Journal Logo

Original Articles: Clinical Transplantation

Serological Viral Testing of Cadaveric Cornea Donors

Challine, Dominique1; Roudot-Thoraval, Françoise2; Sabatier, Patrick3; Dubernet, Fabienne1; Larderie, Patrick1; Rigot, Pierrette1; Pawlotsky, Jean-Michel1,4

Author Information
doi: 10.1097/
  • Free


Tissue graft recipients are exposed to a risk of viral transmission from the donor, and virological screening is thus compulsory before transplantation. Hepatitis B virus (HBV) infection has been observed in cornea recipients (1), but not human immunodeficiency virus (HIV), human T-cell leukemia virus (HTLV) or hepatitis C virus (HCV) infection (2). In France it is obligatory to test cornea donors' blood for the following markers of viral infections: anti-HIV1 and anti-HIV2 antibodies (by means of two different assays), HIV p24 antigen, anti-HTLV-I antibodies, HBs antigen, anti-HBc antibodies, and anti-HCV antibodies (by means of two different assays until January 2004, a single assay since this date).

Cornea grafts can originate from brain-dead heart-beating multiple organ donors hospitalized in intensive care units, or from cadavers. In France, cornea donors with HIV or HTLVI markers, anti-HCV antibodies, HBs antigen and/or anti-HBc antibodies have been excluded from donation, whereas cornea donors with a past HBV infection profile (both anti-HBs and anti-HBc antibodies) or with an HBV vaccination profile (anti-HBs antibodies only) were not until the end of 2005. The presence of anti-HBc antibodies, with or without anti-HBs antibodies, is now sufficient to exclude from donation. Heart-beating donors undergo emergency virological screening before organ harvest, and serum samples are of good quality, apart from possible hemodilution (3, 4). In contrast, cadaveric donors are tested with blood samples collected at variable times after death. These sera are often of poor quality and frequently yield falsely positive results in serological assays (5). This can result in grafts being needlessly discarded, which may aggravate the shortage of corneal grafts.

The objective of this study was to document the influence of the timing of blood collection for serological testing after death, and the relationship between the macroscopic aspect of serum and serological test results in cadaveric cornea donors.


Cornea Donors

From May 2000 to December 2002, 565 consecutive cadaveric cornea donors were systematically screened for serological markers of HIV1 and HIV2, HTLV1, HBV and HCV infections. They comprised 328 men and 237 women, and their mean age was 71 years (range 16 to 102 years). The following data were collected: gender; age; macroscopic aspect of the donor's serum (defined as “normal” or “abnormal” if hemolyzed, icteric, or cloudy); and the time of blood collection after death.

Virological Testing of Donor Blood

Donor blood samples were centrifuged immediately on their arrival in the laboratory, and serological tests were usually performed the same day (always within 48 hr). Anti-HIV1 and -HIV2 antibodies were detected with two assays, namely Access HIV1-2 New EIA (Pasteur-Biorad, Marnes- la-Coquette, France) and Vidas HIV Duo EIA (Biomérieux, Lyon, France). These assays belong to the most recent generations, i.e. 3rd- and 4th-generation, respectively. HIV p24 antigen was detected with the Vidas HIV p24 II method (Biomérieux) and its presence was systematically confirmed by neutralization. Anti-HTLV-I antibodies were detected with the Murex HTLV I+II EIA (Abbott Laboratories, Abbott Park, Illinois). HBs antigen, anti-HBc antibodies and anti-HBs antibodies were detected with the respective Vitros assays (Ortho-Clinical Diagnostics, Raritan, New Jersey), and the presence of HBs antigen was systematically confirmed by neutralization. Anti-HCV antibodies were detected with the Vitros anti-HCV assay (Ortho-Clinical Diagnostics) and the Access HCV Ab Plus method (Pasteur-Biorad). Serological results were interpreted according to the manufacturers' instructions. An equivocal result was defined as a result ranging within the gray zone of the assay, i.e. generally ±10%–20% of its cut-off value.

