Journal of Medical Virology 44:187-191 (1994)

Application of Commercial Assays to Detect IgG Antibodies to Hepatitis C Virus in Urine: A Pilot Study From Autopsy Cases

Xiang Zhang, Niel T. Constantine, Ling Li, M. Abdel-Hamid, Jaya Bansal, and John E. Smialek Department of Pathology, University of Maryland School of Medicine, Baltimore (X.Z., N.T.C., M.A.-H., J.B.), and Office of the Chief Medical Examiner (L.L., J.E.S.), State of Maryland; Faculty of Forensic Medicine, Tongii Medical University, Wuhan, Peoples Republic of China (X.Z.); Department of Microbiology and Immunology, El-Minia University Cairo, Egypt (M.A.-H.)

Hepatitis C virus (HCV), the main cause of non-A, non-B hepatitis in the United States and possibly in the world, is believed to be transmitted pri- marily through parenteral exposure. Many screening and supplemental tests are available to detect antibodies to HCV in serum. The ability to use commercial assays to detect antibodies to HCV in urine was investigated in this study. A total of 229 serumhrine matched samples were collected sequentially from forensic autopsy cases examined at the Office of the Chief Medical Examiner, State of Maryland. Testing was per- formed using the Ortho, Innogenetics, and Ab- bott second generation HCV screening tests and the INNO-LIA HCV Ab supplemental assay. Sam- ple volumes were increased for urine testing. Forty-six of 229 serum samples were positive by screening and confirmed by supplemental tests. The urine samples produced positive results on 44-45 of the same 46 by screening tests and all 46 positives by the supplemental test. There were no false positive samples using urine when compared with the serum pairs. The one false negative sample using urine was still nonreac- tive when the urine volume was increased to 200 pI using the screening tests. Generally, five times the serum volume was required for the screen- ing tests to be optimal for urine samples. The urine samples were stored under different condi- tions prior to testing to determine the influence on antibody stability in urine. 0 1994 Wiley-Liss, Inc.

KEY WORDS urinary antibodies, diagnosis, HCV prevalence

19891. Patients infected with HCV develop antibodies to various structural and nonstructural HCV viral pro- teins. Detection of antibodies to HCV is the easiest method to identify patients who are, or have been, in- fected. Antibodies to HCV have been detected in most patients presenting with posttransfusion NANB hepa- titis [Van der Poel et al., 19891. HCV antibodies have not only been found in patients with acute or chronic forms of hepatitis C, but also in many asymptomatic donors after seroconversion of the recipient [Esteban et al., 19891. The prevalence of antibodies to HCV was found to be high in hemophilia patients, intravenous drug users (IVDU), inmates of prisons, and some organ transplantation recipients [Nicholson et al., 1991; Al- lain et al., 19911.

A number of sensitive and specific serologic tests for HCV are available commercially, but the relative in- convenience of obtaining blood samples and the risk of disease transmission associated with needlestick inju- ries make serologic testing somewhat unattractive, es- pecially in epidemiologic or large seroprevalence stud- ies. Urine has been investigated as an alternative specimen for the detection of viral antibodies, in partic- ular for anti-HIV, anti-hepatitis B core antigens, and anti-HCV [Cao et al., 1989; Perry et al., 1992; Constan- tine et al., 19921. We have explored this area further, not only to confirm the previous reports but also to examine the adaptability of commercially available se- rologic assays to test urine samples and to study the seroprevalence of HCV infections in a forensic autopsy population. In addition, we investigated the effect of storage conditions of urine samples prior to testing in order to determine the optimal temperatures necessary to produce accurate results.

INTRODUCTION The hepatitis C virus (HCV) is believed to be the

major etiologic agent for non-A, non-B (NANB) hepati- tis [Alter et al., 1990; Choo et al., 1990; Kuo et al., 0 1994 WILEY-LISS, INC.

Accepted for publication February 22,1994. Address reprint requests to Dr. Niel T. Constantine, Depart-

ment of Pathology, the University of Maryland, Room 7-58, MSTF, 10 S. Pine St., Baltimore, MD 21201.

