 I-22) 79 I-4456; Fax: (4 I-22) 79 1-4 165: Email: [email protected]
Tu.B.540
RESPONSIVENESS OF THE MQOL-HIVTO QUALITY OF LIFE CHANGES IN A
COHORT OF HIV+ MEN
Smith, KevinW*, Avis N*, Mayer, K**. *New England Research Institutes,Watertown, MA,
USA; **Fenway Community Health Center and Brown Univ.
Objective: Responsiveness, or sensitivity to change, is an important indicator of the validity of
a measurement instrument.The purpose of this study was to evaluate the responsiveness of
the Multidimensional Quality of Life Questionnaire for Persons with HIV/AIDS (MQC)OI -HIV).
Methods: The MQOL-HIV is a 40-item questionnaire measuring 10 domains that are
important to the quality of life of persons infected with HIVThe MQOL-UIV Index derived
from these domains provides an overall quality of life score.This instrument was administered three times (baseline, 2-week follow-up, and 6-month follow-up) to a longitudinal
cohort of 92 HIV+ men at a Boston community health center: At the 6-month follow-up,
subjects were asked to assess changes in their quality of life over the previous 5.5 months
using a 7-point scale ranging from a great deal worse to a great deal better. Changes in
CD4 counts (baseline mean=350 cells/mm3; 33%<200 cells/mm3), hemoglobin (baseline
mean 14.4 gm/dl) and symptom severity (Whalen's HIV Symptom Sc.le) were also
tracked dunng the follow-up interval.
Results: After 5.5 months, 46% of the cohort reported that their quality of life had
improved, 32% reported it was about the same as before, and 23% had gotten worse.
These changes were highly correlated with changes in MQOL-HIV Index scores (r =.52,
p<.0001). Index score changes were also associated with changes in synptom severity (r-.44). Self-assessed quality of life changes were unrelated to changes in CD4 counts (r--.0 ),
and only weakly related to differences in hemoglobin levels (r. 10, p-.35). Correlations
between self-reported changes in individual domains and MQOL- IV domain scores
ranged f6-om r=.44 for financial status to r=-.20 for cognitive functioning.
Conclusions: These results provide strong evidence for the ability of the MQOL-HIV
Index to detect changes in both quality of life and symptom severity in HIV+ men.
Changes in hematologic parameters over time do not appear to be acciurate markers of
quality of life differences.
K.W Smith, 9 Galen Street, Watertown, MA, 02 172 USA
Telephone: [phone], x325 Fax: [phone] email: kevius%[email protected]

O
e)
0
ea
v
csa
cnO
C)
0
Q)
cc
0)
U
c2
r
cO
x
238