an international collaborative study to establish the 1st international standard for hiv-1 rna for...

10
Journal of Virological Methods 92 (2001) 141 – 150 An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques Harvey Holmes a, *, Clare Davis a , Alan Heath b , Indira Hewlett c , Nico Lelie d a Di6ision of Retro6irology, National Institute of Biological Standards and Control, Blanche Lane, South Mimms, Potters Bar, Herts EN63QG, UK b Di6ision of Informatics, National Institute of Biological Standards and Control, Blanche Lane, South Mimms, Potters Bar, Herts EN63QG, UK c Center for Biological E6aluation and Research, Food and Drug Administration, Building 29a, Room 2d -16, Bethesda, MD, USA d Central laboratory of the Netherlands Blood Tranfusion Ser6ice, Viral Diagnostics Laboratory, Plesmanlaan 125, 1066 CX Amsterdam, The Netherlands Received 12 October 2000; received in revised form 7 November 2000; accepted 8 November 2000 Abstract Twenty-six laboratories from 10 different countries participated in a collaborative study to establish the 1st International Standard for HIV-1 RNA for use in nucleic acid-based techniques (NAT). Three candidate preparations were tested all based on genotype B viruses. The candidates were tested by each laboratory at a range of dilutions in four independent assays and the results collated and analysed statistically. All three candidates gave results that were tightly grouped, with little difference between the results from different laboratories or from the use of different assays. Studies of relative potency showed good agreement between laboratories. There were no significant differences between five commercial assay types, except that candidate XX showed a slightly lower potency compared to YY and ZZ with a single commercial assay. The reason for this was not established. Degradation studies showed that the freeze-dried preparations were stable at -20, 4 and 20°C for 26 weeks, the longest period studied, but that they became difficult to reconstitute after 3 weeks at 45°C and 9 weeks at 37°C. As a result of the study, the World Health Organisation (WHO) Expert Committee on Biological Standardisation (ECBS) established the preparation referred to as candidate YY (NIBSC Code No. 97/656) as the 1st International Standard for HIV-1 RNA for use with NAT with an assigned potency of 100 000 International Units per vial. © 2001 Elsevier Science B.V. All rights reserved. Keywords: HIV-1 RNA; Nucleic acid-based techniques; NAT; International standard www.elsevier.com/locate/jviromet 1. Introduction Over the past decade, PCR and other nucleic acid-based techniques (NAT), have improved dra- matically our ability to detect viruses such as the * Corresponding author. Tel.: +44-1707-654753; fax: +44- 1707-649865. E-mail address: [email protected] (H. Holmes). 0166-0934/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII:S0166-0934(00)00283-4

Upload: harvey-holmes

Post on 15-Sep-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

Journal of Virological Methods 92 (2001) 141–150

An international collaborative study to establish the 1stinternational standard for HIV-1 RNA for use in nucleic

acid-based techniques

Harvey Holmes a,*, Clare Davis a, Alan Heath b, Indira Hewlett c, Nico Lelie d

a Di6ision of Retro6irology, National Institute of Biological Standards and Control, Blanche Lane, South Mimms, Potters Bar,Herts EN6 3QG, UK

b Di6ision of Informatics, National Institute of Biological Standards and Control, Blanche Lane, South Mimms, Potters Bar,Herts EN6 3QG, UK

c Center for Biological E6aluation and Research, Food and Drug Administration, Building 29a, Room 2d-16, Bethesda, MD, USAd Central laboratory of the Netherlands Blood Tranfusion Ser6ice, Viral Diagnostics Laboratory, Plesmanlaan 125,

1066 CX Amsterdam, The Netherlands

Received 12 October 2000; received in revised form 7 November 2000; accepted 8 November 2000

Abstract

Twenty-six laboratories from 10 different countries participated in a collaborative study to establish the 1stInternational Standard for HIV-1 RNA for use in nucleic acid-based techniques (NAT). Three candidate preparationswere tested all based on genotype B viruses. The candidates were tested by each laboratory at a range of dilutions infour independent assays and the results collated and analysed statistically. All three candidates gave results that weretightly grouped, with little difference between the results from different laboratories or from the use of differentassays. Studies of relative potency showed good agreement between laboratories. There were no significant differencesbetween five commercial assay types, except that candidate XX showed a slightly lower potency compared to YY andZZ with a single commercial assay. The reason for this was not established. Degradation studies showed that thefreeze-dried preparations were stable at −20, 4 and 20°C for 26 weeks, the longest period studied, but that theybecame difficult to reconstitute after 3 weeks at 45°C and 9 weeks at 37°C. As a result of the study, the World HealthOrganisation (WHO) Expert Committee on Biological Standardisation (ECBS) established the preparation referred toas candidate YY (NIBSC Code No. 97/656) as the 1st International Standard for HIV-1 RNA for use with NAT withan assigned potency of 100 000 International Units per vial. © 2001 Elsevier Science B.V. All rights reserved.

