|
|
|
Although much of this
conference has focused on testing for antiviral resistance, the main problem
confronting many less industrialized countries continues to be diagnosis of HIV
with antibody tests. Great progress has been made in screening blood to prevent
transfusion-associated HIV transmission. However, in many countries, HIV
antibody testing is not widely available. The demand for antibody tests
continues to increase, for example, as perinatal prevention programs are
introduced and as more and more people perceive the value of voluntary
counseling and testing for HIV.
Cost of HIV testing
The cost of HIV antibody testing remains a
major consideration in prevention interventions. Perinatal prevention programs
provide a useful example. Several cost-effectiveness analyses illustrate this
point. The first analysis examined costs related to implementing short-course
zidovudine in a sub-Saharan African country where HIV prevalence is 12.5
percent.1 For every 100,000
pregnant women, HIV counseling and testing would cost $1,540,000--nearly three
times the cost of the zidovudine ($580,000) for HIV-infected women.
A second analysis estimated the costs of
perinatal prevention programs in which the mother and the infant received a
single dose of nevirapine.2 Costs
were calculated for 20,000 women in a country where HIV prevalence is 15
percent. Providing nevirapine to all pregnant women (regardless of HIV status)
would cost $83,333--33 percent less than a targeted program in which therapy
would be administered only to women found to be infected after HIV counseling
and testing. Of course, neither of these analyses reflects the collateral
benefits of testing so that people will learn their HIV status, but both
highlight the effect of the cost and availability of HIV testing for
implementing programs known to prevent infection.
Recommended HIV testing strategies
The World Health Organization (WHO) has developed several strategies to facilitate HIV diagnosis in less industrialized countries by reducing reliance on the Western blot. The choice of testing strategy, in which combinations of screening assays are used, depends on the testing objectives, the sensitivity and specificity of the tests, and the prevalence of HIV infection among the persons to be tested (Table 1).3
|
Strategy 1 uses a single sensitive screening assay: a reactive test is considered HIV positive. This strategy is recommended primarily for blood screening and for surveillance in areas where prevalence is high. Strategy 2 uses two screening assays. If the initial test is reactive, testing is repeated with the second assay. A specimen is considered positive only when both assays are reactive. This strategy is recommended for surveillance when HIV prevalence is 10 percent or less, for the diagnosis of HIV infection in symptomatic persons when prevalence is 30 percent or less, and for diagnosis of HIV infection in asymptomatic persons when prevalence is more than 10 percent. The third testing strategy requires the use of three assays. A specimen is considered HIV-positive only when all three assays are reactive. Strategy 3 is recommended for diagnosing HIV infection in asymptomatic persons when HIV prevalence is 10 percent or less.
To assist the public sector in selecting
appropriate screening assays to implement these strategies, the WHO Programme on
Health Technology has assessed the operational characteristics of many HIV tests
available for bulk purchase (see http://www.who.int.pht).
In addition to this testing, the WHO recommends that test combinations should
always be evaluated in the context in which they will be used before they are
implemented on a wide scale for screening and diagnosis.
Simplifying HIV diagnosis
Traditionally, the development of
high-technology HIV testing was driven by the high-volume needs of blood
screening programs. Most assays were technically demanding, required complex
equipment, and were best suited to centralized laboratories. Such high
technology presents several problems for routine diagnostic testing and for
voluntary HIV counseling and testing. The availability of testing is often
limited by the high cost of equipment and the need for technical expertise.
Because of logistics and specimen transport, centralized testing was not timely.
Ultimately, many persons never received their test results because of problems
with transportation or communication.
