Histocompatibility typing of the human leukocyte antigen (HLA) genes is a standard test to assess living kidney donor-recipient compatibility. Because of HLA inheritance patterns, all biologically related parents and children must share at least a one-haplotype HLA match (i.e., one complement of HLA genes along a chromosome). When a transplant team evaluates a father who wishes to donate a kidney to his biological child (or vice versa), they may discover a less than one haplotype HLA match, precluding the stated biological paternity. To our knowledge, only two case reports have described the discovery of misattributed paternity in living kidney donation (1, 2). Whether transplant professionals should disclose this information is undoubtedly a complex question (3). The physician's obligations to be forthright, keep confidences, and do no harm are often presented as a single duty. In fact these obligations are distinct and potentially conflicting.
To prompt discussion, we conducted this three part study on misattributed paternity in living kidney donation. We used national transplant databases to determine the prevalence of misattributed paternity in the United States and Canada. We considered the preferences of those involved in living donation for the disclosure of incidentally discovered paternity information to potential donors and recipients. Finally, we determined current practices at Canadian transplant centers for handling such situations.
The United Network for Organ Sharing (UNOS) administers the only Organ Procurement and Transplantation Network in the United States (4). We obtained data from the UNOS/Organ Procurement and Transplantation Network Standard Transplant Analysis and Research files from 1987 to January 2008. Similarly, the Canadian Institute for Health Information provides a broad range of national health data, including information on vital organ transplantation, which is stored within the Canadian Organ Replacement Registry (CORR) (5). Data come from participating transplant centers and organ procurement organizations. We obtained data from CORR on living kidney donors from 1992 to 2006.
Case Identification and Analysis
Misattributed paternity was defined as a donor-recipient pair comprising a father and child with less than a one-haplotype HLA match. A one-haplotype match required at least one common allele shared between donor and recipient on each of the HLA-A, HLA-B, and HLA-DR loci.
For each country, transplant databases were filtered to include all cases of living kidney donor transplants with a documented biological, father-child relationship between the donor and recipient. Father-child relationship was ascertained by matching the donor's relationship to the recipient, accounting for donor or recipient gender. Accordingly, all biological parent donors had to be male (a father) and all donors described as biological children had to have a male recipient (their father). Relationship between donor and recipient was explicitly coded as “biological, blood related” (UNOS) or “living, biologically related” (CORR), and excluded any pairs coded as “nonbiological” or “living, biologically unrelated” (e.g., adoptive father, step-father). Relationship status was coded independent of HLA serology.
To determine the prevalence of misattributed paternity, HLA-A, HLA-B, and HLA-DR alleles for each father-child donor-recipient pair was examined, with a minimum and maximum match of 0 and 2, respectively, at each locus. When tissue typing was performed more than once, retyped alleles were preferentially analyzed. The validity of each allele present in the databases was confirmed using a list of all recognized serological and cellular HLA specificities (6). Clear-cut antigen splits were considered matches to their original broad specificities; associated antigen variants of original broad specificities were also regarded as matches (7). For example, HLA-A23 is a split of the original broad specificity HLA-A9. If HLA-A23 and the broad specificity HLA-A9 were both present, such was considered a match at the HLA-A locus. The number of matched alleles across the three HLA-A, HLA-B, and HLA-DR loci could range from 0 to 6. The 95% confidence interval for each proportion was calculated using the Wilson Score Method (8).
The haplotype match obtained from direct analysis of HLA data for each case was compared with the center reported haplotype match from donor histocompatibility data collection forms. We examined the data on biological mother- child kidney donor-recipient pairs. Because mothers can be virtually certain of their maternity, and oocyte donation remains exceedingly rare, we used the prevalence of observed misattributed maternity as an indirect measure of possible data error in the estimates of misattributed paternity. Adoptive mothers were excluded from this analysis. Chi-square and Fisher's exact tests were used to determine if there were significant differences in the prevalence of misattributed paternity by UNOS region and across Canadian provinces. Finally, we assessed ABO blood type compatibility between all father-child donor-recipients; a father and child with AB and O blood type respectively, or vice versa, precludes biological paternity. All analyses were conducted using SPSS 16.0 for Windows, Release 16.0.1, 2007 (SPSS Inc., Chicago, IL).
