Prevalence of antibody to hepatitis C virus in a blood donor population

C. RICHARDS, P. HOLLAND, K. KURAMOTO, C. DOUVILLE, AND R. RANDELL

Blood samples from 2000 accepted blood donors and 343 deferred donors with antibody to hepatitis B core antigen (anti-HBc) and/or an alanine aminotransferase ALT) elevation were evaluated for antibody to hepatitis C virus (anti-HCV). Sixteen

(ELISA); 12 (0.6%) were repeatably reactive. One repeatabl reactive sample had

repeatably reactive E L S A samples were tested by an immunoblot assa , four re- acted, three were indeterminate, and five did not react. Among the 3 4 l deferred donors, HCV antibodies were detected in 8 (3.8%) of 21 0 anti-HBc-reactive samples, 12 (11.8%) of 104 elevated-ALT samples, and 15 (52%) of 29 combined elevated- ALT and anti-HBc-reactive samples: 25 of 28 reacted on immunoblot. The anti-HBc- reactive sera were subdivided into groups according to strength of anti-HBc reactivity (weak or strong) and antibody to hepatitis B surface antigen status and then were compared for anti-HCV reactivity rates. The group of samples showing the greatest frequency of anti-HCV had strong anti-HBc reactivity. For blood donors, the anti-HCV test correlates with the surrogate tests for non-A,non-B hepatitis (anti-HBc and ALT); however, most anti-HCV-reactive units remain undetected by surrogate tests, so that implementation of anti-HCV screening should further reduce the transmission of HCV via transfusion. TRANSFUSION 1991 ;31:109-113.

of the 2000 sera initially reacted on enzyme-linked immunosorbent assay

an elevated ALT; two reacted on anti-HBc testing and had AL # elevations. When the

Abbrevlatlons: ALT = alanlne amlnotransferase; antl-HBc = antlbody to hopatltls B coro antlgen; antl-HBs = antibody to hepatitls B surface antlgen; ELSA = enzymallnked lmmunosorbent assay; HBV = hepatitls B vlrus; HCV = hrpatltlr C vlrus; HIV-1 = human lmmunodeflclency vlrus type 1; HTLV-I = human lymphotroplc vlrus typo I; NANBH = non-A,non-B hepatitis; RlBA = rscomblnant lmmunoblot assay; RNA = rlbonuclolc acld; S:C ratlo = sample-totutoff ratlo.

UNTIL RECENTLY, THERE HAS been no specific assay to identify individuals infected with the non-A,non-B hepa- titis (NANBH) virus, the most common cause of transfu- sion-associated hepatitis. Scientists at Chiron Corporation, in collaboration with the Centers for Disease Control, cloned part of the genome of the hepatitis C virus (HCV),' the major causative agent of NANBH. They were able to de- rive a complementary DNA clone using RNA isolated from the sera of chimpanzees infected with human NANBH. An HCV-derived recombinant polypeptide produced from se- lected yeast complementary DNA clones of HCV reacted with sera from NANBH patients but not with sera from patients with hepatitis A virus, hepatitis B virus (HBV), or cytomegalovirus infection.' The HCV antigen (a non- structural protein of HCV) provides the basis for a test to detect virus-specific antibody (anti-HCV) in human serum or plasma using an enzyme-linked immunosorbent assay (ELISA) .2

From thc Sacramento Medical Foundation Center for Blood Re-

Received for publication February 21, 1990; revision receivcd July search, Sacramento, California.

2, 1990. and acccpted July 3, 1990.

Studies have suggested that one-third to one-half of transfusion-associated NANBH cases might be pre- vented by excluding blood donors who had either an elevated ALT value or antibody to the hepatitis B core antigen (anti-HB~).~-~ This paper describes results with the anti-HCV ELISA on sera from accepted volunteer blood donors and deferred donors found to have anti- HBc and/or an elevated ALT. Donor samples that re- acted with anti-HBc were further divided into categories on the basis of antibody to the hepatitis B surface antigen (anti-HBs) status and strength of the anti-HBc sample- to-cutoff (S:C) absorbance ratio.

