- Email: [email protected]
Validation and standardisation of nucleic acid amplification technology (NAT) assays for the detection of viral contamination of blood and blood products
Journal of Clinical Virology 20 (2001) 7 – 13 www.elsevier.com/locate/jcv
Review
Validation and standardisation of nucleic acid amplification technology (NAT) assays for the detection of viral contamination of blood and blood products John Saldanha * Di6ision of Virology, National Institute for Biological Standards and Control, Blanche Lane, South Mimms, South Mimms, Hertshire, EN6 3QG, UK
Abstract Standardisation of NAT assays is necessary before the introduction of such assays for routine screening of blood and blood products for viral contaminants such as HBV, HCV, and HIV-1. Standardisation can be achieved by the use of well-characterised reference materials (working reagents) to validate each assay run. Working reagents for HCV, HIV-1, HBV, HAV, and human parvovirus B19 have been established by the NIBSC. Such reagents and reference panels are also available from other official medicinal control laboratories and commercial organisations. However, the nucleic acid content of these reagents are expressed in many different units, e. g. genome equivalents/ml, copies/ml, PCR detectable units/ml, making comparisons of results from laboratories using different reagents difficult. The establishment of internationally accepted standards against which all working reagents could be calibrated, using a common standard unit of measurement, IU, would overcome this major problem. The first International Standard for HCV RNA assays was established in 1997. This reagent, 96/790, is a lyophilised preparation of a genotype 1 isolate and the concentration of the standard is 105 IU/ml. Two further International Standards have since been established; for HIV-1 and HBV, containing 105 IU/ml and 106 IU/ml respectively. The establishment of the HCV International Standard has been critical in the introduction of mandatory testing. Since 1st July 1999, all batches of blood products marketed in Europe have to be prepared from plasma pools tested and found non-reactive for HCV RNA using a validated assay which can detect a sample containing 100 IU/ml of HCV RNA. In Germany, screening of blood donations for HCV RNA by NAT has been mandatory since 1st April 1999. The minimum sensitivity of assays should be 5000 IU/ml for a single donation. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Nucleic acid testing; Standardisation; Validation; Blood; Blood products; Reference panel
Abbre6iations: HAV, hepatitis A virus; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCV, hepatitis C virus; HIV-1, human immunodeficiency virus-1; IU, international units; NAT, nucleic acid testing; NIBSC, National Institute for Biological Standards and Control; PCR, polymerase chain reaction. * Corresponding author. Tel.: +44-1707-654753; fax: +44-1707-646730. E-mail address: [email protected] (J. Saldanha). 1386-6532/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 6 5 3 2 ( 0 0 ) 0 0 1 4 9 - 9
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
8
1. Introduction Currently, the safety of blood and blood products is ensured by careful selection of donors, screening of donations and the use of validated viral inactivation/removal steps during the manufacture of blood products. All donations are screened for anti-HCV, anti-HIV-1, and HBsAg. However, occasional transmissions of virus still occur despite these measures due to the inclusion of ‘window-period’ donations (i.e., blood from recently infected donors who are antibody negative but still viremic). Screening of plasma pools and donations by sensitive NAT, such as the PCR assay, would identify these ‘window-period’ donations. The projected residual risk of transmission by donors in the ‘window-period’ could be reduced by 45–72% for HIV-1, HBV and HCV by NAT screening (Schreiber et al., 1996). NAT screening reduces the ‘window-period’, but does not necessarily completely eliminate it. NAT assays are very sensitive; a single nucleic acid molecule may be amplified a million times. The high sensitivity of these assays often results in false positive results and conversely, small variations in assay runs are amplified and could result in false negative results. Thus the sensitivity and specificity of NAT assays can vary from laboratory to laboratory, making comparison of results difficult. Between 1992 and 1994, the Eurohep group ran three NAT quality control studies, two for HCV RNA NAT assays and one for HBV DNA NAT assays (Zaaijer et al., 1993; Gerlich et al., 1995; Damen et al., 1996). The results showed Table 1 NIBSC collaborative studies Date
Study
1990 1992 1994 1996 1996 1996 1997 1998 1998
HIV-1 DNA (Study 1) HIV-1 DNA (Study 2) HCV RNA HIV-1 RNA Human parvovirus B19 DNA HAV RNA HCV RNA (International Standard) HBV DNA (International Standard) HIV-1 RNA (International Standard)
that 55–61% of participants were not proficient in NAT assays. Similarly, between 1991 and 1994 collaborative studies organised by NIBSC for HIV-1 and HCV NAT assays showed great variation in sensitivity and specificity of assays from different laboratories (Bootman and Kitchin, 1992, 1994; Saldanha and Minor, 1996). One discrepant result between, for example, a blood product manufacturer and an official medicinal control laboratory (OMCL), could lead to a delay in the release of a product. Therefore, before routine NAT screening can be introduced, it is necessary to standardise and validate the assays. Standardisation can be achieved by the use of calibrated working reagents (run controls) to ensure reproducibility between assay runs. In order to be able to compare results from different laboratories, it is necessary to establish a commonly accepted international standard against which all reference reagents can be calibrated.
2. Development of working reagents At NIBSC, the development of a working reagent involves a collaborative study including 15–25 laboratories (blood product manufacturers, kit manufacturers, OMCLs, diagnostic laboratories and reference laboratories) in order to determine interlaboratory variations in sensitivity and specificity. A working reagent is then prepared, based on the results of the collaborative study. The working reagent is usually equivalent to the highest dilution of sample detected by the majority of participants in the collaborative study. The NIBSC collaborative studies to establish working reagents are listed in Table 1. Based on these collaborative studies, several working reagents have been established (Table 2). These reagents are sent to laboratories on request. Laboratories are requested to test three dilutions of the reagent in each assay; neat, 1:10 and 1:100 dilutions.
