Performance of HIV-1 DNA or HIV-1 RNA Tests for Early Diagnosis of Perinatal HIV-1 Infection during Anti-Retroviral Prophylaxis

Marianne Burgard, MD1, St ephane Blanche, MD2,3, Carine Jasseron, MS4,5, Philippe Descamps, MD1, 7 Marie-Christine Allemon, MD , Nicole Ciraru-Vigneron, MD8, Corinne Floch, MD9, Brigitte Heller-Roussin, MD10, Eric Lachassinne, MD11, Fabienne Mazy, MD9, Josiane Warszawski, MD, PhD4,5,6, and Christine Rouzioux, PharmD, PhD1,3, on behalf of the Agence Nationale de Recherche sur le SIDA et les Hepatites virales French Perinatal Cohort* Objective To compare performance of testing for human immunodeciency virus (HIV)-1 DNA and HIV-1 RNA for diagnosis of HIV-1 infection in infants receiving preventive antiretroviral therapy. Study design This substudy of the French multicenter prospective cohort of neonates born to HIV-infected mothers, included 1567 infants tested for HIV with polymerase chain reaction (PCR) in a single laboratory, receiving post-natal prophylaxis, not breastfed, and having simultaneous HIV-1 DNA and RNA results before 45 days. The performance of PCR was assessed in reference to the 6-month HIV-1 RNA result. Results Specicity of both HIV-1 RNA and HIV-1 DNA PCR was 100% at all ages (except 99.8% for DNA at birth); sensitivity was 58% (RNA) and 55% (DNA) at birth, and 89% at 1 month, 100% at 3 months for both, and 100% at 6 months (DNA). Concordance between HIV-1 DNA and RNA results was 0.78 and 0.81 (Kappa) at birth and 1 month and 100% at 3 and 6 months. Type of maternal and neonatal prophylaxis had no effect on sensitivity, but inuenced viral load. Conclusion The performances of testing for HIV-1 DNA and RNA were similar with 100% sensitivity at 3 months. At 1 month during prophylaxis, 11% of infected children had negative PCR results. (J Pediatr 2012;160:60-6).
arly diagnosis of human immunodeciency virus (HIV)-1 infection in babies born to seropositive mothers is essential for preventing AIDS and death by allowing early initiation of appropriate antiretroviral therapy (ART).1,2 Recently, the World Health Organization (WHO) recommended systematic HIV-1 diagnosis for all exposed infants, at 4 to 6 weeks of age, with tests having a sensitivity of at least 95%, and initiation of triple-drug ART, immediately after the diagnosis of infection.3,4 It has been estimated that, in the absence of antiretroviral prophylaxis, 95% of HIV-1 infections could be detected by polymerase chain reaction (PCR) at 2 to 4 weeks of age, with the exception of those cases transmitted by breastfeeding.5-7 However, postnatal prophylaxis for at least 4 weeks is currently recommended universally by WHO,8 and the optimal age for HIV-1 diagnosis in such cases is less well documented. Detection of HIV-1 DNA in peripheral blood mononuclear cells (PBMC) and HIV-1 RNA in plasma both have been used to diagnose HIV-1 infection in neonates.9,10 Testing for HIV-1 RNA has been reported to have a better sensitivity than testing for HIV-1 DNA for infants <2 months old receiving zidovudine prophylaxis,11,12 although another study found  From the Laboratoire de Virologie, and Unite matologie et Rhumatologie dImmunologie, He no difference.13 ^ pital Necker, AP-HP, Paris, France; EA diatriques, Ho pe  Paris Descartes, Paris, France; Equipe 3620, Universite The goal of this study was to compare the diagnostic performance of PCR tests ^ pital VIH et IST, CESP, INSERM U1018; AP-HP, Ho ^tre, Service dEpidemiologie and Sante  Publique; Bice for HIV-1 DNA in PBMCs with that of PCR for HIV-1 RNA in plasma in a large  de Me decine Paris-Sud, Le Univ Paris-Sud, Faculte series of infants followed prospectively from birth as part of the Agence Nationale ^tre, France; Service de Pe diatrie, Ho ^ pital Kremlin-Bice Delafontaine, Saint-Denis, France; Service de de Recherche sur le SIDA et les Hepatites virales French Perinatal Cohort (EPFcologie-Obste trique, Ho ^ pital Lariboisie re, AP-HP, Gyne diatrie, Ho ^ pital LouisParis, France; Service de Pe CO1) and receiving prophylaxis for 4 to 6 weeks.14
1 2 3 4 5 6 7 8 9

Mourier, AP-HP, Colombes, France; 10Service de onatalogie, Ho ^ pital de Montreuil, Montreuil, France; Ne diatrie, Ho ^ pital Jean Verdier, AP-HP, and 11Service de Pe Bondy, France

ART EPF-CO1 HIV PBMC PCR WHO

Antiretroviral therapy French Perinatal Cohort Human immunodeciency virus Peripheral blood mononuclear cell Polymerase chain reaction World Health Organization

*A list of members of the Agence Nationale de Recherche sur le SIDA et les Hepatites virales (ANRS) French Perinatal Cohort is available at www.jpeds.com (Appendix). Supported by Agence Nationale de Recherche sur le patites virales, Paris, France. S.B., J.W., SIDA et les He and C.R. serve on the Scientic Committee of the ANRS CO1 Study. The authors declare no conicts of interest.
0022-3476/$ - see front matter. Copyright 2012 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2011.06.053