Keywords: HIV-1 RNA; Nucleic acid-based techniques; NAT; International standard

www.elsevier.com/locate/jviromet

1. Introduction

Over the past decade, PCR and other nucleicacid-based techniques (NAT), have improved dra-matically our ability to detect viruses such as the

* Corresponding author. Tel.: +44-1707-654753; fax: +44-1707-649865.

E-mail address: [email protected] (H. Holmes).

0166-0934/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.

PII: S 0166 -0934 (00 )00283 -4

Page 2: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150142

human immunodeficiency virus (HIV) and hepati-tis C virus (HCV) at increasingly earlier stages ofinfection as well as facilitating an accurate mea-surement of virus load (Busch et al., 1995). Theirapplication in the field of HIV patient manage-ment has provided clinicians with a powerfulprognostic indicator. It has enabled clinicians tomonitor the benefits of combined drug therapysuch as HAART, with the result that viral loadcan be reduced to very low or even undetectablelevels (Carpenter et al., 1996; Mellors et al., 1996;Saag et al., 1996). In the area of blood and bloodproduct safety, the use of NAT can reduce thelikelihood of donations given during the ‘windowperiod’, that period between infection and demon-strable seroconversion, from being administeredor used in the manufacture of blood products andthereby resulting in transmissions of blood-borneviruses. As a result of the transmission of HCV byintravenous immunoglobulin in 1994 (Schneiderand Geha, 1994; James and Mosley, 1995) the USFDA introduced a requirement for the testing forHCV RNA in all immunoglobulin preparationsthat had not been inactivated virally (Scribner,1994; Zoon, 1995). From mid-1999, the EuropeanUnion has required that plasma pools used formanufacturing plasma derivatives must be testedand found non-reactive for HCV RNA (Mueller-Breitkreutz and Allain, 1999). Although the test-ing of plasma pools for HIV-1 RNA by NAT isnot yet a mandatory requirement, with the im-provement in assay sensitivity that has beenachieved recently and the desire to reduce the riskof virus transmission by blood transfusion andblood products as much as is practically possible,a number of organisations have already imple-mented voluntary screening for HIV-1 by NATand more are sure to follow.

The WHO International Working Group onStandardisation of Gene Amplification Tech-niques for the Virological Safety Testing of Bloodand Blood Products (SoGAT) recommended thatthe establishment of a WHO International Stan-dard for HIV-1 RNA based on a genotype B virusbe given a high priority (Rogers et al., 1997). Sucha standard would be given a defined unitage andcould be used to calibrate the range of workingreagents currently in use. This would facilitate

inter-laboratory comparisons and help reducevariations between different laboratories and be-tween different commercial and ‘in-house’ assays(Coste et al., 1996; Revets et al., 1996; Schuurmanet al., 1996; Fransen et al., 1998; Bootman et al.,1999). It would also provide an independentquantitative standard that could be used to moni-tor the sensitivity of assays used in the manufac-ture of blood products.

In response to this need, we identified threecandidate standards for evaluation in an interna-tional collaborative study, with the aim of estab-lishing an International Standard for HIV-1RNA.

2. Materials and methods

2.1. Preparation of the candidate standards

Three candidate standards were evaluated inthe collaborative study, two of which (XX andYY) were freeze-dried and the third (ZZ) wasfrozen at −70°C. Details of each preparation isgiven below:1. Candidate XX was based on a genotype B (env

V3, gag) field isolate of HIV-1 provided by DrP Simmonds, University of Edinburgh, UK. Itwas isolated post-mortem from a patient whohad died from an AIDS-defining illness (pa-tient 4 in Donaldson et al., 1994). The viruswas supplied as a low-passage PBMC culturethat was propagated further on humanPBMCs and a stock of cell-free culture super-natant stored under vapour phase liquid nitro-gen. To prepare the candidate standard, thevirus was diluted 1 in 250 in pooled humancryosupernatant (provided by Dr P Harrison,Bio Products Laboratory, Elstree, UK). Thecryosupernatant had been pre-screened forHBsAg and for antibodies to HCV and HIV-1/2 and was found to be negative. A total of2200 vials were prepared, the cv of the fillvolume was 0.2%.

