Abstract
Background
Transmission of human immunodeficiency virus (HIV) and hepatitis C to transplant recipients has drawn attention to the use of allografts from seronegative donors at increased risk for viral infection (DIRVI).
Methods
We performed a cohort study of 7,803 kidney transplant recipients whose kidneys were recovered through one of two organ procurement organizations (OPO) from 1996 to 2007. Detailed OPO data on donor risk factors were linked to recipient data from the Organ Procurement and Transplantation Network.
Results
Median recipient follow-up was 3.9 years. 368 (5%) patients received DIRVI kidneys, a third of which were procured from donors with a history of injection drug use or commercial sex work. Compared to standard criteria kidney recipients, DIRVI kidney recipients were more likely to be HIV-positive or Black. In multivariable Cox regression, using DIRVI recipients as the reference, recipients of standard criteria donor kidneys had lower mortality (HR 0.71, p<0.01) and no difference in death-censored allograft failure (HR 1.09, p=0.62), whereas recipients of expanded criteria donor kidneys had no significant difference in mortality (HR 0.98, p=0.83), but a higher allograft failure rate (HR 1.93, p<0.01). High-quality data on post-transplant recipient viral testing were not available.
Conclusions
DIRVI kidney recipients experienced higher mortality than standard criteria kidney recipients. This finding could be explained if sicker patients received DIRVI kidneys (i.e., residual confounding) or the less likely possibility of undetected transmission of viral infections. Given the limitations of registry data used in this analysis, prospective studies are needed to further elucidate these findings.
Keywords: Kidney transplantation, Viral transmission, Human immunodeficiency virus, High Risk Donor
Introduction
The lengthy wait for a kidney transplant has driven interest in expanding the criteria for acceptable donor kidneys. Efforts to enlarge the pool of kidney allografts have included the increased use of expanded criteria donor (ECD) kidneys, which are recovered from older donors, as well as kidneys donated after cardiac death (DCD). Compared to so-called standard criteria donor (SCD) kidneys, ECD kidneys have shorter allograft survival, while DCD kidneys have high rates of delayed graft function.(1, 2) Also within the enlarged pool are kidneys from donors at increased risk of blood-borne viral infection (DIRVI)(3, 4)—donors with negative viral serologic tests, but behavioral and clinical risk factors suggesting greater likelihood of undetected infection. Some DIRVIs could be in a “window period” when serologies have not yet become positive, or their viral serologies might be falsely negative due to hemodilution. Limited data about recipient outcomes are available to guide the use of kidneys from these donors.(5–7)
A 1994 report from the Centers for Disease Control and Public Health Service (CDC/PHS) defined criteria for deceased donors at increased risk for infection with human immunodeficiency virus (HIV) (Table 1).(8) The report recommended that organs from donors meeting these criteria should be discarded unless the risk to the recipient of not undergoing transplantation is considered greater than the risk of transplantation and infection.(8) Since 1994, these criteria have also been used to identify donors at increased risk of hepatitis B (HBV) and C (HCV). The precise risk of viral infection with acceptance of a DIRVI organ is uncertain. However, the transmission of HIV and HCV to four patients undergoing solid organ transplantation from a DIRVI in 2007 drew public attention to the possible negative consequences of accepting organs from these donors.(9–11)
Table 1.
Criteria for a seronegative Centers for Disease Control/Public Health Service high behavioral risk adult organ donor
|
Because of concerns about viral transmission or the stigma associated with HIV, kidneys from DIRVIs may be not be recovered, or may be declined by centers. The short-term nature of outcome data for DIRVI kidney recipients and the lack of detailed information about donor risk factors have made it difficult to contextualize the potential hazards of accepting DIRVI organs within the framework of other transplant-related risks – such as accepting an ECD kidney or remaining on dialysis.(5, 12).
We sought to better characterize DIRVIs and their recipients and to compare longer-term outcomes for recipients of DIRVI, SCD, ECD, and DCD kidneys. We hypothesized that recipients of DIRVI kidneys would have similar rates of mortality compared to SCD recipients, and lower rates of allograft failure and mortality compared to ECD recipients.
Results
The cohort comprised 7,803 kidney transplant recipients from 1996 – 2007, among whom 5,484 (70%) received kidneys recovered by the Gift of Life Donor Program (GOL) in Philadelphia, Pennsylvania, and 2,319 (30%) received kidneys recovered by the New Jersey Sharing Network (NJSN) in Springfield, New Jersey. In the cohort, 5,228 individuals (67%) received kidneys from SCDs, 1,621 (21%) from ECDs, 586 (8%) from DCDs, and 368 (5%) from DIRVIs. DIRVI allografts were transplanted at 57 centers; the majority (56%) were used by six centers in Pennsylvania and New Jersey. Median follow-up was 1,430 days.
The mean age of DIRVI kidney recipients was 48.7 years versus 47.7 years for SCD recipients. A higher proportion of DIRVI kidney recipients were black and HIV-infected compared to SCD recipients. Median waiting time was slightly greater for DIRVI recipients (645 days) versus SCD recipients (578 days) (Table 2).
Table 2.