60

Vol. 160, No. 1  January 2012 insufcient plasma volume, the entirety of plasma available was used to be as close as possible to an ultrasensitive test. The limit of detection depended on the volume of plasma available (50-1000 mL), but never exceeded 1000 copies/ mL. The median detection limit for tests giving negative results for plasma samples collected from infected infants at birth was 50 copies/mL (IQR, 50-200), and at 1 month was 50 copies/mL (IQR, 50-400). PCR testing for HIV-1 DNA was performed with puried PBMCs until April 2005, and then on whole blood18 with 3 different methods used successively: rst, amplication of pol, long-terminal-repeat or gag gene sequences from PBMC lysate with radioactive Southern blotting as already described19; second, modied Roche Monitor 1.5 HIV-1 RNA kits with DNA internal standard (Roche Diagnostic Systems, Branchburg, New Jersey) as already described20; and third, since the year 2000, real-time PCR in the LTR region as already described.21,22 All these methods were optimized carefully so that their limit of detection was 5 copies per PCR (ie, 35 copies/106 PBMCs), with 8E5 cells for reference. One microgram of DNA from PBMCs was tested in each assay. Samples giving positive PCR test results before the year 2000 were re-tested with real-time PCR when possible to get quantitative results, as described22 (HIV-1 DNA quantication was available for 67% of HIV-1 DNA-positive samples). Statistical Analysis We rst evaluated the sensitivity, specicity and predictive values of PCR tests for HIV-1 DNA and of PCR tests for HIV1-RNA performed at birth (day 1 to day 7), 1 month (day 30 15), and 3 months (day 91 30), with PCR tests for HIV-1 RNA at 6 months (day 183 61) as the reference. We then estimated the Kappa concordance between tests for HIV-1 DNA and HIV-1 RNA for each period. Specicity, predictive values, and kappa value were estimated for the period 2002 to 2006 because PCR testing for HIV-1 RNA in plasma was not performed for HIV-1 diagnosis before 2002, except for children scoring positive with PCR for HIV-1 DNA. We evaluated sensitivity in the whole period of the study (1994-2006). Exact 95% CIs were calculated for sensitivity, specicity, and positive- and negative-predictive values. Sensitivity was compared according to the type of maternal antiretroviral therapy, the type of post-natal prophylaxis, and geographic origin. The c2 or 2-sided Fisher exact tests were used to compare categorical variables, and Student t test or the Wilcoxon test were used for continuous variables. Statistical analyses were performed with SAS, version 9.1 (SAS Institute, Cary, North Carolina).

Methods
We included in this study a subgroup of children born to HIV-infected women enrolled in the prospective multicenter Agence Nationale de Recherche sur le SIDA et les Hepatites virales French Perinatal Cohort (EPF-CO1), delivering on mainland France between 1994 and 2006. The study design of the EPF-CO1 has been described elsewhere.14 No specic recommendation for HIV treatment and obstetric care was made for women in the cohort, and investigators were expected to follow French national guidelines, as regularly updated.15 These guidelines recommend screening for HIV by PCR at birth and at ages 1, 3, and 6 months. All positive HIV PCR results have to be conrmed on a sequential sample. Informed consent was obtained from all mothers. This cohort study was approved ^pital Cochin (Comit by the ethics committee of Ho e de Protection des personnes) and the French computer database watchdog commission (Commission Nationale de lInformatique et des Libert es). The analysis included all children whose HIV testing was performed in the virology laboratory of Necker Hospital. This laboratory received samples from 30 clinical participating centers. All cases fullling these criteria were included: birth to an HIV-1 infected mother; post-natal prophylaxis; at least one blood sample taken before 45 days of age with PCR results for both HIV-1 DNA and HIV-1 RNA for the same sample; HIV-1 status established at 6 months on the basis of PCR testing for HIV-1 RNA at 6 months (day 183 61), the most widely used criterion to date16 or of death from unambiguous AIDS before 6 months; absence of breastfeeding. We excluded 4 infected infants receiving fully suppressive ART at 6 months and 3 additional infected children with horizontal, late post-natal infection as previously described.17 Virologic Methods HIV-1 diagnostic tests were performed prospectively, by testing for HIV-1 DNA in PBMCs on the basis of PCR detection of two different genes until 2001, and with PCR tests for both HIV-1 DNA in PBMCs and HIV-1 RNA in plasma from 2002 to 2006. PCR tests for plasma HIV-1 RNA were used prospectively and systematically for HIV-1 diagnosis in all infants since 2002, and in infected infants after diagnosis of infection since 1996. All earlier samples were retrospectively tested with PCR for HIV-1 RNA. Blood samples were obtained in tubes with venipuncture. Plasma HIV-1 RNA was assayed with Roche tests (Amplicor HIV-1 Monitor version 1 with add-in primers, version 1.5 or Cobas Taqman HIV-1 tests according to availability, Branchburg, New Jersey). High-speed centrifugation of plasma was used to pellet the virus and eliminate heparin from some plasma samples collected in heparinized tubes before the year 2000, and internal controls were used to verify the absence of inhibition. Ultrasensitive tests were used whenever possible. In case of

Results
A total of 1293 children fullled the inclusion criteria: 65 were infected (HIV-1 RNA positive at 6 months in 62, death from unambiguous AIDS before 6 months in 3), and 1228 were uninfected (HIV-1 RNA negative at 6 months; Table I).
61