3. Development of WHO International Standards At present, working reagents and reference panels are available from many different institutes
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13 Table 2 NIBSC working reagents for NAT assays Reagents
Units
HCV 98/576 HAV 97/540 Human parvovorus B19 99/736 HBV 98/780 HIV-1 PWS-1 HIV-1 PWS-2 Multipex 99/732
710 IU/ml 2000 genome equivalents/ml (approx.) 1000 genome equivalents/ml (approx.)
500 IU/ml 2600 genome equivalents/ml (approx.) 26 000 genome equivalents/ml (approx.) HIV-1, HCV, HBV, HAV, human parvovirus B19
Table 3 WHO International Standards for nucleic acid testing assays HCV International Standard 96 /790 Lyophilised preparation of plasma containing HCV genotype 1a isolate Each vial contains 50 000 IU HBV International Standard 97 /746 Lyophilised preparation of plasma containing HBV genotype A, HBsAg subtype adw Each vial contains 500 000 IU HIV-1 International Standard 97 /656 Lyophilised preparation of HIV-positive donation diluted in defibrinated plasma (FDA/CBER material) Each vial contains 100 000 IU
and laboratories [Paul Ehrlich Institute (PEI), Germany; Istituto Superiore di Sanita (ISS), Italy; CLB, The Netherlands; Center for Biologics Evaluation and Research (CBER), USA; European Pharmacopoeia (EP), Europe; BBI, USA; Bioclinical Partners, USA]. The RNA content of these reagents is expressed in many different units; genome equivalents/ml, copies/ml, PCR detectable units/ml, making comparison of results very difficult. This problem may be overcome by establishing an International Standard, which provides a common biological unit of measurement, the International Unit (IU). All other working reagents can then be calibrated against the International Standard and assigned a concentration in IU/ml. The establishment of an International Standard involves a collaborative study to assess the suitability of candidate materials as the International Standard and to determine the nu-
9
cleic acid content of the chosen material. In general International Standards should be lyophilised preparations with proven long-term stability. There should also be an adequate supply of the material (to last at least 5–10 years) and the collaborative study done to establish an International Standard should include as many laboratories worldwide as possible. The first International Standard was established for HCV RNA NAT assays (96/790) in 1997 (Table 3) (Saldanha et al., 1999). This standard consists of a lyophilised preparation of a HCV genotype 1 isolate at a concentration of 105 IU/ml. Each vial contains 50 000 IU. In 1999, two further International Standards were established; the first International Standard for HBV DNA assays, 97/746 and the first International Standard for HIV-1 RNA assays, 97/656 (Table 3) (WHO International Working Group on the Standardisation of Genomic Amplification Techniques, 1999). The concentrations of these two reagents are 106 and 105 IU/ml respectively. For the establishment of the HBV International Standard, 97/746, 22 laboratories, using several different NAT assays, participated in a collaborative study (Saldanha et al.). These laboratories included reference laboratories, blood product manufacturers, kit manufacturers, regulatory authorities and research laboratories and covered nine different countries. Three candidate materials were analysed; two lyophilised preparations, AA and BB, and a liquid preparation, CC. Laboratories were requested to assay the materials by end-point dilution and the results of four separate assays were used to determine the titre of the three materials for each laboratory using the method of maximum likelihood (Collet, 1991). The calculations were done using the GLIM statistical package (Francis et al., 1993). Laboratories mainly used in-house NAT methods although a few laboratories used the commercial Roche Amplicor™ assay. Three laboratories performed quantitative NAT assays and these were also included in the final analysis. The variation in estimates (log10 PCR detectable units/ml) for the three materials covered 1.25 log10 to 2.75 log10. The overall mean estimates for the samples from
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
10
all the assays, both qualitative and quantitative, are shown in Table 4. Sample AA was proposed as the HBV International Standard and assigned a titre of 106 IU/ml. The proposal was accepted at the meeting of the WHO Expert Committee on Biological Standardisation (ECBS) in October 1999, and material AA was established as the first International Standard for HBV DNA assays. The HBV International Standard, 97/746, consists of a batch of approximately 2000 vials containing a lyophilised preparation of a positive HBV donation (genotype A, HBsAg subtype adw) diluted in HBV-negative cryosupernatant. Each vial contains the equivalent of 0.5 ml of material, which is 500 000 IU. The stability of the International Standard was demonstrated determining the titers of vials incubated at +4, + 20, +37 and + 45°C for different periods of time and comparing these titers with the titers of vials stored at − 20°C (Table 5). The preliminary results indicated that the material is stable at + 4 and +20°C for 18 weeks and at + 37°C for at least 8 weeks, after which it was not possible to redissolve the material. Similarly, after 1 week at + 45°C, the material could not be redissolved.
4. Calibration of HCV working reagents against the HCV International Standard With the establishment of the first International Standard for HCV RNA assays in 1997, it was possible to calibrate several HCV working reagents against this standard. A collaborative study involving 19 laboratories was done in 1999 in order to calibrate working reagents from NIBSC (reagent 96/586), PEI (reference 75), CLB (Pelispy), ISS (ISS 0498), and CBER (panel member no.1) against the International Standard (Saldanha et al., 2000). The titers of the materials and the International Standard were determined in parallel by end-point dilution assays as described previously for the HBV International Standard. The titers of the reagents in IU/ml are shown in Table 6. The titers varied from 250 IU/ml (CBER panel member no. 1) to 25 000 IU/ml (PEI reference 75). Three laboratories in the collaborative study submitted quantitative data and it was possible from these data to estimate the ratio of genome equivalents to IU (Table 7). This ratio varied from 8.3:1 to 3.3:1 and is dependent on the type of assay used. Thus, although the titer of the
Table 4 Overall mean estimates (log10 ‘equivalents’/ml) for the three HBV candidate samples Sample
AA end-point BB end-point CC end-point
Number of assays
26 26 25
Log10 ‘equivalents’/ml Mean
Maximum
Minimum
Range
6.42 6.30 5.03
7.04 7.06 5.65
5.67 4.79 2.91
1.34 2.27 2.74
Table 5 Calculated mean titers of samples (log10 IU/ml) (mean of at least four assays) in the accelerated degradation study compared with the mean titer of the HBV candidate standard AA (stored at −20°C) Incubation time (weeks)
1 week 4 weeks 8 weeks 18 weeks a
ND, not done.