Although no gold standard, or even a confirmatory test, for anti-HCV is available, supplemental assays have been developed. One such assay, an immunoblot test, allows for the qualitative detection of antibodies to HCV recombinant proteins. In the blood donor setting, ap- proximately 51 percent of samples that were repeatably reactive on anti-HCV ELISA have been found to react with a supplemental immunoblot assay (Nelles M. Ortho Diagnostic Systems, written communication, April 1990). Another manufacturer reported9 that approxi- mately 40 percent of blood donor sera that reacted in

109

110 RICHARDS

their anti-HCV ELISA were neutralized by the c100-3 HCV antigen and thus presumably were "true" pos- itives. Although not ideal, anti-HCV testing, along with supplemental assays and surrogate test results, should provide physicians with important information for coun- seling blood donors referred to them, patients with chronic hepatitis not due to hepatitis A or B viruses, and recip- ients of prior and anticipated blood transfusions.

Materials and Methods We evaluated three prelicensure reagent lots of an HCV

antibody ELISA test system (Ortho Diagnostic Systems, Inc., Raritan, NJ) as part of a clinical trial in a blood center setting. The test was used according to the manufacturer's directions for use. Any clinical sample initially found to be greater than or equal to the cutoff value (initially reactive) or to have an S:C ratio between 0.85 and 1.0 (grey zone) was tested again in duplicate. If a sample had at least two values greater than the cutoff, we considered it repeatably reactive and thus to react for anti-HCV on ELISA. All others were considered non- reactive. We determined the cutoff by adding 0.4 to the mean of the negative control values.

The study evaluated fresh serum samples from 2000 quali- fied volunteer blood donors who underwent phlebotomy at the Sacramento Medical Foundation Blood Center in early 1989. We obtained informed consent from all donors before perform- ing the anti-HCV assay. Information collected on these donors included age, gender, race, and history of hepatitis exposure, transfusion, and blood donation. In addition, all donations were tested for anti-HBc, ALT level, antibody to human immuno- deficiency virus-type 1 (anti-HIV-1), antibody to human T-cell lymphotropic virus-type I (anti-HTLV-I), hepatitis B surface antigen, and syphilis. We determined the cutoff for the ALT test (>53 IUiL) by the method of AuBuchon et al.;'" i t was comparable to the upper limit of normal, established as 2 stan- dard deviations above the log mean normal. Units from blood donors with repeatably reactive anti-HCV ELISA results were not released for transfusion; all were sent to Ortho Diagnostic Systems for further evaluation.

ET AL. TRANSFUSION

Vol. 31, No. 2-1591

Serum and plasma specimens from donors with a known elevated (>53 IU/L) ALT value (n = 104; all frozen), a re- active anti-HBc test (n = 210; 135 frozen), and both a reactive anti-HBc test and an elevated ALT result (n = 29; 17 frozen) were also tested with the anti-HCV assay. All 343 donors had been deferred because of the results of their ALT and/or anti- HBc tests; none reacted with hepatitis B surface antigen.

The anti-HBc-reactive samples were divided into four groups. We assigned the samples on the basis of the strength of anti- HBc and the presence or absence of anti-HBs. A weak anti- HBc result, defined as an S:C ratio of approximately 0.45 or greater (and without other markers of HBV infection), may represent a false-positive reaction; these weak reactions tend to have inconsistent test results when tested sequentially by ELISA or in parallel by radioirnmunoas~ay.~.~ A strong anti- HBc result, S:C ratio less than 0.45, probably more often represents a true-positive reaction; these samples tend to be repeatable in sequential tests by ELISA or in parallel by ra- dioimmunoassay and are more often accompanied by anti-HBs. We then compared these four groups of samples to each other for frequency of reactivity with the anti-HCV assay, to deter- mine whether strong anti-HBc reactivity represents a risk factor for HCV infection as well. We used chi-square analysis with Yates' correction to evaluate differences between the defined groups. Differences were considered significant if p values were less than 0.05.

We later tested sera from the repeatably reactive anti-HCV donations by the recombinant immunoblot assay (RIBA), an in vitro test for the detection of antibodies to HCV recombinant antigens, which was purchased from the Chiron Corporation (Emeryville, CA) through Ortho Diagnostic Systems. This RIBA- HCV test system uses three recombinant antigens: c100-3, pro- duced in yeast; 5-1-1, produced in Eschen'chia coli; and human superoxide dismutase. The c100-3 and the 5-1-1 antigens are superoxide dismutase-fusion polypeptides of HCV (5-1-1 is a smaller portion of c100-3). Human superoxide dismutase is a control to detect antibodies against the protein, as i t is fused to c100-3 on the strip and microwell ELISA. We used the immunoblot assay according to manufacturer's directions.