THE JOURNAL OF PEDIATRICS

www.jpeds.com

Vol. 160, No. 1

Table I. Characteristics of the infants

Variable n Maternal characteristics Birth of the mother in sub-Saharan Africa Plasma HIV-1 viral load at delivery Median (IQR) log copies/mL Viral load <400 copies/mL CD4 count at delivery Median (IQR) CD4cells/mL CD4 count <200 cells/mL Infant characteristics Sex ratio (male/female) Age at rst blood sample sent to Necker laboratory 1-3 days 4-7 days 8-45 days Anti-retroviral prophylaxis Prophylaxis during pregnancy None Zidovudine Dual therapy Triple therapy Intrapartum prophylaxis None Intravenous zidovudine alone Single dose nevirapine + IV ZDV Single dose nevirapine alone or other Neonatal prophylaxis Zidovudine Dual therapy Triple therapy Infected infants % (n) 65 64.6 (42) n = 47 3.8 (2.9-4.5) 21.3 (10) n = 63 276 (157-482) 36.5 (23) 0.63 (25/40) 46.1 (30) 38.5 (25) 15.4 (10) n = 65 20.0 (13) 36.9 (24) 20.0 (13)* 23.1 (15) n = 64 14.1 (9) 81.3 (52) 4.7 (3) 0.0 (0) n = 65z 84.6 (55) 13.8 (9) 1.5 (1) Uninfected infants % (n) 1502 76.6 (1050) n = 1418 1.5 (1.4-2.4) 79.1 (1121) n = 1301 450 (310-620) 10.4 (136) 0.98 (735/750) 29.4 (441) 59.1 (888) 11.5 (173) n = 1485 2.1 (31) 8.4 (124) 9.4 (140) 80.1 (1190) n = 1485 5.3 (79) 92.3 (1370) 2.2 (32) 0.2 (4) n = 1502 93.7 (1407) 2.7 (41) 3.6 (54)

*Zidovudine + lamivudine for 92% and didanosine + stavudine for 8%. Two nucleoside reverse transcriptase inhibitors + a protease inhibitor for 93% and 3 nucleoside reverse transcriptase inhibitors for 7%. zOf the 56 infected infants tested at 1 month, 83.9% (47) received zidovudine, 14.3% (8) received zidovudine + lamivudine, and 1.8% (1) a triple therapy including protease inhibitor.

Sensitivity of HIV-1 RNA PCR and HIV-1 DNA PCR Of the 65 infected children born between 1994 and 2006, both methods were used for 55 at birth, for 56 at 1 month, for 38 at 3 months and for 12 at 6 months (Table II).

At 3 Months. At age 3 months, whereas, prophylaxis had

been stopped (at age 4-6 weeks) and multidrug curative ART had not been initiated, the sensitivity of both assays was 100%. Samples collected after the end of prophylaxis were available for all the 7 infected infants who had given one or more negative PCR test results at 1 month. These samples all tested positive for HIV-1 DNA and HIV-1 RNA, and the viral loads were high (range, 4.1-6.4 log HIV RNA copies/mL).

At Birth. The sensitivity of PCR tests for HIV-1 RNA in plasma and for HIV-1 DNA in PBMCs with samples collected up to age 7 days was 58% and 55%, respectively. The sensitivity of the combined use of the two methods (ie, diagnosing infection when one or both methods scored positive) was 62%. The performance of the two tests did not differ when considering age of sampling (before day 3 or from day 4 to 7). At 1 Month. At 1 month, during post-natal prophylaxis, PCR tests for plasma HIV-1 RNA, and PBMC HIV-1 DNA had the same sensitivity, 89%. The sensitivity of the combined use of the two methods was 91%. All blood samples giving false-negative PCR results had been taken between 26 days and 33 days of age. Four of these 6 negative HIV RNA PCR results had a limit of detection at a level of 50 copies/mL. Of the 30 infected infants with at least one positive PCR test at birth, 90% had a positive PCR result in both PCR tests at 1 month, and 97% had a positive result in one or both PCR tests. Of the 17 infected infants with negative PCR results at birth, 76% had positive results in both PCR tests at 1 month, and 24% gave negative results in both tests.
62

At 6 Months. The sensitivity of PCR testing for HIV-1 DNA at 6 months was 100% for the 12 infants who were not receiving multidrug curative ART. Sensitivity According to the Type of Maternal and Post-Natal Prophylaxis. Neither the presence nor type
of maternal antiretroviral therapy, nor post-natal prophylaxis were signicantly associated with positivity of PCR tests for HIV-1 RNA or DNA at birth or at 1 month. The sensitivity of PCR for RNA at birth was 64% when the mother did not receive antiretrovirals, 53% when she received one drug, 58% when she received two drugs, and 62% when she received $3 drugs (P = .9). The sensitivity of PCR for HIV-1 DNA for these same groups was: 45%, 42%, 75%, and 62% respectively (P = .3). At 1 month, both PCR tests were positive for 89% of infants receiving zidovudine and for 78% of those receiving intensied combination
Burgard et al