2. Candidate YY was based on an HIV-1 PCR-positive, antibody-negative plasmapheresis do-nation. Env V3 sequencing had shown this to

Page 3: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150 143

be a genotype B virus. To prepare the candi-date standard, the virus was diluted 1 in 312.5in Base Matrix (defibrinated plasma fromBoston Biomedica Inc, USA). The Base Ma-trix had been pre-screened for HBsAg and forantibodies to HCV and HIV-1/2 and wasfound to be negative. Two thousand four hun-dred vials were filled, the cv of the fill volumewas 0.59%.

3. Candidate ZZ was based on a genotype B (envV3) field isolate. The virus was propagated onMT2 cells for 8 weeks until an infectivity titreof 104.52 TCID50/ml was reached. The culturesupernatant was diluted 1:10 in a pooledclarified plasma that was negative for HBsAg,HBV-DNA, HCV-RNA, HIV-RNA and anti-HTLV I/II. The viral preparation was ho-mogenised and thoroughly mixed aliquotswere stored at −70°C. This HIV batch wasfurther diluted 1:100 000 in the pooled plasmadiluent described above. The diluted ho-mogenised pool was dispensed in 2000 aliquotsof 0.5 ml, snap frozen and stored at −70°C.

2.2. Freeze drying

The following procedure was followed for thefreeze drying of candidates XX and YY. Afterthorough mixing, 1 ml volumes were dispensedinto 2 ml Adelphi vials (Cat No. VC002-13C) thathad been washed in the absence of detergent andoven sterilised. The bulk was kept on ice duringdispensing. Rubber seals (Adelphi Cat No.FD13), that had been immersed in 95% ethanol,5% methanol for a minimum of 1 h, were seatedon top of the vials and the vials loaded into thefreeze-drier, the shelves of which had been pre-cooled to −40°C. The shelf temperature wasmaintained at −40°C for at least 3 h in theabsence of any vacuum. After this initial period, amaximum vacuum was applied, whilst maintain-ing shelf temperature at −40°C, for at least 90 h.The condenser temperature was at or below −70°C. After this period, maximum vacuum wasmaintained, whilst the shelf temperature wasramped from −40 to +20°C over a period of 20h. This temperature was maintained for at least 2h. The freeze-drier was then back filled with N2

and the vials sealed within the freeze-drier. Amoisture trap was used to ensure that the N2 hadless than 5 ppm O2. The vials were then removedfrom the freeze-drier, crimp-sealed with alu-minium tear-off overseals and placed in individualaluminium cans containing absorbent cottonwool. The vials were stored at −20°C.

2.3. Organisation of the study

Twenty-six laboratories participated in the col-laborative study and each was given a randomisedcode number. Laboratories were sent five vials ofeach of the three candidate standards on dry iceand the materials stored at or below −70°C untilassayed. Participants were requested to test thethree candidates in four independent assays onseparate occasions with a fresh vial of each sam-ple for each assay. Freeze-dried preparations werereconstituted using 1.0 ml of deionised RNAse-free water immediately before use and were agi-tated gently for a minimum of 20 min to fullydissolve the contents. The liquid preparation wasthawed quickly in a 37°C water bath. For the firstassay, participants were asked to test 10-fold (1-log) dilution steps and for the subsequent assayshalf-log dilutions around the end-point deter-mined from the first assay. Participants wereasked to prepare the dilutions in the diluent nor-mally used in their assay system. Results andassay details were recorded on a set of formsprovided to each participant and the data wasreturned to NIBSC for collation and analysis.

2.4. Statistical methods

2.4.1. Quantitati6e assaysLaboratories using quantitative assays reported

the number of ‘detectable units per ml’ for eachdilution tested. For each laboratory and assaymethod, the linearity of these estimates waschecked by plotting the log10 estimate against thelog10 dilution. Ideally, a straight line with a slopeof 1.0 should be observed and deviations fromlinearity are readily observed. Results that wereclearly aberrant or deviated markedly from linear-ity were discounted. A single estimate of copiesper ml for the undiluted material was obtained by

Page 4: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150144

correcting each individual estimate by the dilutionfactor, and calculating an overall arithmetic meanof the log estimates.