Incubation temperature −20°C
+4°C
+20°C
+37°C
+45°C
6.04 6.04 6.04 6.04
NDa 5.92 ND 5.92
ND 6.06 ND 5.99
6.15 6.01 6.25 –
6.20 – – –
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
11
Table 6 Calibration of HCV working reagents against the WHO HCV International Standard Samplea
Log10 PCR detectable units or genome equivalents/ml
Log10 IU/ml
International standard PEI reference 75 NIBSC 96/586 CLB Pelispy ISS 0498 CBER panel member no.1
5.32 4.72 3.15 3.30 3.53 2.68
5.00 4.40 2.85 3.01 3.23 2.39
a b
IU/mlb 10 0000 25 000 710 1000 1700 250
PEI, Paul Ehrlich Institute; ISS, Istituto Superiore di Sanita; CBER, Center for Biologics Evaluation and Research. to 2 significant figures.
Table 7 Ratio of log10 (genome equivalents/ml to IU/ml) for individual HCV samples assayed by quantitative assays Samplea
Log10 (genome equivalents/ml to IU/ml) Lab. 3 (bDNA)
Lab. 12 (COBAS)
Lab. 9 (in-house)
Mean
International standard PEI reference 75 NIBSC 96/586 CLB Pelispy ISS 0498 CBER panel member no.1
0.83 1.00 – – – –
0.71 0.66 0.92 0.74 0.83 0.76
0.56 0.44 0.17 0.80 0.62 −0.25
0.70 0.70 0.55 0.77 0.73 0.26b
Mean
0.92
0.77
0.39c
a
PEI, Paul Ehrlich Institute; ISS, Istituto Superiore di Sanita; CBER, Center for Biologics Evaluation and Research. Mean= log10 0.76 if result from lab. 9 excluded. c Mean =log10 0.52 if result for CBER panel no. 1 excluded. b
International Standard has been fixed at 105 IU/ ml, the titers obtained by different assays, and expressed in different units, will vary, depending on the assay. Therefore, the use of IU/ml for the concentration of HCV RNA in a sample will overcome the problem of differences in titers due to different assays and is to be recommended.
5. Quality control of NAT assays NAT assays should be validated before implementation of routine testing. Routine quality control is achieved by the use of calibrated working reagents in each assay run, while participation in proficiency studies will ensure the reproducibility of assays and the ability of a laboratory in performing such assays. Qualitative NAT assays can be considered a limit test for impurities and there-
fore the main characteristics to be defined are the sensitivity (limit of detection), the specificity and robustness of the assay. Guidelines for determining these characteristics have been published (European Department for the Quality of Medicines 1999). Each NAT assay should include an internal control to control for the extraction, reverse transcription, amplification and detection steps. Addition of a different virus e.g. bovine viral diarrhoea virus (BVDV) will control all steps of the assay, whereas addition of an RNA transcript to the extraction mixture will not control for the initial virus lysis step. In addition to the internal control, each assay run should also include an external control, which usually is a calibrated working reagent (run control). For the assay to be valid, the run control should be positive. Proficiency panels are available from many different sources, control agencies such as NIBSC, CBER and
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
12
EDQM, as well as commercial organisations, e. g. BBI, CLB and Bioclinical Partners. Proficiency studies are performed once or twice a year and can be considered as quality control studies in which the consistency of performances of a particular assay and laboratory are assessed. Proficiency panels usually consist of a set of samples containing a dilution series of HCV RNA and several HCV-negative samples. The first NIBSC proficiency panel of 20 samples consisted of two dilution series of HCV genotypes one and three and four HCV-negative samples (Table 8). Six UK laboratories participated in this study and the results show that all laboratories (except laboratory 6) could detect 50 IU/ml of either genotype one or three. None of the laboratories obtained a false positive result.
Table 8 Results of the NIBSC HCV proficiency panel from six laboratories Sample
Genotype 1 10 000 IU/ml 2000 IU/ml 200 IU/ml 200 IU/ml 100 IU/ml 50 IU/ml 10 IU/ml 2 IU/ml 0.2 IU/ml Genotype 3 96 IU/ml 96 IU/ml 48 IU/ml 24 IU/ml 10 IU/ml 1 IU/ml 0.1 IU/ml Diluent A B C D
Laboratory 1
2
3
4
5
6
+ + + + + + + +a −
+ + + + + + +a − −
+ + + + + + − − −
+ + + + + + − + −
+ + + + + + + − −
+ + + + + + − − −
+ + + + + +a −
+ + +a + − +a −
+ + + + − − −
+ + + + + + −
+ + + + + − −
+ + − − − + −
− − − −
− − − −
− − − −
− − − −
− − − −
− − − −
a Laboratories 1 and 2 performed the assay in duplicate and one of the replicate samples was negative.