In addition to the recombinant antigens on the RIBA strip, there are two gamma globulin bands (Levels I and 11); readings are graded as follows: no visible band, -; <Level I but with discernible band, f ; = Level I, 1 +; >Level I but <Level

Table 1. Hepatitis-related test results on 72 blood donors with anti-HCV by €LISA*

Anti-HCV Donor ELISA lmmunoblot ALT

number S:Ct 5-1-1 C100-3 Interpretation* (I U/L) Anti-HBc Anti-HBs

01 43 >5.0 4 f 4 f R 56 R R 01 56 > 5.0 4 f 4 f R 48 0947 > 5.0 - 4 f I 1 a5 R R

- N a 0923 4.8 - 1994 4.6 1 + 3 f R 72 - R

13 4.2 - 2 + I 0015 4.0 3 + 2 + R 10 01 a2 2.8 - - N 7 1830 1.6 - - N 12 1999

0695 1.5 - 1 + I 5 1.2 - - N 14 - R 021 8

0595 1 .o - - N 10

- - - -

- - - -

- - - - - -

- - 'HCV = hepatitis C virus; ELISA = enzyme-linked irnmunosorbent assay. tsarnple-to-cutoff ratio. *R = reactive; I = indeterminate; N or - = nonreactive.

TRANSFUSION 1991-Vol. 31, No. 2 PREVALENCE OF ANTI-HCV IN BLOOD DONORS 111

Table 2. Donor demographics (n = 2000)

Number with anti-HCV reacting repeatably

Number with anti-HCV that did not react

on E L S A Variable total donors (n=12) (n=l988)

Percentage of on ELISA

Race WhRe 90.4 11 Hispanic 4.4 0 Asian 3.3 1 Other 1.9 0

Male 55.0 4 Female 45.0 8

Gender

History Hepatitis exposure 0 0

Blood donation 8 1597 Transfusion-associated 0 0

(last 6 months) Transfusion 0 163

heDatitis

Table 3. Anti-HCV results on elevated-ALT andlor anti-HBc-reactive samples ELISA' results RlBA results (number)

repeatably reactive Percentage R t I* N§ QNSll Number

Anti-HBc 81210 Elevated ALT 1211 04 Antl-HBclelevated ALT 15/29

3.8 4 4 11.8 11 1 52 10 1 1 3

'Enzyme-linked irnmunosorbent assay. Weactive. +Indeterminate.

11, 2 + ; = Level 11, 3 + ; and >Level 11, 4 + . A response of 1 + or greater indicates sample reactivity to a given antigen.

We interpreted the antigen band patterns for the RlBA as follows: no bands of 1 + or greater reactivity present, non- reactive; 1 + or greater reactivity to both ~ 1 0 0 - 3 and 5-1-1. reactive; and 1 + or greater bands present, but pattern does not meet criteria for interpretation as reactive, interminate. In the event that the superoxide dismutase band shows 1 + or greater reactivity, the sample is rated indeterminate irrespective of band pattern.

Results Of the 2000 blood donors tested for anti-HCV by ELISA,

1984 (99.2%) of the samples did not react, 16 (0.8%) initially reacted, and 12 (0.6%) reacted repeatably. Of the 12 anti- HCV-positive samples that were repeatably reactive on ELISA, 3 had an elevated ALT level; 2 of these donor sera also reacted for anti-HBc. No samples reacted for hepatitis B surface an- tigen, anti-HIV-1, or anti-HTLV-I. The results from the RlBA and a comparison of surrogate and other hepatitis markers on the 12 sera are shown in Table 1.

The donor demographics for all 2000 donors are summarized in Table 2. Anti-HCV ELISA results on 343 elevated-ALT and/or anti-HBc-reactive deferred donor specimens are detailed in Table 3. Sufficient material remained of only 28 of the 35 donor specimens with repeatably reactive anti-HCV ELISA tests for testing by RIBA; 25 reacted on that test, 2 were indeterminate, and only 1 did not react.

§Nonreactive. IlQuantity not sufficient.

The frequency of anti-HCV ELISA repeatable reactivity in donor samples with both ALT elevations and anti-HBc reac- tivity is 4 times that in donor samples with only an elevated ALT value (11.8%) and 14 times that in donor samples with only anti-HBc reactivity (3.8%). These results are significant at ~ ~ 0 . 0 5 .

