close

Вход

Забыли?

вход по аккаунту

?

Preliminary analysis of recent HIV infection in Kenya, KAIS 2007

код для вставкиСкачать
Preliminary analysis of recent
HIV infection in Kenya,
KAIS 2007
Where are new infections occurring in
Kenya?
Tom Oluoch on behalf of the KAIS technical working
group
Background
•
Two distinct applications of serologic incidence assays
– Application 1: to estimate HIV incidence rates
– Application 2: to distinguish recent from long-standing infection
•
•
Current incidence assays overestimate both recent infection
and incidence rates due to misclassification of long-term
infection as recent
To minimize error:
– Application 1 (incidence estimation): Calibrate incidence
estimate by incorporating correction factors based on expected
misclassification rates into the mathematical formula for
calculating incidence
– Application 2 (distinguishing recent infection): Predictive value
can be improved using case-based exclusions to identify and
exclude known long-term infections, based on data on ART, CD4
count, and HIV testing history
•
Focus of this presentation is on Application 2 – the use of
the BED assay and case-based exclusions to detect recent
infections. Incidence rates will not be reported.
2
Methods
• BED assay applied to frozen HIV antibody positive serum from the
2007 Kenya AIDS Indicator Survey (KAIS)
• BED results linked to the KAIS questionnaire using unique study
identification number
• Case-based exclusions: Specimens that classified as BED recent and
reported the following were re-classified as long-term infections and
excluded from the incidence analysis
– 1) Currently using ART
– 2) Last HIV positive test that was >1 year ago, or
– 3) CD4 cell counts < 500
• A CD4 cut-off of 500 was based on data from Uganda demonstrating a
lower median baseline CD4 cell counts among HIV negatives
(approximately 500 cells/mm3 using 90% ranges)
Methods
• Weighted descriptive analysis conducted to characterize
the distribution of recent infection by select demographic
variables
• Weighted multivariate analysis conducted to assess
potential risk factors for recent infection
– Outcome: Recent infection compared to HIV negative
– Significant correlates represent risk factors for recent
infection in the past 6 months
The number of recent infection using
case-based exclusions, KAIS 2007
Number HIV antibody + in KAIS: 1,073
Number BED recent: 181
Number on ART: 21
Number CD4<500: 41
Number self-reported HIV infection> 1 year: 1
Excluded:
63 BED
recent
Final number classified as recent
infection: 118
11% of
all HIV
Ab+
Distribution of recent infection by gender
and age group, KAIS 2007, N=118
50
47
43
45
Males (N=41)
Females (N=77)
40
Percent (%)
35
36
30
25
21
16
20
15
13
11
10
5
3
0
15-24
25-34
35-44
Age group
45-54
9
2
55+
Distribution of recent infection by residence,
KAIS 2007, N=118
Urban,
22.0%
Rural,
78.0%
The vast majority of participants in the KAIS 2007 were from rural
residences. Similarly, of all recent infections in KAIS, most (78%) were
observed among rural participants compared to urban participants.
Distribution of recent infection among rural
participants by gender and age, KAIS 2007 (N=86)
2.8
3.9
100
90
80
11.2
15.4
14.1
Percent (%)
70
60
21.1
40.9
50
40
37.5
30
20
10
37.1
16.1
0
Males (N=29)
Females (N=57)
Gender
55+
45-54
35-44
25-34
15-24
Distribution of recent infection by province
and gender, KAIS 2007 (N=118)
45
40
35
Percent (%)
40
Males (N=41)
Females (N=77)
30
26
28
26
25
20
15
10
5
11
9
3
7
11
6
8
7
2
3
1
0
Nairobi
Central
Coast
Eastern
North
Eastern
Province
Nyanza
Rift
Valley
Western
Distribution of recent infection by marital status,
KAIS 2007 (N=118)
15%
Married/Cohabitating
Formerly Married/Cohabitating
9%
Never married/cohabitating
76%
Serodiscordant couples in married/cohabitating relationships may be driving new
infection in Kenya. Among all HIV-infected persons in KAIS that were married or
cohabitating with another person, 44% had an HIV-uninfected partner
Distribution of recent infection by selfreported HIV testing status, KAIS 2007
(N=118)
Accurate
knowledge
of HIV status
1%
HIV-infected
but thought
be HIVnegative
based on
last HIV test
38%
Never tested
for HIV
56%
Missing
5%
Distribution of recent infection by circumcision
status, Kenya and Nyanza Province, KAIS 2007
15%
Circumcised
Uncircumcised
37%
63%
85%
Kenya (N=118)
Nyanza province (N=13)
Distribution of recent infection by HSV-2 status,
KAIS 2007 (N=118)
100
Percent (%)
80
5
2.3
20.3
Indeterminate
59.3
60
40
70.4
20
35.7
0
Males (N=41)
Females (N=77)
Gender
HSV-2 uninfected
HSV-2 infected
Multivariate model: Risk factors for recent
infection among females
Variable
Adjusted Odds Ratio
95% Confidence Interval
Age category
Age<30 years
Age>=30 years
1.0
0.5
0.3 – 0.8
Province
Nairobi
Nyanza
1.0
3.3
1.9 – 5.7
Condom use
Ever used a condom
Never used a condom
1.0
1.7
1.0 – 3.3
HSV2 status
HSV-2 negative
HSV-2 positive
1.0
4.1
2.2 – 7.5
*Comparison group is HIV negative persons. Model controlled for age, education, marital
status, HSV-2, condom use with last sex partner, ever tested, ever used condom, number
of partners in past 12 months, STD symptoms
Multivariate model: Risk factors for recent
infection among males
Variable
Adjusted Odds Ratio
95% Confidence Interval
Circumcision status
Circumcised
Not circumcised
1.0
2.7
1.0 – 7.8
Genital ulcer past 12 months
No
Yes
1.0
4.5
0.9 – 22.2
*Comparison group is HIV negative persons. Model controlled for age,
education, marital status, HSV-2, condom use with last sex partner, ever tested,
ever used condom, number of partners in past 12 months, STD symptoms, GUD
(males only), circumcision (males only)
Conclusion
• The combination of BED results and case based
exclusions using ART and CD4 data improved the
predictive value for recent infection.
• Analysis of recent infection in 2007 KAIS support a
mixed epidemic in Kenya:
– 1) New infections concentrated in Nyanza province fueled
by lack of circumcism in males.
– 2) New infections continue to be high in young people,
especially women. HSV-2 appears to be a major risk factor.
– 3) Recent infections were most notable in rural areas and
higher among older rural men compared to older rural
women.
– 4) Most recent infections were found in married/cohabitating
couples. Serodiscordancy and low condom use in these
relationships are of major concern.
Conclusion
• Data on recent infection in 2007 KAIS has
provided critical information for prevention
program planning in Kenya, including
targeted programs and expansion of HIV
testing.
• Next step: application of dual incidence
assay algorithm in KAIS (BED and Axsym
Avidity Index assay). Results will be
compared to case-based exclusion
method.
Extra slides
Distribution of recent infection among
married/cohabitating persons by gender and age group
KAIS 2007 (N=86)
60
51.3
45.7
Percent (%)
50
Males (N=30)
Females (N=56)
40
30
31.5
22.9
20
15.8
14.2
10
0
4
2.6
15-24
25-34
35-44
Age group
45-54
9
2.9
55+
Документ
Категория
Презентации
Просмотров
5
Размер файла
2 340 Кб
Теги
1/--страниц
Пожаловаться на содержимое документа