188 Zhang et al.

TABLE I. Demographic Characteristics of Autopsy Cases and Seroprevalence of HCV Antibody in Different Groups

Sample HCV antibody (%) positive (%)

Sex Male 185 (80.8) 40 (21.6) Female 44 (19.2) 6 (13.6)

Black 117 (51.1) 32 (27.4) White 110 (48.0) 14 (12.7) Others 2 (0.9) 0 (0)

Natural 76 (33.2) 12 (15.8) Accident 46 (20.1) 8 (17.4) Homicide 41 (17.9) 5 (12.2) Suicide 19 (8.3) 3 (15.8) Undetermined" 47 (20.5) 18 (38.3)

IVDU 50 26 (52.0) Total 229 46 (20.1) "Eighty-five percent (40147) cases were drug related.

Race

Manner of death

MATERIALS AND METHODS From January to February 1993, blood and paired

urine samples were collected sequentially from 229 au- topsy cases examined at the Office of Chief Medical Examiner, State of Maryland. Following centrifugation at 2,500 rpm for 10 minutes at ambient temperature, the sera were recovered and stored frozen at -20°C until testing. About 10% of the blood samples were slightly hemolyzed. Aliquots of the corresponding urine specimens were collected in sterilized plastic tubes and kept at -20°C, 4"C, room temperature (20"C), and 37°C in order to study the storage conditions which may in- fluence antibody stability. Details of the demographic data for the study population can be found in Table I. The average time post-mortem of collection was 16 hours (range 8-26 hours).

Testing was carried out using the Ortho HCV second generation screening assay (Raritan, NJ), a 96-well en- zyme immunoassay (EIA) format that utilizes antigen- coated plates containing the recombinant HCV pro- teins c100-3, c200, and c22-3. Screening was also performed using a second generation EIA, INNOTEST HCV Ab (Innogenetics, Antwerpen, Belgium) that in- corporates synthetic peptides of HCV core, NS4, and NS5 proteins as the coated antigens in the 96-well plate. All testing was carried out according to the man- ufacturer's instructions, except that the urine sample volumes were changed from 20 p1 to 40,80, and 160 p1 for the Ortho test, and from 10 pl to 20,40,80, and 100 p1 for the INNOTEST. The sample diluent was de- creased correspondingly in order to keep the total well volume at 200 pl as recommended by the manufacturer. All the initially reactive samples were repeated in du- plicate to be considered as true reactive by the screen- ing assays. Results were expressed as the ratio of the absorbance value of the sample divided by the cutoff value (ODXO). A sample was considered reactive for anti-HCV antibodies if the ratio was 31. Any samples that were reactive or disconcordant between serum and

TABLE 11. Distribution of Seropositive Cases bv Age

Age (vear) Population (%) HCV positive (%)

20-29 72 (31.4) 4 (5.6) 30-39 69 (30.1) 24 (34.8) 40-49 49 (21.4) 14 (28.6) 50-59 39 (17.0) 4 (10.3) Total 229 46 (20.1)

urine were further tested by the Abbott HCV EIA 2.0 (Abbott Labs, IL). The Abbott HCV EIA 2.0 has HCV recombinant antigens bound to the solid phase and has been developed to detect antibodies to proteins ex- pressed by putative structural (HC-34) and nonstruc- tural (HC-31, c100-3) regions of the HCV genome.

Serahrine samples that were repeatedly reactive by either screening test were tested further by a supple- mental assay (INNO-LIA HCV Ab I1 and 111, Innoge- netics). This supplemental test contains four levels of IgG controls and a total of six discrete HCV antigens representing epitopes from both structural and non- structural proteins. The criteria for positivity of the INNO-LIA HCV Ab I11 were at least two HCV antigen lines with a reactivity rating of 1 + minimum, or a 2 + reactivity to any one antigen, as recommended by the manufacturer. For urine sample testing, the volumes used were changed to 100 pl instead of 10 pl as recom- mended for serum.

pH was determined for the urine samples in order to ascertain if variations would influence the results. In addition, 20 selected positive and 10 negative urine samples were frozedthawed for up to 10 times to study antibody stability of the urine samples.

RESULTS Of the 229 autopsy specimens tested, 46 (20.1%) had

confirmed serologic evidence of HCV infection. The se- roprevalence of HCV in the different groups is shown in Tables I, 11. Increased seroprevalence rates of HCV were found in males, those who were 30-39 years old, blacks, and IVDU-related cases of death.