We compared the frequency of repeatable reactivity with the anti-HCV ELISA for the four groups of anti-HBc-reac- tive samples. The results are shown in Fig. 1. Chi-square analysis shows a significant difference (p <0.05) between anti-HCV frequency in Group 2 samples and those from Groups 1 or 3. There was also a significant difference (p <0.05) between anti-HCV frequency in samples with strong anti-HBc results (Groups 2 and 4) and those with weak anti- HBc results (Groups 1 and 3). The numbers of repeatably reactive anti-HCV ELISA and reactive RlBA results in each group are too small for statistical comparison.

Discussion

This is the first report of the use of anti-HCV ELISA screening and RIBA evaluation of American blood do- nors. Twelve (0.6%) of 2000 qualified blood donor sera tested with the anti-HCV ELISA were found to be re- peatably reactive. Whereas 3 (25%) of these 12 donors would have been deferred because of a positive surrogate marker, 9 (75%) would have continued to be accepted

RICHARDS ET AL. TRANSFUSION Vol. 31, No. 2-1991 112

26

20

s A

7 16 I 4

V

E

T

V E

n 6 10

5

0 I 1 2 3 4

2/47 16/72 2/47 6/73

ANTI-HBc GROUP FIG. 1. Anti-HCV in 239 anti-HBc-rcactive donor specimens. Group

1: weak anti-HBc, anti-HBs absent; Group 2: strong anti-HBc, anti- HBs absent; Group 3: weak anti-HBc, anti-HBs present; and Group 4: strong anti-HBc, anti-HBs prescnt.

and their blood might be released for transfusion. Be- tween 2.1 and 8.5 of 1000 additional units (95% con- fidence interval) would be eliminated because of reactive anti-HCV ELISA results in this blood donor population.

Seven donor sera showed reactions on immunoblot with one or both recombinant HCV antigens. Sera from only 4 (33%) of the 12 donors who were repeatably reactive on anti-HCV ELISA were “confirmed” by the immunoblot technique. The indeterminate RIBA results of three sera are less clear-cut; however, a repeatably reactive anti-HCV ELISA result and RIBA reactivity at the c100-3 band alone may indicate the presence of spe- cific antibodies against HCV, because the 5-1-1 antigen is a smaller portion of the c100-3 antigen. For the five donor sera with a repeatably reactive ELISA and a non- reactive RIBA, it is likely that the individuals do not really have antibodies against HCV and thus have false- positive ELISA results. Further refinements of these sup- plemental assays are being developed and promise to provide more certainty in differentiating true- from false- positive results on the ELISA.

Transfusion-associated NANBH due to HCV has al- ready been reduced by ongoing hepatitis surveillance, implementation of surrogate testing, and deferral and testing of individuals at risk for HIV infection.”*’* The anti-HCV test should further reduce the risk of transfu- sion-associated NANBH because of the correlation be-

tween the presence of anti-HCV in blood donors and the occurrence of HCV in recipients of their blood.I3

Age, gender, race, and history of hepatitis exposure or transfusion seemed to be unimportant as indicators of prior HCV infection among the blood donors tested. All 12 donors who were repeatably reactive on anti-HCV ELISA had no history of hepatitis exposure or transfu- sions; at least 163 (8%) of the 1988 participants whose sera did not react on anti-HCV ELISA reported that they had been transfused in the past. These findings support the contention that most cases of NANBH are not trans- fusion related and are most likely sporadic or community acquired.I4

When comparing the efficacy of surrogate tests to identify anti-HCV-reactive donors, we found a higher percentage of anti-HCV among donors with an elevated ALT plus a reactive anti-HBc test than among those with only an elevated ALT. The lowest frequency of anti- HCV was in the group of individuals with anti-HBc but without an elevated ALT. These findings have been ob- served by others.” Our results also agree with findings from the Transfusion-Transmitted Viruses Study,’J6 which showed that donor blood with both an elevated ALT and anti-HBc carried the highest risk of transmitting NANBH to recipients. For donor sera with an elevated ALT and/ or a reactive anti-HBc result plus repeatably reactive anti-HCV ELISA results, 27 (96%) of 28 reacted with one or both of the HCV recombinant proteins on the RIBA. Thus, such samples would appear to identify do- nors who have been, and may still be, infected with HCV.