6. Introduction of HCV RNA testing In 1994, CBER introduced testing of all immunoglobulins that had not undergone one or more viral inactivation steps, in response to reports on the transmission of HCV by intravenous immunoglobulins (IVIGs) (Center for Disease Control 1994; Scribner, 1994; Zoon, 1995). In March 1994, the Committee for Proprietary Medicinal Products (CPMP) in Europe urged marketing authorisation holders to develop and validate additional viral inactivation/removal steps for plasma products (Committee for Proprietary Medicinal Products). After careful consideration, the CPMP recommended that from 1st July 1999, only batches of blood products derived from plasma pools tested and found non-reactive for HCV RNA, using validated NAT methods of suitable sensitivity and specificity, should be batch released by the Marketing Authorisation holder. Each assay run should be able to detect a run control with an HCV RNA content equivalent to 100 IU/ml (Committee for Proprietary Medicinal Products, 1997). The introduction of HCV RNA testing in plasma pools also required the establishment of NAT assays for individual donations as the positive donation(s) responsible for a reactive pool had to be identified with consequent implications for release of cellular components. Many blood banks initiated testing of donations (Schottstedt et al., 1997; Roth et al., 1998; Da Silva Cardoso et al., 1998; Wolff et al., 1998) and the Paul Ehrlich Institute, Germany, proposed that from 1st April 1999, HCV RNA testing should be introduced for the release of erythrocyte and thrombocyte concentrates in Germany. The required minimal sensitivity for single donation testing is 5000 IU/ml (Nu¨bling et al., 1998). The Austrian Red Cross introduced similar regulations. Testing of single donations (in a mini-pool format) has begun in the UK, but is not yet mandatory.
7. Conclusions Working reagents and standards are required to validate individual NAT assay runs and are essen-
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
tial for the introduction of mandatory testing of samples by NAT and for mutual acceptance of results. Several working reagents are available from different laboratories and three WHO International Standards, for HCV, HIV-1 and HBV NAT assays, have been established. The International Standards are calibrated in International Units and should provide a basis for quantification and comparison of results from different laboratories. The introduction of mandatory HCV RNA testing for plasma pools used in the manufacture of blood products and for single donations for the release of cellular components has been made possible by the establishment of the WHO HCV International Standard.
References Bootman JS, Kitchin PA. An international collaborative study to assess a set of reference reagents for HIV-1 PCR. J Virol Methods 1992;37:3 – 42. Bootman JS, Kitchin PA. Reference preparations in the standardisation of HIV-1 PCR — an international collaborative study. J Virol Methods 1994;49:1–8. Center of Disease Control. Outbreak of hepatitis C associated with intravenous immunoglobulin administration. US October 1993 – June 1994, MMRW, 1994; 43: 505-509. Collet D. Modelling binary data. London: Chapman & Hall, 1991. Committee for Proprietary Medicinal Products. Intramuscular immunoglobulins; nucleic acid amplification tests for HCV RNA detection. Recommendation CPMP/117/95. Committee for Proprietary Medicinal Products. Introduction of Genomic Amplification Technology (GAT) for the Detection of Hepatitis C Virus RNA in Plasma Pools. Addendum to Notes for Guidelines on Plasma derived products 1997. CPMP/BWP/390/97. Damen M, Zaaijer H, Reesink H, Schaaberg W, Gerlich W, Niesters H, Lelie N. International collaborative study on the second Eurohep HCV-RNA reference panel. J Virol Methods 1996;58:175 –85. Da Silva Cardoso M, Ko¨rner K, Kubanek B. PCR screening in the routine of blood banking of the German Red Cross Blood Transfusion service of Baden-Wu¨rttemberg. Infus Ther Trans Med 1998;25:116–20. European Department for the Quality of Medicines. Biological Substances 1999. Validation of nucleic acid amplification
.
13
technology for the detection of hepatitis C virus RNA in plasma pools, 1999; PA/PH/OMCL(98) 22, DEF. Francis B, Green M, Payne C. In: Francis B, Green M, Payne C, editors. The GLIM system, release 4 manual, Oxford Science Publications. Oxford: Clarendon Press, 1993. Gerlich WH; Heermann KH; Thomssen R; Eurohep group. Quantitative assays for hepatitis B virus DNA: standardisation and quality control, Viral Hep Rev, 1995; 1: 53-57. Nu¨bling CM, Seitz R, Lo¨wer J. Application of nucleic acid amplification techniques for blood donation screening. Infus Ther Trans Med 1998;25:86 – 90. Roth WK, Weber M, Buhr S, Seifried E. Rapid routine PCR screening system for blood banks. Infus Ther Trans Med 1998;25:115. Saldanha J, Minor P. Collaborative study to assess the suitability of an HCV RNA reference sample for detection of HCV RNA in plasma pools by PCR. Vox Sang 1996;70:148 – 51. Saldanha J, Lelie N, Heath A, Collaborative Study Group WHO. Establishment of the first International Standard for nucleic acid amplification technology assays for HCV RNA. Vox Sang 1999;76:149 – 58. Saldanha J; Heath A; Lelie N; Pisani G; Nu¨bling M; Yu M; Collaborative Study Group. Calibration of HCV Working Reagents for NAT assays against the HCV International Standard, Vox Sang, 2000 (in press). Saldanha J;, Gerlich W; Lelie N; Dawson P; Heermann K; Heath A. An International Collaborative Study to Establish a WHO International Standard for Hepatitis B Virus, DNA NAT Assays (submitted). Schottstedt V, Tuma W, Bunger G. Identification of a single viraemic donation with PCR-minipool testing in a large blood bank. Trans Med 1997;7:14. Schreiber GB, Busch MP, Kleinman SH, Korelitz JJ. The risk of transfusion-transmitted viral infections. New Eng J Med 1996;334:1685– 90. Scribner, C., 1994. Dear manufacturer letter. Center of Biologics Evaluation and Research, December 27, 1994. WHO International Working Group on the Standardisation of Genomic Amplification Techniques for the Virological Safety Testing of Blood and Blood Products. Minutes of the tenth meeting, National Institute for Biological Standards and Control, South Mimms, UK, 29th October 1999. Wolff C, Boomgaarden M, Ko¨ster-Eiserfunke W, Kleesiek K. Two years of experience with PCR screening of individual blood donations. Infus Ther Trans Med 1998;25:128 – 34. Zaaijer H, Cuypers H, Reesink H, Winkel I, Gerken G, Lelie P. Reliability of polymerase chain reaction for detection of hepatitis C virus. Lancet 1993;341:722 – 4. Zoon KC. Dear manufacturer letter. Center of Biologics Evaluation and Research, March 3, 1995.