Clinical and demographic characteristics of kidney transplant recipients *
| DIRVI kidney recipient (n=368) |
SCD kidney recipient (n=5228) |
ECD kidney recipient (n=1621) |
DCD kidney recipient (n=586) |
p- value |
|
|---|---|---|---|---|---|
| Mean age in yrs (s.e.) | 48.7 (0.6) | 47.7 (0.2) | 57.2 (0.3) | 49.9 (0.5) | <0.01 |
| Male gender (%) | 223 (60.6) | 3151 (60.3) | 1003 (61.9) | 348 (59.4) | 0.64 |
| Race (%) | <0.01 | ||||
| White | 198 (53.8) | 3095 (59.2) | 838 (51.7) | 278 (47.4) | |
| Black | 125 (34.0) | 1524 (29.2) | 591 (36.5) | 234 (39.9) | |
| Hispanic | 33 (9.0) | 388 (7.4) | 92 (5.7) | 40 (6.8) | |
| Asian | 6 (1.6) | 168 (3.2) | 77 (4.8) | 26 (4.4) | |
| Other | 6 (1.6) | 53 (1.0) | 23 (1.4) | 8 (1.3) | |
| Dialysis prior to transplant (%) | 323 (87.8) | 4594 (87.9) | 1494 (92.2) | 533 (91.0) | <0.01 |
| Hepatitis C seropositive (%) | 27 (7.7) | 287 (5.8) | 85 (5.6) | 32 (5.9) | 0.52 |
| CMV seropositive | 167 (45.4) | 1991 (38.1) | 724 (44.7) | 273 (46.6) | <0.01 |
| Education | 0.50 | ||||
| High school or less | 170 (46.2) | 2162 (41.4) | 698 (43.1) | 259 (44.2) | |
| Some college | 53 (14.4) | 781 (14.9) | 232 (14.3) | 76 (13.0) | |
| Bachelor’s degree or higher | 54 (14.7) | 811 (15.5) | 238 (14.7) | 79 (13.5) | |
| Unknown | 91 (24.7) | 1474 (28.2) | 453 (27.9) | 171 (29.3) | |
| Primary payor | <0.01 | ||||
| Private insurance | 117 (31.8) | 1760 (33.7) | 480 (29.6) | 207 (35.3) | |
| Medicare | 222 (60.3) | 3129 (59.9) | 1058 (65.3) | 340 (58.0) | |
| Medicaid | 17 (4.6) | 241 (4.6) | 55 (3.4) | 31 (5.3) | |
| Other | 12 (3.3) | 98 (1.9) | 28 (1.7) | 8 (1.4) | |
| HIV seropositive (%) | 8 (2.4) | 9 (0.2) | 4 (0.3) | 5 (1.0) | <0.01 |
| Hepatitis B seropositive | 6 (1.6) | 74 (1.4) | 22 (1.4) | 16 (2.7) | 0.09 |
| Median days on the wait-list | 645 | 578 | 640 | 751 | <0.01 |
| Prior kidney transplant | 48 (13.0) | 707 (13.5) | 148 (9.1) | 77 (13.1) | <0.01 |
| Multi-organ transplant | 64 (17.4) | 791 (15.1) | 18 (1.1) | 10 (1.7) | <0.01 |
| Pancreas | 51 | 621 | 6 | 8 | |
| Liver | 10 | 117 | 9 | 2 | |
| Heart | 3 | 52 | 3 | 0 | |
| Intestine | 0 | 1 | 0 | 0 | |
| Lung | 0 | 0 | 0 | 0 | |
| Diabetes | 135 (36.7) | 1742 (33.3) | 579 (35.7) | 168 (28.7) | 0.01 |
| Mean percent peak panel reactive antibody | 14.3 (1.5) | 15.1 (0.4) | 9.6 (0.6) | 10.4 (1.0) | <0.01 |
| High peak panel reactive antibody (%)** | 30 (8.9) | 462 (9.6) | 80 (5.1) | 39 (6.8) | <0.01 |
| Blood group (%) | 0.77 | ||||
| A | 130 (35.3) | 1967 (37.6) | 585 (36.1) | 215 (36.7) | |
| B | 52 (14.1) | 681 (13.0) | 217 (13.4) | 70 (12.0) | |
| AB | 24 (6.5) | 314 (6.0) | 83 (5.1) | 35 (6.0) | |
| O | 162 (44.0) | 2266 (43.3) | 736 (45.4) | 266 (45.4) |
DIRVI: donor at increased risk for viral infection according to the Centers for Disease Control categorization; SCD: Standard criteria donor; ECD: Expanded criteria donor; DCD: donation after cardiac death; HIV: human immunodeficiency virus
Recipients of live donor kidneys, recipients of kidneys from deceased donors seropositive for hepatitis C or hepatitis B, and recipients of kidneys from deceased donors that met criteria as ECD-DCD, ECD-DIRVI, DCD-DIRVI, or ECD-DCD-DIRVI were excluded. P-values represent comparisons across all groups. Percentages were calculated among individuals without missing data for that variable.
