The Solid Phase Radioimmunoassay Technique for Hepatitis B (Surface) Antigen: Effects of Progressive Modification L. Y. NI, R. Y. DODD, N. BROWN,R. SALMON, A N D T. J. GREENWALT From thespecial Projects Laborarory. American National Red Cross. Washington. D.C. and the Blood Research Laboratory. Bethesda. Maryland
Serum specimens submitted for confirmation of hepatitis B antigen (HB,Ag) reactivity were tested by counterelectrophoresis (CEP)and by three modifications of the solid phase radioimmunoassay “sandwich” technique (SPRIA). In addition, the specificity of each reaction was determined for two of the SPRIA test methods. One hundred seventy-five of the 688 specimens were reactive by all techniques and a further 60 were specifically reactive by two of modifications of SPRIA. An additional five specimens were positive by a third SPRlA modification. The primary SPRlA technique, using a homologous guinea pig antibody sandwich, detected 184 nonspecific reactive samples. With the same test done at 45C with a reduced incubation time, there were only 37 nonspecific reactive samples. None of the nonspecific samples was detected when the SPRIA technique used a heterologous guinea pig-human antibody sandwich. The results provide a basis for estimating the effectiveness of the SPRlA technique for HBsAg testing.
to false positive reactions caused by antibodies against guinea pig globulins, but it was generally felt that such reactivity would be rare in the general population. However, initial experience with the AusRIA technique did reveal unacceptable levels of false positive reactions, possibly caused by nonspecific antiglobulin r e a ~ t i v i t y .This ~ problem was largely eliminated by addition of normal guinea pig serum to the radiolabeled reagent; nevertheless, a significant proportion of nonspecifically reactive specimens was still This paper compares the specificity of reactions obtained with three versions of AusRIA technique. The results are discussed in the context of a large volunteer blood donor population.
BLOOD CONTAINING hepatitis B surface antigen (HB,Ag) is more likely to transmit type B viral hepatitis than is blood which does not have detectable HBsAg.3*4Therefore, blood collecting agencies routinely test all blood donations for the presence of HB,Ag and discard any that are reactive. HBs.4g reactive blood donors are considered to remain potentially infectious and are permanently deferred. These considerations require that a routine test for HBsAg be sensitive and specific, providing maximum safety for the recipient without unnecessary wastage of the blood resource. Solid phase radioimmunoassay (SPRIA) is currently the most sensitive routine test for HBsAg and a commercially available version* is widely Sgouris’o reported that the AusRIA technique was susceptible
Materials and Methods The 688 serum specimens were all derived from the normal operations of the American National Red Cross Blood Program and represent samples submitted by the Red Cross Regional Blood Centers (RCBCs) during October 1973 for confirmatory testing in our laboratory. Each specimen had been suspected to be reactive on the basis of at least one test by the S P R l A I methodology. The SPRIA** m e t h o d o l ~ g yinvolves ~~~ incubation of 100 ul of serum at room temperature for 16 hours in a polypropylene tube coated with guinea pig antibodies to HB,Ag (anti-HBs). The tube is then washed exhaustively and radiolabeled guinea anti-HB, is added, followed by incubation for 90 minutes at room temperature. Unreacted radiolabeled antibody is washed from the tube and bound antibody is determined by measuring the residual radioactivity retained in the tube. A specimen is regarded as reactive if the count exceeds 2.1 times the mean of seven determinations on a known negative control. This cutoff
Received for publication December I , 1974; accepted February 23, 1975. Publication Number 301 from the American National Red Cross. *AusRIA, Abbott Laboratories. Transfusion Nov.-Dec. 1975