Subsequently, rapid HIV assays were developed that are simple to perform and thus facilitate point-of-service testing. Minimal equipment requirements and straightforward interpretation made many of these tests suitable for low-volume small laboratories and made it possible to offer timely HIV testing and to provide test results the same day. The field use of these tests still presents several challenges, such as maintaining the cold chain for sensitive reagents, and minor technical requirements such as separating serum from whole blood samples or accurately interpreting agglutination tests. In addition, shelf life and the stability of different tests with different lot sizes need to be monitored.
 |
 |
 |
In the more recent capillary flow (so-called
dipstick) devices, HIV antigen is conjugated to colloidal gold or selenium (Figure
3.) A drop of serum or whole blood is added to an absorbent pad, which
contains all necessary reagents and conjugate. The sample then flows along the
test strip. If HIV antibody is present, a color reaction develops at the line of
impregnated HIV-antigen. A second procedural control line develops to verify
that flow has been satisfactory and that the test has been performed correctly.
Results can be read within 10 to 15 minutes. A single line of color at the
control location indicates a negative test result; two lines (at the HIV antigen
and the control site) indicate a reactive test.
Selecting the best diagnostic test: two field
study examples
For less industrialized countries, selecting
specific tests from the large number of available candidates poses additional
challenges. Without an established local evaluating body to evaluate tests,
decisions about which tests to use are often based either on the preferences of
specific donor agencies or on price. Optimally, test selection should be based
on local evaluation of performance, including sensitivity for HIV subtypes known
to be present in the country. This can usually be accomplished expeditiously
even with limited resources.
Several studies illustrate how this has been done. In a field evaluation conducted by Stetler and colleagues from the CDC, in cooperation with the Ministry of Health in Honduras,4 seven rapid assays were selected for retrospective masked testing of 600 specimens stored in the national reference laboratory (Table 2). From the results of this first phase, five tests, selected for sensitivity, specificity, and ease of performance, underwent prospective evaluation with 900 additional specimens at three regional testing centers. Three of these five tests then progressed to phase 3 of the study, in which the tests were evaluated with 1255 specimens in the small rural hospitals and clinics where the tests were likely to be used.
| Phase 1 (n = 600) | Phase 2 (n = 900) | Phase 3 (n = 1255) | ||||
|---|---|---|---|---|---|---|
| Tests | Sensitivity | Specificity | Sensitivity | Specificity | Sensitivity | Specificity |
| Retrocell | 100 | 98.6 | 99.7 | 99.3 | 100 | 99.8 |
| HIVCHEK | 99.7 | 100 | 100 | 99.3 | 100 | 99.9 |
| Genie | 100 | 100 | 99.3 | 99.0 | 100 | 100 |
| Testpack | 100 | 98.9 | 100 | 99.7 | -- | -- |
| SUDS | 99.3 | 96.3 | 98.7 | 99.8 | -- | -- |
| Serodia | 100 | 92.8 | -- | -- | -- | -- |
| RTD | 100 | 97.5 | -- | -- | -- | -- |
Positive and negative predictive values were
calculated for the tests used in combination (Table 3).
From this evaluation Honduras selected two assays for the country's routine
testing algorithm--Retrocell for HIV screening, and Genie (now called Multispot)
for confirmation. Costs were calculated for the reagents used in the third phase
of testing. The Retrocell-Multispot combination cost $2531, a savings of $4473
(64 percent) compared with EIA and Western blot testing of the same specimens.
| Low prevalence (1.5%) (n =8 57) | High prevalence (30.5%) (n = 402) | |||
|---|---|---|---|---|
| Test Combination | PPV (%) | NPV (%) | PPV (%) | NPV (%) |
| Retrocell+HIVCHEK | 100 | 100 | 100 | 99.6 |
| Retrocell+Genie | 100 | 100 | 100 | 99.6 |
| HIVCHECK+Genie | 100 | 100 | 100 | 99.3 |
| NPV = negative predictive value; PPV = positive predictive value | ||||
| Test | Sensitivity | Specificity |
|---|---|---|
| Sero-Strip HIV-1/2 | 61/62 (98.4%) | 186/186 (100%) |
| SeroCard HIV | 62/62 (100%) | 185/186 (99.5%) |
| Capillus HIV-1/2 | 62/62 (100%) | 181/186 (97.3%) |
A similar evaluation process was conducted with
tests proposed for voluntary counseling and testing programs in Uganda.5
After standard, centralized testing, 25 percent of the persons tested during
1990-1996 never received their test results. To implement on-site testing, staff
at the AIDS Information Centre in Kampala evaluated three candidate tests for
five working days. Results of the evaluation (Table 4) led
to the selection of Capillus as the screening assay. Reactive specimens were
then retested by using SeroCard. Multispot was selected as a "tie
breaker" for specimens that were reactive on the first test but negative on
the second.