Forty-six serum samples were reactive by both screening and supplemental tests. Of the 46 seroposi- tive samples, 44 paired urine samples produced repeat- edly reactive results by both the INNOTEST and Ortho screening assays. The two urine specimens that were nonreactive by the INNOTEST had OD/CO of 0.16 and 0.40 using a volume of 100 pl and 0.24 and 0.54 using 200 pl, respectively. The corresponding serum ODE0 ratios were 4.13 and 6.19, respectively. Similar results were also produced using the Ortho test when testing these two samples (Table 111). One of the two samples was reactive by the Abbott EIA when the urine sample volume was increased to 200 p1. All 46 of the urine samples were positive when tested by the INNO-LIA. Volumes of the urine samples used in the INNO-LIA assay were adjusted to 100 p1 instead of 10 ~ 1 , and the comparison of different volumes with band intensity was not significantly different. The comparison of dif- ferent antibody intensity using equal volumes of two

HCV Antibody Testing in Urine 189

TABLE 111. Comparison of ODIC0 Ratios by Different Screening Assays in Two Discordant Urine/Serum Pairs

INNOTEST (pl) Ortho ( ~ 1 ) Abbott (p1) Serum Urine Serum Urine Serum Urine

No. 10 100 200 20 100 200 5 100 200

347 4.13 .16 .24 5.16 .24 .88 >3.6 .66 1.89 573 6.19 .40 .54 5.16 .18 .26 >3.6 .ll .19

Fig. 1. INNO-LIA I1 profiles of two (A,B) representative serum (S) / urine (U) pairs at 10, 20, 40 ~1 volumes. PC, positive control; NC, negative control.

Fig. 2. Distribution of OD/CO ratio of 229 serudur ine pairs by the Ortho test. Fig. 3. Different urine volumes and their corresponding O D E 0

values (SD) of 44 specimens by the Ortho test.

representative positive serudur ine samples is shown in Figure 1. The distribution of urinary and serum IgG anti-HCV reactivities using the Ortho test is shown in Figure 2. Different urine volumes were also used in the screening assays and the corresponding ODIC0 are shown in Figure 3. As indicated, OD/CO increased se- quentially as volumes were increased. The mean pH in positive urine samples were 6.1 2 0.6 and the negative

urine samples were 6.5 2 0.5. Figure 4 shows the effect of different storage conditions on OD/CO ratios of 20 samples by the Ortho test. Ten negative urine samples were stored at 37°C for 20 days and their O D E 0 ratios remained nonreactive. There was no significant change of OD/CO ratios when using different storage tempera- tures for 20 positive urine samples for 20 days. Freezingithawing urine more than 10 times did not

Zhang et al.

puncture requires trained technicians, special equip- ment, and a compliant population, the collection of urine or saliva samples may be more attractive. An advantage of the use of urine is that it has the potential to reduce the risk of transmission of blood-borne infec- tions arising from the multiple use of instruments and from related needlestick injuries.

Recently, data indicated that HIV-1-infected subjects contained measurable anti-HIV-1 IgG in urine [Con- nell et al., 1990; Desai et al., 1991; Constantine et al., 19941. We have previously reported a preliminary study where we successfully detected antibodies to HCV in urine [Constantine et al., 19921. Our present results showed a high seroprevalence of HCV infection among forensic autopsy cases in the Office of Chief Medical Examiner, State of Maryland. Our previous study [Li et al., 19931 also indicated that the seropreva- lence of HCV in the forensic autopsy population was very high.

The use of urine as an alternative sample for detect- ing antibodies to HCV is encouraging in our study. A symmetric distribution of OD/CO values was seen about a mean which was well to the right of the cut-off, suggesting that most of those in whom reactivity was detected contain specific antibody [Nicholson et al., 19911. The volume used in the EIA for urine antibody testing was at least 10 times higher than recommended by the manufacturer for serum. There was no compro- mise of OD/CO using the urine pair in comparison with serum except one sample which consistently showed a negative reaction by all three EIAs even when urine volume was increased to 200 pl. The INNO-LIA supple- mental test showed that this sample was positive. This phenomenon may imply that the commercial EIA screening tests for HCV may not be as sensitive as supplemental assays such as the INNO-LIA when us- ing a urine specimen.