**AusR IA- 125, Abbott Laboratories.
608
Volume I5 Number6
Volume I5 Number6
represents approximately seven standard deviations above the negative control mean. This methodology will be referred to as SPRIA I . The first modification of this technique will be referred to as SPRIA 45. The same reagents are used but the initial incubation is for two hours a t a temperature of 45C and the second incubation is for one hour at 45C. The radiolabeled antibody used in all experiments of both SPRIA I and S P R l A 45 contained normal guinea pig serum to reduce false-positive reactions.s SPRIA 1 and S P R I A 45 were modified to determine the specificity of each specimen as it was tested, using a specific blocking technique. Each sample was incubated, as described above, in duplicate. After the first incubation and wash, one tube of each pair was incubated with the radiolabeled antibody preparation containing 20 per cent precipitating chimpanzee anti-HBs (titer 1.5 12000 by passive hemagglutination assay). The second tube was used as a control and was incubated with the radiolabeled antibody preparation with 20 per cent normal human serum. A sample was defined as specifically reactive if the chimpanzee antibody reduced the residual count of a reactive sample by at least 50 per cent. The third modification to the S P R I A techniquet will be referred to as S P R l A 2. In this, the solid phase is guinea pig anti-HB, coupled to a polystyrene bead and the radiolabeled antibody is of human origin.E Two hundred microliters of specimen and reagent antibody are used. The reaction is performed in a separate tray and incubating and washing are as described for SPRIA 45. The radioactivity bound to each bead is counted after transferring the washed bead to a clean tube. Results are evaluated in exactly the same fashion a s for SPRIA I. Specimens which were reactive by SPRIA 2 were not tested for specificity. All samples were tested by CEP, using a commercially available kit (Spectra Biologicals) with antibody derived from human sources.
Results The results of the tests performed on the 688 samples are summarized in Table 1. Only 175 of the samples were reactive by CEP. All of these samples that were reactive and specific when tested by S P R l A I and S P R l A 45 were also reactive by SPRIA 2. A total of 419 specimens were reactive by SPRIA I but of these, only 235 were specific for HB,Ag. The same 235 specimens were also reactive and specific when tested by S P R l A 45 and reactive by S P R l A 2. Thirty-seven nonspecifically reactive specimens were detected t A u s R l A 11-125
609
RIA FOR HBJG Table I . Results of Testing 688 Samples for HBsAg by CEPand by Three Modifications of the SPRlA Technique Number of Specimens Reactive Technique
CEP SPRIA 1 SPRlA 45 SPRlA2
Specific Nonspecific Total Nonreactive
N D" 235 235
ND
ND 184 37 ND
175 419 272 240
51 3
269 416 448
* N D : Specificity was not determined.
by SPRIA 45 and 32 of these were also nonspecifically reactive by SPRIA 1. SPRIA 2 detected all samples which were positive and specific by the other two RIA techniques, but did not detect any of the nonspecific samples. However, by S P R l A 2 five additional specimens were reactive. These were nonreactive with the other two SPRIA techniques. When these five samples were retested by S P R l A 1 and SPRIA 45 in the absence of any modification to determine specificity, three of the five specimens were reactive by both techniques. Presumably, the dilution of radiolabeled antibody inherent to the specificity test reduced the sensitivity of the methods to such an extent that HB,Ag in these specimens were virtually undetectable. The specificity of these five sera 'was not determined. It should be noted that the S P R l A technique tends to give nonrepeatable false positives.' Many of the samples had been submitted in order to resolve equivocal test results. These facts account for the finding that many of the sera initially reported as reactive were nonreactive when retested in this study. The results of another study on all samples submitted for confirmation during the year ending July 31, 1974 are presented in Table 2. Each sample was tested for specificity by either SPRIA 1 or S P R l A 45 as described above. Of a total 4,492 specimens, 2,753 were SPRIAreactive. Of these, 1,503 were also CEP-positive. Of the 1,250 samples which were CEP-negative and SPRIA-positive, 497 (39.8%) were nonspecific for H B Ag. I n this series, SPRIA detected 753 (50.2%) more specific HB,Ag positives than did CEP. It should be noted that the 688 specimens summarized in Table I are included in this total.