January 2012

ORIGINAL ARTICLES

Table II. Results of HIV-1 PCR according to age* in the 65 infected infants
Median age (IQR) Birth: day 1-7 Plasma HIV-1 RNA PBMC HIV-1 DNA Day 1-3 Plasma HIV-1 RNA PBMC HIV-1 DNA Day 4-7 Plasma HIV-1 RNA PBMC HIV-1 DNA 1 month: day 15-45 During any ARV Plasma HIV-1 RNA PBMC HIV-1 DNA In relation to infant ARV ZDV Plasma HIV-1 RNA PBMC HIV-1 DNA ARV Combination Plasma HIV-1 RNA PBMC HIV-1 DNA 3 months: Day 61-121zz Plasma HIV-1 RNA PBMC HIV-1 DNA 6 months: day 122-244zz PBMC HIV-1 DNA 3 (2-4) 2 (1-3) 4 (4-5) I n 55 58.2 (44.1-71.4), 32/55 54.5 (40.6-68.0), 30/55 30 56.7 (37.4-74.5), 17/30 56.7 (37.4-74.5), 17/30 25 60.0 (38.7-78.9), 15/25 52.0 (31.3-72.2), 13/25 31 (28-34) 56 89.3 (78.1-96.0), 50/56 89.3 (78.1-96.0), 50/56 30 (28-34) 32 (30-33) 81 (68-92) 160 (150-196) 47 91.5 (79.6-97.6), 43/47 89.4 (76.9-96.5), 42/47 9 77.8 (40.0-97.2), 7/9 88.9 (51.8-99.7), 8/9 38 100.0 (92.4-100.0), 38/38 100.0 (92.4-100.0), 38/38 12 100.0 (77.9-100.0), 12/12 1 3.9 (3.4-4.3) 0 (0/12) 1 5.6 (5.1-6.0) 3.7 (3.3-4.1) 0 (0/38) 0 (0/32) 0.61 2.5 (2.2-4.0) 2.5 (2.2-2.7) 71 (5/7) 25 (1/4) 0.88 (0.64-1.00) 5.4 (4.7-5.9) 3.3 (2.8-4.0) 12 (5/43) 8 (2/26) 0.81 (0.56-1.00) 5.1 (3.7-5.8) 3.0 (2.6-3.8) 20 (10/50) 10 (3/30) 0.68 (0.39-0.97) 3.1 (2.5-4.2) 2.7 (1.7-2.9) 47 (7/15) 20 (1/5) 0.86 (0.68-1.00) 3.8 (3.1-4.8) 2.8 (2.3-3.3) 24 (4/17) 25 (2/8) 0.78 (0.61-0.95) 3.6 (2.8-4.6) 2.8 (2.3-3.2) 34 (11/32) 23 (3/13) Sensitivity (95% CI), n Kappa (95% CI)z Median viral loadx (IQR) Frequency of low viral load{ % (n**)

ARV, antiretrovirals. *For each period, the sample nearest to the middle of the age range was considered. Number of infected infants tested. zKappa was calculated in infected infants. xMedian value was determined for positive samples. HIV-1 RNA is expressed as log copies/mL, PBMC HIV-1 DNA as log copies/106 PBMCs. HIV-1 DNA quantication was available for 67% of HIV-1 DNA-positive samples. {Low viral load is dened as #3.0 log copies/mL for HIV-1 RNA in plasma, and to 2.0 log copies/106 cells for HIV-1 DNA in PBMCs. Frequency was determined in positive samples. **Number of low viral loads in available viral loads for positive samples. Only samples taken during prophylaxis are considered in this study at 1 month. zzOnly samples taken before multiple curative ART are considered in this study at 3 and 6 months.

prophylaxis (P = .3). Of the infants with positive results in both PCR tests at birth, both PCR tests were negative at 1 month in one of the 6 infants receiving combination prophylaxis, and in none of the 19 receiving zidovudine. In neonates with a positive PCR result before day 7 of life, the HIV-1 RNA load in plasma was related to maternal therapy: the median was 3.9 log copies/mL in cases in which the mother received no or zidovudine therapy, and 3.1 in cases in which the mother received two- or 3-drug ART (P = .052). At 1 month, the proportions of infected infants receiving zidovudine prophylaxis scored as having a low viral load by PCR tests for HIV RNA was 12% and with PCR tests for HIV DNA was 8% (#3.0 log copies/mL and 2.0 log copies/ 106 PBMCs, respectively). Infected infants receiving multidrug prophylaxis had lower plasma loads of HIV-1 RNA and lower HIV-1 DNA concentrations in PBMCs than infants receiving zidovudine prophylaxis (medians, 2.5 versus 5.4 log copies/mL of plasma and 2.5 versus 3.3 log copies/ 106 PBMCs; P = .004 and P = .04, respectively). Plasma HIV-1 RNA loads tended to be lower at 1 month of age, during prophylaxis (median, 5.1 log copies/mL), than at 3 months of age, after interruption of prophylaxis (median, 5.6 log copies/mL, P = .065) in the 27 infants with positive PCR results for HIV-1 RNA at both 1 and 3 months.

Sensitivity According to Geographic Origin of the Mother. The frequency of positive PCR results did not differ according to the geographic origin of the mother (subSaharan Africa origin versus other origins). Both PCR tests were positive for 54% of children born to African mothers and 44% of the others at birth (P = .5), and for 86% and 90% at 1 month, respectively (P = 1.0).

Sensitivity According to Birth Period. The sensitivity of

PCR at 1 month did not differ according to birth years of study. Both PCRs were positive at 1 month for 100%, 79%, 86%, and 86% of the children born during these periods (quartiles) 1994 to mid 1996, mid 1996 to 1998, 1999 to 2001, and 2002 to 2006, respectively (P = .4). Similar results were obtained at birth: both PCRs were positive for 55%, 50%, 43%, and 54% of children from the oldest period to the most recent one, respectively (P = .9).