Defined as peak panel reactive antibody >80%
In multivariable logistic regression for the outcome of receiving a DIRVI versus a SCD kidney, recipient HIV infection (OR 7.72, OR 3.95 – 15.1, p<0.01) and transplantation in the most recent era (OR 2.25, CI 1.64 – 3.10, p<0.01 for the years 2004 – 2007, and OR 2.11, CI 1.65 – 2.70, p<0.01 for the years 2000 – 2003, compared to the reference years 1996 – 1999) were strongly associated with DIRVI kidney use. Among the 26 HIV-infected recipients, DIRVI kidney use (31% of recipients) was nearly as common as SCD kidney use (35% of recipients).
The mean age of DIRVIs (32.9 years) was similar to that of SCDs (33.9 years) and DCD donors (33.7 years), but much lower than the mean age of ECDs (62.1 years). Compared to SCD kidneys, DIRVI kidneys were more likely to be from black (p<0.01) and male (p<0.01) donors (Table 3).
Table 3.
Characteristics of HIV seronegative donor allografts classified as increased risk for viral infection, standard criteria, expanded criteria, and donation after cardiac death
| DIRVI (n=368) |
SCD (n=5228) |
ECD (n=1621) |
DCD (n=586) |
p- value |
|
|---|---|---|---|---|---|
| Donor age (s.e.) | 32.9 (0.6) | 33.9 (0.2) | 62.1 (0.2) | 33.7 (0.6) | <0.01 |
| Donor male gender (%) | 297 (80.7) | 3141 (60.1) | 788 (48.6) | 396 (67.6) | <0.01 |
| Donor race (%) | <0.01 | ||||
| White | 224 (60.9) | 3768 (72.1) | 1271 (78.4) | 451 (77.0) | |
| Black | 92 (25.0) | 750 (14.4) | 185 (11.4) | 92 (15.7) | |
| Hispanic | 48 (13.0) | 556 (10.6) | 128 (7.9) | 32 (5.5) | |
| Asian | 2 (0.5) | 100 (1.9) | 30 (1.9) | 9 (1.5) | |
| Other | 2 (0.5) | 54 (1.0) | 7 (0.4) | 2 (0.3) | |
| Median cold ischemia in hours | 16 | 17 | 18 | 17 | <0.01 |
| Zero antigen mismatch (%) | 32 (8.7) | 636 (12.2) | 92 (5.7) | 43 (7.3) | <0.01 |
| Share type non-local (%) | 45 (12.2%) | 970 (18.6) | 128 (7.9) | 43 (7.3) | <0.01 |
| CMV donor | 226 (61.4) | 3134 (60.0) | 1093 (67.4) | 312 (53.2) | <0.01 |
SCD: Standard criteria donor; DIRVI: designation of donor as increased risk for viral infection according to the Centers for Disease Control categorization; ECD: Expanded criteria donor; DCD: donation after cardiac death
We also examined patterns of kidney usage. DIRVI kidneys were less likely to be shared outside the local geographic area (12% for DIRVI versus 19% for SCDs, p<0.01). In multivariable logistic regression, DIRVIs were no less likely to have 1 versus 2 kidneys used for transplantation (OR 0.90, CI 0.51 – 1.60, p=0.73).
Kidneys were most likely to have been characterized as from a DIRVI due to donor history of incarceration (59%). The next most common reasons were injection drug use (26%), exchanging sex for drugs or money (10%), having a partner with a DIRVI risk factor (10%), and being a male with a history of sex with another male (5%) (Table 4).
Table 4.
Among DIRVI kidney recipients: reasons why donors were categorized as elevated risk for blood-borne viral infection *
| DIRVI criterion | Total (n=368) |
Number from GOL (n=257) |
Number from NJSN (n=111) |
|---|---|---|---|
| Inmates of correctional systems (%) ** | 218 (59.2) | 143 (55.6) | 75 (67.6) |
| Injection drug use (%) ** | 95 (25.8) | 79 (30.7) | 16 (14.4) |
| Commercial sex work (i.e. history of exchanging sex for money or drugs) (%) | 37 (10.1) | 22 (8.6) | 15 (13.5) |
| Donor partner had DIRVI risk factor (%) | 35 (9.5) | 23 (9.0) | 12 (10.8) |
| Man who had sex with another man (%) ** | 19 (5.2) | 19 (7.4) | 0 (0.0) |
| Persons with hemophilia or related clotting disorders (%) | 8 (2.2) | 6 (2.3) | 2 (1.8) |
| Exposure to HIV infected blood (%) ** | 8 (2.2) | 0 (0.0) | 8 (7.2) |
DIRVI: designation of donor as increased risk for viral infection according to the Centers for Disease Control/Public Health Service categorization; HIV: human immunodeficiency virus
Columns do not sum to total because some donors had met multiple criteria for increased risk for viral infection
Indicated difference between proportions across the two organ procurement organizations (p<0.05).
Recipient infections
In this cohort, no donor-to-recipient transmissions of HIV, HCV, or HBV were reported to either organ procurement organization (OPO).
Multivariable analyses of mortality and allograft failure (Table 5)
Table 5.