Discussion The results presented here suggest t h a t progressive modification of t h e A u s R I A
610
NI, ET AL
Table 2. Confirmatory Studies upon Reactive or Suspect Samples Received During the Year Ending July 31, 1974 Samples in Each Category Test Results CEP neg, R I A neg CEP pos, R I A pos CEP neg, R I A pos (specific) CEP neg, R I A pos (nonspecific) CEP pos, R I A neg (nonspecific CEP) Totals
Number
Per Cent
1,726 1,503
38.42 33.46
753
16.76
497
11.06
13 4,492
0.28 99.98
~
technique has resulted in a considerable increase in the specificity of this test without any loss of sensitivity. Indeed, the results in Table 1 show that 44 per cent of all SPRIA 1 reactive samples were nonspecific while the comparable figure for SPRIA 45 was 14 per cent. These figures do not represent the true specificity or sensitivity of the SPRIA techniques since the sera were submitted as reactive on the basis of SPRIA 1 test results. The American National Red Cross Blood Program collects approximately four million units of blood per year and all donations have been tested for HB,Ag since April 1971. The SPRIA technique was introduced in January 1973 and by July 1973, the majority of centers were utilizing SPRIA methodology. After December 1973, all donor samples have been tested in this fashion. The incidence of HBsAg as detected by CEP remained stable at approximately 0.84 per 1,000 donors during the period April 1972 to December 1973.2 The incidence as detected by RIA has altered materially. Red Cross Regional Blood Centers generally report only the specimens which are repeatably positive; however, some are reported which have not been retested. Therefore, the reported incidence may be somewhat higher. Of the 700,000 samples tested during the period January-June 1973, there were 5.35 times as many SPRIA reactives reported as CEP reactives. During this period, the SPRIA reagents did not include
Transfusion Nov.-Dec. 1975
normal guinea pig serum. Approximately 1.4 million tests were performed by SPRIA 1 methodology during the period JulyDecember 1973 and the reported incidence of RIA reactivity was 2.9 times that reported for CEP. Between January and June 1974, 1,500,000 samples were tested by the SPRIA 45 method and the reported incidence of reactivity was approximately 1.45 per 1,000. This is approximately 1.7 times the figure previously found by CEP.2 With the technique, there has been a significant increase in the detection of HB,Ag carriers, as compared to CEP. The SPRIA 45 result2 indicates that this gain can not be more than 70 per cent and the data of specificity testing suggests that the SPRIA methodology may detect about 50 per cent more true HB,Ag positives than does CEP in our donor popuIat ion. The basis for the lack of specificity of SPRIA 1 is not readily explainable, although it is evident that some component in the test serum must be able to bind radiolabeled antibody to the solid phase. The reactivity cannot be toward native guinea pig serum components, as the large excess of normal guinea pig proteins present in the antibody reagent would be expected to neutralize any such reactivity. It is possible that the material responsible for the nonspecific reactivity does bind guinea pig proteins which have been physically or chemically altered by the processes involved in coupling to the polypropylene tube and in iodination. The reduction of such nonspecific reactives in the SPRIA 45 technique suggests that this reaction is heat-sensitiveg and may be due to a component of complement which binds to aggregated or structurally modified gamma globulins. As might be expected, a further degree of specificity is introduced when the radiolabeled antibody reagent is derived from human sources. In this study, none of our samples which were reactive but nonspecific with SPRIA 1 and SPRIA 45 was reactive when tested with SPRIA 2. The SPRIA 2 technique is claimed to be
Volume 15
RIA FOR HB,AG
Number6
more sensitive than either SPRIA 1 or SPRIA 45 and our results do indicate that this may be the case, although our highly selected test panel does not permit any firm conclusions. Indeed, it is quite possible that the net gain of two SPRIA 2 positive samples merely indicates the redetection of specimens which were originally found reactive by reagents of unusually high sensitivity. Certainly, the AusRIA 11-125 methodology demonstrates a very high degree of specificity when compared with techniques using the earlier AusRIA-125 test materials. Hopefully, this specificity will prove beneficial in the economical selection of healthy donors. However, specificity testing of all reactive samples should be continued on a routine basis, at least until the specificity of newer techniques is established. Acknowledgment We are indebted to Ms. Irene Feller for compiling data, Ms. Janet Shoff and Dr. Edwin Steane for reading the manuscript, and Dr. Jay Hoofnagel of BOB, FDA, for providing chimpanzee anti-HB,. Sincere thanks are to be extended to 58 Red Cross Regional Blood Centers for making available serum specimens in this study.
References I.
2.