Study Setting, Participants, and Design
We performed a cross-sectional survey on a convenience sample of individuals directly involved in living kidney donation. The sample consisted of potential donors, potential recipients, and transplant professionals. We determined their preferences for sharing incidentally discovered paternity information. Potential donors were recruited from a single center in Canada through a database of individuals who had contacted the transplant program with an interest in living kidney donation. Potential recipients with no contraindications to transplantation were recruited from dialysis units. Transplant professionals (nephrologists, surgeons, and living donor coordinators) were recruited from transplant centers across Canada. By using recommended methods (9), participants were recruited in person or by mail.
The survey questions were developed by a group of nephrologists and epidemiologists involved in living kidney donation. Participants provided demographic data and completed two validated scales measuring social curiosity and self disclosure. The Social Curiosity Scale is a 10-item validated scale which assesses participants' interest in how other people behave, think, and feel (10). Agreement with various statements (e.g., “I find it fascinating to get to know new people”) were assessed on a 4-point Likert scale (“1—strongly disagree” to “4—strongly agree”). Higher scores were indicative of greater social curiosity. The self-disclosure situations survey is a 6-item validated scale (11). Participants indicate their willingness to disclose information (“1—I would be willing to discuss only certain topics and on a superficial level only” to “6—I would be willing to express, in complete detail, personal information about myself”) in a variety of situations. These scales helped gauge the personal context in a decision to disclose misattributed paternity. Participants were given a general case of a father who wished to donate a kidney to his presumed biological child. HLA testing revealed that the father and child were not biologically related. Participants were asked to indicate their degree of agreement with the following statement: “The transplant team should tell the father and child that they are not biologically related,” on a 5-point Likert scale (1—“strongly disagree” to 5—“strongly agree”). We hypothesized that, overall, opinions on this issue would be variable. The questions were pilot tested on several potential donors, recipients, and transplant professionals, and were modified to ensure they were being interpreted accurately.
We first considered the preferences of all the surveyed participants, each of whom would be in a good position to understand this issue. We then compared the opinions of potential donors, potential recipients, and transplant professionals separately. To test for baseline differences between the three groups, we used a one-way analysis of variance and Fisher's exact test. A Kruskal-Wallis test was used to determine if our three groups differed in their median response for having misattributed paternity information shared. The survey was approved by the research ethics board at the University of Western Ontario.
Through telephone interview, we communicated with transplant professionals from Canadian transplant centers performing living kidney donor transplant evaluations during the last year. We determined whether they let their patients know about the possibility of misattributed paternity before HLA testing is performed. We spoke mostly with living donor coordinators at each center. In some cases, additional information was provided by transplant nephrologists. We also asked about how past cases of misattributed paternity, if any, were handled. To protect patient privacy, we consciously limited these descriptions to key details of previous practice.
In the United States, there were 87,490 living kidney transplants performed from October 1987 to January 2008, of which 12,895 were between a living kidney donor and recipient with a reported biological, father-child relationship (Fig. 1). Analysis was performed on 12,260 (95%) eligible cases with sufficient HLA data. In Canada, there were 4994 living kidney transplants performed from 1992 to 2006. Of these, 701 were between a biological father and child. HLA data were available for 441 (63%) cases (Fig. 1).
In the United States, there were 367 (2.8%; 95% confidence interval [CI] 2.6–3.1%) cases of misattributed paternity among father-child living donor kidney transplants. The frequency of matched alleles among father-child living kidney donors and recipients is presented in Table 1. Cases of misattributed paternity included 316 (2.4%) pairs with matches on two alleles or less, 45 (0.3%) pairs with three matching alleles on two HLA loci (complete mismatch on the other locus), and six (0.1%) pairs with four matching alleles on two HLA loci (complete mismatch on the other locus). These estimates conservatively assume no misattributed paternity among all 635 eligible pairs with inadequate data to assess HLA-A, HLA-B, and HLA-DR match.
In Canada, there were 40 (5.7% [95% CI 4.2–7.7]) cases of misattributed paternity. There were 34 (4.9%) cases of father-child donor-recipient pairs with less than three matching alleles, 5 (0.7%) cases with three matching alleles on two HLA loci, and 1 (0.1%) case with four matching alleles on two HLA loci. Conservatively, these estimates assume there were no cases of misattributed paternity among the additional 260 pairs with incomplete HLA data.