Sensitivity According to PCR Method. PCR sensitivity

did not differ according to the method of HIV DNA PCR (44%, 56%, and 54% at birth; 88%, 82%, and 88% at 1 month in cases of PCR with Southern blotting, Monitor technique, and real-time PCR, respectively; P = .7 and P = .9).
63

Performance of HIV-1 DNA or HIV-1 RNA Tests for Early Diagnosis of Perinatal HIV-1 Infection during Anti-Retroviral Prophylaxis

THE JOURNAL OF PEDIATRICS

www.jpeds.com

Vol. 160, No. 1

Table III. Results of HIV-1 PCR according to age* during the period 2002 to 2006
Median age (IQR) I n Uz n Kappax (95% CI) Specicity (95% CI), n PPV (95% CI), n 100.0 (68.8-100.0), 8/8 80.0 (44.4-97.5), 8/10 100.0 (47.3-100.0), 4/4 75.0 (19.4-99.4), 3/4 100.0 (47.3-100.0), 4/4 83.3 (35.9-99.6), 5/6 100.0 (77.9-100.0), 12/12 100.0 (77.9-100.0), 12/12 100.0 (71.7-100.0), 9/9 100.0 (71.7-100.0), 9/9 NPV (95% CI), n 99.6 (99.1-99.9), 1333/1338 99.6 (99.1-99.9), 1331/1336 99.3 (98.0-99.9), 440/443 99.1 (97.7-99.8), 439/443 99.8 (99.2-99.9), 893/895 99.9 (99.4-99.9), 892/893 99.8 (99.5-99.9), 1301/1303 99.8 (99.5-99.9), 1301/1303 100.0 (99.8-100.0), 1203/1203 100.0 (99.8-100.0), 1203/1203

Birth: day 1-7 4 (3-5) 13 1333 Plasma HIV-1 RNA 0.78 100.0 (99.8-100.0), 1333/1333 PBMC HIV-1 DNA 0.56-0.99 99.8 (99.5-99.9), 1331/1333 Day 1-3 3 (2-3) 7 440 Plasma HIV-1 RNA 0.75 100.0 (99. 3-100.0), 440/440 PBMC HIV-1 DNA 0.40-1.00 99.8 (98.7-99.9), 439/440 Day 4-7 4 (4-5) 6 893 Plasma HIV-1 RNA 0.8 100.0 (99.7-100.0), 893/893 PBMC HIV-1 DNA 0.52-1.00 99.9 (99.4-99.9), 892/893 1 month: day 15-45 33 (30-36) 14 1301 During any ART{ Plasma HIV-1 RNA 1 100.0 (99.8-100.0), 1301/1301 PBMC HIV-1 DNA 100.0 (99.8-100.0), 1301/1301 93 (90-98) 9 1203 3 months: day 61-121k Plasma HIV-1 RNA 1 100.0 (99.8-100.0), 1203/1203 PBMC HIV-1 DNA 100.0 (99.8-100.0), 1203/1203 6 months: day 122-244k 185 (179-194) 10 1112 PBMC HIV-1 DNA 1 100.0 (99.7-100.0), 1112/1112
*For each period, the sample nearest to the middle of the age range was considered. Number of infected infants tested. zNumber of uninfected infants tested. xKappa was calculated for all infants born between 2002 and 2006. {Only samples taken during prophylaxis are considered in this study at 1 month. kOnly samples taken before multiple curative ART are considered at 3 and 6 months.

100.0 (74.1-100.0), 10/10 100.0 (99.7-100.0), 1112/1112

When only the plasma samples with limit of detection at 50 copies/mL were considered, the sensitivity of HIV RNA PCR would be 75% at birth and 93% at 1 month. When HIV RNA PCRs with results <400 copies/mL were considered negative (as in non-ultrasensitive HIV RNA assays), the sensitivity would be 45% at birth and 79% at 1 month.

Samples that were positive for HIV-1 RNA and negative for HIV-1 DNA had low plasma viral load (median value, 2.0 log copies/mL; maximum value, 3.4 log copies/mL). Samples that tested negative for HIV-1 RNA and positive for HIV-1 DNA were tested with a median limit of detection of HIV-1 RNA PCR at 2.1 log copies/mL (maximum value 2.6 log copies/mL).

Specicity of HIV-1 PCRs. Among the 1502 uninfected children born between 2002 and 2006, both methods were used to test 1333 at birth, 1301 at 1 month, 1203 at 3 months, and 1112 at 6 months. The specicity of PCR testing for HIV-1 DNA was 99.8% at birth (two false-positive results of real-time PCR on days 3 and 4) and 100% at other ages. The specicity of PCR testing for HIV-1 RNA was 100% at all ages (Table III). Predictive Values. The positive predictive value during
the rst week of PCR testing for HIV-1 RNA was 100% and that of PCR testing for HIV-1 DNA was 78%. The positive predictive value of both tests was 100% at 1 month and 3 months. The negative predictive value for both tests was 99.5% at birth, 99.8% at 1 month, and 100% at 3 months. The predictive value of PCR testing for HIV-DNA at 6 months was 100%.

Discussion
This study compared the performance, including specicity, of PCR testing for HIV-1 DNA in PBMCs and HIV-1 RNA in plasma on such a large scale with recruitment from a multicenter prospective cohort of children. We show that the two methods had similar diagnostic performance in the postnatal period with a very high specicity and similar sensitivity; sensitivity was 100% for children aged 3 months, but was only 89% for those aged 1 month and receiving prophylaxis. One limitation of our study is caused by successive use of different PCR techniques in the 12 years of the study, but no difference was detected between techniques, all optimized to give the lowest detection limits and checked for correct quantication of non-B viruses.21 Another limitation is the use of in-house PCR tests for HIV-1 DNA; however, most HIV-1 DNA detection techniques are of this type and are less standardized than tests for HIV-1 RNA. Indeed, this may explain the lack of sensitivity of PCR tests for HIV-1 DNA observed in some studies,11,12 although this was not the case in ours and other studies.13 Third, unfortunately, all plasma samples could not be tested with a limit of detection at 50 copies/mL (ultrasensitive assay) caused by insufcient plasma volume, but the entirety of available plasma was used so as to have the lowest limit of detection as possible. The 1-mL plasma volume
Burgard et al