2.4.2. Qualitati6e assaysLaboratories using qualitative assays reported

the number of samples positive out of the numbertested at various dilutions. These were treated as adilution series and used to provide a single esti-mate of ‘detectable units per ml’ in the undilutedsample using the method of maximum likelihoodfor ‘dilution assays’ (Collet, 1991). This modelassumes that the probability of a positive result ata given dilution follows a Poisson distribution(with mean given by the expected number of‘copies’ in the sample tested), and that a single‘copy’ will lead to a positive result. Calculationswere carried out using the statistical packageGLIM (Francis et al., 1993). No correction ismade for transcription efficiency however and theestimates are not necessarily directly equivalent toa genuine genome equivalence number.

2.4.3. Relati6e potenciesThe relative potency of, for example, sample

XX relative to sample YY was calculated as thedifference in estimated log ‘detection units per ml’(XX−YY).

2.4.4. Comparison of assay typesResults from different assay types were com-

pared using a Wilcoxon rank test.

3. Results

From a total of 26 laboratories which partici-pated in the study, 27 data sets were reported, 20(74%) of which were in a quantitative format and7 (26%) qualitative. Of these, 18 and 3, respec-tively were from commercially available kits. Thequantitative assays included the Abbott LCx HIVQuantitative RNA Assay (LCx), the Bayer (for-merly Chiron) Quantiplex HIV RNA bDNA As-say version 3.0 (bDNA), the Organon TeknikaNASBA NucliSens HIV-1 QT Assay (Nuclisens)and the Roche Amplicor HIV-1 Monitor Assayversion 1.5 (with or without the Ultrasensitive

modification) (Monitor). Qualitative assays in-cluded the Gen-Probe HIV-1 TMA Assay (TMA),the Roche Ampliscreen HIV-1 Assay (Amplis-creen) and the Roche Amplicor HIV-1 MonitorAssay in a qualitative format. In house assaysincluded both quantitative (two assays) and quali-tative formats (four assays). A variety of differentmethodologies were used including initial ultra-centrifugation (two laboratories); extraction withRNAzol (two laboratories), sodium iodide (onelaboratory), proteinase K/SDS (one laboratory)or guanidine thiocyanate/silica gel (one labora-tory); nested (three laboratories) or single (twolaboratories) PCR; gag, pol or LTR primers; tran-scription mediated amplification (TMA) (one lab-oratory); and use of an internal calibrator (fourlaboratories).

The results of testing the three candidate stan-dards at a range of dilutions in a number ofcommercial and in-house quantitative assays aresummarised graphically in Fig. 1. At the higherconcentrations, the results from all the assaysclustered closely together and showed an overalllinear relationship between dilution and copynumber estimation. At the highest dilutions, thegreatest number of deviations from linearity wereapparent, but these occurred mostly at a low copynumber (B500 copies/ml), usually below theworking range of the assay. In most cases this hadlittle effect on the overall estimates of copynumber.

Results from the qualitative and quantitativeassays were treated separately in the calculation ofthe overall mean values and these are summarisedin Table 1. All three candidate standards gaveresults that were tightly grouped, with only smalldifferences between results from laboratories us-ing the same assay and between assays. The agree-ment between quantitative assays was particularlygood with the means for each assay type within0.27–0.40 log 10 of each other. For individualassays, the difference between the highest andlowest value for each sample was 50.2 log 10 forthe bDNA, in-house and Nuclisens assays, al-though the most widely used assay, the RocheMonitor, ranged up to 0.79 log 10. This may,however, be due, at least in part, to the highernumber of assays carried out, as statistically, the

Page 5: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150 145

observed range is expected to increase with thenumber of assays undertaken.

The results clearly show that the three candi-dates contain different amounts of viral RNAwith sample XX containing the highest and ZZthe lowest amount. The quantitative assays gave amean titre for candidates XX, YY and ZZ of 5.34

log 10 [219 000 geq/ml], 4.75 log 10 [56 000 geq/ml] and 4.17 log 10 [15 000 geq/ml], respectively.

Qualitative assays gave overall mean valuesthat were 0.45–0.63 log 10 detectable units per mllower than those from the quantitative assays,with the results from one laboratory using anin-house assay giving a particularly low value for

Fig. 1. RNA copy number (log10 genome equivalents/ml) estimated for dilutions (log10) of candidate standard XX (a), candidate YY(b) and candidate ZZ (c) by the Abbott LCx HIV Quantitative RNA Assay (a), the Bayer Quantiplex HIV RNA bDNA Assayversion 3.0 (b3), ‘in-house’ assays (ih), Roche Amplicor HIV-1 Monitor Assay version 1.5 (m) and the Organon-Teknica NASBANuclisens HIV-1 QT Assay (nu).

