NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

WHO Recommendations on the Diagnosis of HIV Infection in Infants and Children. Geneva: World Health Organization; 2010.

Cover of WHO Recommendations on the Diagnosis of HIV Infection in Infants and Children

WHO Recommendations on the Diagnosis of HIV Infection in Infants and Children.

Show details

ANNEX 2GRADE profiles for clinical algorithms

A2.1. Methodology

Based on the new WHO paediatric ART guidelines, all HIV-infected children should be initiated on ART during the first year of life and should, therefore, be identified as early as possible. Since virological tests are still unaffordable and/or unavailable in many settings, it was necessary to investigate:

  • How well HIV-exposed infants (needing special follow up) can be identified by using antibody tests
  • How well HIV-infected infants (needing ART in the first year of life) can be identified by using clinical algorithms (possibly combined with a rapid test)
  • How well HIV-uninfected infants or HIV-exposed infants who have seroreverted (not needing special follow up) can be identified by using rapid tests
  • What would be the best age to conduct such screening?

These questions were reframed according to the PICOT framework, a format and structure for developing research questions (1).

  1. Is the accuracy of applying a clinical algorithm (with or without performing a rapid test) to identify HIV infection in a child younger than 18 months comparable with that of the gold standard infant test (PCR or retrospectively determined serostatus)?
  2. Is the accuracy of one rapid test performed to exclude HIV infection in an infant less than 18 months comparable with that of the gold standard infant test (PCR or retrospectively determined serostatus)?
  3. Is the accuracy of one rapid test performed to identify HIV-exposure in certain1 young infants comparable with that of the gold standard?2
  4. What would be the best age to conduct such screening (also taking into account information about seroreversion in untreated infants)?

A review of the literature was performed, using the following sites and search engines to identify suitable literature: PubMed, Bandolier, the Journal of Family Practice, BMJ Clinical Evidence, The Cochrane Library, SUM search, National Guideline Clearinghouse, Institute for Clinical Systems Improvement and Medadvocates. Various combinations of the following terms were used: serological, rapid1, antibody, test, algorithm, diagnosis1, seroreversion, AIDS or SIDA, HIV or VIH and children, enfants, pediatr1, infants. A total of 2096 citations (published before 10 October 2008) were screened. Subsequently, hand searches of related and referenced articles were done. There were no restrictions on language, age of the children or date of publication. Abstracts and reports were included only if they were referenced.

All identified studies reporting on the accuracy of clinical algorithms or serological tests in infants and children or allowing the calculation of some measure of accuracy, as well as studies providing information on seroreversion in infants and children, were listed and described.

For the pooled analysis, the figures from the 2×2 contingency tables of each individual study were entered into Excel spreadsheets. The information about seroreversion was also entered in spreadsheets to facilitate further calculations. Studies with insufficient or unclear information were excluded from this analysis, as well as studies with incorrect comparisons due to very poor methodology (e.g. a comparison of the performance of a clinical algorithm against children identified by the CDC definition of AIDS) and studies that had remained unpublished for more than three years. In studies where the performance of different algorithms was evaluated in the same study population, only the data from the groups with the best accuracy were used for further analyses. In studies where the performance of different tests was evaluated in the same study population, only the data for the groups including the highest number of HIV-infected infants were used for further analyses.

Certain studies were pooled to compare and evaluate the accuracy of a diagnostic tool in a larger group of infants. GRADE PRO analysis was prepared for a selected number of case scenarios, based on the most interesting statistics from the pooled analysis (23). Information from the Newcastle–Ottawa quality assessment scale for cohort studies and the Standards for the Reporting of Diagnostic accuracy studies (STARD) initiative were kept in mind during evaluation of the quality of evidence (3).

A2.2. GRADE PRO analyses of studies on clinical algorithms

Key outcomes to be considered for recommendations include:

  • Factors related to the performance of the tool:
    Sensitivity, specificity, PPV, NPV, LR+, LR− and interobserver agreement
  • Factors related to possible task-shifting from paediatricians to non-paediatricians, mainly PHCWs:
    Competence/quality assurance/agreement in conclusion between paediatricians and non-paediatricians, acceptability of clinical algorithms (among PHCWs and the population), PHCWs' self-confidence in applying an algorithm and time needed to apply the algorithm and the cost of applying it.