Alter, H. J., P. J. Holland, R. H. Purcell, and J. L. Gerin: The Ausria test: critical evaluation of sensitivity and specificity. Blood 423947, 1973. Dodd, R. Y., L. Y. Ni, W. S. Mallin, and T. J. Greenwalt: Hepatitis B (surface) antigen testing by radioimmunoassay: experience in a very
3.
4.
5.
6. 7.
8.
9.
10.
61 1
large volunteer donor population. Am. J. Clin. Pathol. 63:847, 1975. Gocke, D. J.: A prospective study of posttransfusion hepatitis. The role of Australia antigen. JAMA 219:1165, 1972. Goldfield, M., J . Bill, H. Black, W. Pizzati, and S. Srihongse: Hepatitis associated with the transfusion of HBAg-negative blood. In Hepatitis and Blood Transfusion. G. N. Vyas, H. A. Perkins and R. Schmid, Eds. New York, Grune and Stratton, 1972, p. 353. Ling, C. M., and L. R. Overby: Prevalence of hepatitis B virus antigen as revealed by direct radioimmune assay with '251-antibody. J. Immunol. 109:834, 1972. Overby, L. R., R. H. Decker, W. W. Sultan, and C. M. Ling: Personel communication. -, J. P. Miller, I. D. Smith, R. H. Decker, and C. M. Ling: Radioimmunoassay of hepatitis B virus-associated (Australia) antigen employing '251-antibody. Vox Sang. Supp. 24:102, 1973. Prince, A. M., B. Brotman, D. Jass, and H. Ikram: Specificity of the direct solid-phase radioimmunoassay for detection of hepatitis-B antigen. Lancet 1:1346, 1973. Prince, A. M., and D. Jass: Kinetic studies on the direct solid phase radioimmunoassay for hepatitis B antigen. Vox Sang. 26:209, 1974. Sgouris, J. T.: Limitations of the radioimmunoassay for hepatitis B antigen. N. Engl. J. Med. 288:160. 1973.
L. Y. Ni, B.A., M.S., Assistant Technical Director, Special Projects Laboratory, American National Red Cross, Washington, D.C. 20006. R. Y. Dodd, BSc., Research Scientist. N. Brown, B.S., Research Technologist. R. Salmon, B.S., Research Technologist. T. J. Greenwalt, M.D., Medical Director.
Serum specimens submitted for confirmation of hepatitis B antigen (HB,Ag) reactivity were tested by counterelectrophoresis (CEP)and by three modifications of the solid phase radioimmunoassay “sandwich” technique (SPRIA). In addition, the specificity of each reaction was determined for two of the SPRIA test methods. One hundred seventy-five of the 688 specimens were reactive by all techniques and a further 60 were specifically reactive by two of modifications of SPRIA. An additional five specimens were positive by a third SPRlA modification. The primary SPRlA technique, using a homologous guinea pig antibody sandwich, detected 184 nonspecific reactive samples. With the same test done at 45C with a reduced incubation time, there were only 37 nonspecific reactive samples. None of the nonspecific samples was detected when the SPRIA technique used a heterologous guinea pig-human antibody sandwich. The results provide a basis for estimating the effectiveness of the SPRlA technique for HBsAg testing.
to false positive reactions caused by antibodies against guinea pig globulins, but it was generally felt that such reactivity would be rare in the general population. However, initial experience with the AusRIA technique did reveal unacceptable levels of false positive reactions, possibly caused by nonspecific antiglobulin r e a ~ t i v i t y .This ~ problem was largely eliminated by addition of normal guinea pig serum to the radiolabeled reagent; nevertheless, a significant proportion of nonspecifically reactive specimens was still This paper compares the specificity of reactions obtained with three versions of AusRIA technique. The results are discussed in the context of a large volunteer blood donor population.