Review
Validation and standardisation of nucleic acid amplification technology (NAT) assays for the detection of viral contamination of blood and blood products John Saldanha * Di6ision of Virology, National Institute for Biological Standards and Control, Blanche Lane, South Mimms, South Mimms, Hertshire, EN6 3QG, UK
Abstract Standardisation of NAT assays is necessary before the introduction of such assays for routine screening of blood and blood products for viral contaminants such as HBV, HCV, and HIV-1. Standardisation can be achieved by the use of well-characterised reference materials (working reagents) to validate each assay run. Working reagents for HCV, HIV-1, HBV, HAV, and human parvovirus B19 have been established by the NIBSC. Such reagents and reference panels are also available from other official medicinal control laboratories and commercial organisations. However, the nucleic acid content of these reagents are expressed in many different units, e. g. genome equivalents/ml, copies/ml, PCR detectable units/ml, making comparisons of results from laboratories using different reagents difficult. The establishment of internationally accepted standards against which all working reagents could be calibrated, using a common standard unit of measurement, IU, would overcome this major problem. The first International Standard for HCV RNA assays was established in 1997. This reagent, 96/790, is a lyophilised preparation of a genotype 1 isolate and the concentration of the standard is 105 IU/ml. Two further International Standards have since been established; for HIV-1 and HBV, containing 105 IU/ml and 106 IU/ml respectively. The establishment of the HCV International Standard has been critical in the introduction of mandatory testing. Since 1st July 1999, all batches of blood products marketed in Europe have to be prepared from plasma pools tested and found non-reactive for HCV RNA using a validated assay which can detect a sample containing 100 IU/ml of HCV RNA. In Germany, screening of blood donations for HCV RNA by NAT has been mandatory since 1st April 1999. The minimum sensitivity of assays should be 5000 IU/ml for a single donation. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Nucleic acid testing; Standardisation; Validation; Blood; Blood products; Reference panel
Abbre6iations: HAV, hepatitis A virus; HBsAg, hepatitis B surface antigen; HBV, hepatitis B virus; HCV, hepatitis C virus; HIV-1, human immunodeficiency virus-1; IU, international units; NAT, nucleic acid testing; NIBSC, National Institute for Biological Standards and Control; PCR, polymerase chain reaction. * Corresponding author. Tel.: +44-1707-654753; fax: +44-1707-646730. E-mail address: [email protected] (J. Saldanha). 1386-6532/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 6 5 3 2 ( 0 0 ) 0 0 1 4 9 - 9
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
8
1. Introduction Currently, the safety of blood and blood products is ensured by careful selection of donors, screening of donations and the use of validated viral inactivation/removal steps during the manufacture of blood products. All donations are screened for anti-HCV, anti-HIV-1, and HBsAg. However, occasional transmissions of virus still occur despite these measures due to the inclusion of ‘window-period’ donations (i.e., blood from recently infected donors who are antibody negative but still viremic). Screening of plasma pools and donations by sensitive NAT, such as the PCR assay, would identify these ‘window-period’ donations. The projected residual risk of transmission by donors in the ‘window-period’ could be reduced by 45–72% for HIV-1, HBV and HCV by NAT screening (Schreiber et al., 1996). NAT screening reduces the ‘window-period’, but does not necessarily completely eliminate it. NAT assays are very sensitive; a single nucleic acid molecule may be amplified a million times. The high sensitivity of these assays often results in false positive results and conversely, small variations in assay runs are amplified and could result in false negative results. Thus the sensitivity and specificity of NAT assays can vary from laboratory to laboratory, making comparison of results difficult. Between 1992 and 1994, the Eurohep group ran three NAT quality control studies, two for HCV RNA NAT assays and one for HBV DNA NAT assays (Zaaijer et al., 1993; Gerlich et al., 1995; Damen et al., 1996). The results showed Table 1 NIBSC collaborative studies Date
Study
1990 1992 1994 1996 1996 1996 1997 1998 1998
HIV-1 DNA (Study 1) HIV-1 DNA (Study 2) HCV RNA HIV-1 RNA Human parvovirus B19 DNA HAV RNA HCV RNA (International Standard) HBV DNA (International Standard) HIV-1 RNA (International Standard)
that 55–61% of participants were not proficient in NAT assays. Similarly, between 1991 and 1994 collaborative studies organised by NIBSC for HIV-1 and HCV NAT assays showed great variation in sensitivity and specificity of assays from different laboratories (Bootman and Kitchin, 1992, 1994; Saldanha and Minor, 1996). One discrepant result between, for example, a blood product manufacturer and an official medicinal control laboratory (OMCL), could lead to a delay in the release of a product. Therefore, before routine NAT screening can be introduced, it is necessary to standardise and validate the assays. Standardisation can be achieved by the use of calibrated working reagents (run controls) to ensure reproducibility between assay runs. In order to be able to compare results from different laboratories, it is necessary to establish a commonly accepted international standard against which all reference reagents can be calibrated.
2. Development of working reagents At NIBSC, the development of a working reagent involves a collaborative study including 15–25 laboratories (blood product manufacturers, kit manufacturers, OMCLs, diagnostic laboratories and reference laboratories) in order to determine interlaboratory variations in sensitivity and specificity. A working reagent is then prepared, based on the results of the collaborative study. The working reagent is usually equivalent to the highest dilution of sample detected by the majority of participants in the collaborative study. The NIBSC collaborative studies to establish working reagents are listed in Table 1. Based on these collaborative studies, several working reagents have been established (Table 2). These reagents are sent to laboratories on request. Laboratories are requested to test three dilutions of the reagent in each assay; neat, 1:10 and 1:100 dilutions.