An unexplained finding in this study was the high frequency of anti-HCV reactivity in donor sera with strong anti-HBc results. These are the samples that we believe show true-positive reactions for anti-HBc. There is a significant difference between the anti-HCV frequency in the strong anti-HBc only group (Group 2) and that in either Group 1 (weak anti-HBc, no anti-HBs) or Group 3 (weak anti-HBc, anti-HBs present), as well as between that in the strong anti-HBc groups (Groups 2 and 4) and in those with weak anti-HBc (Groups 1 and 3). If weak anti-HBc results are, in fact, false-positive reactions, these data then support the belief that true anti-HBc-reactive donors are at increased risk of transmitting HCV infec- tion, perhaps because of similar epidemiology for HCV and HBV infe~t ion .~ .~ The high percentage of anti-HCV- reactive samples in Group 2 could also be a reflection of some crossreactivity of the antibodies measured (an- tibodies reacting not only with an antigen in the anti- HCV assay but also with antigens in the anti-HBc assay). These crossreacting antibodies may be the result of in- fection with an aberrant HBV (cryptic HBV or HBV infection without HBV marker^),".'^ but this remains to be proven. Ten of the 15 anti-HCV Group 2 samples (strong anti-HBc) also had ALT elevations. This could

TRANSFUSION 1991-Vol. 31. No. 2 PREVALENCE OF ANTI-HCV IN BLOOD DONORS 113

be a reflection of liver injury from ongoing HCV infec- tion, chronic HBV infection (with a subdetectable hep- atitis B surface antigen level), or both.

The 0.6 percent repeatably reactive rate on anti-HCV ELISA in this study of 2000 qualified blood donors may be lower than expected, on the basis of studies prior to the institution of ALT and anti-HBc testing.l5 The vast majority of the donors in our series have been screened since 1986 for an elevated ALT and the presence of anti- HBc. Only 12 of the 2000 currently accepted blood do- nors were repeatably reactive on anti-HCV ELISA; how- ever, 35 (10%) of 343 previously deferred donors with an elevated ALT and/or anti-HBc reactivity in their sera were repeatably reactive on anti-HCV ELISA (a rate 17 times higher).

It is apparent that false-positive hepatitis test results occur, especially in the testing of blood donors who gen- erally deny a history of hepatitis. At this point, there are no confirmatory tests available for anti-HCV, although supplemental assays provide valuable additional infor- mation regarding the significance of anti-HCV ELISA results. Even if 42 percent, or 67 percent, of our 12 donors with a repeatably reactive ELISA for anti-HCV are considered falsely positive, the ELISA for anti-HCV will detect many donors capable of transmitting NANBH via transfusions but who are not identified by surrogate tests.I3 ALT elevation and the presence of anti-HBc should continue to provide useful information because of their combined value as surrogate tests, either prior to evi- dence of hepatitis C infection or for non-A,non-B (non- C) hepatitis agents that may exist. Furthermore, for iden- tifying donors with anti-HCV who have active disease due to HCV, ALT is useful as an indicator of ongoing liver injury. Regardless of the limitations of the current anti-HCV ELISA and supplemental assays, such testing is an important first step in the specific detection of carriers of the elusive HCV and should further reduce the risk of transfusion-associated NANBH (now hepatitis C) when used to screen blood donors.

Acknowledgments The authors acknowledge the assistance of the following individuals

in the prcparation of this manuscript: Joyce Davis and Drs. Petra Krauledat, Mitchell Nelles, Jerome Zeldis, and Teresa Paglieroni.

References 1. Choo QL, Kuo G . Weiner AJ, et at. Isolation of a cDNA clone

derived from a blood-borne non-A, non-B viral hcpatitis gcnomc. Science 1989;244:359-62.

2. Kuo G, Choo QL, Alter HJ, et al. An assay for circulating anti- bodies to a major etiologicvirus of human non-A, non-B hepatitis. Science 1989;244:362-4.

3. Aach RD, Szmuness W, Mosley JW, et at. Serum alaninc ami- notransferase of donors in relation to the risk of non-A,non-B

hepatitis in recipients: the transfusion-transmitted viruses study. N Engl J Med 1981;304:989-94.