A2.3. Review of the literature on clinical algorithms for infants and children

Table A–19 summarizes the 22 studies that assessed the accuracy of a clinical algorithm. While a variety of clinical algorithms was assessed, most studies reported on the accuracy of a WHO clinical case definition or a modified version of it.

Table A–19. List of studies on clinical algorithms in infants and children.

Table A–19

List of studies on clinical algorithms in infants and children.

All these observational studies, except one, were performed in African settings (with HIV-1 subtypes A and C predominantly). Other subtypes (B or HIV-2), as well as differences in environmental factors, have been associated with different patterns of disease progression, possibly resulting in better outcomes of clinical algorithms.

Most of the studies (14/22) selected only inpatients and some included only very ill infants (e.g. those admitted to the intensive care unit). For 17/22 studies, the algorithm was applied by paediatricians or study physicians of tertiary-level hospitals, rather than by primary and secondary health-care workers among outpatients, such as children under five years of age attending clinics.

While the study population included children from birth to the age of 18 years, all studies included infants younger than 18 months but the age distribution, which is closely related to the HIV prevalence rate, however, was most often not described. Three studies included only infants younger than 18 months.

The HIV-prevalence rates among the study populations ranged between 4% and 84%. The studies included different percentages of young infants, used different inclusion criteria, and were performed in different countries and in different decades. Only 8/22 studies used a gold standard that is currently an acceptable gold standard for infants younger than 18 months.

The reported sensitivities ranged between 9% and 89%, specificities between 42% and 99%, and PPVs between 3% and 95%.

Summary of search findings

  • 22 observational studies
  • All but one in African settings, 1981–2005
  • Investigator: paediatrician at tertiary level (17/22)
  • Study population: mainly inpatients (14/22)
  • Age group: 0–18 years (percentage of infants unclear), 3/22 in infants only
  • HIV prevalence: 4–84%
  • Not always acceptable gold standard (8/22)
  • Variety of clinical algorithms
  • Not equal stratifications
  • Wide variation in reported test performance:
  • Sensitivity: 9–89%, specificity: 42–99%, PPV: 3–95%

Table A–20 provides the GRADE profile for clinical algorithms without antibody testing. This was based on the pooled information of 15 observational studies performed in children up to 18 years of age.

Table A–20. GRADE profiles for clinical algorithms without antibody testing.

Table A–20

GRADE profiles for clinical algorithms without antibody testing. Question: Should a clinical algorithm (without rapid testing) versus a gold standard infant test be used for the identification of HIV infection in infants <12 months of age? Settings: (more...)

Table A–21 provides the GRADE profile for the evaluation of clinical symptoms in seropositive infants 6 months of age. This was based on the pooled information of two observational studies (26, 27). This profile is used as a proxy for ‘the use of an algorithm after performing antibody testing’.

Table A–21. GRADE profile for clinical algorithms with antibody testing.

Table A–21

GRADE profile for clinical algorithms with antibody testing. Question: Should a clinical algorithm with rapid testing versus a gold standard infant test be used to identify HIV infection in infants <12 months of age? Settings: In resource-limited (more...)

The evidence for possible task-shifting was derived from one observational study that provided information on the interobserver agreement, agreement between the conclusions made by paediatricians and those by non-paediatricians (even though this included highly qualified IMCI trainers, namely, one professional nurse and one general practitioner, rather than PHCWs); as well as the difference in some indicators of test performance.

A2.4. GRADE PRO analyses of studies on antibody testing

Key outcomes to be considered for recommendations include:

  • Factors related to the performance of the tool:
    Sensitivity, specificity, PPV, NPV, LR+, LR− and interobserver agreement
  • Factors related to possible task-shifting from laboratory to non-laboratory personnel (mainly PHCWs):
    Competence/quality assurance/agreement between rapid test results reported by non-laboratory and those by laboratory personnel, competence in phlebotomy, PHCWs' self-confidence in phlebotomy and performing rapid tests, acceptability of phlebotomy and rapid testing in infants among PHCWs and the population, costs and time needed for phlebotomy and performing rapid tests, and occupational risks.