Prevalence estimates from our analysis were comparable with the center reported (vs. calculated) mismatch prevalence of 3.3% (426 cases of 12,895 pairs, 95% CI 3.0%–3.6%) recorded in the UNOS database.
Analysis of 15,014 mother-child donor-recipient pairs in the United States found 288 (1.9% [95% CI 1.7%–2.1%]) cases with a less than a one haplotype match. Assuming this level of data error is present in the father-child estimates, the true prevalence of misattributed paternity may be 2.8% minus 1.9%, or 0.9%. In Canada, there were 745 documented transplants between mothers and children. Of these, 24 (3.2% [95% CI 2.2%–4.7%]) cases were not haploidentical. Assuming this level of data error is present in the father-child estimates, the true prevalence of misattributed paternity may be 5.7% minus 3.2%, or 2.5%.
There were no significant differences in the prevalence of misattributed paternity across UNOS regions (P=0.35) or across Canadian provinces (P=0.22).
Analysis of ABO blood types of biological, father-child donor-recipient pairs only revealed four cases of paternal incompatibility in the United States; however, all four pairs shared at least a one haplotype match. There were two cases of ABO paternal incompatibility in Canada; one was confirmed to have a one haplotype mismatch, whereas the other did share at least a one haplotype match.
A total of 102 participants completed the survey (35 potential recipients, 35 potential donors, and 32 transplant professionals). Baseline characteristics of the study participants are described in Table 2. Of the transplant professionals, 17 (53%) were nephrologists, 9 (28%) were living donor nurse coordinators, 5 (16%) were surgeons, and 1 (3%) was a social worker. Potential recipients were slightly older than donors and transplant professionals. Of the three groups, potential donors were more often female. Donors and recipients were predominantly white. There were no differences in social curiosity (P=0.73) or willingness to self-disclose (P=0.53) across the three groups.
Disclosure of Misattributed Paternity Information
Overall, there were variable opinions on whether the transplant team should share incidentally discovered misattributed paternity information with the father and child. Opinions were almost evenly divided: 24% of the group strongly disagreed with information disclosure, whereas an almost equal proportion (23%) strongly agreed that the information should be shared (Fig. 2a). There were statistically significant differences in the opinions by group (P<0.001). In general, potential recipients were more likely to agree with the disclosure of such information; 60% of the group somewhat agreed or strongly agreed. Among transplant professionals, 57% somewhat disagreed or strongly disagreed with the disclosure of paternity information. Potential donors' opinions were evenly divided on the issue (Fig. 2b).
We contacted 13 of 16 adult Canadian transplant programs that performed living kidney donor evaluations including HLA testing at their center within the last year. One center indicated that they were in the process of implementing a policy where donors and recipients decide before any HLA testing whether or not the transplant team discloses misattributed paternity information should it be found. This center allows adult donors and recipients to independently make their own decision regarding disclosure. The remaining centers did not have any policy on the issue, and before testing do not inform potential donors and recipients that such information may be found.
We identified three general scenarios for how transplant centers currently handle misattributed paternity information. Transplant professionals from one center reported sharing the discovery of misattributed paternity information with the family. This was performed without any discussion before HLA testing. The information was disclosed to donors and recipients simultaneously. Transplant professionals from another center described a case-by-case approach to handling discovered cases. In one case, a potential donor who knew he was not the biological father of the potential recipient, of his own accord, informed the center that misattributed paternity would be found during the evaluation process. He requested that this information not be disclosed to the recipient. The center respected his request. Transplant professionals from another center reported misattributed paternity would never be shared, citing this discovered information had no bearing on the success of the transplant.
By using primary HLA data from American and Canadian transplant registries, we provide the first national estimates of the prevalence of misattributed paternity in the living kidney donor population. Among living donations between a father and presumed biological child, and accounting for coding errors and missing data, we conservatively estimate the true prevalence to be approximately 1% to 3% in the United States and Canada. These results are consistent with a reported median level of 3.7% in the general population (range 0.8%–30%), an estimate from 17 studies conducted in eight different countries (3). Opinions were variable about disclosing the discovery of misattributed paternity to the donor-recipient pair: 23% strongly believed such information should be disclosed, whereas another 24% strongly believed it should not. Current practices are variable across transplant centers; some centers disclosed this information, whereas other centers did not.
