“Sex, Drugs & Circumcision: What

Does a Person Have to Do to Survive

HIV Infection Anyway?”

Kenrad E. Nelson, MD
Professor, Department of Epidemiology
Bloomberg School of Public Health,
Johns Hopkins University
 HIV infection rates among IV drug users in
Bangkok and national median rates, 1985-1991
 Seroprevelence of Elisa Positivity to HIV by Month Among
Female Prostitutes Attending Health Department STD Clinics
Chiang Mai, Thailand, Jan,1988 - Aug, 1989
Median provincial seroprevalence rates for selected
groups in national surveys

Data source: Division of Epidemiology

 The 100% Condom Program in Thailand

1. Designed and implemented by the Ministry of Public Health.

2. Is targeted at promoting 100% condom use in commercial sex

establishments.

3. Thailand has developed a national codom production capability.

4. Ministry of Public Health distributes free condoms to brothels and

other commercial sex establishments.

5. Publicity campaign to promote condom use in commercial sex.

6. Enforcement for non-compliance, i.e. closure of establishments.

 AIDS Policy Under Prime Minister
Anand Panyarachun (1991-1992)

1. Continued semiannual sentinal surveillance and make results public.

2. Increase mass media education about AIDS.

3. Enact legislation to protect the rights of those infected

4. Control the commercial sex industry

-100% condom use
-combat under age CSW’s
5. Expand the role and no. of participants in HIV/AIDS control.

6. Prime minister became chair of national AIDS committee.

7. AIDS recognized officially as one of the most important public health

problems.
HIV Seroprevalence by RTA Cohort,
Upper North, 1991-2001

Nelson et al., JAIDS 2002; AFRIMS 2001

Attributable risk due to sexual risk and IDU
Thailand army recruits, 1991-1998
Infection rates in high-risk groups
 Chiang Mai Couples Study
-Population-

1. Men who tested HIV-positive when donating

blood between 1992 and 1998 were
contacted.
2. Men who were married (or had a regular
female sex partner) were asked to bring her to
the blood bank for counseling & enrollment in
the study.
3. Couples were enrolled if woman had no other
risks for HIV except contact with male blood
donor.
 Results

1. From 1993-1999, 269 (49.4%) of men

and 65 (25.7%) of women had died.
2. Median time to death from
seroconversion was 7.8 years (95%
CI=7.0-9.1 yrs).
3. Subjects with enrollment.
- CD4+ counts < 200 cell/ul=RR=11.0
- CD4+ counts 200-500 cell/ul=RR=2.0
- CD4+ counts >500 cells/ul=RR1.0
Studies of Highly Exposed Persistent Seronegative
(HEPS) Women in Thai Couples Study

1. HEPS women (N=18)

a. Over 100 unprotected sex contacts with
HIV-pos, man, past year
b. Women remained HIV seronegative
2. Rapid seroconverters women (N=14)
a. Seroconveted <1 year after <100 sex
contacts with HIV pos men
3. Low risk seronegative women (N=12)
a. HIV negative, monogamous with HIV
negative man
 IFN-G ELISPOT Responses
(Peripheral Blood)

ENV GAG NEF POL ANY

HEPS 18% 44% 28% 28% 50%
N=18
RS 88% 75% 0% 56% 100%
N=14
CW 0% 0% 0% 0% 0%
N=12
The kindest cut: Male circumcision
for HIV and STI prevention
in men and women
Ron Gray, Johns Hopkins

A relief on the tomb of Ankh-Mahor

 Circumcision and HIV prevalence in adults
in sub-Saharan Africa

2001

Caldwell; Sci Am; 1996

 Ecologic Relationship Between HIV and Male
Circumcision
HIV Seroprevalence (%)

25

20

15

10

0
0 20 40 60 80 100

% Circumcised males

Bongaarts AIDS 1989

 MALE CIRCUMCISION AND POPULATION BASED HIV
PREVALENCE IN AFRICA

Senegal

Sierra Leone

Guinea

Burkina Faso

Ghana

Cameroon

Kenya High (>80%) male circumcision

Tanzania Low (<20%) male circumcision
Zambia

Lesotho

Botswana

0 10 20 30 40
Sources: ORC/MACRO, 2005, USAID, 2002
Cochrane Review of HIV and Circumcision
in High Risk Heterosexual Men

Siegfried et al. Lancet Infect Dis 2005

 HIV Prevalence by age in circumcised and
uncircumcised men. Rakai, Uganda
Kelly et al AIDS 1999,

Current Circumcised Uncircumcised PRR (CI)

Male age HIV (%) HIV (%)

15-24 0.8 3.5 0.22

(0.07-0.70)

25-34 14.6 26.2 0.56

(0.41-0.76)

35+ 9.1 19.2 0.47

(0.30-0.94)

Effects of MC appear long lasting

 Limits to observational studies
„ Circumcised men are often highly selected:
„ Religion (Islam, Judaism).