Summary of search findings

  • 12 observational studies (1989–2007) on the use of serological tests in infants and children, studies performed for different purposes
  • Investigators: laboratory personnel
  • Study population: Children 0–13 years, variable HIV prevalence rates
  • Not always compared with acceptable gold standard. Mixed range of serological assays used, not all currently accepted and available in the market
  • Samples: fresh/frozen samples/DBS – oral fluid/plasma
  • Reported assay performance differed by type of rapid test and age group.

Table A–22 provides the GRADE profile for the identification of HIV exposure in newborns using one rapid test (Determine).

Table A–22. GRADE profile for the identification of HIV exposure in newborns.

Table A–22

GRADE profile for the identification of HIV exposure in newborns. Question: Should one rapid test in ‘certain’ newborns versus gold standard test in the mothers be used for the identification of HIV exposure? Settings: Resource-limited (more...)

Table A–23 provides the GRADE profile for the exclusion of HIV infection using EIA in infants aged 12–18 months.

Table A–23. GRADE profile for the exclusion of HIV infection in infants aged 12–18 months.

Table A–23

GRADE profile for the exclusion of HIV infection in infants aged 12–18 months. Question: Should one rapid test at 12–18 months versus gold standard infant tests be used for the exclusion of HIV infection? Settings: Resource-limited settings (more...)

Table A–24 provides the GRADE profile for the performance of HIV antibody testing in infants aged 6–9 months compared with a gold standard infant diagnostic test. This was based on three observational studies.

Table A–24. GRADE profile for the exclusion of HIV infection in infants aged 6–9 months.

Table A–24

GRADE profile for the exclusion of HIV infection in infants aged 6–9 months. Question: Should one rapid test at 6–9 months versus gold standard infant test be used for the exclusion of HIV? Settings: Resource-limited settings without access (more...)