4. Alter HJ, Purcell RH, Holland PV, Alling DW, Koziol DE. Donor transaminasc and recipient hepatitis. Impact on blood transfusion services. JAMA 1981;246:630-4.

5 . Stevens CE, Aach RD, Hollinger FB, et al. Hepatitis B virus antibody in blood donors and the Occurrence of non-A, non-B hepatitis in transfusion recipients. An analysis of the Transfusion- Transmitted Viruses Study. Ann Intern Med 1984;101:733-8.

6. Koziol DE, Holland PV, Alling DW, et al. Antibody to hepatitis B core antigen as a paradoxical marker for non-A, non-B hepatitis agents in donated blood. Ann Intern Med 1986;104:488-95.

7. Schmidt PJ, Leparc GF, Samia CT. Comparison of assays for anti-HBc in blood donors. Transfusion 1988;28:389-91.

8. Hanson MR. Polesky HF. Evaluation of routine anti-HBc screen- ing of volunteer blood donors: a questionable surrogate test for non-A, non-B hepatitis. Transfusion 1987;27:107-8.

9. Zuck T. quoted by Goldsmith MF. Blood bank officials hope donor altruism will pass new (anti-HCV) test (news). JAMA

10. AuBuchon JP. Wilkinson JS, Kassapian SJ, Edwards GC. Estab- lishmcnt of a system to standardize acceptability criteria for ala- nine aminotransferase activity in donated blood. Transfusion

11. Alter MJ, Hadler SC, Margolis HS, et al. The changing epide- miology of non-A, non-B hepatitis in the United States: relation- ship to transfusions. Presentcd at the Transfusion-Associated Infections and Immune Response Seminar, University of Califor- nia at San Francisco, San Francisco, California, March 1988.

12. Chambers LA, Popovsky MA. Decrease in post-transfusion hep- atitis (PTH): contribution of changes in incidence of hepatitis in the general population (abstract). Transfusion 1989; 29(Suppl):43S.

13. Alter HJ, Purcell RH, Shih JW, et al. Detection of antibody to hepatitis C virus in prospectively followed transfusion recipients with acute and chronic non-A, non-B hepatitis. N Engl J Med

4. Alter MJ, Coleman PJ, Alexander JA, et al. Importance of het- erosexual activity in the transmission of hepatitis B and non-A, non-B hepatitis. JAMA 1989;262: 1201-5.

5 . Stcvens CE, Taylor PE, Pindyck J, et al. Epidemiology of hcp- atitis C virus. A preliminary study in volunteer blood donors.

6. Hollingcr FB, Moslcy JW, Szmuness W, et al. Non-A, non-B hepatitis following blood transfusion: risk factors associated with donor characteristics. In: Szmuness W. Alter HJ, Maynard JE, eds. Viral hepatitis: 1981 international symposium. Philadelphia: Franklin Institute Press; 1982:361-76.

17. Trepo C. Vitvitski L, Hantz 0, et al. Characterization and detec- tion of a virus related to HBV in NANB hepatitis. In: Szmuness W, Altcr HJ, Maynard JE, cds. Viral hepatitis: 1981 international symposium. Philadelphia: Franklin Institute Press, 1982:339-54.

18. Hantz 0, Vitvitski L, Trcpo C. Non-a, non-b hepatitis: identifi- cation of hepatitis-B-like virus particles in serum and liver. J Med Virol 1980;5:73-86.

199O;263: 1749-51.

1989;29: 17-22.

1989;321:1494-500.

JAMA 1990;263:49-53.

Cheryl A. Richards, MT(ASCP)SBB, Technical DircctorDirector of Education, Sacramento Medical Foundation Center for Blood Research; current address: Medical Marketing Consultants, Bethesda, MD.

Paul V. Holland, MD, Medical DirectorKEO, Sacramento Medical Foundation, Blood Center/Center for Blood Research, 1625 Stockton Boulevard, Sacramento, CA 95816-7089. [Reprint requests]

Isamu K. Kuramoto, MT(ASCP), Research Technologist, Sacra- mento Medical Foundation Center for Blood Research.

Claire M. Douville. MT(ASCP), Research Technologist, Sacra- mento Medical Foundation Center for Blood Research.

Robert L. Randell, MD, Associate Medical Director, Sacramento Medical Foundation Center for Blood Research; current address: Dc- partment of Pathology. Los Robles Regional Medical Center, Thou- sand Oaks, CA.