Page 6: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150146

Table 1Mean values and range for each assay typea

Method Numberof assaysSample MeanType Minimum Maximum Range

Ampliscreen 1 4.91XX 4.91Qualitative 4.91 –In-house 4 4.58 4.04 4.84 0.80Monitor 1 5.01 5.01 5.01 –All assays 6 4.71 4.04 5.01 0.97

Quantitative LCx 1 5.51 5.51 5.51 –bDNA V3 4 5.32 5.22 5.42 0.20In-house 2 5.38 5.29 5.48 0.19Monitor 10 5.37 5.08 5.83 0.75Nuclisens 3 5.24 5.10 5.31 0.20All assays 20 5.35 5.08 5.83 0.75

Ampliscreen 1 4.55YY 4.55Qualitative 4.55 –In-house 4 4.24 3.80 4.48 0.68Monitor 1 3.91 3.91 3.91 –All assays 6 4.24 3.80 4.55 0.75LCx 1 4.99Quantitative 4.99 4.99 –bDNA V3 4 4.61 4.51 4.69 0.18In-house 2 4.71 4.70 4.72 0.02Monitor 10 4.71 4.37 5.16 0.79Nuclisens 3 5.01 4.94 5.12 0.17All assays 20 4.75 4.37 5.16 0.79

Ampliscreen 1 4.02ZZ 4.02Qualitative 4.02 –In-house 4 3.59 3.25 3.88 0.62Monitor 1 3.91 3.91 3.91 –All assays 6 3.72 3.25 4.02 0.77LCx 1 4.39Quantitative 4.39 4.39 –bDNA V3 4 4.06 4.01 4.08 0.07In-house 2 4.30 4.13 4.46 0.34Monitor 10 4.11 3.87 4.50 0.63Nuclisens 3 4.36 4.00 4.61 0.60All assays 20 4.17 3.87 4.61 0.73

a Values expressed as log 10.

samples XX and ZZ (not shown). For samplesXX, YY and ZZ, the mean estimates from thedifferent assays ranged from 0.43–0.64 log 10,with the ‘in-house’ assay giving the lowest esti-mate for samples XX and ZZ, but not YY.

The potency of each candidate standard wasdetermined relative to the other candidate stan-dards. Overall, there was good agreement betweenlaboratories and no significant differences werefound between the five commercial assay types,with one exception. For the Nuclisens assay, thepotency of sample XX relative to either sampleYY or ZZ was lower (PB0.05) than for the otherassay types. There was no significant differencefor YY relative to ZZ. However, looking at the

actual estimates for XX, YY and ZZ, the Nu-clisens results were not significantly lower for XX(P\0.05).

The two freeze-dried candidates (XX and YY)were analysed for their stability over time inaccelerated degradation studies. Samples were in-cubated in temperature controlled environments,withdrawn at specified times, reconstituted withwater and tested in duplicate using the RocheAmplicor HIV-1 Monitor assay version 1.5. Thepotencies relative to concurrently tested samplesstored at −20°C are shown in Table 2. Theresults show that when incubated at −20, +4and +20°C, both candidates were stable for peri-ods up to 26 weeks, the longest period tested.

Page 7: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150 147

Table 2Stability of freeze-dried candidates at higher storagetemperaturesa

Candidate Storage temperature (oC)Week

4 20 37 45

93.33XX 102.331 67.61 60.2657.54 64.573 47.86XX 35.48

6XX 70.79 77.62 54.95 –7XX 89.13 93.33 95.5 44.67

93.33 95.59 29.51XX 38.02XX 11 109.65 128.82 · –

95.5 75.8615 23.44XX –19XX 100.93 75.86 17.22 –26XX 134.9 190.55 · –

107.15 186.21YY 102.331 141.25YY 3 45.71 70.79 77.62 35.48

104.71 114.826 109.65YY 56.23154.88 141.25YY 56.237 56.23134.9 117.499 61.66YY –

13YY 79.43 74.13 32.36 –74.13 70.7915 40.74YY –

19YY 180.3 155.24 31.92 –95.5YY 10026 · –

a Potency relative to −20 sample expressed as percentage.

plasma is being used extensively in HIV patientmanagement as a prognostic marker and inmonitoring the effects of antiviral chemotherapy(Carpenter et al., 1996; Mellors et al., 1996;Saag et al., 1996), and increasingly by manufac-turers of blood products. A number of studieshave performed comparative evaluations of a va-riety of clinical samples with different NAT as-says and differences in sensitivity as well asdiscordant results have been identified (Coste etal., 1996; Schuurman et al., 1996; Fransen et al.,1998). Such results highlight the need for readilyavailable, internationally recognised quantitationstandards (Berry and Tedder, 1999). We there-fore embarked upon the current study with theaim of establishing an international standard forHIV-1 RNA for use in NAT.