Concordance of PCR Testing for HIV-1 RNA and HIV-1 DNA. Overall, during the period 2002 to 2006,
the value of kappa between the two methods was 0.78 at birth. At 1, 3, and 6 months, concordance was 100%. For infected infants during the period 1994 to 2006, kappa was 0.78 at birth (89% of concordant results) and 0.81 at 1 month (96% of concordant results). At 3 and 6 months, concordance was 100%.
64

January 2012 required in ultrasensitive assays often is difcult to obtain at this age. At 1 month of age, 11% of infected infants were not diagnosed as infected, and this value was higher, at 24%, for infected infants with negative PCR results at birth. This sensitivity is lower than the value of at least 95% described in several studies,5,11,12,23 but is consistent with what has been reported in one study.13 Our study is very representative of routine practice and also avoided selection biases by prospectively recruiting from birth as part of the national EPFCO1; it also uses as the reference technique PCR tests for HIV-1 RNA at age 6 months, universally considered to be the gold standard for HIV diagnosis in non-breastfed children. Our results suggest that it may be difcult to respect the WHO recommendation to use assays with a sensitivity of at least 95%, ideally 98%, at 6 weeks of age.3 The children with negative PCR results at 1 month while taking zidovudine or combination prophylaxis had positive PCR results with high viral load at 3 months, after interruption of prophylaxis. Such viral kinetics are consistent with a delay in HIV detection induced by the 4 to 6 weeks of prophylaxis administered to children, as described in a prevention trial,24 but not in two historical comparisons.5,13 The overall trend toward increased HIV-1 RNA loads at 3 months of age, after interruption of prophylaxis, supports this notion. Our study included no children without post-natal prophylaxis, and it is therefore difcult to draw conclusions concerning this issue. The sensitivity of HIV detection was not signicantly associated with the type or existence of maternal or post-natal prophylaxis; however, HIV-1 RNA loads at birth and HIV1 RNA and DNA levels at 1 month were lower in children receiving multi-drug prophylaxis, suggesting diagnostic sensitivity may be affected by the type of prophylaxis if PCR tests less sensitive than ours for HIV-1 RNA and DNA were to be used. Our ndings of low viral loads at birth and 1 month of age further emphasize the need for use of highly sensitive assays, whatever the prophylaxis. Dried blood/plasma samples would have to be used if early diagnosis was to be generalized early diagnosis in low-income countries. However, dried blood/plasma samples result in a higher detection limit because of lower volume of blood tested.25-28 In our series, a PCR test for HIV-1 RNA with a detection limit of 2000 copies/mL would have a sensitivity of 77% for 1-month-old children receiving zidovudine. Because our study covers a 12-year period, it includes mothers receiving one- and two-drug prophylaxis, as may still be administered in resource-limited settings. The wide circulation of various HIV-1 non-B subtypes in France29 may make our results relevant to African countries; however, any PCR test for HIV must be validated with local HIV strains before use for diagnostic purposes. This description of the performance of PCR testing for HIV during post-natal combination prophylaxis is recommended in infants born after non-suppressive maternal therapy.15 We did not nd that testing for HIV-1 DNA

ORIGINAL ARTICLES
presented better results during post-natal combination than testing for HIV-1 RNA as had been speculated,16 but our data are limited. The viral kinetics in one of these infants was very unusual, with both PCR tests for HIV repeatedly being negative at 1 month, despite having positive results at birth and at age 3 months. This type of prole, with tests for HIV-1 DNA becoming negative so rapidly during anti-retroviral administration, is uncommon during treatment of adult HIV-1 primary infection, possibly because of the longer delays between infection and initiation of therapy in adults.30 This type of prole complicates the conrmation of diagnosis, and the prevalence of such cases needs to be assessed. Similarly, the performance of diagnostic testing during administration of nevirapine prophylaxis, for 4 to 6 weeks, has not been described.8 Our study shows that PCR testing for HIV-1 RNA and HIV-1 DNA perform similarly with 100% detection rates at 3 months, but failure to detect HIV in 11% (95% CI, 4%22%) of infections at 1 month. The age at which the WHO recommends blood sampling is 4 to 6 weeks, so this nding is worrying.3 Further studies may clarify the contribution of post-natal prophylaxis to HIV non-detectability at 1 month as long-course post-natal prophylaxis is becoming universally recommended.8 n
We thank Val erie Benhammou, Leila Boufassa, Corinne Laurent, J er^ ome Le Chenadec, Marl ene Peres, and Jean-Paul Teglas for their participation in the Agence Nationale de Recherche sur le SIDA et les Hepatites virales Cohort Study Group.
Submitted for publication Mar 20, 2011; accepted Jun 28, 2011. ^ pital Reprint requests: Marianne Burgard, MD, Laboratoire de Virologie, Ho vres, 75015 Paris, France. E-mail: marianne.burgard@ Necker, 149 rue de Se nck.aphp.fr