In 1997, when rapid testing was expanded to several counseling and testing centers, 35,000 persons were tested (Figure 4). Approximately 8000 specimens (22 percent) were reactive on the initial rapid test, 90 percent of which were confirmed to be HIV-positive after the subsequent rapid tests were performed.6 All these persons were able to receive their test results the same day, without the need for a return visit to the testing site. For quality assurance 5 percent of the specimens tested by the rapid test algorithm were submitted for retesting by standard EIA and Western blot at the Nakasera Blood Bank. Fewer than 0.5 percent of these tests did not correspond with the results obtained on site with the rapid algorithm.
 |
Keys to successful local HIV diagnosis
Accurate and timely HIV diagnosis remains a major challenge for many less industrialized countries. However, numerous technologies now make it possible to perform prompt, accurate, and affordable HIV testing in many settings that lack complex laboratory equipment or technical sophistication. Studies have demonstrated that the combination of several screening assays can achieve diagnostic accuracy comparable to that of the widely used EIA and Western blot algorithm.
Choosing the specific tests should be based on an
in-country evaluation with specimens from the population that will be tested, in
the laboratories where the tests will be used. This requires several steps.
Staff in local reference labs must develop the skill and confidence necessary to
establish testing panels in which local specimens are used, and they must
evaluate several tests against the same panel. This step is followed by a field
assessment at representative testing sites to validate that specific
combinations of tests are both accurate and practical. Finally, continued
quality assurance is essential. Preferably, 5 percent of samples from testing
sites should be sent for masked retesting at the reference laboratory. In
addition, expiration dates and storage requirements of the testing products
require painstaking attention, to ensure that accuracy is not compromised.
References
1. Mansergh
G, Haddix AC, Steketee RW, et al. Cost-effectiveness of short-course zidovudine
to prevent perinatal HIV type 1 infection in a sub-Saharan African developing
country setting. JAMA 1996;276:139-145.
2. Marseille E, Kahn JG, Mmiro F, et al. Cost effectiveness of single-dose nevirapine regimen for mothers and babies to decrease vertical HIV-1 transmission in sub-Saharan Africa. Lancet 1999;354:803-809
3. Joint United Nations Programme on HIV/AIDS (UNAIDS)--World Health Organization. Revised recommendations for the selection and use of HIV antibody tests. Wkly Epidemiol Rec 1997; 72: 81-88.
4. Stetler HC, Granade TC, Nunez CA, et al. Field evaluation of rapid HIV serologic tests for screening and confirming HIV-1 infection in Honduras. AIDS 1997; 11: 369-75.
5. Downing RG, Otten RA, Manun E. Optimizing the delivery of HIV counseling and testing services: The Uganda experience using rapid HIV antibody test algorithms. J Acquir Immune Defic Syndr 1998; 18: 384-8.
6. Mujurizi T, Alwano-Edyegu MG, Biryahwaho B, et al. Performance of a rapid HIV testing algorithm for same day results at the AIDS Information Centre, Uganda. Presented at: 12th World AIDS Conference; June 28-July 3, 1998; Geneva . Abstract 43109.
Correspondence
Bernard M. Branson, MD
Division of HIV/AIDS Prevention Centers for Disease Control and Prevention
1600 Clifton Road, Mailstop E-46
Atlanta, GA 30333 USA
Email: Bbranson@cdc.gov
© 2000 Medical Advocates
for Social Justice
Email: info@medadvocates.org