BLOOD CONTAINING hepatitis B surface antigen (HB,Ag) is more likely to transmit type B viral hepatitis than is blood which does not have detectable HBsAg.3*4Therefore, blood collecting agencies routinely test all blood donations for the presence of HB,Ag and discard any that are reactive. HBs.4g reactive blood donors are considered to remain potentially infectious and are permanently deferred. These considerations require that a routine test for HBsAg be sensitive and specific, providing maximum safety for the recipient without unnecessary wastage of the blood resource. Solid phase radioimmunoassay (SPRIA) is currently the most sensitive routine test for HBsAg and a commercially available version* is widely Sgouris’o reported that the AusRIA technique was susceptible
Materials and Methods The 688 serum specimens were all derived from the normal operations of the American National Red Cross Blood Program and represent samples submitted by the Red Cross Regional Blood Centers (RCBCs) during October 1973 for confirmatory testing in our laboratory. Each specimen had been suspected to be reactive on the basis of at least one test by the S P R l A I methodology. The SPRIA** m e t h o d o l ~ g yinvolves ~~~ incubation of 100 ul of serum at room temperature for 16 hours in a polypropylene tube coated with guinea pig antibodies to HB,Ag (anti-HBs). The tube is then washed exhaustively and radiolabeled guinea anti-HB, is added, followed by incubation for 90 minutes at room temperature. Unreacted radiolabeled antibody is washed from the tube and bound antibody is determined by measuring the residual radioactivity retained in the tube. A specimen is regarded as reactive if the count exceeds 2.1 times the mean of seven determinations on a known negative control. This cutoff
Received for publication December I , 1974; accepted February 23, 1975. Publication Number 301 from the American National Red Cross. *AusRIA, Abbott Laboratories. Transfusion Nov.-Dec. 1975
**AusR IA- 125, Abbott Laboratories.
608
Volume I5 Number6
Volume I5 Number6
represents approximately seven standard deviations above the negative control mean. This methodology will be referred to as SPRIA I . The first modification of this technique will be referred to as SPRIA 45. The same reagents are used but the initial incubation is for two hours a t a temperature of 45C and the second incubation is for one hour at 45C. The radiolabeled antibody used in all experiments of both SPRIA I and S P R l A 45 contained normal guinea pig serum to reduce false-positive reactions.s SPRIA 1 and S P R I A 45 were modified to determine the specificity of each specimen as it was tested, using a specific blocking technique. Each sample was incubated, as described above, in duplicate. After the first incubation and wash, one tube of each pair was incubated with the radiolabeled antibody preparation containing 20 per cent precipitating chimpanzee anti-HBs (titer 1.5 12000 by passive hemagglutination assay). The second tube was used as a control and was incubated with the radiolabeled antibody preparation with 20 per cent normal human serum. A sample was defined as specifically reactive if the chimpanzee antibody reduced the residual count of a reactive sample by at least 50 per cent. The third modification to the S P R I A techniquet will be referred to as S P R l A 2. In this, the solid phase is guinea pig anti-HB, coupled to a polystyrene bead and the radiolabeled antibody is of human origin.E Two hundred microliters of specimen and reagent antibody are used. The reaction is performed in a separate tray and incubating and washing are as described for SPRIA 45. The radioactivity bound to each bead is counted after transferring the washed bead to a clean tube. Results are evaluated in exactly the same fashion a s for SPRIA I. Specimens which were reactive by SPRIA 2 were not tested for specificity. All samples were tested by CEP, using a commercially available kit (Spectra Biologicals) with antibody derived from human sources.
Results The results of the tests performed on the 688 samples are summarized in Table 1. Only 175 of the samples were reactive by CEP. All of these samples that were reactive and specific when tested by S P R l A I and S P R l A 45 were also reactive by SPRIA 2. A total of 419 specimens were reactive by SPRIA I but of these, only 235 were specific for HB,Ag. The same 235 specimens were also reactive and specific when tested by S P R l A 45 and reactive by S P R l A 2. Thirty-seven nonspecifically reactive specimens were detected t A u s R l A 11-125
609
RIA FOR HBJG Table I . Results of Testing 688 Samples for HBsAg by CEPand by Three Modifications of the SPRlA Technique Number of Specimens Reactive Technique
CEP SPRIA 1 SPRlA 45 SPRlA2
Specific Nonspecific Total Nonreactive
N D" 235 235
ND
ND 184 37 ND
175 419 272 240
51 3
269 416 448
* N D : Specificity was not determined.
