1. Challenges for Blood Donor Confirmatory Testing Algorithms
Susan L. Stramer, Ph.D.
American Red Cross, for
HIV Diagnostics:
New Developments and Challenges Meeting,
Orlando, FL; Feb 28-March 1, 2005

2. Background
HIV-1 algorithms utilize immunoassays for confirmation of repeat reactive blood donors
Most commonly used: Western Blot (WB)
Contains electrophoresed whole HIV lysate
HIV-1 Immunofluorescence (also uses viral lysate)
Not widely used in blood centers due to difficulty in interpretation
HIV-2 confirmation also required for HIV-1 WB ind/neg samples
Licensed HIV-2 EIA => Research based HIV-2 WB (CA)
6 or fewer HIV-2 confirmed pos blood donors in the US reported since 1992
No confirmation for HIV-1 type O

3. Background Issues with HIV confirmatory assays Poor performance
Unreadable, uninterpretable or invalid results generated
Misclassification of donors
False positive and false negative results
High rates of indeterminate results (neither pos or neg) in healthy individuals
Donor deferral, anxiety, and use of complicated donor reentry algorithms, when available
Mixed message, “You are healthy and not HIV infected, but you cannot donate blood”

4. Background Issues with HIV confirmatory assays, cont’d
High costs
Inconsistent availability
No new products for “antibody confirmation” on the market for HIV-1 or HIV-2
Screening tests ongoing improvements since 1985
Costs associated with new products extremely high
Can alternate algorithms be validated?
HIV RNA (NAT), use of high/low screening S/CO ratios, dual EIA algorithm

5. Can Alternate Algorithms be Implemented
Kits will require labeling for intended use
21 CFR Subpart E
“You must further test each donation, including autologous donations, found to be reactive by a screening test …whenever a supplemental (additional, more specific) test has been approved for such use by FDA….”
FDA’s current thinking is that an antibody RR must be confirmed by an antibody-specific assay
Variances have been granted to exempt RNA pos samples from requiring WB
Minipool or single donor RNA neg samples still require WB

6. Reactivity with 50 HIV-1 Commercial Seroconversion Panels Genetic SystemsTM HIV-1/HIV-2 plus O EIA May 2004, Package Insert (FDA License 1109)
HIV-1/HIV-2 plus O
Equivalent
> Sensitive
< Sensitive
vs Licensed Kit #1 (EIA)
12/46*
(26%)
34/46*
(74%)
(0%)
vs Licensed Kit #2 (EIA)
35/50
(70%)
9/50
(18%)
6/50
(12%)
vs Licensed Western Blot
13/50
37/50
* 4 panels not tested (no longer available)

7. Cambridge Biotech Human Immunodeficiency Virus Type 1 (HIV -1) Western Blot Kit
March 16, 2000 BPAC
Any bands present but pattern does not meet criteria for POSITIVE = INDETERMINATE
Non-viral bands have been observed with certain specimens. These bands are not usually accompanied by any of the major viral bands of diagnostic significance (p24, gp41/120/160). The non-viral bands appear to be cell related with the most common in the molecular weight range of 70K, 51-55K (possible HLA-DR) and 43K (possible HLA-ABC).

8. HIV-1 Blood Donor Screening/Supplemental Test Results (ARC); March 16, 2000 BPAC
12.4 Million Donations Screened
11,080 Repeat Reactive (0.09%) HIV-1 WB (Cambridge Biotech)
Positive
791 (7.1%)
Indeterminate
5,161 (46.6%)
Negative
5,128 (46.3%)

9. HIV-1 Blood Donor Screening/Supplemental Test Results (ARC) (Cont
HIV-1 Blood Donor Screening/Supplemental Test Results (ARC) (Cont.); March 16, 2000 BPAC
5,161 Indeterminate (46.6%)
Multiple Viral Bands
(p24, gp41, gp120/160,
but +/- intensity)
46 (1%)
One
Viral
Band
Multiple
Viral
Bands
Background Only
(obscures
reading)
Non-
Viral Bands
2,752 (53.3%)
589 (11.4%)
724 (14%)
1,050 (20.3%)
1,925 (70%) p24 “GAG” only
464 (79%) “GAG” reactivity with or without other banding
569 (79%) “p70”
16 (35%) “ENV” only
65.7% Viral Banding