If the results were equivocal, or if a discrepancy arose between the two serological tests for anti-HIV and anti-HTLVI antibodies, the samples were systematically centrifuged and retested with the initial assays plus Vitros Anti-HIV 1+2 (Ortho-Clinical Diagnostics), Innotest™ HIV Antigen (Innogenetics, Gent, Belgium), and Serodia™ HTLV-I (Fujirebio Inc., Tokyo, Japan). The Vidas HBsAg, Vidas anti-HBs Total and Vidas anti-HBc Total II methods (BioMérieux) were used to check equivocal results for HBs antigen, anti-HBs antibody and anti-HBc antibody, respectively, and the Ortho HCV 3.0 Elisa (Ortho-Clinical Diagnostics) or Innotest™ HCV Ab IV (Innogenetics) was used for anti-HCV antibodies.

The final result for a given marker was only considered negative if all tests were negative.

Statistical Analysis

The results are reported as means±1 standard deviation or as percentages with their 95% confidence intervals. Relationships between categorical data were identified by using the chi square test or Fisher's exact test, as appropriate. Quantitative data were compared by using a t test and ANOVA or a nonparametric test such as the Mann-Whitney or Kruskal-Wallis test, depending on the number of groups.


Global Results of Serological Testing

Table 1 shows the prevalence of virological markers, based on the results of both initial and confirmatory testing. The prevalence rates were higher than reported in the French general population and in our population of living donors of organs and tissues other than corneas (Table 2) (6, 7). Not enough serum was available for retesting in 12 of the 565 cadaveric donors, and the grafts were discarded for this reason. Among the remaining 553 donors, 367 (66.4%) were negative in all serological tests at the first determination, and the corneas could thus be released for transplantation without confirmatory testing. Thirteen donors (2.4%) were unequivocally positive for at least one marker. Corneas from 67 donors (12.1%) were accepted after at least one marker had been controlled to resolve an equivocal or discrepant initial result. One hundred six donors (19.1%) were rejected because control procedures gave an unequivocally positive result, or because issues raised by the initial assays could not be resolved. Overall, among the 553 donors with sufficient available blood for full testing, 434 (78.5%) were accepted and 119 (21.5%) were rejected on the basis of virological testing.

Prevalence of serological virological markers that tested positive (according to the manufacturers' criteria)
Prevalence of virological markers in the French general population, and in our populations of heart-beating organ donors and living donors of tissues other than corneas

Relationship Between the Macroscopic Aspect of Serum and the Results of Virological Testing

The macroscopic aspect of the serum samples was noted in 554 of the 565 cases. It was considered normal in 299 cases (54.0%) and abnormal (i.e. hemolyzed, icteric or cloudy) in 255 cases (46.0%).

The prevalence of the virological markers is shown in Table 1 according to the macroscopic aspect of the sera. Overall, sera with an abnormal macroscopic aspect were significantly more likely to test positive for anti-HIV, p24 antigen, anti-HTLV1, and anti-HBs, and tended to be more frequently positive for anti-HBc. The results of anti-HCV detection did not vary according to the macroscopic aspect of serum, with either of the assays used in this study. Thus, overall, macroscopically abnormal sera generated false-positive results with all the tests used in this study, except for the HCV tests and, maybe, the anti-HBc test used in this work.

The macroscopic aspect was recorded for 545 of the donor samples that were sufficiently abundant for retesting: 295 were macroscopically normal (54.1%) and 250 were abnormal (45.9%). At least one marker had to be retested in 170 of these samples (see Methods), comprising 46 macroscopically normal sera and 124 macroscopically abnormal sera (15.6% versus 49.6%, P<0.001).

In order to assess the practical implications of our findings, we examined the relationship between the macroscopic aspect of the sera and the final decision on whether to use the corresponding corneas for transplantation. This information was available for 550 donors (297 macroscopically normal sera, 253 abnormal). One hundred twenty of these donors (21.8%) were excluded from donation because of the results of virological testing, of whom 84 (70.0%) had macroscopically abnormal serum and 36 (30.0%) macroscopically normal serum. Overall, 36 of 297 cornea donors with macroscopically normal serum and 84 of 253 donors with macroscopically abnormal serum were excluded from donation because of the results of serological testing for viral infections (12.1% versus 33.2%, P<0.001).