References
1. Faye A, Le Chenadec J, Dollfus C, Thuret I, Douard D, Firtion G, et al. Early versus deferred antiretroviral multidrug therapy in infants infected with HIV type 1. Clin Infect Dis 2004;39:1692-8. 2. Violari A, Cotton MF, Gibb DM, Babiker AG, Steyn J, Madhi SA, et al. Early antiretroviral therapy and mortality among HIV-infected infants. N Engl J Med 2008;359:2233-44. 3. WHO. Recommendations on the diagnosis of HIV infection in infants and children Available at: http://www.who.int/hiv/pub/paediatric/ diagnosis/en/index.html. Accessed Aug 11, 2011. 4. WHO. Antiretroviral therapy for HIV infection in infants and children: toward universal access. Recommendations for a public health approach. 2010 revision. Available at: http://www.who.int/hiv/pub/paediatric/ infants2010/en/index.html. Accessed Aug 11, 2011. 5. Dunn DT, Simonds RJ, Bulterys M, Kalish LA, Moye J, Jr., de Maria A, et al. Interventions to prevent vertical transmission of HIV-1: effect on viral detection rate in early infant samples. Aids 2000;14:1421-8. 6. Rouzioux C, Costagliola D, Burgard M, Blanche S, Mayaux MJ, Griscelli C, et al. Estimated timing of mother-to-child human immunodeciency virus type 1 (HIV-1) transmission by use of a Markov model. The HIV Infection in Newborns French Collaborative Study Group. Am J Epidemiol 1995;142:1330-7. 7. Tournoud M, Ecochard R. Age at which HIV infection can be detected in infants: place of the Yakovlev model. J Acquir Immune Dec Syndr 2006; 42:362-7. 65

Performance of HIV-1 DNA or HIV-1 RNA Tests for Early Diagnosis of Perinatal HIV-1 Infection during Anti-Retroviral Prophylaxis

THE JOURNAL OF PEDIATRICS

www.jpeds.com

Vol. 160, No. 1

are consistent markers for estimating viral load in patients undergoing long-term potent treatment. AIDS Res Hum Retroviruses 2000;16: 1939-47. Rouet F, Chaix ML, Nerrienet E, Ngo-Giang-Huong N, Plantier JC, Burgard M, et al. Impact of HIV-1 genetic diversity on plasma HIV-1 RNA quantication: usefulness of the Agence Nationale de Recherches sur le SIDA second-generation long terminal repeat-based real-time reverse transcriptase polymerase chain reaction test. J Acquir Immune Dec Syndr 2007;45:380-8. Viard JP, Burgard M, Hubert JB, Aaron L, Rabian C, Pertuiset N, et al. Impact of 5 years of maximally successful highly active antiretroviral therapy on CD4 cell count and HIV-1 DNA level. Aids 2004; 18:45-9. Sherman GG, Cooper PA, Coovadia AH, Puren AJ, Jones SA, Mokhachane M, et al. Polymerase chain reaction for diagnosis of human immunodeciency virus infection in infancy in low resource settings. Pediatr Infect Dis J 2005;24:993-7. Prasitwattanaseree S, Lallemant M, Costagliola D, Jourdain G, Mary JY. Inuence of mother and infant zidovudine treatment duration on the age at which HIV infection can be detected by polymerase chain reaction in infants. Antivir Ther 2004;9:179-85. Beck IA, Drennan KD, Melvin AJ, Mohan KM, Herz AM, Alarcon J, et al. Simple, sensitive, and specic detection of human immunodeciency virus type 1 subtype B DNA in dried blood samples for diagnosis in infants in the eld. J Clin Microbiol 2001;39:29-33. Comeau AM, Pitt J, Hillyer GV, Landesman S, Bremer J, Chang BH, et al. Early detection of human immunodeciency virus on dried blood spot specimens: sensitivity across serial specimens. Women and Infants Transmission Study Group. J Pediatr 1996;129:111-8. Lofgren SM, Morrissey AB, Chevallier CC, Malabeja AI, Edmonds S, Amos B, et al. Evaluation of a dried blood spot HIV-1 RNA program for early infant diagnosis and viral load monitoring at rural and remote healthcare facilities. Aids 2009;23:2459-66. Monleau M, Montavon C, Laurent C, Segondy M, Montes B, Delaporte E, et al. Evaluation of different RNA extraction methods and storage conditions of dried plasma or blood spots for human immunodeciency virus type 1 RNA quantication and PCR amplication for drug resistance testing. J Clin Microbiol 2009;47: 1107-18. Descamps D, Chaix ML, Andre P, Brodard V, Cottalorda J, Deveau C, et al. French national sentinel survey of antiretroviral drug resistance in patients with HIV-1 primary infection and in antiretroviral-naive chronically infected patients in 2001-2002. J Acquir Immune Dec Syndr 2005;38:545-52. Ngo-Giang-Huong N, Deveau C, Da Silva I, Pellegrin I, Venet A, Harzic M, et al. Proviral HIV-1 DNA in subjects followed since primary HIV-1 infection who suppress plasma viral load after one year of highly active antiretroviral therapy. Aids 2001;15:665-73.