3. Development of WHO International Standards At present, working reagents and reference panels are available from many different institutes
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13 Table 2 NIBSC working reagents for NAT assays Reagents
Units
HCV 98/576 HAV 97/540 Human parvovorus B19 99/736 HBV 98/780 HIV-1 PWS-1 HIV-1 PWS-2 Multipex 99/732
710 IU/ml 2000 genome equivalents/ml (approx.) 1000 genome equivalents/ml (approx.)
500 IU/ml 2600 genome equivalents/ml (approx.) 26 000 genome equivalents/ml (approx.) HIV-1, HCV, HBV, HAV, human parvovirus B19
Table 3 WHO International Standards for nucleic acid testing assays HCV International Standard 96 /790 Lyophilised preparation of plasma containing HCV genotype 1a isolate Each vial contains 50 000 IU HBV International Standard 97 /746 Lyophilised preparation of plasma containing HBV genotype A, HBsAg subtype adw Each vial contains 500 000 IU HIV-1 International Standard 97 /656 Lyophilised preparation of HIV-positive donation diluted in defibrinated plasma (FDA/CBER material) Each vial contains 100 000 IU
and laboratories [Paul Ehrlich Institute (PEI), Germany; Istituto Superiore di Sanita (ISS), Italy; CLB, The Netherlands; Center for Biologics Evaluation and Research (CBER), USA; European Pharmacopoeia (EP), Europe; BBI, USA; Bioclinical Partners, USA]. The RNA content of these reagents is expressed in many different units; genome equivalents/ml, copies/ml, PCR detectable units/ml, making comparison of results very difficult. This problem may be overcome by establishing an International Standard, which provides a common biological unit of measurement, the International Unit (IU). All other working reagents can then be calibrated against the International Standard and assigned a concentration in IU/ml. The establishment of an International Standard involves a collaborative study to assess the suitability of candidate materials as the International Standard and to determine the nu-
9
cleic acid content of the chosen material. In general International Standards should be lyophilised preparations with proven long-term stability. There should also be an adequate supply of the material (to last at least 5–10 years) and the collaborative study done to establish an International Standard should include as many laboratories worldwide as possible. The first International Standard was established for HCV RNA NAT assays (96/790) in 1997 (Table 3) (Saldanha et al., 1999). This standard consists of a lyophilised preparation of a HCV genotype 1 isolate at a concentration of 105 IU/ml. Each vial contains 50 000 IU. In 1999, two further International Standards were established; the first International Standard for HBV DNA assays, 97/746 and the first International Standard for HIV-1 RNA assays, 97/656 (Table 3) (WHO International Working Group on the Standardisation of Genomic Amplification Techniques, 1999). The concentrations of these two reagents are 106 and 105 IU/ml respectively. For the establishment of the HBV International Standard, 97/746, 22 laboratories, using several different NAT assays, participated in a collaborative study (Saldanha et al.). These laboratories included reference laboratories, blood product manufacturers, kit manufacturers, regulatory authorities and research laboratories and covered nine different countries. Three candidate materials were analysed; two lyophilised preparations, AA and BB, and a liquid preparation, CC. Laboratories were requested to assay the materials by end-point dilution and the results of four separate assays were used to determine the titre of the three materials for each laboratory using the method of maximum likelihood (Collet, 1991). The calculations were done using the GLIM statistical package (Francis et al., 1993). Laboratories mainly used in-house NAT methods although a few laboratories used the commercial Roche Amplicor™ assay. Three laboratories performed quantitative NAT assays and these were also included in the final analysis. The variation in estimates (log10 PCR detectable units/ml) for the three materials covered 1.25 log10 to 2.75 log10. The overall mean estimates for the samples from
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
10
all the assays, both qualitative and quantitative, are shown in Table 4. Sample AA was proposed as the HBV International Standard and assigned a titre of 106 IU/ml. The proposal was accepted at the meeting of the WHO Expert Committee on Biological Standardisation (ECBS) in October 1999, and material AA was established as the first International Standard for HBV DNA assays. The HBV International Standard, 97/746, consists of a batch of approximately 2000 vials containing a lyophilised preparation of a positive HBV donation (genotype A, HBsAg subtype adw) diluted in HBV-negative cryosupernatant. Each vial contains the equivalent of 0.5 ml of material, which is 500 000 IU. The stability of the International Standard was demonstrated determining the titers of vials incubated at +4, + 20, +37 and + 45°C for different periods of time and comparing these titers with the titers of vials stored at − 20°C (Table 5). The preliminary results indicated that the material is stable at + 4 and +20°C for 18 weeks and at + 37°C for at least 8 weeks, after which it was not possible to redissolve the material. Similarly, after 1 week at + 45°C, the material could not be redissolved.
4. Calibration of HCV working reagents against the HCV International Standard With the establishment of the first International Standard for HCV RNA assays in 1997, it was possible to calibrate several HCV working reagents against this standard. A collaborative study involving 19 laboratories was done in 1999 in order to calibrate working reagents from NIBSC (reagent 96/586), PEI (reference 75), CLB (Pelispy), ISS (ISS 0498), and CBER (panel member no.1) against the International Standard (Saldanha et al., 2000). The titers of the materials and the International Standard were determined in parallel by end-point dilution assays as described previously for the HBV International Standard. The titers of the reagents in IU/ml are shown in Table 6. The titers varied from 250 IU/ml (CBER panel member no. 1) to 25 000 IU/ml (PEI reference 75). Three laboratories in the collaborative study submitted quantitative data and it was possible from these data to estimate the ratio of genome equivalents to IU (Table 7). This ratio varied from 8.3:1 to 3.3:1 and is dependent on the type of assay used. Thus, although the titer of the
Table 4 Overall mean estimates (log10 ‘equivalents’/ml) for the three HBV candidate samples Sample
AA end-point BB end-point CC end-point
Number of assays
26 26 25
Log10 ‘equivalents’/ml Mean
Maximum
Minimum
Range
6.42 6.30 5.03
7.04 7.06 5.65
5.67 4.79 2.91
1.34 2.27 2.74
Table 5 Calculated mean titers of samples (log10 IU/ml) (mean of at least four assays) in the accelerated degradation study compared with the mean titer of the HBV candidate standard AA (stored at −20°C) Incubation time (weeks)
1 week 4 weeks 8 weeks 18 weeks a
ND, not done.