Influence of the Time of Blood Sampling After Death on the Results of Virological Testing

The time of blood collection after death was noted for 420 of the 565 donors. It was on average 22 hr and 4 min ± 12 hr and 25 min (range: 25 min to 71 hr).

The prevalence of positive serological results did not differ significantly between samples collected less than 12 hr, 12 to 24 hr, 24 to 36 hr and more than 36 hr after death (Table 1). Among the 420 donors for whom the time of blood sampling was known, enough serum was available to control equivocal results in 409 cases. Among these donors, 126 (30.8%) had to be retested for at least one marker (with the initial and additional assays) in order to draw firm conclusions. These samples had been taken significantly later than those that did not need checking (on average 25 hr and 14 min versus 20 hr and 38 min, P<0.001). As shown in Figure 1, the proportion of donor samples that needed to be retested because of equivocal or discrepant results increased significantly with the interval between death and sampling (P<0.001). Cornea grafts that were not used because of virological positivity had been taken from donors who were sampled significantly later than donors whose corneas were used (24 hr and 48 min versus 21 hr and 12 min, respectively; P=0.015).

Proportion of postmortem cornea donor samples that needed to be retested because of an equivocal or discrepant serological result on initial testing, according to the interval between death and blood collection and the macroscopic aspect of serum (all sera: black; macroscopically normal sera: light gray, macroscopically abnormal sera: dark gray). The actual numbers are given in parenthesis.

Relationship Between the Macroscopic Aspect of Serum and the Time of Blood Sampling After Death; Respective Influence on Virological Results

Macroscopically abnormal blood samples were collected significantly later than macroscopically normal samples (24 hr and 54 min versus 19 hr and 36 min, P<0.001). As shown in Figure 2, about 25% of blood samples collected less than 12 hr after death were macroscopically abnormal, compared to more than 50% of samples collected more than 12 hr after death.

Proportion of macroscopically abnormal cornea donor sera according to the interval between death and blood collection. The actual numbers are given in parenthesis.

As the macroscopic aspect of serum and the time of blood collection were significantly related, we examined which of these two parameters had most influence on the serological results. When macroscopically normal and abnormal sera were considered separately, the time of blood collection after death did not influence the number of samples that needed to be retested because of equivocal or indeterminate results (data not shown). In contrast, within each time interval (<12 hr, 12–24 hr, 24–36 hr and >36 hr), significantly more samples needed to be retested when the serum was macroscopically abnormal than when it was normal (Fig. 1). If we take the example of anti-HIV antibody detection, the number of initially positive samples did not differ according to the sampling time. However, within each time interval after death (except after 36 hr), significantly more macroscopically abnormal than normal samples were anti-HIV-positive (Fig. 3).

Proportion of anti-HIV antibody-reactive cornea donor samples on first-line assay, according to the macroscopic aspect of serum (normal: light gray, abnormal: dark gray) and the interval between death and blood collection. The actual numbers are given in parenthesis.


False-positive virological results have been reported for donor samples taken postmortem (5, 8–10). The results of the present study, based on a large series of cadaveric cornea donors, show: (i) a consistently higher proportion of sera positive for at least one viral marker than in the general population and in heart-beating donors (6, 7); (ii) a very high proportion of samples that needed to be retested with the initial and additional assays because the initial results were equivocal or discrepant (in the case of anti-HIV and anti-HCV antibodies); (iii) as a result, a high proportion of cornea grafts that were discarded; (iv) a significant relationship between the macroscopic aspect of serum at the time of testing and a positive, equivocal or discrepant result for all markers except anti-HCV antibodies (this latter result contrasts with previous reports for second-generation anti-HCV assays (11, 12) and suggests that the third-generation anti-HCV assays used here are more specific); (v) a significant relationship between the macroscopic aspect of serum and the exclusion of potential cornea grafts; and (vi) a significant relationship between the macroscopic aspect of serum and the time of blood sampling after death. The longer the interval between death and blood sampling, the more likely the serum was to have an abnormal aspect (hemolyzed, icteric or cloudy). In addition, the clear and direct relationship between the macroscopic aspect of serum and the results of serological testing establish for the first time a clear relationship between the quality of postmortem blood samples and the results of virological testing.