Multivariable Cox analyses of patient mortality and death-censored allograft failure
| Mortality | Allograft failure | ||||||
|---|---|---|---|---|---|---|---|
| HR | CI | p-value | HR | CI | p-value | ||
| Recipient of: | |||||||
| DIRVI kidney | Reference | Reference | |||||
| SCD kidney | 0.71 | 0.58–0.87 | <0.01 | 1.09 | 0.78–1.51 | 0.62 | |
| ECD kidney | 0.98 | 0.83–1.17 | 0.83 | 1.93 | 1.33–2.81 | <0.01 | |
| DCD kidney | 0.67 | 0.54–0.83 | <0.01 | 1.19 | 0.82–1.73 | 0.36 | |
| Recipient age (years) | |||||||
| Age<40 | Reference | Reference | |||||
| 40≤age<50 | 1.34 | 1.16–1.55 | <0.01 | 0.83 | 0.70–0.98 | 0.03 | |
| 50≤age<60 | 2.44 | 2.12–2.81 | <0.01 | 0.76 | 0.64–0.89 | <0.01 | |
| 60≤age<70 | 3.23 | 2.65–3.92 | <0.01 | 0.79 | 0.64–0.98 | 0.03 | |
| Age≥70 | 5.69 | 4.41–7.35 | <0.01 | 0.60 | 0.46–0.78 | <0.01 | |
| Recipient black race | 0.94 | 0.84–1.04 | 0.21 | 1.56 | 1.36–1.78 | <0.01 | |
| Recipient male | 1.13 | 1.05–1.23 | <0.01 | 1.06 | 0.96–1.16 | 0.24 | |
| Recipient prior kidney transplant | 1.20 | 0.99–1.45 | 0.06 | 1.29 | 1.12–1.47 | <0.01 | |
| Recipient on dialysis | 1.30 | 1.09–1.56 | <0.01 | 1.16 | 0.96–1.42 | 0.13 | |
| Recipient diabetes | 1.74 | 1.54–1.96 | <0.01 | 0.98 | 0.88–1.08 | 0.67 | |
| Recipient hepatitis C sero-positive | 1.63 | 1.37–1.94 | <0.01 | 1.40 | 1.22–1.61 | <0.01 | |
| Recipient HIV sero-positive | 1.24 | 0.69–2.21 | 0.46 | 1.39 | 0.95–2.03 | 0.09 | |
| Recipient high PRA* | 1.53 | 1.25–1.88 | <0.01 | 1.30 | 1.08–1.55 | <0.01 | |
| CMV positive recipient | 1.01 | 0.91–1.11 | 0.87 | 0.95 | 0.87–1.03 | 0.20 | |
| CMV positive donor | 1.09 | 1.00–1.20 | 0.05 | 1.15 | 1.07–1.24 | <0.01 | |
| Wait-list time>1000 days | 1.13 | 0.97–1.32 | 0.11 | 0.99 | 0.89–1.11 | 0.89 | |
| Dual organ transplant | 1.16 | 0.89–1.50 | 0.28 | 0.72 | 0.57–0.91 | <0.01 | |
| Donor black race | 1.12 | 0.96–1.29 | 0.15 | 1.10 | 0.95–1.29 | 0.21 | |
| Donor male | 0.92 | 0.84–1.01 | 0.07 | 0.95 | 0.84–1.06 | 0.36 | |
| Allograft cold ischemia (hours) | |||||||
| <12 | Reference | Reference | |||||
| 12≤time<24 | 1.04 | 0.91–1.18 | 0.59 | 1.09 | 0.94–1.26 | 0.28 | |
| ≥24 | 0.96 | 0.81–1.14 | 0.63 | 1.12 | 0.91–1.37 | 0.30 | |
| Non-local transplant | 0.93 | 0.76–1.14 | 0.50 | 1.00 | 0.81–1.23 | 0.99 | |
| Zero mismatch | 0.89 | 0.73–1.07 | 0.22 | 0.71 | 0.57–0.89 | <0.01 | |
| Steroid induction | 1.08 | 0.96–1.20 | 0.19 | 1.12 | 0.96–1.30 | 0.17 | |
| Antibody induction | Reference | Reference | |||||
| Lymphocyte depleting | 1.08 | 0.96–1.22 | 0.22 | 1.09 | 0.95–1.23 | 0.21 | |
| Non- Lymphocyte depleting | 1.14 | 0.97–1.33 | 0.11 | 1.08 | 0.92–1.27 | 0.37 | |
| Initial immuno-suppression | |||||||
| Tacrolimus based | Reference | Reference | |||||
| Cyclosporine based | 1.19 | 0.98–1.44 | 0.09 | 1.12 | 0.98–1.49 | 0.08 | |
| Rapamycin based | 1.27 | 1.05–1.52 | 0.01 | 1.54 | 1.22–1.93 | <0.01 | |
| Other | 1.92 | 1.59–2.32 | <0.01 | 2.81 | 2.39–3.31 | <0.01 | |
| Transplant era | |||||||
| 1996 – 1998 | Reference | Reference | |||||
| 1999 – 2002 | 0.94 | 0.79–1.12 | 0.48 | 1.09 | 0.98–1.23 | 0.12 | |
| 2003 – 2006 | 0.87 | 0.73–1.05 | 0.15 | 1.05 | 0.83–1.34 | 0.68 | |
Defined as ≥80% peak panel reactive antigen
DIRVI: donor at increased risk for viral infection according to the Centers for Disease Control categorization; SCD: Standard criteria donor; ECD: Expanded criteria donor; DCD: donation after cardiac death; HIV: human immunodeficiency virus
Among DIRVI kidney recipients, 86 (23%) individuals died, compared with 1,032 (20%) SCD kidney recipients, 500 (31%) ECD kidney recipients, and 93 (16%) DCD recipients. In Cox regression, compared to DIRVI recipients, SCD (HR 0.71, CI 0.58 – 0.87, p<0.01) and DCD kidney recipients (HR 0.67, CI 0.54 – 0.83, p<0.01) had a lower mortality rate, while ECD recipients (HR 0.98, CI 0.83 – 1.17, p=0.83) had no significant difference in mortality.
