Indian Journal of Medical Microbiology Medknow Publications on behalf of Indian Association of Medical Microbiology ISSN: 0255-0857 EISSN: 1998-3646 Vol. 26, Num. 4, 2008, pp. 327-332

Indian Journal of Medical Microbiology, Vol. 26, No. 4, October-December, 2008, pp. 327-332

Original Article

An evaluation of saliva as an alternative to plasma for the detection of hepatitis C virus antibodies

Moorthy M, Daniel HD, Kurian G, Abraham P

Department of Clinical Virology, Christian Medical College, Vellore, Tamil Nadu - 632 004
Correspondence Address:Department of Clinical Virology, Christian Medical College, Vellore, Tamil Nadu - 632 004
priyaabraham@cmcvellore.ac.in

Date of Submission: 31-Mar-2008
Date of Acceptance: 05-May-2008

Code Number: mb08103

Abstract

Keywords: Gingivo-crevicular transudate, Hepatitis C virus Antibody, Saliva

Hepatitis C virus (HCV) is a parenterally transmitted virus that is responsible for 170 million cases of chronic hepatitis in the world. About 75-80% of those persons infected with HCV tend to become chronic carriers and the majority of these patients are asymptomatic. The natural history of hepatitis C virus infection has a prominent latent phase where the patient is infected but does not manifest disease. Infected patients may remain undiagnosed due to lack of testing facilities in low resource setting or due to an unwillingness of the patient to provide a blood sample for testing. In such situations, HCV may continue to spread in the community. Therefore, it is of paramount importance to any health care system to detect latently infected persons in order to prevent spread of infection. Much of the data on HCV seroprevalence in India comes from larger health-care facilities with established clinical and laboratory facilities for the management of HCV infection while data from the community is very limited. Seroepidemiological studies are hampered by the reluctance of the patients to undergo a venipuncture to provide blood for testing. This problem can be circumvented by the use of alternative specimens like saliva, gingival-crevicular transudate and urine for the detection of antibodies to infectious agents.

Saliva-based screening assays have been extensively evaluated for screening of antibodies to human immunodeficiency virus (HIV). Standardization of saliva testing for HCV-Ab detection and its adoption into routine practice would greatly improve the ability to determine the prevalence of HCV infection in our country.

The study aimed to compare HCV-antibody (HCV-Ab) screening in saliva with that of serum/plasma among high-risk chronic liver disease individuals in a tertiary care hospital in south India.

Materials and Methods

Institutional review board (IRB) approval was sought before the commencement of the study. Consent was sought from chronic liver disease (CLD) patients attending the liver clinic of Christian Medical College, Vellore for recruitment into this study. CLD patients, as part of their management, are sent to the department of Clinical Virology to provide a blood sample for DNA/RNA detection of hepatitis B virus (HBV) and HCV respectively. After informed consent, a visual examination of the oral cavity was performed to exclude patients with evidence of ulceration, bleeding or inflammation of the gingival mucosa. Patients with evidence of immunosuppression were also excluded from participation in this study. A sample of blood and saliva (designated as a sample-pair) was collected from each recruited patient. Blood drawn under sterile conditions was collected in sterile tubes with EDTA.

Saliva fluid was collected using a commercial collection device - OmniSal (Saliva Diagnostic systems Inc., NY, USA). The device consists of an absorbent cotton pad attached to a plastic stem containing a window that is white in color. The pad was placed under the tongue enabling absorption of fluid from the floor of the mouth by capillary action. The collected fluid travels up the stem and causes the window to turn blue in color. This serves as a "volume adequacy indicator". The pad was placed in a tube of 1.1 mL transport buffer containing phosphate-buffered saline, protease inhibitors, antimicrobial agents, and 0.2% sodium azide as a preservative. ^[1] Sample pairs were refrigerated at 4°C until processing.