Incubation temperature −20°C
+4°C
+20°C
+37°C
+45°C
6.04 6.04 6.04 6.04
NDa 5.92 ND 5.92
ND 6.06 ND 5.99
6.15 6.01 6.25 –
6.20 – – –
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
11
Table 6 Calibration of HCV working reagents against the WHO HCV International Standard Samplea
Log10 PCR detectable units or genome equivalents/ml
Log10 IU/ml
International standard PEI reference 75 NIBSC 96/586 CLB Pelispy ISS 0498 CBER panel member no.1
5.32 4.72 3.15 3.30 3.53 2.68
5.00 4.40 2.85 3.01 3.23 2.39
a b
IU/mlb 10 0000 25 000 710 1000 1700 250
PEI, Paul Ehrlich Institute; ISS, Istituto Superiore di Sanita; CBER, Center for Biologics Evaluation and Research. to 2 significant figures.
Table 7 Ratio of log10 (genome equivalents/ml to IU/ml) for individual HCV samples assayed by quantitative assays Samplea
Log10 (genome equivalents/ml to IU/ml) Lab. 3 (bDNA)
Lab. 12 (COBAS)
Lab. 9 (in-house)
Mean
International standard PEI reference 75 NIBSC 96/586 CLB Pelispy ISS 0498 CBER panel member no.1
0.83 1.00 – – – –
0.71 0.66 0.92 0.74 0.83 0.76
0.56 0.44 0.17 0.80 0.62 −0.25
0.70 0.70 0.55 0.77 0.73 0.26b
Mean
0.92
0.77
0.39c
a
PEI, Paul Ehrlich Institute; ISS, Istituto Superiore di Sanita; CBER, Center for Biologics Evaluation and Research. Mean= log10 0.76 if result from lab. 9 excluded. c Mean =log10 0.52 if result for CBER panel no. 1 excluded. b
International Standard has been fixed at 105 IU/ ml, the titers obtained by different assays, and expressed in different units, will vary, depending on the assay. Therefore, the use of IU/ml for the concentration of HCV RNA in a sample will overcome the problem of differences in titers due to different assays and is to be recommended.
5. Quality control of NAT assays NAT assays should be validated before implementation of routine testing. Routine quality control is achieved by the use of calibrated working reagents in each assay run, while participation in proficiency studies will ensure the reproducibility of assays and the ability of a laboratory in performing such assays. Qualitative NAT assays can be considered a limit test for impurities and there-
fore the main characteristics to be defined are the sensitivity (limit of detection), the specificity and robustness of the assay. Guidelines for determining these characteristics have been published (European Department for the Quality of Medicines 1999). Each NAT assay should include an internal control to control for the extraction, reverse transcription, amplification and detection steps. Addition of a different virus e.g. bovine viral diarrhoea virus (BVDV) will control all steps of the assay, whereas addition of an RNA transcript to the extraction mixture will not control for the initial virus lysis step. In addition to the internal control, each assay run should also include an external control, which usually is a calibrated working reagent (run control). For the assay to be valid, the run control should be positive. Proficiency panels are available from many different sources, control agencies such as NIBSC, CBER and
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
12
EDQM, as well as commercial organisations, e. g. BBI, CLB and Bioclinical Partners. Proficiency studies are performed once or twice a year and can be considered as quality control studies in which the consistency of performances of a particular assay and laboratory are assessed. Proficiency panels usually consist of a set of samples containing a dilution series of HCV RNA and several HCV-negative samples. The first NIBSC proficiency panel of 20 samples consisted of two dilution series of HCV genotypes one and three and four HCV-negative samples (Table 8). Six UK laboratories participated in this study and the results show that all laboratories (except laboratory 6) could detect 50 IU/ml of either genotype one or three. None of the laboratories obtained a false positive result.
Table 8 Results of the NIBSC HCV proficiency panel from six laboratories Sample
Genotype 1 10 000 IU/ml 2000 IU/ml 200 IU/ml 200 IU/ml 100 IU/ml 50 IU/ml 10 IU/ml 2 IU/ml 0.2 IU/ml Genotype 3 96 IU/ml 96 IU/ml 48 IU/ml 24 IU/ml 10 IU/ml 1 IU/ml 0.1 IU/ml Diluent A B C D
Laboratory 1
2
3
4
5
6
+ + + + + + + +a −
+ + + + + + +a − −
+ + + + + + − − −
+ + + + + + − + −
+ + + + + + + − −
+ + + + + + − − −
+ + + + + +a −
+ + +a + − +a −
+ + + + − − −
+ + + + + + −
+ + + + + − −
+ + − − − + −
− − − −
− − − −
− − − −
− − − −
− − − −
− − − −
a Laboratories 1 and 2 performed the assay in duplicate and one of the replicate samples was negative.
6. Introduction of HCV RNA testing In 1994, CBER introduced testing of all immunoglobulins that had not undergone one or more viral inactivation steps, in response to reports on the transmission of HCV by intravenous immunoglobulins (IVIGs) (Center for Disease Control 1994; Scribner, 1994; Zoon, 1995). In March 1994, the Committee for Proprietary Medicinal Products (CPMP) in Europe urged marketing authorisation holders to develop and validate additional viral inactivation/removal steps for plasma products (Committee for Proprietary Medicinal Products). After careful consideration, the CPMP recommended that from 1st July 1999, only batches of blood products derived from plasma pools tested and found non-reactive for HCV RNA, using validated NAT methods of suitable sensitivity and specificity, should be batch released by the Marketing Authorisation holder. Each assay run should be able to detect a run control with an HCV RNA content equivalent to 100 IU/ml (Committee for Proprietary Medicinal Products, 1997). The introduction of HCV RNA testing in plasma pools also required the establishment of NAT assays for individual donations as the positive donation(s) responsible for a reactive pool had to be identified with consequent implications for release of cellular components. Many blood banks initiated testing of donations (Schottstedt et al., 1997; Roth et al., 1998; Da Silva Cardoso et al., 1998; Wolff et al., 1998) and the Paul Ehrlich Institute, Germany, proposed that from 1st April 1999, HCV RNA testing should be introduced for the release of erythrocyte and thrombocyte concentrates in Germany. The required minimal sensitivity for single donation testing is 5000 IU/ml (Nu¨bling et al., 1998). The Austrian Red Cross introduced similar regulations. Testing of single donations (in a mini-pool format) has begun in the UK, but is not yet mandatory.