A total of 26 international laboratories tookpart in the collaborative study and of the 27data sets that were reported, seven were in aqualitative format. In calculating the number ofdetectable units per ml for the undiluted sample,no correction was made for the overall efficiencyof the assays, particularly with regard to theefficiency of extraction, reverse transcription,amplification and detection. The values, there-fore, may not relate directly to copy numberand it was decided to analyse the two sets ofdata separately.

The characteristics of the three candidate sam-ples are summarised in Table 3. All three werebased on HIV-1 subtype B strains. The decisionto use a virus of this subtype was arrived at by

After 3 weeks at +37 or +45°C, the samplesbecame progressively more difficult to reconsti-tute and this coincided with a loss of potency.

4. Discussion

The measurement of HIV-1 RNA in the

Table 3Characteristics of candidate standards

Candidate YYCandidate XX Candidate ZZCriteria

Virus Field isolatePrimary isolate Plasmapheresis donationB B BSubtypePBMCsPropagated No T Cell lineHigh (219 K)Copy number (quantitative assays) Medium (56 K) Low (15 K)

FrozenFreeze-driedPresentation Freeze-driedCryosupernatant Defibrinated plasma Plasma poolDiluent1 mlVolume 1 ml 0.5 ml

Not testedGoodGoodStabilityNoProposed International Standard Yes No

Page 8: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150148

the WHO Working Group SoGAT after someconsideration. Most NAT that are currently inuse were designed initially to detect the subtype Bstrains that predominated in North America andEurope, although with newer assays and modifi-cations of earlier assays, considerable efforts havebeen made to ensure that they detect satisfactorilyas broad a range of subtypes as possible. As theInternational Standard would primarily be usedfor the calibration of working standards and inestablishing the levels of sensitivity of individualassays, it was decided that a single standard basedon a subtype B strain would be appropriate.Although the ability of NAT assays to detect thefull range of HIV-1 subtypes is of great impor-tance, especially in regions of the world such assouthern Africa and India where the AIDS pan-demic is at unprecedented levels and non-B sub-types predominate, this cannot be addressed by asingle preparation. Reference preparations such ascharacterised panels of different HIV-1 subtypeswill be needed for this purpose and such panelsare in preparation.

Candidate XX was a freeze-dried, PBMC-grown, primary isolate diluted in plasma cryosu-pernatant and had the highest mean viral RNAconcentration. Candidate YY was also freeze-dried, was based on a human plasmapheresis do-nation diluted in defibrinated plasma and had anintermediate mean viral RNA concentration.Candidate XX was based on a T-cell adaptedstrain of HIV-1 diluted in pooled plasma, wasfrozen and had the lowest mean viral RNAconcentration.

The results of both the quantitative and thequalitative assays showed that all three candidatesperformed well and that each could serve as theInternational Standard. Although the viral RNAconcentration differed in each preparation, allwere considered within the range that would beacceptable for use as an International Standard.There was some evidence that the potency ofsample XX relative to sample YY or ZZ waslower in the NASBA Nuclisens assay. The reasonfor this is unclear, although it may be associatedwith primer/probe mismatches. Although no evi-dence of primer mismatches were observed, twomismatches were found between the Nuclisens

detection probe and the viral sequence of sampleXX, but it is unclear whether this would be suffi-cient to account for the observed results.

To serve as an International Standard, it isdesirable that a preparation should be stable formany years and for this reason freeze-dried prepa-rations are preferred. From the Arrhenius modelfor accelerated degradation (Kirkwood, 1977;Kirkwood and Tydeman, 1984), the estimates ofpotency for the samples stored at elevated temper-atures can be used to predict the long-term stabil-ity at various temperatures. The results suggestgood stability for the two freeze-dried candidates,XX and YY, with the loss per year at −20°Cpredicted to be less than 0.2% and the predictedloss per month at B0.6% at +4°C and around3–4% at +20°C. These results show that a shorttime at elevated temperatures during shipmentshould not cause unacceptable loss of potency.