The blood was centrifuged at 2000 rpm for 10 minutes and multiple aliquots of the separated plasma were stored at -70 °C for future use. Saliva was processed by vortexing the tube briefly to detach the pad from its plastic stem. The detached pad was squeezed against the wall of the tube, using a sterile forceps and then discarded. The contents of the tube were transferred to a microcentrifuge tube that was centrifuged at 2000 rpm for 10 minutes and multiple aliquots of the supernatant were stored at -20 °C until use.

Detection of HCV-Ab was carried out in sample pairs using a commercially available enzyme immunoassay (Hepanostika HCV Ultra, UBI Diagnostics, Beijing, China). For saliva testing, the manufacturer′s protocol was modified such that, 100 µL of saliva fluid was added without sample diluent. Sample pairs were tested in the same plate to minimize inter-assay variation. The cut-off absorbance value (COV) above which plasma samples were declared as positive was calculated as per the manufacturer′s instructions. Since saliva is not routinely used to screen patients, there are no standard guidelines to calculate the cut-off absorbance value for saliva. The absorbance value for saliva samples above which samples were considered positive, COV, was calculated by four methods: In the first method, the manufacturer′s recommendation for calculation of cut-off was used (COV1). In the second method, 3 standard deviations above the mean saliva absorbance of HCV seronegative samples (mean + 3SD) was chosen as cut-off (COV2). In the third method, a receiver operating characteristic (ROC) curve analysis was done for saliva absorbance values using the MedCalc statistical software (version 9.2.1.0, MedCalc Software, Belgium). The absorbance value which yielded the maximum sensitivity and specificity was chosen as the cut-off value (COV3). The ROC curve for the saliva absorbance is shown in the figure. Using the ROC method, the cutoff was lowered to obtain 100% sensitivity i.e., cutoff 4 (COV4). The accuracy indices for COV determined by each of the methods are shown in [Table - 1]. COV 1, 2 and 4 were found to be suboptimal when compared to COV 3. The accuracy indices obtained using COV3 and COV4 were compared. Sample pairs where saliva was negative and plasma was positive were termed discordant positive. The plasma samples of these pairs were tested with a commercially available recombinant immunoblot assay (RIBA, CHIRON RIBA HCV 3.0 SIA - Chiron Corporation, Emeryville, USA) to confirm plasma HCV-Ab reactivity. The results were interpreted as per the manufacturer′s instructions and plasma samples found positive in RIBA were deemed HCV seroreactive. The corresponding negative saliva sample were declared false negative.

HCV viral load was determined using a commercial real time RT-PCR assay based on Taqman chemistry (HCV RG RT-PCR, Qiagen GmBh, Hilden, Germany). To determine the status of infection, in the case of an indeterminate RIBA and additionally for comparison of saliva reactivity with viral load in plasma, HCV RT-PCR was performed.

The infecting genotype was determined in 27 out of 49 HCV seropositive plasma samples. This was performed using genotype specific primers directed against the core region of the HCV genome. ^[2],[3] The saliva antibody reactivity as determined by COV 3 and 4, were compared with infecting genotype in plasma.

Results

A total of 142 sample pairs were collected in this study. HCV-Ab testing revealed a total of 49 positive and 93 negative plasma samples. Using COV3, saliva antibody was positive for 47 samples of which 40 had a corresponding HCV-Ab positive plasma sample. The remaining saliva samples (n=7) had no corresponding HCV-Ab positive plasma sample and were declared false positives. Out of the 49 HCV-Ab positive plasma samples, 9 samples did not have corresponding saliva HCV-Ab positivity. These discordant positive (saliva negative, plasma positive) sample pairs were subjected to a RIBA. The RIBA was positive for 7 of 9 plasma samples tested and the corresponding saliva samples were declared as false negative for HCV-Ab. The remaining two samples with an indeterminate RIBA result were positive for HCV RNA by RT-PCR and their corresponding saliva samples were also declared false negative. Using COV4, a total of 89 saliva samples were found to be positive, but only 49 had a corresponding positive plasma sample and were declared true positives. Of the remaining saliva samples (n=40), none had a corresponding plasma HCV-Ab positivity, hence these were declared as false positive. The remaining saliva samples (n=53) were found to be negative and were declared as true negative as all corresponding plasma samples were negative for HCV-Ab.

