Final Program and Oral Abstracts [International Conference on AIDS (8th: 1992: Amsterdam, Netherlands)]

WeC 1091-WeC 1095

TRACK C: EPIDEMIOLOGY

WeC 1091 AGE, PERIOD AND COHORT MODEL TO ESTIMATE
PREVALENCE AND INCIDENCE OF HIV INFECTION. Verdecchia. Arduino:
Mariotto, A.; Capocaccia, R.; Mariotti, S. Istituto Superiore di Sanita, Rome, Italy
Objective. To present an age, period, and cohort (APC) approach both to estimating and analyzing HIV epidemics and to a
reconstruction of the HIV epidemics in Italy.
Methods. In a back-calculation framework, a system of integral equations is defined, relating AIDS counts to the probability
of new infections among the population, to the AIDS incubation time distribution and to the differential force of mortality
from other causes for HIV infected subjects. The equations provide a regression model to fit age and period specific AIDS
counts assuming HIV infection incidence as a function of age at the time of infection, period, and/or birth cohort in the
logistic scale. Maximum likelihood estimates of the HIV infection incidence are then calculated. A two-stage AIDS
incubation time distribution, estimated from data on an Italian cohort of 420 subjects who seroconverted between 1982 and
1990, was found to be dependent on age at infection and on period of infection due to the introduction of AZT therapy in mid
1987; this two-stage distribution also allowed for the estimation of the number of infected individuals in the ARC stage.
AIDS counts were corrected for reporting delay.
Results. The estimated number of new infections in Italy decreased after 1987, though the number of new AIDS cases has
continued to rise, albeit less sharply in recent years. When delay in the progression to AIDS due to therapy is taken into
account, the estimated number of persons infected in mid-1990 increases from 52,000 to 67,000, with approximatively 12%
of subjects already in the ARC stage. Approximatively 68% of HIV infections in Italy are among intravenous drug users, and
15% are among homosexual males. The age at maximum risk of infection is 25 years in males and 23 years in females.
Based on a hypothesis thatthe rate of new infections will level-off at the 1989 level, it is predicted that ARC counts will
continue to rise until 1993, and that AIDS counts will continue to rise beyond 1995.
Conclusions. Although the rate of new infections has been rapidly decreasing in recent years, it is still unacceptable (10-20
new infections every day), especially considering the very young age at which they occur. Efforts discouraging drug use and
promoting safe sex are still needed to prevent HIV infection among young people. Special attention should be paid to
monitoring trends of new infections in the heterosexual population, in order to prevent a possible spread of the HIV epidemic
among the general population.
Arduino Verdecchia. Laboratorio di Epidemiologia e Biostatistica, Istituto Superiore di Sanita,
V. Regina Elena, 299, 00161- Roma, Italy. Tel: [phone] Fax: [phone].
W eC 1093 AIDS IN THE WORLD 1992 - ESTIMATING HIV/AIDS MORBIDITY AND
MORTALITY SINCE THE BEGINNING OF THE PANDEMIC. Mantel, Carsten;
Tarantola, D.; Lepisto, E.; Jain, N.; Shahi, G.; Mann, JM. Global AIDS Policy
Coalition, International AIDS Center, Harvard School of Public Health, Boston, USA.
Objectives: The reported number of persons with AIDS gives an incomplete picture of the global situation. We
used a simple multi-stage modeling approach to better determine the dynamics of the pandemic.
Methods: (1) From a Delphi survey, estimates of AIDS reporting efficiency were produced in each of ten
defined regions in the world and crude corrections for the reporting delays were made. This provided adjusted
AIDS incidence figures for the period 1984-1991. (2) From a review of all available recent seroprevalence data
for antenatal women, blood donors and commercial sex workers, a current number of HIV infected persons was
computed for each region. Adjustment was made for type of sentinel group and gender and urban/rural
differences in HIV prevalence. (3) Adult HIV incidence and prevalence curves were modeled independently for