„ May be correlated with lower risk behaviors, alcohol use

etc.
„ Mainly neonatal

„ Traditional/tribal,

„ Often puberty rituals, sexual initiation,

„ safety of surgery

„ Medical indications (phimosis, GUD),

„ correlated with higher risk behaviors

„ Possible confounding between reasons for circumcision and

potential HIV effects
 Age at circumcision
„ Infancy or childhood (usually religious in Africa)
„ Puberty (ritual and sexual initiation)
„ Adult (often for medical indications)
„ Age of circumcision and reasons for circumcision can
confound observational studies

Circumcision for puberty ritual and sexual initiation,

 Traditional pubertal circumcision

„ Pubertal initiation rite (e.g., Zulu)

„ Poor surgical procedures, lack of asepsis and pain

control, possible HIV risk with contaminated
instruments

„ Often associated with sexual initiations of boys

before wound healing, could increase risk of HIV
 The apparent HIV protection afforded by
circumcision is greater in high risk populations
than in general populations
„ General population (10 studies)
„ RR = 0.57 (0.47-0.70),

„ Heterogeneity p = 0.23

„ High risk populations (10 studies)

„ RR = 0.31 (0.23-0.42)

„ Heterogeneity p = 0.005

(Weiss Geneva 05)

HIV Acquisition among HIV- Male Partners of HIV+
Female Partners By Circumcision Status In Rakai
30 27.7 27.7
Circumcised
25 Uncircumcised
Acquisition/100py

20

15

10 8.2

5
0 0 0
0
<10,000 10,000-49,999 >50,000
HIV+ Female viral load

Quinn et al NEJM 2000, Gray et al AIDS 2001

Meta-analysis of circumcision and
symptomatic male GUD

nasio

gray
telzak

reynolds

seed
new ell

donovan
bw ayo

simonsen

agot Pooled OR = 0.63 (0.56-0.72)

Heterogeneity p = 0.08
Combined
.1 1 10
Odds ratio

M Thoma, R Gray 2006

Circumcision Status and HIV Transmission to Women

30
Circumcised 25.6
25
25 Uncircumcised
Transmission/100py

20

15 12.6

10 6.9

5
0 0
0
<10,000 10,000-49,999 >50,000
Male Viral load
47 couples in which circumcised male partner was HIV+ AND whose viral load was
<50,000 particles, 0 of female partners were infected after two years, vs. 26 of 143
female partners of uncircumcised HIV+ men (9.6/100 py) (p = 0.02).
Gray et al AIDS 2000
Circumcision and Male-to-Female HIV & STI
Transmission in Rakai (Gray et al CROI 2006)

Female Infection RR (CI)

Prevalent HIV 0.76 (0.62-0.92)*
HSV-2 0.75 (0.54-1.03)*
BV 0.79 (0.69-0.91)*
Trichomonas 0.65 (0.55-0.77)*
Chlamydia 1.06 (0.61-1.84)
Gonorrhea 1.19 (0.51-2.79)
Syphilis 0.93 (0.76-1.13)
HPV 0.72 (0.46-1.12)

*P<0.05
Biological Rationale for circumcision & HIV

„ Biological plausibility
„ Inner mucosa of foreskin is rich in HIV target cells (9x
cervix)
„ External foreskin/ shaft keratinized and not vulnerable

„ After circumcision, only vulnerable mucosa is meatus

„ Foreskin is retracted over shaft during intercourse

„ Large inner mucosal surface exposure

„ Micro-tears, especially of frenulum

„ Intact foreskin associated with infections

„ GUD

„ Balanitis/phimosis

„ Possible increase HIV entry or shedding

Possible circumcision protective mechanisms

Circumcision Anatomic effect by removal

of foreskin

Reduced GUD,
Reduced HIV STI cofactor
Target cells effects
KERATIN THICKNESS ON THE EXTERNAL AND INTERNAL
MUCOSAL SURFACES OF HUMAN FORESKIN

Keratin

External Surface Mucosal Surface

Patterson et al. Am J Pathol 2002
 HUMAN FORESKIN INFECTED WITH HIV-1Bal IN
EXPLANT CULTURE

Infected T-cells

Mucosal Surface External Surface

Red-uninfected cells
Green-infected T-cells
Patterson et al. Am J Path 2002
Yellow-HIV-1 bound to Langerhan’s cells
 Paradox: Why is circumcision most protective in the
most highly exposed?
(Wawer et al AIDS 2005)

„ Circumcision more protective in men with repeated HIV

exposures (e.g., high risk populations), than men with
infrequent HIV exposures (general populations)

„ Possible induced mucosal immunity following repeated

subinfectious HIV antigen stimulation of urethral
mucosa

„ Analogous to reduced infection in highly exposed but

persistent seronegative commercial sex workers (Devito
et al AIDS 2000, Kaul et al J Immunol 2001)

„ Circumcision reduces GUD which is a cofactor for HIV.