Comparing the accuracy indices of serum versus salivary HCV-Ab detection using COV1, COV2, COV3 and COV4 [Table - 1], the use of COV3, was found to yield the best results. Samples pairs were classified into concordant positive (saliva positive, plasma positive), concordant negative (saliva negative, plasma negative), discordant positive (saliva negative, plasma positive) and discordant negative (saliva positive, plasma negative) based on their reactivity. The range and median absorbance values of saliva and plasma samples each group is shown in [Table - 2].

HCV viral load was estimated for all plasma samples that were HCV-Ab positive (n=49). The detection limit of the HCV RG RT-PCR assay used for viral load estimation was 300 IU/ml. The viral loads varied from < 300 to 8x106 IU/mL. HCV RNA was detected in a total of 34 plasma samples. HCV RNA positivity rate among concordant positive sample pairs (n=40), was 72.5% (29/40), while that among discordant positive pairs (n=9) was 55.5% (5/9). The range and median saliva absorbance for different ranges of viral load is depicted in [Table - 3]. Using COV3, the detection of HCV-Ab in saliva was maximal when viral load was between 300 and 10,000 IU/mL (100%). The detection rate of HCV-Ab in saliva when HCV RNA was not detected in plasma was 73.3% (11/15) while it was 85.3% (29/34) when RNA was detectable. This difference in detection rate was not statistically significant. Our data was contrary to reports by de Cock et al ^[4] and Cameron et al ^[5] who demonstrated a statistically significant difference in salivary HCV-Ab detection rate in the presence of HCV RNA in serum as compared to absence of HCV viraemia.

Of a total of 27 plasma samples genotyped, genotype 3 was most frequently detected (n=20). The other genotypes included genotype 1 (n=3), genotype 4 (n=3) and genotype 6 (n=1). The median HCV-Ab absorbance in saliva for genotype 3 was 0.283 while that in plasma was 1.405. The median absorbance for the non-3 genotypes in saliva was 0.553 and that in plasma was 1.453. The difference in saliva absorbance values between the genotype 3 and non-3 samples was not statistically significant ( p >0.20). Among the genotype 3 infected patients, a poor correlation was observed between saliva and plasma absorbance (n=20, Spearman′s coefficient of rank correlation ρ (Rho) = 0.349, p >0.12). There was a moderate correlation between saliva and plasma absorbance among the non-3 genotypes (n=7, ρ=0.571, p >0.16).

Discussion

Saliva has been suggested as a convenient specimen for the detection of antibodies to various infectious disease agents. Human immunodeficiency virus (HIV) occupies a prominent place in this regard with numerous studies reporting a favorable sensitivity of saliva for HIV antibody detection. ^[1] Saliva sample collection is easy, cheap, non-invasive and does not require specialized measures for transport. Saliva has the potential to replace serum/plasma based screening in community-based seroprevalence studies. This study aimed to adapt a routinely used immunoassay towards salivary HCV-Ab detection, using a commercially available saliva collection device.

Saliva contains a mixture of gingivo-crevicular transudate (GCT), glandular saliva, particulate debris, bacteria, epithelial cells and mucus. GCT is an ultra-filtrate of plasma that enters the oral cavity by transudation from capillaries present in the mucosa of the gingival crevice. This fluid is known to contain similar composition of immunoglobulins as found in plasma but at a concentration that is 800-1000 times lower. GCT contains higher amounts of IgG and IgM but lower amounts of IgA that is enriched in the glandular fraction of saliva. ^[6] Different components of saliva can be used for the detection of antibodies. These include stimulated or unstimulated whole saliva, glandular duct saliva and GCT. GCT is collected by the use of specialized collection devices. These devices are optimized to collect mainly GCT and minimize the amount of the glandular saliva in the specimen.