10. Calypte Biotech HIV-1 Western Blot
gp160
gp120
p66
p55/51
gp41
p31
p24
p17

11. Calypte HIV-1 Western Blot
gp160
gp120
p66
p55/51
gp41
p31
p24
p17

12. HIV-1/HIV-2 EIA Repeat Reactives (ARC testing of 25
HIV-1/HIV-2 EIA Repeat Reactives (ARC testing of 25.6 million donations from 9/8/99 to 6/30/03)
Perform WB N = 17,090
Pos
Ind
Neg
HIV NAT
HIV NAT
HIV NAT
Reactive Individually
NonRx in Pool or Individually
Reactive Individually
NonRx in Pool or Individually
Reactive Individually
NonRx in Pool or Individually
HIV Infected
HIV Infected? Viral load < NAT cutoff False Pos WB
HIV Not Infected
HIV Not Infected
4.4%
51%
44.2%
0.4%
757
HIV Infected? Early seroconversion
8,704
HIV Infected? Early seroconversion False Rx NAT
7,562 61
<0.1% 0% 6

13. HIV-1/HIV-2 EIA Repeat Reactives (ARC testing of 25
HIV-1/HIV-2 EIA Repeat Reactives (ARC testing of 25.6 million donations from 9/8/99 to 6/30/03)
Review HIV-1 NAT Results N = 17,090
Pool or Individual Unit NAT NonRx (16 dtns ® HIV-1/HCV TMA)
Individual Unit NAT Rx (dHIV TMA)
96% 4%
16,327
763
Perform WB
No Further Testing HIV Infected
Pos
Ind
Neg
99.2% PPV
0.4%
53.3%
46.3%
757 WB Pos 61
8,704
7,562
92.5% Sensitivity

14. Correlation of NAT with Supplemental HIV Serological Data (17,791 RR donations of which 17,090 (96.1%) had EIA and WB data)
9/8/99 to 6/30/03
Western Blot Result
NAT Result
Pos
Ind
Neg
Total
dHIV Rx
NonRx
Total
757 (99.2%)
61 (0.4%)
818 (4.8%) 6
8,704
8,710
(0.8%)
7,562
763
16,327
17,090
Of HIV WB neg or ind, 6/16,272 (1:2712 or 0.037%) infected with HIV

15. Characteristics of HIV-1 WB Pos/TMA Nonreactive Samples
9/8/99 – 8/31/00
Sample Pool Neat HIV-1/HIV-2 S/CO WB HIV PCR
1 NR , 120, 160 Neg
2 NR , 160 Neg
3 NR , 160 Neg
4 NR , 55, 160 Neg
5 NR , 41, 51, 61, 160 Neg
6 NR all bands Neg
7 NR , 41, 160 Neg
8 NR , 41, 120, 160 Neg
9 NR , 160 Neg
10 NR , 41, 160 Neg
11 NR , 24, 41, 51, 160 Neg
12 NR all bands Pos (200 copies/mL)
13 NR all bands Neg
All samples p24 Ag negative; lack of HIV infection in those weakly EIA Rx with follow up

16. Characteristics of HIV-1 WB Pos/TMA Nonreactive Samples
9/1/00 – 12/29/01
Sample Pool Neat HIV-1/HIV-2 S/CO WB HIV PCR
14 NR , 24, 160 Neg
15 NR , 55, 120, 160 Neg
16 NR all bands Neg
17 NR all bands Pos (200 copies/mL)
18 NR , 66, 120, 160 Neg
19 NR , 160 Neg
20 NR , 41, 160 Neg
21 NR , 41, 160 Neg
22 NR , 160 Neg
23 NR all bands Neg
24 NR all bands Neg
All samples p24 Ag negative; lack of HIV infection in those weakly EIA Rx with follow up

17. Relationship between HIV-1/HIV-2 EIA and Western Blot NAT Reactive Samples 9/8/99 to 6/30/03
N = 763 (4% of total)
Neg N = 0
Ind N = 6 0.8%
Pos N = %
0.00
5.00
10.00
15.00
20.00
25.00
19.298
91.5% of WB pos/
NAT rxs had an
S/CO ³ 15
18.182
17.837
16.373
EIA S/CO
11.208
5.683
3.632 –
8.349 –
95% Range
Western Blot Results

18. Relationship between HIV-1/HIV-2 EIA and Western Blot NAT Nonreactive Samples 9/8/99 to 6/30/03
N = 16,327 (96% of total)
Neg N = 7, %
Ind N = 8, %
Pos N = %
0.00
5.00
10.00
15.00
20.00
25.00
98.5% of WB neg/ind
NAT nonrxs had an S/CO < 15
16.282
EIA S/CO
2.600
2.667
2.496
1.602
1.598
1.401
1.225
1.230
1.018 –
1.018 –
1.015 –
95% Range
Western Blot Results

19. Relationship between HIV-1/HIV-2 EIA and Western Blot NAT Nonreactive Samples 9/8/99 to 6/30/03
N = 16,327 (96% of total)
Neg N = 7, %
Ind N = 8, %
Pos, no p31 N = 39
0.2%
Pos, with p31 N = %
0.00
5.00
10.00
15.00
20.00
25.00
18.867
98.5% of WB neg/ind
NAT nonrxs had an S/CO < 15
16.675
16.137
EIA S/CO
2.600
2.667
2.390
1.602
1.598
1.523
1.225
1.230
1.228
1.018 –
1.018 –
1.000 –
6.023 –
95% Range
Western Blot Results