each region using a gamma-distribution (EPIMODEL software). The model was simultaneously adjusted to fit
the AIDS incidence curve derived previously, incorporating differential rates of progression from HIV to AIDS.
The adult AIDS mortality distribution was generated, and the number of children with AIDS, the pediatric HIV
incidence and pediatric AIDS mortality were calculated.
Results: It was estimated that the cumulative global number of adults with AIDS was between 1.6 and 1.9
million in early 1992. About 63% were from Sub-Saharan Africa, 15% from North America and 11% from Latin
America and the Caribbean. Up to 11 million adults have been infected with HIV since the beginning of the
pandemic. More than 500.000 children are estimated to have developed AIDS and subsequently died. The overall
number of adult deaths from AIDS is thought to have reached 1.5 million. Projections based on refined models
were made for the years 1995 and 2000.
Conclusions: The projections derived from Delphi-adjusted AIDS incidence estimates combined with data from
HIV serosurveys indicate that the pandemic will continue to expand during the 90's with: (a) further spread in
areas which already experienced high HIV/AIDS incidence in the 80's; and (b) a rapid progression in areas where
incidence so far had remained low. Global and regional estimates will be regularly updated and published in
"AIDS in the World".
Mantel Carsten, MD, MPH, Harvard School of Public Health,
Department of Epidemiology, 677 Huntington Ave., Boston, MA, 02115, USA,
Telephone: [phone], FAX: [phone]
W eC 1 095 AIDS PREVENTION IN THE NETHERLANDS: 1992-2000
Reinking, Dick P.', C.E.S. Albers, MJ. Postma", J.C. Jager", F.M.L.G. van den Boom'
* Netherlands Institute of Mental Health (NcGv), Utrecht, The Netherlands; " National
Institute of Public Health and Environmental Protection (RIVM), Bilthoven, The Netherlands
Objectives: To assess the necessity, purposiveness and effectiveness of AIDS-prevention activities for the years 1992-2000.
Methods: (1) integrative research (literature study and consultation of experts) on developments in HIV-infections, size
of risk groups, riskful behavior and the impact of AIDS-prevention so far; (2) elaboration of scenarios, exploring the
future spread of HIV in the Netherlands among varying risk groups, based on simulations with mathematical models;
(3) assessment of the required level and intensity of AIDS-prevention for the period 1992-2000
Results: In a reference scenario, exploring continuation of existing epidemiological trends, the AIDS-incidence in the
period 1992-2000 is estimated at 5,700. This trendextrapolation corresponds to an HIV-incidence in the period 1990-2000
of more than 3,000 infections. Thus the necessity of continued and intensified AIDS-prevention the next decade is obvious,
especially when the underlying assumptions of this scenario are considered: (1) improved effectiveness of prevention,
(2) no substantial spread of HIV outside known risk groups; (3) no effects of increased mobility.
Scenarios exploring developments in riskful behavior give additional support for the conclusions from the reference
scenario. Among relatively small groups of (a) 20,000 men with homosexual contacts, (b) 8,000 ivdu and their
heterosexual partners and (c) 15,000 prostitutes and their clients, scenario analysis shows a possible HIV-incidence on
a short term up to 8,000 HIV-infections among group (a) due to relapse, up to 1,000 HIV-infections among female non
ivdu heterosexual partners of ivdu and up to 700 HIV-infections among group (c). Furthermore scenario analysis
indicates that a substantial spread of HIV among the heterosexual population seems unlikely within a few years, but
might not be excluded on a long term.
Conclusions: For known risk groups continuation of targeted AIDS-prevention is essential and intensification of it must
be considered. For the heterosexual population at large a reflection is recommended on an effective AIDS-prevention,
combining factual information of current risk and the consequences of neglect on the long run.