Whole saliva for detection of antibodies to HCV has been employed by Elsana et al ^[7] and Yaari et al ^[8] to yield fair results. In the light of reports of HCV RNA detectable in saliva, the collection of dribbled saliva as performed by Elsana et al ^[7] , was not chosen as the method of saliva collection as it could potentially pose a risk to the person collecting the sample. Other studies have used various collection devices including Salivette ^[5],[9] Orasure ^{[10],[11],[12]} Oracol ^[13] and OmniSal. ^[9] The OmniSal device was chosen over other collection devices in view of previous experience of our laboratory with salivary fluid for HIV antibody detection (unpublished data).

Development of an assay for detection of antibodies in saliva requires careful optimization of numerous parameters to maximize sensitivity and specificity. The various approaches include - decreasing the dilutional effect of the sample buffer, ^[5],[7],[13] exclusion of dilution step entirely, ^[12] increasing the sample input, ^{[10],[11],[14]} increasing the sample incubation time, ^{[5],[9],[10],[11],[14],[15]} increasing the time for conjugate incubation ^[9] and modification of the conjugate to detect antibodies besides IgG. ^[16] In this study, the testing protocol was modified to exclude the dilution of sample with sample diluent provided in the assay. Instead, 100 µL of saliva was directly added to the well of the microtitre plate. Saliva and plasma samples from a patient (sample pair) were tested in the same microtitre plate to exclude inter-plate and inter-assay variability. Since sample pairs were tested in the same plate, an increase in sample or conjugate incubation times could have lead to abnormally high absorbance values for plasma samples. The objective of this study was to adapt a commercial serum based antibody detection assay to testing with saliva for the purpose of large-scale community based screening for HCV infection. In order to enable this assay to blend into our existing HCV screening service, a change in sample dilution alone was effected with sample incubation times left unaltered. This was as opposed to methods adopted by other researchers where sample dilution, sample incubation temperature and duration, and conjugate incubation temperature were also modified.

Another approach to assay optimization was the use of an altered cut-off value for saliva. The different methods used to determine cut-off absorbance included - (i) reduction to an absorbance of 0.200, ^[5] (ii)calculation of standard deviations from the mean saliva absorbance of HCV seronegative samples (Mean + xSD, where x varied from 1 to 6) ^{[10],[15],[16]} (iii) lowering the value of reactivity rate (sample/cut-off value) by 20%, ^[13] (iv) use of different dilutions of the controls, ^[11] (v) use of a formula based on mean ODs for negative and positive samples ^[7],[14] and ROC analysis. ^[16] In our study we calculated the COV by 4 different methods as outlined earlier. The accuracy indices returned for each of these cut-off levels is depicted in [Table - 1]. The ROC method (using COV3) provided reasonable sensitivity and specificity desirable for a screening assay. ROC curve analysis could be used for further assay optimization to obtain 100% sensitivity albeit at the cost of specificity. The ROC method of cut-off estimation has been suggested as an effective alternative to other methods of estimation of cut-off absorbance. ^[17] Using the ROC method, the assay can be optimized to suit any need - either screening (100% sensitivity) or confirmation (100% specificity).

The accuracy indices for the various studies from different regions of the world are depicted in [Table - 4]. The indicators of test performance obtained in our study were comparable to or better than a few reports, but fared somewhat poorly compared to some others. The commonly used indicators of test performance for the evaluation of an assay are sensitivity, specificity, PPV and NPV. The disadvantage of using the latter two as indicators is that they vary with the prevalence of disease in the population. Other indicators including likelihood ratios, diagnostic odds ratio, accuracy and Youden′s J index which are not affected by disease prevalence ^[18] and can be considered as true indicators of test performance, but these are only infrequently reported in medical literature.

It may be speculated that the reason for favorable performance as reported by some investigators [Table - 4], could be due to extensive modifications of the testing protocol that was undertaken. The relatively lower accuracy indices obtained in our study could be due to the following: firstly, our stringent inclusion criteria enabling collection of saliva samples that were truly transudative in nature. Secondly, among the various reported methods of assay optimization attempted by other researchers, only a change in sample dilution / input was effected in this study, primarily to enable this assay to blend into our existing plasma HCV-Ab screening service. The modification of the conjugate in the commercial ELISA to additionally detect IgA over and above the detection of IgM and IgG as performed by Zmuda et al ^[16] was not attempted in this study. Thirdly, the OmniSal device used in this study required the use of a transport buffer that contains stabilizing agents as well as preservative. Saliva fluid collected with this device gets diluted by this buffer, leading to a further reduction in the concentration of antibodies.