The pH of the urine samples did not seem to influence antibody testing. Similarly, one study showed that changes in urine pH had no significant effect on the absorbance value of the enzyme-linked immunosorbent assay (ELISA) for antibody detection [Alemohammad et al., 19931.

It has been reported that urine samples should be stored at 4°C with no freezing or freezekhaw cycles [Mortimer and Parry, 19911. Our results indicate that freezing urine samples will not affect the stability of antibodies when compared with 4°C storage. Further- more, freezing urine may prevent or at least delay bac- terial growth and biochemical changes of the samples. One study found that four of six anti-HCV negative sera became reactive after heating at 56°C for 5 min- utes [Mortimer et al., 19891. Our results indicate that storing urine at 37°C for 20 days does not produce any false positivity.

Different screening assays for detecting HCV anti- bodies may produce different results. We have shown previously that the use of different kinds of screening assays simultaneously may be a good strategy for pre- venting false negatives or loss of sensitivity [Bansal et

190

5

4

0 3 ? a 0

2

1

0

" - 2 o c - room temperature ---- 3 7 c

0 5 10 1 5 20 25 3 0

Days

Fig. 4. Different storage conditions and corresponding mean OD/CO ratios from 20 urine reactive samples by the Ortho test.

120

100

80

- 60 a L E 40

20

0 NS4 NS5 core1 core2 core3 Core4

Fig. 5. Frequency (percent) of reaction with HCV antigens by anti- bodies in 46 serumiurine pairs using the INNO-LIA I1 assay.

affect OD/CO and there were no false positive reactions (data not shown). The OD/CO ratios for the positive samples were decreased only slightly after multiple freezeithaw cycles. A comparison of reactivity with an- tigens using the INNO-LIA is depicted in Figure 5. For both the serum and urine samples, the highest frequen- cies of reactivity were to NS4 and core 2 antigens and the lowest to core 3. It was also noted that the control bands (anti-human IgG positive control) were more in- tense in the urine samples than in sera. There were five serum/urine pairs in which the urine samples produced strong reactions on one of the six antigens (two NS5; two C2; one Cl), while the serum pair showed no reac- tion or a very weak reaction.

DISCUSSION One of the difficulties of epidemiologic studies is to

collect blood samples from the population. Since veni-

HCV Antibody Testing in Urine 191

al., 19931. In the present study, the data indicate that the Abbott second generation HCV EIA may be the most sensitive screening test for detecting HCV anti- bodies in urine, perhaps due to the increased surface area of the solid phase beads.

For confirming antibodies in urine, the INNO-LIA assay produced similar results in serum and urine. However, the frequency of different antibody bands that were positive was relatively higher in serum than that of urine. Of the 46 positive samples confirmed by the INNO-LIA, we found that in 5 paired serumhrine samples, in which urine produced at least one antibody positive band, the paired sera were negative or showed only a very weak reaction to these antigens. It is very interesting that the IgG antibody intensity control bands on the INNO-LIA were more strongly reactive in urine than in those of the paired serum even when using the same volumes. It is unclear if the concentra- tions of the antibodies in urine are higher than those of serum. But the stronger reaction of IgG controls on the strips by most urine samples and the five discrepant urinekerum paired samples may indicate that a t least in some cases, the concentrations of antibody in urine might be higher. This phenomenon has also been inves- tigated by others for HIV urine antibody testing [Mon- tagnier, 19931.

In summary, the use of urine as an alternative speci- men for the dignosis of some virus infections such as HIV and HCV is encouraging. The advantages of using urine as a specimen include better safety, easier han- dling, and cost saving, especially for epidemiologic studies. Our data indicate that the commercial HCV antibody test kits are adaptable to detect antibodies against HCV in urine. However, the occurrence of one or two false negative results using urine samples in our study by three screening tests indicates the need to further optimize assays and to perform additional stud- ies before a firm conclusion regarding the routine use of urine can be made.

REFERENCES Alemohammed MM, Foley TJ, Cohen H (1993): Detection of immuno-

globulin G antibodies to Helicobmterpylori in urine by an enzyme immunoassay method. Journal of Clinical Microbiology 31:2174- 2177.