7. Conclusions Working reagents and standards are required to validate individual NAT assay runs and are essen-
J. Saldanha / Journal of Clinical Virology 20 (2001) 7–13
tial for the introduction of mandatory testing of samples by NAT and for mutual acceptance of results. Several working reagents are available from different laboratories and three WHO International Standards, for HCV, HIV-1 and HBV NAT assays, have been established. The International Standards are calibrated in International Units and should provide a basis for quantification and comparison of results from different laboratories. The introduction of mandatory HCV RNA testing for plasma pools used in the manufacture of blood products and for single donations for the release of cellular components has been made possible by the establishment of the WHO HCV International Standard.
References Bootman JS, Kitchin PA. An international collaborative study to assess a set of reference reagents for HIV-1 PCR. J Virol Methods 1992;37:3 – 42. Bootman JS, Kitchin PA. Reference preparations in the standardisation of HIV-1 PCR — an international collaborative study. J Virol Methods 1994;49:1–8. Center of Disease Control. Outbreak of hepatitis C associated with intravenous immunoglobulin administration. US October 1993 – June 1994, MMRW, 1994; 43: 505-509. Collet D. Modelling binary data. London: Chapman & Hall, 1991. Committee for Proprietary Medicinal Products. Intramuscular immunoglobulins; nucleic acid amplification tests for HCV RNA detection. Recommendation CPMP/117/95. Committee for Proprietary Medicinal Products. Introduction of Genomic Amplification Technology (GAT) for the Detection of Hepatitis C Virus RNA in Plasma Pools. Addendum to Notes for Guidelines on Plasma derived products 1997. CPMP/BWP/390/97. Damen M, Zaaijer H, Reesink H, Schaaberg W, Gerlich W, Niesters H, Lelie N. International collaborative study on the second Eurohep HCV-RNA reference panel. J Virol Methods 1996;58:175 –85. Da Silva Cardoso M, Ko¨rner K, Kubanek B. PCR screening in the routine of blood banking of the German Red Cross Blood Transfusion service of Baden-Wu¨rttemberg. Infus Ther Trans Med 1998;25:116–20. European Department for the Quality of Medicines. Biological Substances 1999. Validation of nucleic acid amplification
.
13
technology for the detection of hepatitis C virus RNA in plasma pools, 1999; PA/PH/OMCL(98) 22, DEF. Francis B, Green M, Payne C. In: Francis B, Green M, Payne C, editors. The GLIM system, release 4 manual, Oxford Science Publications. Oxford: Clarendon Press, 1993. Gerlich WH; Heermann KH; Thomssen R; Eurohep group. Quantitative assays for hepatitis B virus DNA: standardisation and quality control, Viral Hep Rev, 1995; 1: 53-57. Nu¨bling CM, Seitz R, Lo¨wer J. Application of nucleic acid amplification techniques for blood donation screening. Infus Ther Trans Med 1998;25:86 – 90. Roth WK, Weber M, Buhr S, Seifried E. Rapid routine PCR screening system for blood banks. Infus Ther Trans Med 1998;25:115. Saldanha J, Minor P. Collaborative study to assess the suitability of an HCV RNA reference sample for detection of HCV RNA in plasma pools by PCR. Vox Sang 1996;70:148 – 51. Saldanha J, Lelie N, Heath A, Collaborative Study Group WHO. Establishment of the first International Standard for nucleic acid amplification technology assays for HCV RNA. Vox Sang 1999;76:149 – 58. Saldanha J; Heath A; Lelie N; Pisani G; Nu¨bling M; Yu M; Collaborative Study Group. Calibration of HCV Working Reagents for NAT assays against the HCV International Standard, Vox Sang, 2000 (in press). Saldanha J;, Gerlich W; Lelie N; Dawson P; Heermann K; Heath A. An International Collaborative Study to Establish a WHO International Standard for Hepatitis B Virus, DNA NAT Assays (submitted). Schottstedt V, Tuma W, Bunger G. Identification of a single viraemic donation with PCR-minipool testing in a large blood bank. Trans Med 1997;7:14. Schreiber GB, Busch MP, Kleinman SH, Korelitz JJ. The risk of transfusion-transmitted viral infections. New Eng J Med 1996;334:1685– 90. Scribner, C., 1994. Dear manufacturer letter. Center of Biologics Evaluation and Research, December 27, 1994. WHO International Working Group on the Standardisation of Genomic Amplification Techniques for the Virological Safety Testing of Blood and Blood Products. Minutes of the tenth meeting, National Institute for Biological Standards and Control, South Mimms, UK, 29th October 1999. Wolff C, Boomgaarden M, Ko¨ster-Eiserfunke W, Kleesiek K. Two years of experience with PCR screening of individual blood donations. Infus Ther Trans Med 1998;25:128 – 34. Zaaijer H, Cuypers H, Reesink H, Winkel I, Gerken G, Lelie P. Reliability of polymerase chain reaction for detection of hepatitis C virus. Lancet 1993;341:722 – 4. Zoon KC. Dear manufacturer letter. Center of Biologics Evaluation and Research, March 3, 1995.