As depicted in [Table - 3], saliva reactivity was detected among 73.3% of samples when the plasma viral load was below 300 while it was detected in 77.7% and 83.3% of samples with viral loads between 10,000-100,000 IU/mL and above 1,00,000 IU/mL, respectively. The maximum reactivity (100%) was achieved when the viral loads were between 300 and 10,000 IU/mL. Though the numbers are small, these findings suggest that this assay appears to be suited for detection of antibodies in persons with a low viral load.

Our study is the first of its kind attempting to compare saliva reactivity and infecting genotype. There was no significant difference in the saliva absorbance values between genotype 3 and non-3 genotypes. There was also a poor correlation between saliva and plasma absorbance in both the genotype 3 and non-3 groups. Clearly saliva reactivity is not influenced by genotype of the virus as has been seen in our study. This is an encouraging finding because other non-3 genotypes such as genotype 4 and 6 are being increasingly encountered in our country. ^[19],[20]

A review of the published literature on testing for HCV-Ab in saliva revealed many areas of heterogeneity in the data. These included geographic region, patient group studied, device used for saliva collection, ELISA kit used for testing, modification in testing protocols and methods of determining cut-off absorbance. It is noteworthy to mention that these sources of heterogeneity may hamper comparison between studies and may lead to an unclear overall picture on the performance of saliva as a substitute to blood-based screening for HCV-Ab. Studies ensuring comparable experimental conditions are needed.

In summary, salivary detection of HCV-Ab had a moderate sensitivity and specificity as compared to plasma in the present study. Salivary antibody detection was not influenced by circulating genotype or virus load. The assay needs further optimization before it can be recommended as a screening test in the general population. However, saliva may yet be an important substitute to blood-based screening for antibodies to HCV in an epidemiological setting.