Sixty-one (17%) DIRVI kidney recipients experienced death-censored allograft failure, compared with 1,103 (21%) SCD kidney recipients, 504 (31%) ECD kidney recipients, and 121 (21%) DCD kidney recipients. In Cox regression, compared to DIRVI recipients, SCD (HR 1.09, CI 0.78 – 1.51, p=0.62) and DCD recipients (HR 1.19, CI 0.82 – 1.73, p=0.36) had no statistically significant difference in allograft failure, while ECD recipients had a higher rate of allograft failure (HR 1.93, CI 1.33 – 2.81, p<0.01).
Additional analyses
Sensitivity analyses revealed results consistent with the primary analyses.
Secondary analyses of two subgroups of DIRVI kidneys showed that, compared to SCD kidney recipients, recipients of kidneys from DIRVIs with only an incarceration history had no difference in mortality (HR 1.25, CI 0.92 – 1.71, p=0.15), while recipients of kidneys from DIRVIs with other risk factors had elevated mortality (HR 1.46, CI 1.10 – 1.94, p<0.01).
We also examined Organ Procurement and Transplantation Network (OPTN) cause of death data. Compared to SCD recipients, DIRVI recipients were more likely to have died of infectious (OR 1.99, CI 1.23 – 3.21, p<0.01) or cardiovascular causes (OR 1.60, CI 1.04 – 2.48, p=0.03), but not of malignancy (OR 1.64, CI 0.78 – 3.46, p=0.19).
Discussion
The lengthening waiting list for kidney transplants has motivated efforts to maximize the use of good quality donor organs. We found that DIRVI kidney recipients had rates of patient survival lower than among SCD and DCD recipients, but similar to ECD recipients. DIRVI kidney recipients had allograft survival that was similar to SCD and DCD recipients and better than for ECD recipients. The higher mortality among DIRVI kidney recipients suggests that DIRVI kidney allografts are received by transplant candidates who are sicker than SCD recipients. A less likely possibility is that these findings are attributable to undetected, donor-derived transmission of viral infections.
The DIRVIs in this study were a heterogeneous group. Most had an incarceration history, whereas a third had used injection drugs and/or engaged in commercial sex work—behaviors associated with the highest risk for HIV infection. The risk of viral transmission from a DIRVI depends on the risk factor, the testing modality, and the accuracy of the donor’s history. In a decision model that employed data from non-organ donor populations, Schweitzer et al. calculated that, if nucleic acid testing (NAT) and antibody testing were used, the chance of HIV and/or HCV infection would be approximately 1 in 107 if a donor was a commercial sex worker or as low as 1 in 10,000 if the donor was an inmate. The risks would be higher if only antibody testing was used, as was the case during most years of our study.(3) These estimates, and data from the OPTN/UNOS Disease Transmission Advisory Committee (DTAC), indicate that viral transmission should be a rare event.(7)
Recipients of DIRVI kidneys had excellent allograft survival, but contrary to our hypothesis, they had higher mortality compared to SCD kidney recipients. The proportion who died over 4 years was 3.7% higher among DIRVI recipients compared to SCD recipients, a small but clinically important difference. This finding would be explained if sicker patients received these organs; unfortunately, their co-morbidities were not fully captured by our data. For instance, during the study period, OPTN data on cardiovascular disease, vascular access, and functional status lacked granularity and were not included in our multivariable models. Additionally, our analyses of cause of death revealed that, compared to SCD kidney recipients, recipients of DIRVI kidneys were more likely to die of infectious as well as cardiovascular causes, which is consistent with the hypothesis that DIRVI kidney recipients were a sicker group.
HIV-positive patients were much more likely to receive DIRVI organs, suggesting that they were more willing to accept DIRVI organs or that transplant staff directed kidneys to this group due to concerns about their outcomes while receiving chronic dialysis. The acceptance of organs from DIRVIs requires that patients overcome the stigma associated with the possibility of an HIV-infected donor, and this stigma may be less important to candidates who already have HIV. Another strong possibility is that HIV-positive patients go to centers where the transplant clinicians are comfortable with more clinically complex recipients and are, therefore, willing to accept kidneys from higher risk donors such as DIRVIs.