GUD is more common in high risk populations
(possible indirect effects)
 Randomized trials of male circumcision for HIV
prevention in men

„ Three trials:
„ South Africa (ANRS),
„ Kenya (NIH),
„ Uganda (NIH)

„ Similar designs:
„ Enroll HIV- uncircumcised men, randomize to:
„ Immediate circumcision (Intervention)

„ Circumcision delayed 21-24 months (Control)

„ Endpoints:

„ HIV incidence

„ Safety

„ Behavioral disinhibition

„ STIs
 Trial Site Community Characteristics

South Africa Uganda Kenya

Trial setting Semi-urban Rural Urban

Study sites 3 clinics Community 1 clinic

based
Circumcision (%) ~20% 16% <10%

HIV incidence 2.7-5.2 1.1-2.1 2.5

 The Three Trials: Study Characteristics

South Africa Uganda Kenya

Age 18-24 15-49 18-24

Final N 3,520 ~5,000 2,784

Enrollment completed Jan 2005 Jul 2005 Sept 2005

Completion date Mar 2005 Jul 2007 Sept 2007

 South African ANRS 1265 Trial

„ Trial stopped by DSMB (Nov, 2004) at interim analysis

„ Intervention n= 1538, Control n= 1590
„ Enrollment comparability good
„ Cross-over: Intervention 4.8%, Control 8.4%
„ Loss to follow up: Intervention 6.8%, Control 9.7%

Auvert et al PLoS Med 2005

 South African ANRS Trial
Incident cases M0-M3 M4-M12 M13-M21 Total
Intervention 2 7 11
20
Control 9 15 25
49
Total 11 22 36
69
Incidence rates :
Intervention : 0.9 (0.6 - 1.3) /100 py
Control : 2.1 (1.6 - 2.8) /100 py

Unadjusted RR : 0.40 (0.24 – 0.68) p=0.0006

Protection (1-RR): 60% (32% - 76%)

Auvert et al PLoS Med 2005
 Effect of behavioral disinhibition on HIV incidence, by
circumcision IRR in both sexes combined, 75% coverage

2.4
H IV Incidence/100 py

RR 0.5
1.9 RR 0.4
RR 0.3
1.4
Current incidence
1.24/100 py
0.9

0.4 1 2 3 4

Observed = 1.0 1.25 1.50 2.0

Increased number of sex partners

Behavioral disinhibition can offset all benefit, even at high circumcision efficacy
 AIDS Link to Intravenous
Experience (ALIVE II)
„ Prospective cohort study, began 1988.
„ Eligibility: AIDS free, injection drug use < 10
yrs prior, over 18 yrs.
„ Re-opened recruitment in 1994 and 1998
„ Semi-annual HIV antibody screen and interview
„ Endpoints: HIV seroconversion, death
„ HIV seroconverters enter parallel cohort study
of HIV natural history
 ALIVE-II

3. Importance
a. HIV incidence & risk behaviors
b. Control for drug use in analysis of HIV-
related morbidity and morbidity
4. Populations
a. N= 1,532 (1988-9)
b. N= 338 (1994)
c. N= 400 (1998)
 Heterosexual Sex and IDU Risks
for HIV Incidence
Exposures No p/yrs Scs Inc RI (95% CI)

No sex, no IDU 653 8 1.23 1.00

Het sex, no IDU 2,382 43 1.81 1.47 (0.69, 3.13

Het sex, inj <daily 2,282 60 2.63 2.14 (1.02,4.48)

Het sex, inj ≥ daily 2,427 180 4.45 3.63 (1.77,7.44)

No sex, inj < daily 426 20 4.70 3.83 (1.69,8.69)

No sex, inj ≥ daily 504 26 5.16 4.21 91.90, 9.29)

 Temporal Trends in Seroconversion Rates - ALIVE 1988 - 2003

SC Rate/100 PY: All cohorts SC Rate/100 PY: 70's SC Rate/100 PY: 80's
4.5

3.5

2.5

1.5

0.5

0
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Year
 Trends of Reported Drug Use in ALIVE Sero-negative participants: 1988 - 2003

60

50

40
Rate per year (%)

30

20

10

Shooting gallery visits Shared needles

0
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
 HIV/AIDS Cases among IDUs

No Data Reported
0% HIV/AIDS cases among IDUs
< 10% HIV/AIDS cases among IDUs
10-40% HIV/AIDS cases among IDUs
SOURCE: WHO, UNAIDS
40%+ HIV/AIDS cases among IDUs
Proportion of Injecting Drug Users Among
Reported HIV Cases*
11% <20%