Acknowledgement

References

1.	Hodinka RL, Nagashunmugam T, Malamud D. Detection of human immunodeficiency virus antibodies in oral fluids. Clin Diagn Lab Immunol 1998; 5:419-26.  Back to cited text no. 1  [PUBMED]  [FULLTEXT]
2.	Ohno O, Mizokami M, Wu RR, Saleh MG, Ohba K, Orito E, et al . New hepatitis C virus (HCV) genotyping system that allows for identification of HCV genotypes 1a, 1b, 2a, 2b, 3a, 3b, 4, 5a, and 6a. J Clin Microbiol 1997; 35:201-7.  Back to cited text no. 2  [PUBMED]  [FULLTEXT]
3.	Raghuraman S, Shaji RV, Sridharan G, Radhakrishnan S, Chandy G, Ramakrishna BS, et al . Distribution of the different genotypes of HCV among patients attending a tertiary care hospital in south India. J Clin Virol 2003; 26:61-9.  Back to cited text no. 3  [PUBMED]  [FULLTEXT]
4.	Cock De L, Hutse V, Vranckx R. Correlation between detection of antibodies against hepatitis C virus in oral fluid and hepatitis C virus RNA in serum. Eur J Clin Microbiol Infect Dis 2005; 24:566-8.  Back to cited text no. 4  [PUBMED]  [FULLTEXT]
5.	Cameron SO, Wilson KS, Good T, McMenamin J, McCarron B, Pithie A, et al . Detection of antibodies against hepatitis C virus in saliva: a marker of viral replication. J Viral Hepat 1999; 6:141-4.  Back to cited text no. 5  [PUBMED]  [FULLTEXT]
6.	Parry JV, Perry KR, Mortimer PP. Sensitive assays for viral antibodies in saliva: An alternative to tests on serum. Lancet 1987; 2:72-5.  Back to cited text no. 6  [PUBMED]  [FULLTEXT]
7.	Elsana S, Sikuler E, Yaari A, Shemer-Avni Y, Abu-Shakra M, Buskila D, et al . HCV antibodies in saliva and urine. J Med Virol 1998; 55:24-7.  Back to cited text no. 7  [PUBMED]  [FULLTEXT]
8.	Yaari A, Tovbin D, Zlotnick M, Mostoslavsky M, Shemer-Avni Y, Hanuka N, et al . Detection of HCV salivary antibodies by a simple and rapid test. J Virol Methods 2006; 133:1-5.  Back to cited text no. 8  [PUBMED]  [FULLTEXT]
9.	van Doornum GJ, Lodder A, Buimer M, van Ameijden EJ, Bruisten S. Evaluation of hepatitis C antibody testing in saliva specimens collected by two different systems in comparison with HCV antibody and HCV RNA in serum. J Med Virol 2001; 64:13-20.  Back to cited text no. 9  [PUBMED]  [FULLTEXT]
10.	Judd A, Parry J, Hickman M, McDonald T, Jordan L, Lewis K, et al . Evaluation of a modified commercial assay in detecting antibody to hepatitis C virus in oral fluids and dried blood spots. J Med Virol 2003; 71:49-55.  Back to cited text no. 10  [PUBMED]  [FULLTEXT]
11.	Sherman KE, Creager RL, O'Brien J, Sargent S, Piacentini S, Thieme T. The use of oral fluid for hepatitis C antibody screening. Am J Gastroenterol 1994; 89:2025-7.  Back to cited text no. 11  [PUBMED]
12.	Amado LA, Villar LM, de Paula VS, de Almeida AJ, Gaspar AM. Detection of hepatitis A, B, and C virus-specific antibodies using oral fluid for epidemiological studies. Mem Inst Oswaldo Cruz 2006; 101:149-55.  Back to cited text no. 12
13.	De Cock L, Hutse V, Verhaegen E, Quoilin S, Vandenberghe H, Vranckx R. Detection of HCV antibodies in oral fluid. J Virol Methods 2004; 122:179-83.  Back to cited text no. 13  [PUBMED]  [FULLTEXT]
14.	Bello PY, Pasquier C, Gourney P, Puel J, Izopet J. Assessment of a hepatitis C virus antibody assay in saliva for epidemiological studies. Eur J Clin Microbiol Infect Dis 1998; 17:570-2.  Back to cited text no. 14  [PUBMED]  [FULLTEXT]
15.	McIntyre PG, Laszlo J, Appleyard K, Ogden GR. Modified enzyme immunoassay to detect hepatitis C virus antibodies in oral fluid. Eur J Clin Microbiol Infect Dis 1996; 15:882-4.  Back to cited text no. 15
16.	Zmuda JF, Wagoneer B, Liotta L, Whiteley G. Recognition of multiple classes of hepatitis C antibodies increases detection sensitivity in oral fluid. Clin Diagn Lab Immunol 2001; 8:1267-70.  Back to cited text no. 16  [PUBMED]  [FULLTEXT]
17.	Obuchowski NA. ROC analysis. AJR Am J Roentgenol 2005; 184:364-72.  Back to cited text no. 17  [PUBMED]  [FULLTEXT]
18.	Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PM. The diagnostic odds ratio: A single indicator of test performance. J Clin Epidemiol 2003; 56:1129-35.  Back to cited text no. 18  [PUBMED]  [FULLTEXT]
19.	Raghuraman S, Abraham P, Sridharan G, Ramakrishna BS. Hepatitis C virus genotype 6 infection in India. Indian J Gastroenterol 2005; 24:72-3.  Back to cited text no. 19  [PUBMED]
20.	Raghuraman S, Abraham P, Sridharan G, Daniel HD, Ramakrishna BS, Shaji RV. HCV genotype 4--an emerging threat as a cause of chronic liver disease in Indian (south) patients. J Clin Virol 2004; 31:253-8  Back to cited text no. 20  [PUBMED]  [FULLTEXT]

Copyright 2008 - Indian Journal of Medical Microbiology

The following images related to this document are available:

Photo images