Allain JP, Dailey SH, Laurian Y, Vallari DS, Rafowicz A, Desai SM, Devare SG (1991): Evidence for persistent hepatitis C virus (HCV)

infection in hemophiliacs. Journal of Clinical Investigation 88: 1672-1679.

Alter MJ, Hadler SC, Judson FN, Mares A, Alexander WJ, Hu PY, Miller JK, Moyer LA, Fields HA, Bradley DW, Margolis HS (1990): Risk factors for acute non-A, non-B hepatitis in the United States and association with hepatitis C virus infection. Journal of the American Medical Association 264:2231-2235.

Bansal J , Constantine NT, Zhang X, Callahan JD, Marsiglia VC, Hyams KC (1993): Evaluation of five hepatitis C virus screening tests and two supplemental assays: Performance when testing a high risk population in the U.S.A. Clinical and Diagnostic Virol- ogy 1:113-121.

Cao Y, Hosein B, Borokowsky W, Mirabile M, Baker L, Baldwin DF, Poiesz BJ, Friedman-Kien AE (1989): Antibodies to human immu- nodeficiency virus type 1 in the urine of HIV-1 seropositive indi- viduals. AIDS Research and Human Retroviruses 5:311-319.

Choo QL, Weiner AJ, Overby LR, Kuo G, Houghton M, Bradley DW (1990): Hepatitis C virus: The major causative agent of viral non-A, non-B hepatis. British Medical Bulletin 46:423441.

Connell JA, Parry JV, Mortimer PP, Duncan RJS, Mclean KA, Johnson AM, Hambling MH, Barbara J , Farrington CP (1990): Preliminary report: Accurate assays for anti-HIV in urine. Lan- cet 335:1366-1369.

Constantine NT, Zhang X, Li L, Smialek J E (1992): Detection of anti- bodies to hepatitis C virus in urine. Lancet 339:1607-1608.

Constantine NT, Zhang X, Li L, Bansal J , Hyams KC, Smialek J E (1994): Application of a rapid assay for detection of antibodies to HIV-1 in urine. American Journal of Clinical Pathology 101:157- 161.

Desai S, Bates H, Michalski FJ (1991): Detection of antibody to HIV-1 in urine. Lancet 337:183-184.

Esteban JI, Esteban R, Viladomiu JC, Lopez TJ, Gonzalez M, Roget V, Vargas J , Genesca MB, Guardia J (1989): Hepatitis C virus anti- bodies among risk groups in Spain. Lancet 2:294-297.

Kuo G, Choo QL, Alter HJ, Gitnick GL, Redeker AG, Purcell RH, Miyamura T, Dienstag JL, Alter MJ, Stevens CE (1989): An assay for circulating antibodies to a major etiologic virus of human non-A, non-B hepatitis. Science 244:362364.

Li L, Zhang X, Constantine NT, Smialek JE (1993): Seroprevalence of parenterally transmitted viruses (HIV-1, HBV, HCV, and HTLV- 1/11] in forensic autopsy cases. Journal of Forensic Sciences 38: 1075-1083.

Montagnier L (1993): Divergence of AIDS virus discloses need for improved diagnostic tests. IXth International Conference on AIDS, Berlin, June 9-16,1993,

Mortimer PP, Parry JV (1991): Non-invasive virological diagnosis: Are saliva and urine specimens adequate substitutes for blood? Medical Virology 1:73-78.

Mortimer PP, Coher BJ, Litton PA (1989): Hepatitis C virus antibody. Lancet ii:798.

Nicholson S, Leslie DE, Efandis T, Fairley CK, Gust ID (1991): Hepa- titis C antibody testing: Problems associated with nonspecific binding. Journal of Virological Methods 33:311-317.

Perry KR, Parry JV, Vandervelde EM, Mortimer PP (1992): The detec- tion in urine specimens of IgG and IgM antibodies to hepatitis A and hepatitis B core antigens. Journal of Medical Virology 38:265- 270.

Van der poel C, Reesink HW, Lelie PN (1989): Anti-hepatitis C anti- bodies and non-A, non-B post-transfusion hepatitis in the Nether- lands. Lancet ii:297-298.