References

1.
Schardt C, et al. Utilization of the PICO framework to improve searching PubMed for clinical questions. BMC Medical Informatics and Decision Making. 2007;7:16. [PMC free article: PMC1904193] [PubMed: 17573961]
2.
GRADE Working Group. Grading quality of evidence and strength of recommendations. British Medical Journal. 2004;328:1490–1498. [PMC free article: PMC428525] [PubMed: 15205295]
3.
Bossuyt PM, et al. for the STARD Group. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. American Journal of Radiology. 2003;181:51–56. [PubMed: 12818829]
4.
Bahwere P, et al. Uptake of HIV testing and outcomes within a community-based therapeutic care (CTC) programme to treat severe acute malnutrition in Malawi: a descriptive study. BMC Infectious Diseases. 2008;8:106. [PMC free article: PMC2536666] [PubMed: 18671876]
5.
Jones SA, Sherman GG, Coovadia AH. Can clinical algorithms deliver an accurate diagnosis of HIV infection in infancy? Bulletin of the World Health Organization. 2005;83:559–560. [PMC free article: PMC2626291] [PubMed: 16175834]
6.
Thurstans S. The application of clinical algorithms as a tool for the identification of HIV symptomatic malnourished children in the NRUs in Malawi. Malawi: Action Against Hunger; 2004.
7.
Joubert G, Schoeman CJ, Bester CJ. Validation of a new clinical case definition for pediatric HIV infection, Bloemfontein, South Africa. Journal of Tropical Pediatrics. 2006;51:387. [PubMed: 15890723]
8.
Horwood C, et al. Diagnosis of paediatric HIV infection in a primary health care setting with a clinical algorithm. Bulletin of the World Health Organization. 2003;81:856–864. [PMC free article: PMC2572386] [PubMed: 14997238]
9.
van Gend CL, et al. Evaluation of the WHO clinical case definition for pediatric HIV infection in Bloemfontein, South Africa. Journal of Tropical Pediatrics. 2003;49:143–147. [PubMed: 12848202]
10.
Yeung S, et al. Paediatric HIV infection in a rural South African district hospital. Journal of Tropical Pediatrics. 2000;46:107–110. [PubMed: 10822937]
11.
Bakaki P, et al. Epidemiologic and clinical features of HIV-infected and HIV-uninfected Ugandan children younger than 18 months. Journal of Acquired Immune Deficiency Syndromes. 2001;28:35–42. [PubMed: 11579275]
12.
Atakouma DY, et al. Sida pédiatrique au CHU-Tokoin (Lomé): place de la malnutrition protéino-énergétique et essai d'élaboration d'un score de diagnostic clinique. Cahiers Santé 1997;7:397–404. [PubMed: 9503498]
13.
Agbèré AD, et al. Sensibilité, spécificité et valeur prédictive positive des critères de diagnostic clinique du SIDA pédiatrique de l'OMS au CHU Tokoin (Lomé, Togo) Archives of Pédiatrics. 1998;5:207–208. [PubMed: 10223150]
14.
Jeena PM, Coovadia HM, Chrystal V. Pneumocystis carinii and cytomegalovirus infections in severely ill, HIV-infected African infants. Annals of Tropical Paediatrics. 1996;16:361–368. [PubMed: 8985536]
15.
Chintu C, et al. Case definitions for paediatric AIDS: the Zambian experience. International Journal of STD & AIDS. 1993;4:83–85. [PubMed: 8476970]
16.
Kline MW, et al. Sensitivity, specificity and predictive value of physical examination, culture and other laboratory studies in the diagnosis during early infancy of vertically acquired human immunodeficiency virus infection. Pediatric Infectious Diseases Journal. 1993;12:33–36. [PubMed: 8417423]
17.
Otieno FA, Mbori-Ngacha DA, Wafula EM, Ndinya-Achola JO. Evaluation of a proposed clinical case definition of paediatric acquired immune deficiency syndrome. East African Medical Journal. 2002;79:111–114. [PubMed: 12389953]
18.
Vetter KM, et al. Projet RETRO-CI. Clinical spectrum of human immunodeficiency virus disease in children in a west African city. Pediatric Infectious Diseases Journal. 1996;15:438–442. [PubMed: 8724067]
19.
Msellati P, et al. Evaluation of the WHO clinical definition of pediatric AIDS in Kigali, Rwanda; 7th International Conference on AIDS; Florence, Italy. 16–21 June, 1991.
20.
European Collaborative Study. Children born to women with HIV-1 infection: natural history and risk of transmission. Lancet. 1991;337:253–260. [PubMed: 1671109]
21.
Lepage P, et al. Evaluation and simplification of the World Health Organization clinical case definition for paediatric AIDS. AIDS. 1989;3:221–225. [PubMed: 2500955]
22.
Colebunders R, et al. Evaluation of a clinical case definition of AIDS in African children. AIDS. 1987;1:151–153. [PubMed: 3126755]
23.
Harms G, et al. Evaluation of the WHO clinical case definition of AIDS in rural South-Rwanda; 7th International conference on AIDS in Africa; Yaoundé, Cameroon. 9–11 December, 1992.
24.
Müller O, et al. Pediatric HIV-1 disease in a Kampala Hospital. Journal of Tropical Pediatrics. 1990;36:283–286. [PubMed: 2280434]
25.
Jonckheer T, et al. AIDS case definitions for African children. Lancet. 1988;2:690. [PubMed: 2901552]
26.
Kline MW, et al. Sensitivity, specificity and predictive value of physical examination, culture and other laboratory studies in the diagnosis during early infancy of vertically acquired human immunodeficiency virus infection. Pediatric Infectious Diseases Journal. 1993;12:33–36. [PubMed: 8417423]
27.
European Collaborative Study. Children born to women with HIV-1 infection: natural history and risk of transmission. Lancet. 1991;337:253–260. [PubMed: 1671109]
28.
Horwood C, et al. Diagnosis of paediatric HIV infection in a primary health care setting with a clinical algorithm. Bulletin of the World Health Organization. 2003;81:858–866. [PMC free article: PMC2572386] [PubMed: 14997238]
29.
Kanal K, et al. Evaluation of the proficiency of trained non-laboratory health staffs and laboratory technicians using a rapid and simple HIV antibody test. AIDS Research and Therapy. 2005;2:5. [PMC free article: PMC1156864] [PubMed: 15907202]
30.
Tegbaru B, et al. Assessment of the implementation of HIV-rapid test kits at different levels of health institutions in Ethiopia. Ethiopian Medical Journal. 2007;45:293–99. [PubMed: 18330330]
31.
Granade TC, et al. Performance of the OraQuick and Hema-Strip rapid HIV antibody detection assays by non-laboratorians. Journal of Clinical Virology. 2004;30:229–232. [PubMed: 15135740]
32.
Ziyambi Z, Osewe P, Taruberekera N. Evaluation of the performance of non-laboratory staff in the use of simple rapid HIV antibody assays at the New Start voluntary counseling and testing (VCT) centers; XIV International AIDS Conference; Barcelona. 7–12 July 2002.
33.
Thurstans S, et al. HIV prevalence in severely malnourished children admitted to nutrition rehabilitation units in Malawi: geographical and seasonal variations a cross-sectional study. BMC Pediatrics. 2008;8:22. [PMC free article: PMC2408573] [PubMed: 18495024]
34.
Akpede GO, Lawal RS, Momoh SO. Perception of voluntary screening for paediatric HIV and response to post-test counselling by Nigerian parents. AIDS Care. 2002;14:683–697. [PubMed: 12419118]
35.
Gumodoka B, Favot I, Berege ZA, Dolmans WM. Occupational exposure to the risk of HIV infection among health care workers in Mwanza Region, United Republic of Tanzania. Bulletin of the World Health Organization. 1997;75:133–140. [PMC free article: PMC2486935] [PubMed: 9185365]
36.
Gounden YP, Moodley J. Exposure to human immunodeficiency virus among healthcare workers in South Africa. International Journal of Gynecology & Obstetrics. 2000;69:265–270. [PubMed: 10854870]
37.
De Baets AJ, Sifovo S, Pazavakavambwa IE. Access to occupational postexposure prophylaxis for primary health care workers in rural Africa: a cross-sectional study. American Journal of Infection Control. 2007;35:545–551. [PubMed: 17936147]
38.
Gulia J, et al. HIV seroreversion time in HIV-1-uninfected children born to HIV-1-infected mothers in Malawi. Journal of Acquired Immune Deficiency Syndromes. 2007;46:332–337. [PubMed: 17786126]
39.
Read JS. Diagnosis of HIV-1 infection in children younger than 18 months in the United States. Pediatrics. 2007;120:e1547–e1562. [PubMed: 18055670]
40.
De Baets AJ, et al. The unique features of pediatric HIV-1 in sub-Saharan Africa. Current HIV Research. 2008;6:351–362. [PubMed: 18691033]