With the current cornea graft shortage, it is important to avoid false-positive virological results and to ensure that corneas which carry no viral risk are not needlessly discarded. Our study suggests that cadaveric donors should not be sampled for serological testing more than 12 hr after death. A study based on a questionnaire sent to 16 tissue banks in the United Kingdom suggested that hemolysis increased with the time of sampling after death and was associated with more frequent serological reactivity (10). Our study shows that the longer the interval between death and sampling, the lower the likelihood that the corneas will be used, owing to the increased probability of a positive, equivocal or discrepant serological test result in the absence of any viral infection. Alternatives include testing of premortem samples, if available, and the use of other virological methods to test for viral replication. Testing of samples taken several days to weeks prior to death may reliably show the absence of specific antibodies in good-quality serum. According to the European Eye Bank Association (EEBA), if the postmortem serum is hemolyzed or hemodiluted, then the results of tests done up to seven days before donation can be used, provided the patient received no transfusions or infusions in the interval (13). This, however, raises issues such as the traceability of sera originating from different laboratories, and the ethical question of sampling, with a view to donation, a patient who may not be aware of his or her imminent demise. In addition, viral infection between premortem sampling and death cannot be ruled out, and transplant teams would likely feel uncomfortable about using tissues from a donor who was seronegative before death but had a positive or equivocal serological result postmortem (5). In this respect, our results suggest that the quality of the postmortem sample and the interval between death and sampling should be taken into account when interpreting apparent seroconversion close to death.

Another option is to use molecular methods to detect viral genomes, either routinely or only in antibody-positive samples. Nucleic acid testing (NAT) has good sensitivity and specificity for direct markers of infection (14, 15). It has proven to be efficient in reducing the residual risk of both HIV and HCV transmission in the blood transfusion setting, although the cost-efficacy of this measure remains debated (16). Only assays based on transcription-mediated amplification (TMA) have been approved for use in cadaveric donor samples, but polymerase chain reaction (PCR)-based methods are currently being validated for this use. Using NAT testing, we recently showed that the viral nucleic acids are detectable in most organ and tissue donors with serological markers of HIV or HCV infection (17). We also identified organ donors positive for HCV RNA but negative for HCV serological markers (17). Such donors have been recently shown to be able to transmit HCV infection to both organ and tissue recipients (18). In addition, none of 912 brain-dead heart-beating organ donors without HBs antigen, anti-HBc antibodies or anti-HBs antibodies were found to be HBV DNA positive, whereas HBV DNA was found in one out of 521 organ donors with anti-HBs antibodies alone in our recent study (19). However, NAT in cadaveric cornea donors is problematic (20). Its sensitivity and specificity remain to be established, and the influence of sample quality will have to be determined in the same type of study as this one. Another issue is the case of positive antibody detection in a donor who is negative for the corresponding viral genome (7).

The need for systematically testing for anti-HBs antibodies in cornea donors is questionable, especially in the context of a 16.7% false-positive result rate. Until recently, corneas from cadaveric donors with isolated anti-HBs antibodies (vaccination profile) or with both anti-HBs and anti-HBc antibodies (recovery profile) were not excluded from donation in France. Since the beginning of 2006, any donor with detectable anti-HBc antibodies, with or without anti-HBs antibodies, is excluded from donation in France. In this context, the relevance of systematic anti-HBs antibody testing is low and the risk of exclusion due to a falsely positive result is not neglectable.