Strengths and Limitations of This Study
National transplant registries are often used to examine determinants of transplant graft survival (4, 5). To our knowledge, this is the first study to use national registries to assess the broader issue of misattributed paternity. We used all available data (including retyped HLA genes when available), confirmed the validity of each allele identified within the databases, and accounted for variations in laboratory testing by considering associated antigens and known antigen splits (6, 7). We confirmed the presence of multiple cases of misattributed paternity with individual transplant programs in Canada. Our data are also the first to assess the preferences of individuals involved in the living donation process for the sharing of such information. We used validated surveys to assess social curiosity (10) and willingness to self disclose (11), and confirmed similarity between the three subgroups of potential donors, potential recipients, and transplant professionals. We pilot tested the survey to ensure accurate understanding of the question posed.
Missing data and inaccuracies in database coding leave some uncertainty about the true prevalence of misattributed paternity. Such an estimate requires complete data for all transplant cases and an accurate capture and coding of donor-recipient relationship, gender, and HLA alleles. We all recognize errors exist, whether during data collection or coding. We confirmed this through additional analysis of misattributed maternity between living donors and recipients. Notwithstanding, our estimated prevalence for misattributed paternity among North American countries exceeds the estimated prevalence of error. For both countries, the prevalence is likely somewhere between 1% and 3% of living kidney donations involving only fathers and children, which is approximately 0.25% to 0.50% of all living kidney donations. This is likely an underestimate of the true prevalence, because the most conclusive determination of paternity necessitates HLA typing of not only the father and child, but the mother as well (12, 13). Without maternal HLA data, some transplants that seem to have a one haplotype HLA match may be misclassified as noncases. Moreover, the discovery of misattributed paternity could potentially lead to coding the actual donor-recipient relationship as “nonbiological” when data are sent to the transplant registries. Such cases would also be missed. The direction of bias, if any, because of genetic mutations of HLA loci could not be determined.
Our survey of individuals involved in the living kidney donation process was based on convenience sampling. Though having the opinions of only father-child donor recipient pairs would have been ideal, our sampling nonetheless involved individuals who we believed would be in a good position to understand the situation of unsought information in living kidney donation. Although purposeful sampling would provide more precise estimates, nevertheless, we were able to clearly demonstrate there are discordant opinions on this issue.
Misattributed Paternity in Living Kidney Donation
The decision to donate a kidney to a friend or loved one is an emotional one. The situation becomes increasingly complicated and ethically challenging if unexpected information, such as misattributed paternity, is identified during the evaluation process. As highlighted in current practice, its potential discovery is usually omitted from discussions of the risks and benefits before the transplant evaluation (2).
Although the discovery of misattributed paternity is uncommon, we have nonetheless established that it can occur. In addition to two case reports in the transplantation literature (1, 2), we confirmed additional cases from individual transplant centers. Whether the prevalence is 0.5% or 5%, the discovery of misattributed paternity could directly affect tens to hundreds of families, because approximately 27,000 living kidney donations are performed annually around the world each year (14). When one considers the number of additional people evaluated for living donation and how the discovery of misattributed paternity may terminate the donation process, the impact may be substantially more. The social, ethical, and legal consequences of disclosure in any setting are wide reaching (3). Many family members beyond the “legal” father and child are clearly affected by this information, including the mother, the true biological father, and his family. Whether identified cases were informed of these findings before transplantation remains unknown.
To Disclose or Not to Disclose?
There are many arguments both for and against the sharing of misattributed paternity information with a potential donor and recipient. Previous studies have described the ethical considerations in detail (1, 2, 15). Briefly, nonmaleficence, maintenance of family harmony and respect for the mother's right to privacy call for nondisclosure. Respect for autonomy and informed consent and the right for a child to not be misled about genetic heritage, and possibly their genetic predisposition to developing kidney failure, favors disclosure. The effects of disclosure on the decision to pursue living kidney transplantation or subsequent family dynamics cannot be predicted.
Principles of medical ethics can certainly help guide the decision making of the transplant team. In addition, knowing the general preferences for information disclosure of those involved in the living kidney donation process is helpful. Herein, we determined that opinions on this issue are variable. In general, transplant professionals were more likely to favor nondisclosure, potential recipients were more likely to favor disclosure, and potential donors were comparatively divided on the issue. Whether transplant professionals have a duty to inform their patients about this truth is debatable. Although 41% of transplant professionals strongly disagreed that they should have to disclose such unsought information, 25% somewhat agreed that they should. There are many practical and theoretical issues surrounding truth telling in the doctor-patient relationship (17). This all speaks to the need for consensus among transplant professionals, particularly the need for a fair approach in the face of varying opinions. Disclosing such information to the 23% of potential donors who are strongly opposed to being told could lead to irrevocable harm. In a society increasingly influenced by genetics, revealing generic HLA match information to a donor and recipient (e.g., “you are a 2 of 6 match”) may encourage, rather than prevent, additional questions from donors and recipients.