The discussions outlined above suggest thatcandidate YY would be the most suitable to serveas the International Standard. It also has anadvantage in being based on plasma from a natu-rally occurring HIV-1 infection and would there-fore simulate most closely the field situation. Thedata from the study were presented to a meetingof the WHO International Working Group (So-GAT) and the selection of candidate YY wasendorsed. In October 1999 the WHO ExpertCommittee on Biological Standardisation (ECBS),noted the results of the collaborative study andestablished candidate YY as the 1st InternationalStandard for HIV-1 RNA and assigned an activ-ity of 100 000 IU per vial.

Acknowledgements

We wish to thank the EU Programme EVA/MRC Centralised Facility for AIDS Reagents,NIBSC, UK, for distributing the panels and JanetBootman for her help and support in setting upthe study. We are especially grateful to all theparticipants in the collaborative study (see Ap-pendix A) without whom it could not have beenundertaken. This work was supported in part by agrant from the World Health Organisation.

Page 9: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150 149

Appendix A. Participants in the collaborative study

Dr F. Araujo, Hospital S. Joao, Porto, Portugal.Dr S. Best/Dr K. McGavin, National Serology Reference Laboratory, Victoria, Australia.Dr S. Bosbach, Swiss National Center for Retroviruses, University of Zurich, Switzerland.Dr A. Cleland/Dr P. Simmonds, University of Edinburgh Medical School, Edinburgh, UK.Dr C. Defer, Etablissement de Transfusion Sanguine, Lille Cedex, France.Dr K. Fransen, Institute of Tropical Medicine, Antwerp, Belgium.Dr T. Gierman, Bayer Corporation, Clayton, NC, USA.Dr T. Hammerle, Biomedical Research Center, Orth/Donau, Austria.Dr I. Held, BSPI, Vienna, Austria.Dr I. Hewlett, FDA/CBER, Rockville, MD, USA.Dr H. Holmes/Ms C Davis, NIBSC, Potters Bar, Hertfordshire, UK.Dr R. Kaiser, Universitat Bonn, Institut fur Medizinische Mikrobiologie, Bonn, Germany.Dr S. Kaye/Dr R. Tedder, UCL Medical School, London, UK.Dr T. Cuypers/Dr N. Lelie, CLB, Plesmanlaan, Amsterdam, The Netherlands.Dr M. McMorrow, B.V. Chiron, Amsterdam Zuidoost, The Netherlands.Dr L. Mimms/Dr C. Giachetti, Gen-Probe Incorporated, San Diego, California, USA.Dr T. Murozuka, Japanese Red Cross Plasma Fractionation Center, Hokkaido, Japan.Dr M. Nubling, Paul-Ehrlich-Institut, Langen, Germany.Dr J. Robinson, Abbott Laboratories, Abbott Park, IL, USA.Dr C. Rouzioux/Dr M. Burgard, Hopital Necker, Paris, France.Dr L. Sawyer/Dr B. McGovern, Chiron Diagnostics, Emeryville, California, USA.Dr R. Sun/Dr J. Novotny, Roche Molecular Systems Inc, Somerville, New Jersey, USA.Dr P. van de Wiel, Organon Teknika BV, Boxtel, The Netherlands.Dr T. Weimer, Centron Pharma GmbH, Marburg, Germany.Dr U. Wirthmueller, ZLB Central Laboratory SRC, Bern, Switzerland.Dr G. Zerlauth/Dr M. Gessner, Immuno AG, Vienna, Austria.

References

Berry, N., Tedder, R., 1999. HIV-1 and HIV-2 moleculardiagnosis. HIV New viruses Second Edn. 13, 207–222.

Bootman, J., Heath, A.B., Hughes, P., Holmes, H.C., 1999.An international collaborative study on the detection of anHIV-1 genotype B field isolate by nucleic acid amplifica-tion techniques. J. Virol. Meth. 78, 21–34.

Busch, M.P., Lee, L.L.L., Satten, G.A., Henrard, D.R.,Farzadegan, H., Nelson, K.E., Read, S., Dodd, R.Y.,Petersen, L.R., 1995. Time course of detection of viral andserological markers preceding human immunodeficiencyvirus type 1 seroconversion: implications for screening ofblood and tissue donors. Transfusion 35, 91–97.