by SPRIA 45 and 32 of these were also nonspecifically reactive by SPRIA 1. SPRIA 2 detected all samples which were positive and specific by the other two RIA techniques, but did not detect any of the nonspecific samples. However, by S P R l A 2 five additional specimens were reactive. These were nonreactive with the other two SPRIA techniques. When these five samples were retested by S P R l A 1 and SPRIA 45 in the absence of any modification to determine specificity, three of the five specimens were reactive by both techniques. Presumably, the dilution of radiolabeled antibody inherent to the specificity test reduced the sensitivity of the methods to such an extent that HB,Ag in these specimens were virtually undetectable. The specificity of these five sera 'was not determined. It should be noted that the S P R l A technique tends to give nonrepeatable false positives.' Many of the samples had been submitted in order to resolve equivocal test results. These facts account for the finding that many of the sera initially reported as reactive were nonreactive when retested in this study. The results of another study on all samples submitted for confirmation during the year ending July 31, 1974 are presented in Table 2. Each sample was tested for specificity by either SPRIA 1 or S P R l A 45 as described above. Of a total 4,492 specimens, 2,753 were SPRIAreactive. Of these, 1,503 were also CEP-positive. Of the 1,250 samples which were CEP-negative and SPRIA-positive, 497 (39.8%) were nonspecific for H B Ag. I n this series, SPRIA detected 753 (50.2%) more specific HB,Ag positives than did CEP. It should be noted that the 688 specimens summarized in Table I are included in this total.
Discussion The results presented here suggest t h a t progressive modification of t h e A u s R I A
610
NI, ET AL
Table 2. Confirmatory Studies upon Reactive or Suspect Samples Received During the Year Ending July 31, 1974 Samples in Each Category Test Results CEP neg, R I A neg CEP pos, R I A pos CEP neg, R I A pos (specific) CEP neg, R I A pos (nonspecific) CEP pos, R I A neg (nonspecific CEP) Totals
Number
Per Cent
1,726 1,503
38.42 33.46
753
16.76
497
11.06
13 4,492
0.28 99.98
~
technique has resulted in a considerable increase in the specificity of this test without any loss of sensitivity. Indeed, the results in Table 1 show that 44 per cent of all SPRIA 1 reactive samples were nonspecific while the comparable figure for SPRIA 45 was 14 per cent. These figures do not represent the true specificity or sensitivity of the SPRIA techniques since the sera were submitted as reactive on the basis of SPRIA 1 test results. The American National Red Cross Blood Program collects approximately four million units of blood per year and all donations have been tested for HB,Ag since April 1971. The SPRIA technique was introduced in January 1973 and by July 1973, the majority of centers were utilizing SPRIA methodology. After December 1973, all donor samples have been tested in this fashion. The incidence of HBsAg as detected by CEP remained stable at approximately 0.84 per 1,000 donors during the period April 1972 to December 1973.2 The incidence as detected by RIA has altered materially. Red Cross Regional Blood Centers generally report only the specimens which are repeatably positive; however, some are reported which have not been retested. Therefore, the reported incidence may be somewhat higher. Of the 700,000 samples tested during the period January-June 1973, there were 5.35 times as many SPRIA reactives reported as CEP reactives. During this period, the SPRIA reagents did not include
Transfusion Nov.-Dec. 1975
normal guinea pig serum. Approximately 1.4 million tests were performed by SPRIA 1 methodology during the period JulyDecember 1973 and the reported incidence of RIA reactivity was 2.9 times that reported for CEP. Between January and June 1974, 1,500,000 samples were tested by the SPRIA 45 method and the reported incidence of reactivity was approximately 1.45 per 1,000. This is approximately 1.7 times the figure previously found by CEP.2 With the technique, there has been a significant increase in the detection of HB,Ag carriers, as compared to CEP. The SPRIA 45 result2 indicates that this gain can not be more than 70 per cent and the data of specificity testing suggests that the SPRIA methodology