20-40%

40-60%

60-80%

>80%
19.5% 3.7%

11.2%
85%
39.8% 92.7%
8% 76.3%
7.3% 10.8%
72.3%
87.1%
6.5% 78%
14.4% 62.5%
16% 82%
4%
22.7%
1.3% 69.4%
11.8% 24.5%
49.7% 1% 60%
2.6% 83% 76%
32%
63.5% 12.5%
<0.1%
24%
59.6% 50.7%*
2.5% 70.9%
47.7%
1.3% 13.6%
49.2%

3.8% 7.5%

10%

*Proportion among AIDS cases in countries not reporting HIV data (Sources:
European Commission;
* EuroHIV; Council of Europe)
12.3%
States (N= 26) with at least one site with
HIV prevalence > 20% in IDU in 2005
„ Belarus „ Burma
„ Estonia „ China
„ Kazakhstan „ India
„ Russia „ Indonesia
„ Ukraine „ Malaysia
„ Iran „ Nepal
„ Thailand
„ Italy „ Vietnam
„ Netherlands
„ Portugal „ Argentina
„ Serbia & Montenegro „ Brazil
„ Spain „ Canada
„ Uruguay
„ Libya „ Puerto Rico
„ USA
* Denotes recent emergence [Adapted: Aceijas, et al, AIDS 2004 18:2295-2302]
Emerging Epidemics: Former Soviet Union (FSU)

Country Adult HIV Prevalence % transmission % transmission

attributable to IDU in persons
<30 y.o.
Belarus .2%-.8% 74%

Estonia 1.1% 90% 75%

Kyrgyzstan .1%-.2% 83% 70%

Russian 1.1% 76%

Federation
Ukraine 1.4% 72%

UNAIDS AIDS Epidemic Update, 2004 available

online:http://www.unaids.org/wad2004/report.html
Afghan Opium Production (metric tons)
1980-2004

Afghanistan Opium Survey, 2004, UNODC

Economic Rationale for Poppy
Cultivation

Afghanistan Opium Survey, 2004, UNODC

HIV Prevention Tools with Evidence of
Efficacy for IDU Transmission
1. Methadone Maintenance Therapy
2. Other Substitution (Buprenorphine)
3. Harm Reduction
¾ Needle and Syringe Exchange (NEP)
¾ Condom provision
¾ Psycho/social counseling
4. Peer Education/Behavioral Interventions
 High HIV Prevalence and Incidence
Observed among African-American Men who
have Sex with Men (MSM) in Baltimore:

The Behavioral Surveillance Research (BESURE) Study

Frangiscos Sifakis, PhD, MPH

Department of Epidemiology - Seminar

March 17, 2006
 The BESURE Study - MSM
• Study design
– Cross-sectional, anonymous, venue-based
– Venues: public/private places frequented by MSM
– Data collected by means of questionnaire administration
and blood drawing for HIV testing
• Study area
– All venues within Baltimore City limits
– Random selection of sampling calendar
– Up to three non-random, seasonal/special events
• Eligibility Criteria
– Residents of Baltimore Metropolitan area
– 18 years of age or older
– No prior participation
 RESULTS - Overview
• Overall, 243 of 645 participants were HIV positive

HIV prevalence = 37.7%

• Overall, 20 of the 243 HIV positive specimens were recent cases

HIV incidence =10.7% per year (95% CI: 5.3, 19.3)

• Overall, 142 of 243 HIV positive participants did not know they
were HIV infected

Prevalence of Unrecognized HIV Infection = 58.4%

 HIV Seropositivity & Associated Risks
Total (%) HIV+ (%) OR (95% C.I.)
N = 645 243 (37.7)
Race/Ethnicity
White, not Hispanic 201 (31.2) 26 (12.9) Reference
African American, not Hispanic 373 (57.8) 192 (51.5) 7.1 (4.6, 11.5) **
Hispanic 34 (5.3) 11 (32.5) 3.2 (1.4, 7.3) *
Other, not Hispanic 37 (5.7) 14 (37.8) 4.1 (1.9, 8.9) **
Age Group
18 to 24 163 (25.3) 41 (25.2) Reference
25 to 34 166 (25.7) 53 (31.9) 1.4 (0.9, 2.3)
35 to 44 197 (30.5) 99 (50.5) 3.0 (1.9, 4.7) **
45 or older 119 (18.5) 50 (42.0) 2.2 (1.3, 3.6) *

* p-value <0.05; **p-value <0.001

 Acknowledgements

Ron Gray Natawan Lanier David Vlahov

Tom Quinn Janet McNicholl Steffanie Strathdee
Caroline Costello Ann Duerr Frank Sifakis