The definition of ‘certain’ young infants includes infants of mothers with unknown serostatus and may include infants from ‘certain’ HIV-negative mothers. In high HIV-prevalence areas or high-risk groups, we may want to also include mothers who tested HIV negative less than 3 months prior to delivery, depending on the need for repeat testing.

The gold standard for the identification of HIV exposure should be the serostatus of the mother. Additional information from the test performance as compared with the infant gold standard diagnostic test will also be provided.

Footnotes

1

The definition of ‘certain’ young infants includes infants of mothers with unknown serostatus and may include infants from ‘certain’ HIV-negative mothers. In high HIV-prevalence areas or high-risk groups, we may want to also include mothers who tested HIV negative less than 3 months prior to delivery, depending on the need for repeat testing.

2

The gold standard for the identification of HIV exposure should be the serostatus of the mother. Additional information from the test performance as compared with the infant gold standard diagnostic test will also be provided.

Copyright © 2010, World Health Organization.

All rights reserved. Publications of the World Health Organization can be obtained from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: tni.ohw@sredrokoob). Requests for permission to reproduce or translate WHO publications – whether for sale or for noncommercial distribution – should be addressed to WHO Press, at the above address (fax: +41 22 791 4806; e-mail: tni.ohw@snoissimrep).

Bookshelf ID: NBK138548
PubReader format: click here to try

Views

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to pubmed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...