may detect about 50 per cent more true HB,Ag positives than does CEP in our donor popuIat ion. The basis for the lack of specificity of SPRIA 1 is not readily explainable, although it is evident that some component in the test serum must be able to bind radiolabeled antibody to the solid phase. The reactivity cannot be toward native guinea pig serum components, as the large excess of normal guinea pig proteins present in the antibody reagent would be expected to neutralize any such reactivity. It is possible that the material responsible for the nonspecific reactivity does bind guinea pig proteins which have been physically or chemically altered by the processes involved in coupling to the polypropylene tube and in iodination. The reduction of such nonspecific reactives in the SPRIA 45 technique suggests that this reaction is heat-sensitiveg and may be due to a component of complement which binds to aggregated or structurally modified gamma globulins. As might be expected, a further degree of specificity is introduced when the radiolabeled antibody reagent is derived from human sources. In this study, none of our samples which were reactive but nonspecific with SPRIA 1 and SPRIA 45 was reactive when tested with SPRIA 2. The SPRIA 2 technique is claimed to be
Volume 15
RIA FOR HB,AG
Number6
more sensitive than either SPRIA 1 or SPRIA 45 and our results do indicate that this may be the case, although our highly selected test panel does not permit any firm conclusions. Indeed, it is quite possible that the net gain of two SPRIA 2 positive samples merely indicates the redetection of specimens which were originally found reactive by reagents of unusually high sensitivity. Certainly, the AusRIA 11-125 methodology demonstrates a very high degree of specificity when compared with techniques using the earlier AusRIA-125 test materials. Hopefully, this specificity will prove beneficial in the economical selection of healthy donors. However, specificity testing of all reactive samples should be continued on a routine basis, at least until the specificity of newer techniques is established. Acknowledgment We are indebted to Ms. Irene Feller for compiling data, Ms. Janet Shoff and Dr. Edwin Steane for reading the manuscript, and Dr. Jay Hoofnagel of BOB, FDA, for providing chimpanzee anti-HB,. Sincere thanks are to be extended to 58 Red Cross Regional Blood Centers for making available serum specimens in this study.
References I.
2.
Alter, H. J., P. J. Holland, R. H. Purcell, and J. L. Gerin: The Ausria test: critical evaluation of sensitivity and specificity. Blood 423947, 1973. Dodd, R. Y., L. Y. Ni, W. S. Mallin, and T. J. Greenwalt: Hepatitis B (surface) antigen testing by radioimmunoassay: experience in a very
3.
4.
5.
6. 7.
8.
9.
10.
61 1
large volunteer donor population. Am. J. Clin. Pathol. 63:847, 1975. Gocke, D. J.: A prospective study of posttransfusion hepatitis. The role of Australia antigen. JAMA 219:1165, 1972. Goldfield, M., J . Bill, H. Black, W. Pizzati, and S. Srihongse: Hepatitis associated with the transfusion of HBAg-negative blood. In Hepatitis and Blood Transfusion. G. N. Vyas, H. A. Perkins and R. Schmid, Eds. New York, Grune and Stratton, 1972, p. 353. Ling, C. M., and L. R. Overby: Prevalence of hepatitis B virus antigen as revealed by direct radioimmune assay with '251-antibody. J. Immunol. 109:834, 1972. Overby, L. R., R. H. Decker, W. W. Sultan, and C. M. Ling: Personel communication. -, J. P. Miller, I. D. Smith, R. H. Decker, and C. M. Ling: Radioimmunoassay of hepatitis B virus-associated (Australia) antigen employing '251-antibody. Vox Sang. Supp. 24:102, 1973. Prince, A. M., B. Brotman, D. Jass, and H. Ikram: Specificity of the direct solid-phase radioimmunoassay for detection of hepatitis-B antigen. Lancet 1:1346, 1973. Prince, A. M., and D. Jass: Kinetic studies on the direct solid phase radioimmunoassay for hepatitis B antigen. Vox Sang. 26:209, 1974. Sgouris, J. T.: Limitations of the radioimmunoassay for hepatitis B antigen. N. Engl. J. Med. 288:160. 1973.
L. Y. Ni, B.A., M.S., Assistant Technical Director, Special Projects Laboratory, American National Red Cross, Washington, D.C. 20006. R. Y. Dodd, BSc., Research Scientist. N. Brown, B.S., Research Technologist. R. Salmon, B.S., Research Technologist. T. J. Greenwalt, M.D., Medical Director.
