The Medical Evaluation of Living Kidney Donors: A Survey of US Transplant Centers

Abstract

The use of living donors for kidney transplantation in the United States is common, and long-term studies have demonstrated the safety of donation by young, healthy individuals. However, transplant programs have little data to guide them in deciding which donors are unacceptable, and which characteristics are associated with kidney disease or poor psychosocial outcomes after donation. To document current practices in evaluating potential donors, we surveyed all US kidney transplant programs. Compared to a survey 12 years ago, medical criteria for donation are more inclusive in several areas. All responding programs now accept living unrelated donors. Most programs no longer have an upper age limit to be eligible. Programs are now more likely to accept donors with treated hypertension, or a history of kidney stones, provided that certain additional criteria are met. In contrast, medical criteria for donation are more restrictive in other areas, such as younger donor age and low creatinine clearance. Overall, significant variability remains among transplant programs in the criteria used to evaluate donors. These findings highlight the need for more data on long-term outcomes in various types of donors with potential morbidities related to donation.

Introduction

Living donors play a critical role in renal transplantation. In the United States, the number of living donors surpassed deceased donors in 2001, and the use of living donors continues to rise (1). Minimal long-term risk has been demonstrated in studies comparing young, healthy individuals after nephrectomy with nondonating siblings (2), nonnephrectomized military recruits (3) or age-matched general populations (4,5). However, most of these studies are limited to the healthiest donors, and provide little information about different subsets of potential donors, such as those who are older, obese, hypertensive or have other characteristics potentially associated with renal disease. Decisions about which kidney donors to accept can affect the health outcomes of both donors and recipients.

In 1995, Bia et al. (6) published a survey of US transplant center practices in evaluating living renal donors. In part because of the continued lack of good data on donor outcomes, this survey about practices in 1992 was widely cited, and helped provide some guidelines about exclusion criteria. Since the publication of this survey, practices at US transplant centers likely have evolved. Reviews of how to evaluate kidney donors have been published (7,8), but actual practices have not been documented since Bia et al.'s 1995 study. To document current practices, and to help guide programs in developing protocols, we conducted an updated and expanded survey of US transplant programs.

Methods

We designed a 56-question web-based survey to gather information about living kidney donor evaluation and selection processes, including living donor types, alternative living donor programs, evaluation procedures, medical and psychosocial contraindications, informed consent policies and decision-making processes. We replicated the Bia et al.'s survey as much as possible, but expanded the questions in several areas. We omitted or modified some items that are addressed in other studies or are no longer relevant in the current era of living kidney donation.

A list of medical and surgical directors for US kidney transplant centers was purchased from the United Network for Organ Sharing (UNOS). Because UNOS did not provide email addresses, we searched the internet and called centers to obtain the current email addresses of each medical and surgical director. We then sent emails describing the study purpose along with a secured hyperlink (http://www.surveymonkey.com) to complete the online survey. We asked respondents to provide their UNOS program code so that we could track whether multiple responses were received from a particular program. If medical and surgical directors from the same program responded independently to the survey, only the medical director's responses were included in the statistical analyses. If a program director (medical or surgical) and another transplant team member completed the survey, the program director's responses were selected.

Survey responses were coded and downloaded into Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL) for analysis. Data are expressed as means and standard deviations, or the percentage of centers with specific responses. Survey responses were examined by program patient volume (median split: larger vs. smaller programs), program type (adult only vs. pediatric only or both) and UNOS region. Analyses included t-tests for continuous variables, the Fisher exact test for variables with two categories or a 2-tailed χ²-test for variables with three or more categories.

Results

Survey responses

Of the 248 programs listed with UNOS, we were unable to obtain valid email addresses for 43 (17%) transplant centers. Consequently, directors at 205 centers were sent an email inviting survey participation and 132 (64%) responded. Seven programs submitted more than one response, so only one set of responses was used in our analysis to avoid giving extra weight to those programs. Email invitations were sent three times prior to October 31, 2006, at which time the survey site was closed. Email invitations to 20 centers were returned as ‘undeliverable’ or ‘error’. The 132 responding programs represented 71% of programs that received the email and 53% of all UNOS-listed programs.

Program characteristics

Table 1 lists program characteristics for both the current survey and the 1995 survey (6). Among those completing the current survey, 51% were nephrologists and 35% were surgeons, representing an increased proportion of surgeon respondents compared to 1995. Sixty-one percent of programs perform both adult and pediatric transplants, more than twice the percentage doing so in 1995. All UNOS regions were represented by multiple centers. Mean number of kidney transplants performed by responding programs in 2005 was as follows: total = 93.7 ± 67.4, deceased donor = 55.0 ± 40.3, living related donor = 26.0 ± 24.0 and living unrelated donor = 12.7 ± 12.9. These mean numbers correspond to 10 965 total transplants, 6436 from deceased donors, 3036 from living related donors and 1484 from living unrelated donors. Using publicly available data (http://www.ustransplant.org), we compared survey responders and nonresponders on patient volume statistics. Responders performed significantly more total, deceased donor, and living donor transplants than nonresponders (55.0 ± 40.4, 31.8 ± 31.3, 23.3 ± 28.2, respectively—all p's < 0.0001).