WeC 1092 NEEDLE EXCHANGE LOWERS HIV INCIDENCE IN NEW HAVEN Kaplan, Edward
1. *; Heimer, Robert* *Yale School of Organization and
Management, and Yale School of Medicine, New Haven, CT USA,
Yale School of Medicine, New Haven, CT USA
Objective: To verify our earlier prediction that HIV incidence has been reduced
among IDUs in the New Haven needle exchange program.
Methods: A syringe tracking and testing system (STT) was established whereby all
needles distributed by the program receive unique tracking codes, and a sample of
returned needles are tested for the presence of HIV proviral DNA using PCR. From
the STT, it is possible to sequence chronologically the needles returned by
individual IDUs in the program, and to statistically determine whether the
infection pattern in the needles shifted from predominantly negative to
predominantly positive during the course of the needle exchange, as would occur
if the IDU in question became infected.
Results: Earlier modeling predicted a drop in HIV incidence from 6.4 per 100
IDUs per year to 4.3 per 100 IDUs per year owing to the operation of the needle
exchange. The data employed here covered the equivalent of 9,062 client days,
thus one would expect 1.07 new infections to be discovered if the incidence of
infection had been reduced as predicted, with a 95% Poisson coverage interval of
0-3 infections. PCR tests of 850 needles covering 69 client IDUs (for an average
of 12.3 needles tested per IDU), were statistically analyzed four ways. Three-of
the analyses determined mean incidence levels falling inside the coverage
interval (0.31, 0.55, 1.31), while the fourth fell just outside (3.55).
Conclusions: These data are consistent with the incidence reduction forecast
earlier. That the technique employed here is completely different from Kaplan's
"Needles That Kill" model on which earlier conclusions were based provides
additional evidence that needle exchange has reduced HIV incidence in New Haven.
Kaplan, Edward Yale School of Organization and Management,
Box 1A, New Haven, CT, USA 06520 Tel. [phone]
FAX [phone]
WeC 1094 INDIRECT ESTIMATION OF THE PREVALENCE OF HIV INFECTION AND ADVANCED
HIV DISEASE IN THE EUROPEAN COMMUNITY
Downs, Angela M.; Ancelle-Park, R.A.; Brunet, J.B. European Centre
for the Epidemiological Monitoring of AIDS, Paris, France.
Objective: To estimate, by back calculation, prevalence levels of HIV infection and
advanced stage HIV-related disease in the European Community (EC).
Methods: AIDS surveillance data reported by December 1991 were adjusted for reporting
delays and under-reporting. Back calculations used a 2-stage incubation period with
overall median 10 years in absence of treatment (stage 1: Weibull, 6.5 years; stage 2:
exponential, 2.5 years). In stage 2 (advanced disease, T4 count < 200), the effects of
treatment were modelled from 1988 onwards. The HIV infection curve was modelled by a
step function with 4 or 5 steps chosen to give the best fit by least squares criteria.
Results: Baseline estimates were obtained using AIDS cases diagnosed before 1988 only
(prior to early treatment). In the absence of effective treatment after 1987, projected
AIDS incidence in 1988-1990 (PROJ-untreated) is significantly higher than reported.
Group HIV prevalence 31.12.1987 ------- AIDS cases 1988-1990 -------
cumulated stage 1 stage 2 PROJ-untreated (lower CI) reported*
ALL 400 000 350 000 34 000 52 000 (47 000) 38 000
Homosex. 140 000 110 000 16 000 22 000 (19 000) 16 000
IDU 140 000 130 000 8 600 15 000 (13 000) 12 000
Heterosex. 51 000 47 000 2 600 4 900 ( 3 800) 3 400
* data at 31.12.91, adjusted for reporting delay and 10% under-reporting
Preliminary results indicate that effective treatment (RR of progression=0.35) of stage
2 patients (with 38% treated by Jan. 1989) could account for the overall discrepancies
in 1988/89, but not from 1990 onwards. Further investigations are in progress.
Conclusion: In the EC, a cumulative total of around 400 000 persons are estimated to
have been infected with HIV by the end of 1987. Recent AIDS incidence figures suggest
significant treatment effects in all transmission groups. Updated prevalence estimates,
taking into account these effects, will be presented.
Downs, Angela M., European Centre for the Epidemiological Monitoring
of AIDS, H6pital Saint-Maurice, 14 rue du Val d'Osne, 94410 SaintMaurice, France. Tel. [phone], FAX: [phone]

NOTES

Dick P. Reinking, Netherlands Institute of Mental Health (NcGv), POBox 5103,
3502 JC Utrecht, The Netherlands. Tel. [phone]; FAX: [phone].

We62