Implications for Clinical Practice
The discovery of misattributed paternity during the evaluation process for living kidney donation is often unanticipated and confronted with great debate. To date, few transplant programs have formalized or even discussed what should be done if cases arise within their own centers. To this effect, it is wholly possible for transplant programs to establish, a priori, what should be done with incidentally discovered information in the face of competing ethical viewpoints. This may include a model similar to one Canadian transplant center, where they acknowledge that unsought information may be found. Consent is then obtained from the donor and recipient before the tissue typing process. Donors and recipients can decide, together or independently, whether the transplant team should disclose any unanticipated information. Certainly, there may be negative repercussions to this approach: acknowledging the possibility of discovering misattributed paternity before testing may distress many potential donors and recipients for whom this will never be an issue. Alternatively, being forthright about the potential for misattributed paternity may encourage some donor-recipient pairs who suspect issues with paternity to seek independent validation of genetic heredity (e.g., through readily available commercial genetic testing ).
If discrepant paternity information is disclosed, a standardized strategy for when and how the information is shared should be considered. Additional services and support will need to be made available to the family. The decision to disclose and the method of disclosure must also be sensitive to cultural differences between different recipient-donor families. The implications of misattributed paternity for the safety of the mother may be dire in some cultures. For centers that adopt a policy of disclosure, it might be reasonable to apply this going forward and not to past cases on record in the tissue typing laboratory.
The concept of misattributed paternity brings up a broader issue of the general utility of HLA testing. In living kidney transplantation, mismatched donor-recipient pairs have similar outcomes to haploidentical pairs (17, 18). Improved outcome has only been observed in pairs with HLA identical kidneys (18), which would be uncommon among parents and children. From an immunologic standpoint, there is little concern about the degree of HLA match, if ABO compatibility and a negative T-cell crossmatch are present. Its value is further reduced if only one living donor is available. Despite this, we acknowledge that eliminating HLA testing will not eliminate the discovery misattributed paternity, as reviewing blood types could also theoretically reveal similar information. However, with only six common alleles that produce ABO blood type and variants that are otherwise rare (19), ABO typing will reveal far fewer cases of misattributed paternity as demonstrated in our analyses. Furthermore, for sensitized recipients, the presence of donor-specific anti-HLA antibodies would be a factor in the evaluation of risk for acute rejection episodes and would help guide clinical decision making on immunosuppressive therapy (20, 21). In such cases, HLA antigens of both the recipient and the donor would be required. However, one possible approach would be to avoid typing the donor when the cell-based and flow-based crossmatches with a nonsensitized recipient are negative. This could be a sizable proportion of the living donor population.
Ultimately, this study approximates the prevalence of misattributed paternity in living kidney donation in North America. It affirms the need for transplant centers to consider their plan for information disclosure (or nondisclosure) and informed consent in the setting of incidentally discovered misattributed paternity information. We encourage transplant centers to increase their dialogue about this issue, and consider strategies adopted by other centers. Consensus on the handling of this sensitive information will help minimize any harm that could come to donors, recipients, and their families.
The investigators of Donor Nephrectomy Outcomes Research (DONOR) Network are as follows: Neil Boudville, Laurence Chan, Christine Dipchand, Mona Doshi, Liane Feldman, Amit Garg, Colin Geddes, Eric Gibney, John Gill, Martin Karpinski, Scott Klarenbach, Greg Knoll, Charmaine Lok, Mauricio Monroy-Cuadros, Norman Muirhead, Christopher Y. Nguan, Chirag Parikh, Emilio Poggio, G.V. Ramesh Prasad, Leroy Storsley, Sudha Tata, Darin Treleaven, Robert Yang, and Ann Young.
The authors thank Dr. Edward Ball, Mr. Steve Leckie, Ms. Diane Smith, and Ms. Jennifer Cross from the London Health Science Centre. They also thank the transplant centre representatives who provided information for this study.