The explanation of greater rates of viral transmission among DIRVI recipients must also be considered. The risk of donor transmission of infections has received far greater scrutiny over the past several years, whereas our cohort spans an earlier period from 1996 – 2007. Recipient infections with hepatitis or HIV might not have been diagnosed. For instance, opportunistic infections such as Pneumocystis pneumonia could have been attributed to immunosuppression rather than HIV, and antibody serologies to diagnose viral infection (the standard of care during much of the study) can produce false negative results in immunosuppressed patients or those with renal disease. We acknowledge the limitation that we were not able to report post-transplant viral testing outcomes for recipients at over 200 centers. Nonetheless, given the low rate of viral transmission expected from DIRVIs and the lack of reports to the OPOs of viral transmission, donor-derived infections are unlikely to explain our results.(3, 7)
These findings of higher mortality among recipients of kidneys from DIRVIs versus SCDs contrast with two earlier studies (one from our group).(5, 6) These previous studies reported outcomes from a national cohort with shorter follow-up, whereas the current study reports outcomes related to kidneys from two OPOs and the majority of transplants took place at six transplant centers in northeastern US, where waiting times are long. It is plausible that in this region (compared to others), sicker patients are more willing to accept DIRVI kidneys, and that a slightly higher rate of mortality only became evident after four years.
For transplant candidates considering the acceptance of a DIRVI kidney, the risks of viral transmission must be weighed against the substantial risks and diminished quality of life with end-stage renal disease.(13) Accepting a DIRVI kidney may provide a way to decrease waiting time. Since 2007, many OPOs have also adopted NAT, which narrows the window period of seronegative infections and could reassure patients. On the other hand, even with NAT, viral transmission will remain a small risk regardless of donor type, and informed consent of recipients remains an essential ethical precondition to transplantation.(14) Further study is needed about the most effective methods for informing transplant candidates about DIRVI organs, particularly since HIV and hepatitis are also associated with social stigma.(4, 15)
Conclusion
In a cohort of kidney transplant recipients over a decade, recipients of kidneys from DIRVIs experienced higher rates of mortality and similar allograft survival compared to SCD kidney recipients, but similar mortality and better allograft survival compared to ECD kidney recipients. With the registry data used in this study, we were unable to determine conclusively if our findings were due to the allocation of DIRVI kidneys to sicker patients versus the consequences of viral transmissions to recipients. Prospective studies are needed to resolve the mechanisms driving these findings.
Methods
Data on donor and recipient characteristics and recipient outcomes came from the OPTN. However, because the OPTN has only recorded DIRVI status since 2004, and has only coded this variable as present or absent, we also obtained information about specific viral infection risk factors and donor testing for HIV, HCV, and HBV from GOL and the NJSN – two geographically contiguous OPOs. We linked these data and performed a non-concurrent cohort study of adult (age≥18 years) kidney transplant recipients from 1996 – 2007 whose allografts were recovered through either OPO. The University of Pennsylvania Institutional Review Board approved the study.
The primary outcome was mortality and the primary exposure was donor type. We defined a DIRVI as any deceased donor with negative test results for HIV, HCV, and HBV, and who was classified as having a risk factor for infection due to social/medical history (Table 1).(8) Although the conventional definition of a DIRVI encompasses donors infected with HCV or HBV, we excluded recipients of kidneys from these donors (n=375) because of the belief that the decision-making by centers and patients about acceptance of organs with known infections is different from acceptance of organs from seronegative donors with only risk factors for infection.
NJSN and GOL employed slightly different criteria to categorize donors as “inmates of correctional systems” (a DIRVI category). The NJSN categorized individuals as DIRVIs if they had ever been incarcerated for >72 hours. GOL categorized donors as increased risk if they had been incarcerated for >72 hours in the last 12 months.
All donors underwent ELISA testing for HIV and measurement of HCV antibody and HBV surface antigen. The NJSN additionally measured HBV core IgM, and since 2004, performed nucleic acid testing for HIV, HCV, and HBV for all donors using Roche Cobas Ampliscreen.
As per OPTN definitions, we defined an ECD as a deceased donor ≥60 years old, or ≥50 years with two of the following: hypertension, terminal serum creatinine>1.5mg/dL, or death by cerebrovascular accident. A DCD was a donor for whom cardiac death was diagnosed prior to kidney recovery.(16) For simplicity, the 180 recipients of kidneys categorized as any combination of ECD, DCD, and DIRVI were excluded.
Recipient viral infections
Surveillance to detect viral transmissions to recipients was not mandated during the study period. However, several recommendations were in place to encourage the detection of such events. First, the CDC/PHS recommended that solid organ transplant recipients undergo HIV testing at 3, 6, and 12 months after transplantation. Second, the CDC/PHS recommended that if a transplant recipient was diagnosed with HIV, “the transplant center or health-care provider should, consistent with state law, immediately notify the state health department and the organization from which the tissue was obtained.”(8)
Notably, during this period, the OPTN did not obtain comprehensive data on infections after transplantation. Specifically, only 33.2% of the recipients in the study cohort ever had any HIV serologic testing results reported to the OPTN after transplantation. Even among DIRVI kidney recipients, only 35.8% had any HIV serologic data reported.