In conclusion, the macroscopic aspect of serum collected postmortem appears to be the best predictor of the specificity of serological testing in cadaveric cornea donors, and serological results should therefore be interpreted cautiously when the serum is macroscopically abnormal. More objective automated markers for bad serum quality, such as lactate dehydrogenase or bilirubin, or markers of hemodilution could be helpful in the context of cornea donor screening. The viral safety of cornea grafts can be improved in the following ways: blood sampling rapidly after death (within 12 hr if possible); use of premortem samples handled and stored according to strictly defined procedures; improvement of assay specificity; and evaluation of viral nucleic acid screening in this setting.


1. Hoft RH, Pflugfelder SC, Forster RK, et al. Clinical evidence for hepatitis B transmission resulting from corneal transplantation. Cornea 1997; 16: 132.
2. Eastlund T. Infection disease transmission through cell, tissue and organ transplantation: reducing the risk through donor selection. Cell Transplantation 1995; 4: 455.
3. Tjabbes H. Donor screening: an overview of current regulation and practices in Europe. Transplantation Proceedings 1996; 28: 2950.
4. Eastlund T. Hemodilution due to blood loss and transfusion and reliability of cadaver tissue donor infectious disease testing. Cell and Tissue Banking 2000; 1: 121.
5. Heim A, Wagner D, Rothamel T, Hartmann U, Flik J, et al. On of serological screening of cadaveric sera for donor selection for cornea transplantation. Journal of Medical Virology 1999; 58: 291.
6. Estimation des taux de prévalence des anticorps anti-VHC et des marqueurs du virus de l'hépatite B chez les assurés sociaux du régime général de France métropolitaine, 2003–2004. Institut National de Veille Sanitaireqq, January 2005.
7. Desenclos J-C, Costagliola D, Commenges D, et al. La prévalence de la séropositivité VIH en France. Bulletin Epidémiologique Hebdomadaire 2005; 11: 41–44.
8. Chung CW, Rapuano CJ, Laibson PR, et al. Human immunodeficiency virus p24 antigen testing in cornea donors. Cornea 2001; 20: 277.
9. Miedouge M, Chatelut M, Mansuy JM, et al. Screening of blood from potential organ and cornea donors for viruses. Journal of Medical Virology 2002; 66: 571.
10. Stanworth SJ, Warwick RM, Ferguson M, et al. A UK survey of virological testing of cadaver tissue donors. Vox Sanguinis 2000; 79: 227.
11. Laycock KA, Essary LR, Delaney S, et al. A critical evaluation of hepatitis C testing of cadaveric corneal donors. Cornea 1997; 16: 146.
12. Cahane M, Bark A, Goller O, et al. The incidence of hepatitis C virus positive serological test results among cornea donors. Cell and Tissue Banking 2000; 1: 81.
13. European Eye Bank association. Agreement on minimum standards.
14. Pawlotsky JM. Molecular diagnosis of viral hepatitis. Gastroenterology 2002; 122: 1564.
15. Tabor E, Epstein JS. NAT screening of blood and plasma donations: evolution of technology and regulatory policy. Transfusion 2002; 42: 1230.
16. Stramer SL, Glynn SA, Kleinman SH, et al. Detection of HIV-1 and HCV infections among antibody-negative blood donors by nucleic acid-amplification testing. N Engl J Med 2004; 351: 760.
17. Challine D, Pellegrin B, Bouvier-Alias M, et al. HIV and hepatitis C virus RNA in seronegative organ and tissue donoes. Lancet 2004; 364: 1611–1612.
18. Tugwell BD, Patel PR, Williams IT, et al. Transmission of hepatitis C virus to several organ and tissue recipients from an antibody-negative donor. Ann Intern Med 2005; 143: 648.
19. Chevaliez S, Challine D, Brillet R, et al. Full-length preS1-preS2-HBs antigen (HBsAg) sequence analysis of HBV surface mutants in HBV DNA-positive organ, tissue and cell donors with various HBV serological profiles. J Hepatol 2006; 44(suppl. 2): S22.
20. Strong DM, Nelson K, Pierce M, et al. Preventing disease transmission by deceased tissue donors by testing blood for viral nucleic acid. Cell Tissue Bank 2005; 6: 255.

Cornea; Transplantation; Viral risk; Serological testing

© 2006 Lippincott Williams & Wilkins, Inc.