A Randomized Controlled Trial of Male Circumcision to Reduce HIV Incidence in Kisumu

1
A Randomized Controlled Trial of Male
Circumcision to Reduce HIV Incidence in Kisumu

• Robert C. Bailey J.O. Ndiny-Achola
• Univ. of Illinois at Chicago Univ. of Nairobi
• Stephen Moses Kawango Agot
• University of Manitoba Moi University
• Ian Maclean John Krieger
• University of Manitoba University of
  Washington
• Corette Parker Richard Campbell
• RTI International Univ. of Illinois at Chicago
• Supported by Division of AIDS, NIAID, and the
  Canadian Institute for Health Research

2
Need for Clinical Trials

• Randomized clinical trials are needed to
  determine the utility of circumcision as an HIV
  preventive measure.
• Reasons
• All epidemiological studies to date have been
  observational (cannot exclude residual
  confounding) and not all results are consistent.
• Risk of too early resumption of sexual activity
  after circumcision or subsequent behavioural
  disinhibition could counteract any protective
  effect.
• Risk of post-surgical complications must be
  balanced against any protective effect.

3
Hypotheses

• Circumcision will reduce HIV incidence among men
  aged 18-24 years by 50.
• Circumcision will result in less than a 2 rate
  of significant post-surgical complications
  requiring follow-up care.
• There will be no difference between circumcised
  men and controls in reported sexual behaviour
  following the circumcision procedure.

4
Study Population

• Approximately 2,600,000 Luo live in Nyanza
  Province, western Kenya.
• Luo do not practice traditional circumcision.
• 9.9 of Luo men are circumcised.
• Kisumu District has approximately 39,000 men aged
  18-24 years.
• 86 live within 30 minutes of central Kisumu
  City.

5
Study Design

• Randomized, unblinded treatment arm is
  circumcision, control arm is no circumcision.
• Both arms HIV testing, behavioural
  questionnaire, medical examination, STD testing
  and HIV prevention counselling at baseline, and
  at 6, 12, 18 and 24 months post-enrolment.
• Additional HIV testing at 1 and 3 months.
• Circumcision arm Post-operative check-ups at 3
  days, 8 days and 30 days after the procedure.

6
Inclusion and Exclusion Criteria

• Inclusion Criteria
• Uncircumcised.
• HIV negative.
• Sexually active.
• Aged 18-24 years.
• Resident of Kisumu District.
• No plans to move for at least the next 2 years.
• Consent to participate.
• Hemoglobin over 9.0 grams per 100 ml.

• Exclusion Criteria
• Foreskin covers less than half of glans.
• Hemophiliac or other bleeding disorder.
• High Prothrombin Time Index.
• Other medical condition contra-indicating
  surgery.
• Absolute indication for circumcision.

7
Sample Size

• 1,388 men in each arm provides 80 power (with
  alpha error 0.05) to detect a 50 protective
  effect of male circumcision
• Assumes 5 (2.5 annual) incidencein the control
  group
• 15 loss to follow-up
• 5 cross-over in each direction
• Plan to screen approximately 4,500 - 5,000 men to
  arrive at the final sample size.

8
Assumed Participant Flow from Screening to
Randomization
Report for VCT (Screening) N 4,525
HIV neg. N 3,653
5
95
Excluded N 183
95 accept
Asked to enroll N 3,470
20
80
Declined N 694
HIV test N 4,298
Enrolled N 2,776
61 of screened
15
85
50
50
Circumcised N 1,388
Controls N 1,388
HIV pos. N 645
HIV neg. N 3,653
9
STD Testing

• Serum for syphilis (RPR, TPHA) and HSV-2
  (type-specific ELISA)
• Urine for gonorrhea and chlamydia by PCR, and
  trichomonas by culture
• Urethral, glans and shaft swabs for HPV by
  nucleic acid amplification
• If urethral discharge, urethral swabs for
  gonorrhea and chlamydia by PCR
• If genital ulcer present, ulcer swab for
  multiplex PCR

10
Circumcision Procedure

• Infection control
• Bleeding control
• Emergency procedures
• Follow-up
• 3 days, 8 days, 30 days

11
Potential Adverse Events

• Excessive bleeding
• Infection
• Excessive pain
• Too much skin removed
• Anesthetic-related problem
• Damage to the penis

• Cosmetic problem
• Erectile dysfunction
• Psycho-behavioural problem
• Penile amputation
• Death

12
Adverse Events

• Categories of Severity
• Mild
• Moderate
• Severe
• Categories of Relationship
• Definitely related
• Probably related
• Possibly related
• Not related

13
Recruitment of Participants

• Clients of STD clinics
• Clients of VCT centres
• Organizations (boda boda, jua kali)
• Peer recruiters
• Fliers, radio, other media
• Drama and musical groups

14
Willingness to Participate

• 86 of 18-25 year-olds said they would prefer to
  be circumcised.
• 97/103 (94) said they want to participate in the
  trial.
• Reasons given
• Free circumcision
• Free medical treatment for two years
• Compensation for time and transport
• Getting free counselling about safe behaviours

15
Actual Participant Proportions from Screening to
Randomization
Report for VCT (Screening) N 4,525
HIV neg. N 3,653
(Actual - 22)
5
95
(78)
Excluded N 183
95 accept
(Actual - 99)
Asked to enroll N 3,470
(Actual - 17)
20
80
(83)
Declined N 694
HIV test N 4,298
Enrolled N 2,776
(Actual- 11)
( 89)
61 of screened
15
85
50
50
Circumcised N 1,388
Controls N 1,388
HIV pos. N 645
HIV neg. N 3,653
(Actual - 60)
16
Progress to Sept 14, 2002

• 1245 men screened and 1205 consented for HIV
  testing (97)
• HIV positive 137 (11.0) 5.4 in 18-20 age
  group, 15.3 in 21-24 age group
• 662 men randomized
• Approximately 12 (3.6) of those in circumcision
  arm have asked that procedure be delayed and are
  not yet re-scheduled (potential cross-overs)
• No reports of cross-over from control to
  circumcision arm

17
Concerns

• Power to detect a 50 reduction in HIV incidence.
• Will HIV incidence be high enough?
• Cross-over (dis-satisfaction with being
  randomized to control group).
• Differential loss to follow-up.
• Behavioural disinhibition in the MC group.
• Minimizing loss to follow-up.
• Use of clinic for non-essential visits.
• Care for HIV positive men.