Statistical analysis
We conducted analyses using Stata (Stata 11.0, Stata Corporation, College Station, Texas). We used ANOVA to compare the means of normally distributed continuous variables across donor types, and the Kruskal-Wallis test to compare the medians of non-normally distributed variables. For categorical variables, we used the chi-square test or Fisher’s exact as appropriate.
We fit multivariable Cox regression models for the outcomes of mortality and allograft failure, adjusted for recipient center. We inspected log-log plots and other graphical displays to confirm model selection. Based on prior studies and clinical judgment, we identified independent variables for these models.(2, 17, 18) Recipient variables included age (defined as <30, ≥30 and <50, ≥50 and <70, and ≥70 years), gender, race (black or non-black), diabetes, history of dialysis, prior kidney transplant, multi-organ transplant, waiting time (<1000 versus ≥1000 days), and high peak panel reactive antibody (PRA, defined by OPTN convention as <80% versus ≥80%), HCV seropositivity, HIV seropositivity, and HBV seropositivity. HCV and HIV are reported as positive, negative or unknown “serostatus” to OPTN. We categorized individuals with hepatitis B surface antigen positive as HBV seropositive. Induction antibody therapy was categorized as lymphocyte-depleting or non-depleting. Initial immunosuppression was a categorical variable (FK-506 with neither cyclosporine nor rapamycin, cyclosporine with neither FK-506 nor rapamycin, Rapamycin used alone or with other agents, and other regimens).
Allograft variables included cold ischemia (<12, ≥12 and <24, ≥24 hours), antigen mismatch (zero mismatch or not), transplant era (the years 1996–1998, 1999–2002, 2003–2007), and share type. Share type, the geographical relationship between the donor service areas where the organ was recovered and accepted, was categorized as local/non-local. Donor characteristics that define ECD status (e.g., age) were not included in multivariable analyses.
A multivariable logistic model was fit for the outcome of receiving a DIRVI kidney; recipient independent variables used in the mortality model were used for this model. Lastly, to assess the willingness of centers to accept DIRVI kidneys, we fit a multivariable logistic model for the outcome of having both donor kidneys used for transplantation using donor independent variables included in the mortality model.
Missing data
A minority of recipients had missing data on variables of interest, including HIV serostatus (n=1,045, 13%), cold ischemia (n=841, 11%), hepatitis C serostatus (n=488, 6%), and peak PRA (n=478, 6%). In our primary multivariable analyses, we employed multiple imputation.(19) Additionally, we performed sensitivity analyses for mortality in which extreme values were assigned to individuals with missing data.
Acknowledgements
The authors wish to thank Katarina Linden at the United Network for Organ Sharing and Sharon West at Gift of Life for their help and advice with the study. Dr. Reese had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Preliminary results were reviewed at an oral presentation during the 2009 American Transplant Congress.
Abbreviations
- CDC/PHS
Centers for Disease Control and Public Health Service
- DCD
Donation after cardiac death
- DIRVI
Seronegative donors at increased risk of viral infection
- DATC
Disease Transmission Advisory Committee
- ECD
Expanded criteria donor
- GOL
Gift of Life Donor Program
- HBV
Hepatitis B virus
- HCV
Hepatitis C virus
- HIV
Human immunodeficiency virus
- NAT
Nucleic acid testing
- NJSN
New Jersey Sharing Network
- OPO
Organ Procurement Organization
- OPTN
Organ Procurement and Transplantation Network
- PRA
Panel reactive antibody
- SCD
Standard criteria donor
- UNOS
United Network for Organ Sharing
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Funding sources and participation:
Dr. Reese is supported by NIH Career Development Award, K23 - DK078688-01. He participated in research design, data analysis and writing of the paper
Dr. Halpern is supported by a Greenwall Foundation Faculty Scholar Award in Bioethics. He participated in research design, data analysis, data interpretation and writing of the paper.
Dr. Asch participated in research design, data interpretation and writing of the paper
Dr. Bloom participated in research design, data interpretation and writing of the paper
Mr. Nathan participated in data interpretation and writing of the paper
Mr. Hasz participated in data interpretation and writing of the paper
Mr. Roth participated in research design and writing of the paper
Mr. Reitsma participated in research design, data collection, data analysis and writing of the paper
Mr. Krefski participated in data collection, data analysis and data interpretation
Mr. Goerlitz participated in data collection, data analysis and data interpretation
Ms. DeLauro participated in data collection and data analysis
Dr. Blumberg participated in research design and writing of the paper
Dr. Weng participated in data analysis, data interpretation and writing of the paper
Dr. Caplan participated in research design and writing of the paper
Ms. Thomasson participated in biostatistical analysis and data interpretation
Dr. Shults participated in biostatistical analysis and data interpretation
Dr. Feldman is supported by NIH grant K24 - DK002651. He participated in research design and writing of the paper
OPTN Disclaimer: "This work was supported in part by Health Resources and Services Administration contract 234-2005-370011C. The content is the responsibility of the authors alone and does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government."
Conflicts of interest: The authors have no financial conflicts of interest to disclose.