Carpenter, C.C.J., Fischl, M.A., Hammer, S.M., Hirsch, M.S.,Jacobsen, D.M., Katzenstein, D.A., Montaner, J.S., Rich-man, D.D., Saag, M.S., Schooley, R.T., Thompson, M.A.,Vella, S., Yeni, P.G., Volderding, P.A., 1996. Antiretrovi-ral therapy for HIV infection in 1996. J. Am. Med. Assoc.276, 146–154.

Collet, D., 1991. Modelling Binary Data. Chapman Hall,London.

Coste, J., Montes, B., Reynes, J., Peeters, M., Segarra, C.,Vendrell, J.-P., Delaporte, E., Segondy, M., 1996. Compar-ative evaluation of three assays for the quantitation ofhuman immunodeficiency virus type 1 RNA in plasma. J.Med. Virol. 50, 293–302.

Donaldson, Y.K., Bell, J.E., Holmes, E.C., Hughes, E.S.,Brown, H.K., Simmonds, P., 1994. In 6i6o distribution andcytopathology of variants of human immunodeficiencyvirus type 1 showing restricted sequence variability in theV3 loop. J Virol. 68, 5991–6005.

Francis, B., Green, M., Payne, C. (Eds.), 1993. The GLIMSystem, Release 4 Manual. Oxford Science Publications,Clarendon Press, Oxford.

Fransen, K., Colebunders, B., Heyndrickx, L., Janssens, W.,van der Groen, G., 1998. How many different HIV-1 viralload test kits do we need? AIDS 12, 230–231.

James, R.C., Mosley, J.W., 1995. Hepatitis C virus transmis-sion by intravenous immunoglobulin. Lancet 346, 374.

Page 10: An international collaborative study to establish the 1st international standard for HIV-1 RNA for use in nucleic acid-based techniques

H. Holmes et al. / Journal of Virological Methods 92 (2001) 141–150150

Kirkwood, T.B.L., 1977. Predicting the stability of biologicalstandards and products. Biometrics 33, 736–742.

Kirkwood, T.B., Tydeman, M.S., 1984. Design and analysis ofaccelerated degradation tests for the stability of biologicalstandards. J. Biol. Stand. 12, 195–224.

Mellors, J.W., Rinaldo, C.R., Gupta, P., White, R.M., Todd,J.A., Kingsley, L.A., 1996. Prognosis in HIV-1 infectionpredicted by the quantity of virus in plasma. Science 272,1167–1170.

Mueller-Breitkreutz, K., Allain, J.P., 1999. Sixth EPFA/NIBSC workshop on NAT in Rome 1999: outcome ofdiscussions. Biologicals 27, 349–352.

Revets, H., Marissens, D., De Wit, S., Lacor, P., Clumeck, N.,Lauwers, S., Zissis, G., 1996. Comparative evaluation ofNASBA HIV-1 RNA QT, AMPLICOR-HIV monitor, andQUANTIPLEX HIV RNA assay, three methods for quan-tification of human immunodeficiency virus type 1 RNA inplasma. J. Clin. Microbiol. 34, 1058–1064.

Rogers, P.M., Saldanha, J., Allain, J.P., 1997. Report of theEPFA/NIBSC workshop ‘nucleic acid amplification tests(NAT) for the detection of blood-borne viruses’ held on 31October 1996 in Amsterdam, The Netherlands. Vox Sang

72, 199–206.Saag, M.S., Holidniy, M., Kuritzkes, D.R., O’Brien, W.A.,

Coombs, R., Poscher, M.E., Jacobsen, D.M., Shaw, G.M.,Richman, D.D., Volderding, P.A., 1996. HIV viral loadmarkers in clinical practice. Nat. Med. 2, 625–629.

Schneider, L., Geha, R., 1994. Outbreak of hepatitis C associ-ated with intravenous immunoglobulin administration —United States, October 1993–June 1994. CDC Morbidityand Mortality Weekly Report 43, 505–509.

Schuurman, R., Descamps, D., Weverling, G.J., Kaye, S.,Tijnagel, J., Williams, I., van Leeuwen, Tedder, R.,Boucher, C., Brun-Vezinet, F., Loveday, C.,1996. Multi-center comparison of three commercial methods for quan-tification of human immunodeficiency virus type 1 RNA inplasma. J. Clin. Microbiol. 34, 3016–3022.

Scribner, C., 1994. Dear manufacturer letter, Centre for Bio-logics Evaluation and Research, US Food and Drug Ad-ministration, December 27.

Zoon, K.C., 1995. Dear manufacturer letter, Centre for Bio-logics Evaluation and Research, US Food and Drug Ad-ministration, March 3.

.