20. Dual EIA Algorithm Feasibility based on the concept:
If two assays with comparable sensitivity are composed of differing rare reagents and have a different format, the false positive populations should have limited cross over; the more unique the tests, the greater the separation of false positive populations
Used successfully for HTLV to eliminate >60% of repeat reactives requiring further testing by expensive, complicated, error prone and unavailable/unlicensed tests

21. HIV Dual EIA Algorithm Qualification
Qualified in both directions based on the two FDA licensed HIV-1/HIV-2 EIAs:
Abbott (EIA-1) ® Genetic Systems (EIA-2) (ARC)
Genetic Systems (EIA-1) ® Abbott (EIA-2) (BSL)
All HIV EIA repeat reactive samples from 1/1/00-3/31/02 having adequate volume for additional testing were evaluated (N=7884); all allogeneic donors
Western Blot and NAT (TMA pools of 16) test-of-record data were used for analysis
All 2nd EIA testing was performed centrally at Blood Systems Laboratories (BSL)

22. Application of HIV NAT/Dual EIA Algorithm in Routine Operations
HIV Repeat Reactives (BSL + ARC; 1/1/00 – 3/31/02) 1, ,227 (7,884 RRs)
Pool or Individual Unit NAT NonRx (16 dtns ® HIV-1/HCV TMA)
1, ,989 (7,567)
Individual Unit NAT Rx (dHIV TMA)
** (317)
Perform 2nd EIA
No Further Testing HIV Infected (4%)
EIA NonRx
1, ,948* (7,445)
EIA RR
(122)
* Contains 16 false pos WB results
** Of 317: 316 WB pos, EIA concordant Rx; 1 WB ind, EIA concordant Rx
No Further Testing HIV Not Infected (94.5%)
Perform WB (1.5%)
Pos
1 +13 (14)
Ind
(46)
Neg
(62)

23. N = 13, Abbott RR, WB Pos, GSC RR, NAT NR
WB Banding Pattern
Abbott S/CO
GSC S/CO 1
24, 31, 41, 51, 66, 120, 160
17.43
5.80 2
17, 24, 120, 160
15.52
2.91 3
all bands
15.93
9.30 4
17.19
8.62 5
19.30
8.87 6
17, 24, 61, 160
2.40
7.06 7
20.95
4.39 8
24, 41, 120, 160
1.40
1.33 9
8.07 10
41, 66, 120, 160
1.04
9.38 11
17.60
8.59 12
17.84
8.93 13
18.80
11.12

24. N = 16, Abbott RR, WB Pos, GSC NR, NAT NR
WB Banding Pattern
Abbott S/CO
GSC S/CO 1
17, 41, 160
1.53
0.17 2
1.50
0.20 3
41, 160
1.34
0.30 4
41, 120, 160
4.90
0.25 5
4.72
0.31 6
24, 120
1.52
0.29 7
4.92 8
24, 41, 160
1.00
0.38 9
1.19 10
24, 160
1.43
0.22 11
1.60
0.32 12
17, 24, 41, 120, 160
2.36
0.33 13
17, 41, 120, 160
1.11 14
24, 55, 120, 160
10.56
0.57* 15
17, 24, 160
8.73
0.62* 16
1.37
0.28
*Both samples PCR and repeat TMA (undilute) NR; one of two reported participation in an HIV vaccine trial;
Repeat undilute PCR and follow up on others demonstrate no HIV infection

25. HIV Dual EIA Algorithm Findings/Conclusions
Feasibility of NAT combined with the dual EIA algorithm for HIV confirmation was demonstrated
Sensitivity = 100% (317/317); 95% CI %
Specificity = 98.4% (7,445/7,567); 95% CI %
No. WBs eliminated = 98.5% (7,762/7,884)
No. indeterminate interpretations eliminated = 98.6% (3,388/3,434)
Majority of WB false pos, selected by the use of one particular EIA, were eliminated
Changes in EIAs to more sensitive/specific versions should not require extensive validations
WB shown to be of little if any value!!!

26. Application of HIV NAT/Dual EIA Algorithm in Routine Operations AABB Proposal
HIV Repeat Reactives
Pool or Individual Unit NAT NonRx (16 dtns ® HIV-1/HCV TMA)
Individual Unit NAT Rx (dHIV TMA)
Perform 2nd EIA and Individual Unit NAT
No Further Testing HIV Infected
EIA NonRx
and NAT Nonrx
EIA RR
Since most testing performed
in pools, additional individual
unit NAT significant
(at max 7,445/7,567)
No Further Testing HIV Not Infected
Perform WB
Pos
Ind
Neg