References
- 1.Merion RM, Ashby VB, Wolfe RA, et al. Deceased-donor characteristics and the survival benefit of kidney transplantation. JAMA. 2005;294(21):2726. doi: 10.1001/jama.294.21.2726. [DOI] [PubMed] [Google Scholar]
- 2.Locke JE, Segev DL, Warren DS, Dominici F, Simpkins CE, Montgomery RA. Outcomes of kidneys from donors after cardiac death: implications for allocation and preservation. Am J Transplant. 2007;7(7):1797. doi: 10.1111/j.1600-6143.2007.01852.x. [DOI] [PubMed] [Google Scholar]
- 3.Schweitzer EJ, Perencevich EN, Philosophe B, Bartlett ST. Estimated benefits of transplantation of kidneys from donors at increased risk for HIV or hepatitis C infection. Am J Transplant. 2007;7(6):1515. doi: 10.1111/j.1600-6143.2007.01769.x. [DOI] [PubMed] [Google Scholar]
- 4.Halpern SD, Shaked A, Hasz RD, Caplan AL. Informing candidates for solid-organ transplantation about donor risk factors. N Engl J Med. 2008;358(26):2832. doi: 10.1056/NEJMsb0800674. [DOI] [PubMed] [Google Scholar]
- 5.Reese PP, Feldman HI, Asch DA, et al. Transplantation of kidneys from donors at increased risk for blood-borne viral infection: recipient outcomes and patterns of organ use. Am J Transplant. 2009;9(10):2338. doi: 10.1111/j.1600-6143.2009.02782.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Duan KI, Englesbe MJ, Volk ML. Centers for Disease Control 'high-risk' donors and kidney utilization. Am J Transplant. 2010;10(2):416. doi: 10.1111/j.1600-6143.2009.02931.x. [DOI] [PubMed] [Google Scholar]
- 7.Ison MG, Hager J, Blumberg E, et al. Donor-derived disease transmission events in the United States: data reviewed by the OPTN/UNOS Disease Transmission Advisory Committee. Am J Transplant. 2009;9(8):1929. doi: 10.1111/j.1600-6143.2009.02700.x. [DOI] [PubMed] [Google Scholar]
- 8.Guidelines for preventing transmission of human immunodeficiency virus through transplantation of human tissue and organs. Centers for Disease Control and Prevention. MMWR Recomm Rep. 1994;43(RR-8):1. [PubMed] [Google Scholar]
- 9.Fishman JA. Informing candidates for transplantation about donor risk factors. N Engl J Med. 2008;359(11):1182. doi: 10.1056/NEJMc081540. author reply 1182. [DOI] [PubMed] [Google Scholar]
- 10.Ahn J, Cohen SM. Transmission of human immunodeficiency virus and hepatitis C virus through liver transplantation. Liver Transpl. 2008;14(11):1603. doi: 10.1002/lt.21534. [DOI] [PubMed] [Google Scholar]
- 11.Ison MG, Friedewald JJ. Transmission of human immunodeficiency virus and hepatitis C virus through liver transplantation. Liver Transpl. 2009;15(5):561. doi: 10.1002/lt.21756. author reply 562. [DOI] [PubMed] [Google Scholar]
- 12.Kucirka LM, Alexander C, Namuyinga R, Hanrahan C, Montgomery RA, Segev DL. Viral nucleic acid testing(NAT) and OPO-level disposition of high-risk donor organs. Am J Transplant. 2009;9(3):620. doi: 10.1111/j.1600-6143.2008.02522.x. [DOI] [PubMed] [Google Scholar]
- 13.Freeman RB, Cohen JT. Transplantation risks and the real world: what does 'high risk' really mean? Am J Transplant. 2009;9(1):23. doi: 10.1111/j.1600-6143.2008.02476.x. [DOI] [PubMed] [Google Scholar]
- 14.Humar A, Morris M, Blumberg E, et al. Nucleic acid testing(NAT) of organ donors: is the 'best' test the right test? A consensus conference report. Am J Transplant. 2010;10(4):889. doi: 10.1111/j.1600-6143.2009.02992.x. [DOI] [PubMed] [Google Scholar]
- 15.Reese PP, Tehrani T, Lim MA, et al. Determinants of the Decision to Accept a Kidney from a Donor at Increased Risk for Blood-Borne Viral Infection. Clin J Am Soc Nephrol. doi: 10.2215/CJN.08251109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Childress J, Liverman C. Organ Donation: Opportunities for Action. Washington: National Academies Press; 2006. [Google Scholar]
- 17.Bunnapradist S, Cho YW, Cecka JM, Wilkinson A, Danovitch GM. Kidney allograft and patient survival in type I diabetic recipients of cadaveric kidney alone versus simultaneous pancreas kidney transplants: a multivariate analysis of the UNOS database. J Am Soc Nephrol. 2003;14(1):208. doi: 10.1097/01.asn.0000037678.54984.41. [DOI] [PubMed] [Google Scholar]
- 18.Ojo AO, Hanson JA, Meier-Kriesche H, et al. Survival in recipients of marginal cadaveric donor kidneys compared with other recipients and wait-listed transplant candidates. J Am Soc Nephrol. 2001;12(3):589. doi: 10.1681/ASN.V123589. [DOI] [PubMed] [Google Scholar]
- 19.Royston P. Multiple imputation of missing values. Stata Journal. 2004;4(3):227. [Google Scholar]