8. WHO. Antiretroviral drugs for treating pregnant women and preventing HIV infections in infants. Recommendations for a public health approach. 2010 version. Available at: http://www.who.int/hiv/pub/mtct/ antiretroviral2010/en/index.html. Accessed Aug 11, 2011. 9. Delamare C, Burgard M, Mayaux MJ, Blanche S, Doussin A, Ivanoff S, et al. HIV-1 RNA detection in plasma for the diagnosis of infection in neonates. The French Pediatric HIV Infection Study Group. J Acquir Immune Dec Syndr Hum Retrovirol 1997;15:121-5. 10. Young NL, Shaffer N, Chaowanachan T, Chotpitayasunondh T, Vanparapar N, Mock PA, et al. Early diagnosis of HIV-1-infected infants in Thailand using RNA and DNA PCR assays sensitive to non-B subtypes. J Acquir Immune Dec Syndr 2000;24:401-7. 11. Cunningham CK, Charbonneau TT, Song K, Patterson D, Sullivan T, Cummins T, et al. Comparison of human immunodeciency virus 1 DNA polymerase chain reaction and qualitative and quantitative RNA polymerase chain reaction in human immunodeciency virus 1exposed infants. Pediatr Infect Dis J 1999;18:30-5. 12. Lambert JS, Harris DR, Stiehm ER, Moye J, Jr., Fowler MG, Meyer WA 3rd, et al Performance characteristics of HIV-1 culture and HIV-1 DNA and RNA amplication assays for early diagnosis of perinatal HIV-1 infection. J Acquir Immune Dec Syndr 2003;34:512-9. 13. Nesheim S, Palumbo P, Sullivan K, Lee F, Vink P, Abrams E, et al. Quantitative RNA testing for diagnosis of HIV-infected infants. J Acquir Immune Dec Syndr 2003;32:192-5. 14. Warszawski J, Tubiana R, Le Chenadec J, Blanche S, Teglas JP, Dollfus C, et al. Mother-to-child HIV transmission despite antiretroviral therapy in the ANRS French Perinatal Cohort. Aids 2008;22:289-99. 15. Minist ere de la sant e et des sports. Prise en charge m edicale des personnes infect ees par le VIH. Rapport 2010 sous la direction du Pr Patrick Yeni. Available at: http://www.sante-sports.gouv.fr/IMG/pdf/Rapport_2010_ sur_la_prise_en_charge_medicale_des_personnes_infectees_par_le_ VIH_sous_la_direction_du_Pr-_Patrick_Yeni.pdf. Accessed Aug 11, 2011. 16. AIDSinfo. Panel. on Antiretroviral Therapy and Medical Management of HIV-Infected Children. Guidelines for the use of antiretroviral agents in pediatric HIV infection. Available at: http://aidsinfo.nih.gov/ ContentFiles/PediatricGuidelines.pdf. Accessed Aug 11, 2011. 17. Frange P, Burgard M, Lachassinne E, le Chenadec J, Chaix ML, Chaplain C, et al. Late postnatal HIV infection in children born to HIV-1-infected mothers in a high-income country. Aids 2010;24:1771-6. 18. Avettand-Fenoel V, Chaix ML, Blanche S, Burgard M, Floch C, Toure K, et al. LTR real-time PCR for HIV-1 DNA quantitation in blood cells for early diagnosis in infants born to seropositive mothers treated in HAART area (ANRS CO 01). J Med Virol 2009;81:217-23. 19. Laure F, Courgnaud V, Rouzioux C, Blanche S, Veber F, Burgard M, et al. Detection of HIV1 DNA in infants and children by means of the polymerase chain reaction. Lancet 1988;2:538-41. 20. Burgard M, Izopet J, Dumon B, Tamalet C, Descamps D, Ruffault A, et al. HIV RNA and HIV DNA in peripheral blood mononuclear cells

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

66

Burgard et al

January 2012

ORIGINAL ARTICLES Appendix

The following institutions (and persons) participated in this substudy of the Agence Nationale de Recherche sur le SIDA et les Hepatites virales French Perinatal Cohort (EPF-CO1): ^pital Nord, Amiens (J. L. Schmit); Ho ^pital Victor DuHo ^pital Robert pouy, Argenteuil (C. Allisy and D. Brault); Ho Ballanger, Aulnay (E. Questiaux, A. Zakaria, and C. Golden^pital Saint Jacques, Besanc stein); Ho on (J. M. Estavoyer and ^pital Jean Verdier, Bondy (L. Benoist, S. Bolie, R. Maillet); Ho ^pital de Creil (B. Carpentier, M. DuvalE. Lachassine); Ho ^pital Intercommunal, Arnould, and C. Kingue-Ekollo); Ho ^pital B Cr eteil (C. Pichon); Ho ecl ere, Clamart (L. Clech); ^pital Louis Mourier, Colombes (L. Mandelbrot, Ho C. Crenn-Hebert, C. Floch-Tudal, F. Mazy, and C. Meier); ^pital denfants, Dijon (S. Martha and I. Reynaud); Ho ^pital Ho ^pital Intercommunal, Evreux (C. Allouche and K. Tour e); Ho Francilien Sud, Evry-Corbeil (A. Devidas, M. Granier, ^pital de Fontainebleau (K. Alissa A. May, and I. Turpault); Ho ^pital de Lagny (A. Chalvon); Ho ^pital de and C. Routier); Ho ^pital de LongBic^ etre, Le Kremlin-Bic^ etre (C. Fourcade); Ho ^pital Franc jumeau (I. Turpault, and H. Seaume); Ho ois Quesnay, Mantes La Jolie (A. Doumet, F. Granier, and ^pital de Meaux (L. Karaoui and V. Lef J. L. Salomon); Ho e^pital Marc Jacquet, Melun (B. Le Lorier); Ho ^pital Invre); Ho tercommunal, Montfermeil (M. Dehlinger and P. Talon); ^pital Intercommunal, Montreuil (B. Heller-Roussin); Ho ^pital Maternit e R egionale A. Pinard, Nancy (C. Hubert); Ho ^pital Lariboisi dOrsay (S. Chanzy, C. De Gennes); Ho ere, Paris (D. Ayral, N. Ciraru-Vigneron, and G. Mouchnino); ^pital Necker, Paris (S. Blanche, M. Burgard, S. Parat, Ho ^pital Ren C. Rouzioux, J. P. Viard, A. Yamgnane); Ho e ^pital Am Dubos, Pontoise (A. Coursol); Ho ericain, Reims ^pital de Saint-Denis (M. C. Allemon, (M. Munzer); Ho ^pital Brabois, Vandoeuvre les P. Bolot, D. Ekoukou); Ho ^pital de Villeneuve Saint Nancy (L. Neimann); and Ho Georges (F. Guillot).

Performance of HIV-1 DNA or HIV-1 RNA Tests for Early Diagnosis of Perinatal HIV-1 Infection during Anti-Retroviral Prophylaxis

66.e1