Table 1.  Demographics

Variable	2007 (%)	1995 (%)*
*From Bia et al. (6).
Respondents
Nephrologist	51	72
Surgeon	35	14
Other	15	14
Types of surgery
Adult only	32	58
Pediatric only	7	12
Both adult and pediatric	61	30
Type of center
University	59	65
Private hospital, affiliated with university	27
Private hospital, not affiliated with university	13	32
Military-based	1	3

No program in 2005 refused to accept living unrelated donors, but 9% performed no such transplants that year. In contrast, the 1995 survey reported 8% of centers refusing to accept living unrelated donors, but 69% transplanted no kidneys from unrelated donors in 1992.

Medical evaluation

Age: One of the most striking changes between 1995 and 2007 is in the exclusion criteria for kidney donation by age. Figure 1A shows the distribution of upper limits by age. Although a similar percentage of programs (21%) exclude candidates over 65 years old, the percentage of programs without a set upper age limit has now more than doubled to 59%. This is reflected in a reduction in programs with other cutoffs by age, including 55, 60 and 70 years old. In contrast to this less strict policy toward upper age limits, programs have become stricter with respect to younger candidates. Figure 1B shows that there are no longer any programs reporting an age cutoff of 14 or 16, and almost none saying that they have no lower age limit at all. This is reflected in an increase in the number of programs in 2007 with higher cutoffs, although most programs continue to use 18-year-old as the lower age limit. A number of comments accompanying answers to this question reported that younger than 18-year-old is an absolute exclusion criterion, but that age 18–21 years is a relative exclusion, and that these candidates must be carefully evaluated for maturity and mitigating circumstances before being accepted.

Figure 1. (A) Exclusion criteria by older age (n = 121). Data from 1995 is from Bia et al. (6). (B) Exclusion criteria by younger age (n = 121).

Diabetes: No clear consensus emerged regarding screening or excluding patients for glucose intolerance. Glucose tolerance testing (GTT) is performed by most programs for elevated fasting glucose, strong family history of diabetes or a history of gestational diabetes (Figure 2A). Screening for glucose intolerance is generally applied more broadly now compared to 1995. Thirteen percent of (n = 132) respondents screen for anti-islet antibodies in potential donors for a family member with diabetes, similar to the practice in 1995. A total of 49% of programs exclude donors based on elevated fasting blood glucose (FBG), but various cutoffs are used to define ‘elevated’ (Figure 2B). Most programs exclude based on abnormal oral glucose tolerance test or Type II diabetes. These criteria were not dramatically different from those reported in 1995, when 61% of centers excluded for ‘mildly elevated FBG’ and 46% for abnormal GTT.

Figure 2. (A) Glucose tolerance testing is done for … (n = 132). (B) Exclusion criteria based on glucose tolerance (n = 122).

Obesity: Because of data suggesting an association between obesity and kidney disease (9,10), we asked programs which BMI they would use to exclude donors (Figure 3). Fifty-two percent of programs use a BMI cutoff of 35, while 10% exclude donors with BMI over 30. In contrast, the 1995 survey reported that only 16% exclude donors for ‘moderate obesity’.

Figure 3. Exclusion criteria based on BMI (n = 109).

Lipids: Unlike the trend toward stricter BMI criteria, practices regarding lipids have changed little since 1995. Eight percent of (n = 114) respondents now exclude donors for hyperlipidemia (9% in 1995), while 34% have no policy and 58% accept donors with hyperlipidemia.

Family cardiovascular history: Practices with respect to family history also changed little, as 4% of (n = 112) respondents exclude donors for strong family history of cardiovascular disease (6% in 1995), 19% have no policy and 78% accept donors with strong family history.

Hypertension: Hypertension exclusion criteria have become more flexible in 2007 compared with 1995 (Figure 4). While 47% of programs exclude donors on any antihypertensive medication, 41% exclude donors if they are taking more than one medication, and 8% exclude donors taking more than two medications. Many respondents commented that blood pressure criteria are looser if the donor is older, or if end organ damage is ruled out. Others commented that the 24-h blood pressure monitoring can be useful in assessing hypertension, especially if candidates might have ‘white coat’ hypertension.

Figure 4. Exclusion criteria by categories of blood pressure (n = 121).

Cardiac testing: Cardiac stress testing has become somewhat more widespread in the donor evaluation. Two percent of (n = 132) respondents perform stress tests for all donors, 59% for donors over 50 years old, 80% for donors with cardiovascular risk factors and 74% for an abnormal EKG. In 1995, cardiac stress testing was not routinely performed unless the donor had increased age (12%) or cardiovascular risk factors (88%). That is, no programs in 1995 reported testing all donors or donors with abnormal EKG. Ten percent of (n = 132) respondents currently perform echocardiograms in all donors, 43% for donors over 50 years old, 61% for cardiac symptoms and 61% for abnormal EKG. The approach was similar in 1995, when an echocardiogram was routinely obtained in 12%, and in 66% of programs was obtained for older donors, those with cardiac symptoms or an abnormal EKG.

Renal evaluation

Anatomy: Because of the evolution of radiological techniques, the evaluation of renovascular anatomy has dramatically changed since 1995 (Figure 5). Seventy-three percent of programs now perform spiral CT and 16% use MR angiography. Renal angiography was universally performed in 1995, but only 9% still use this modality.

Figure 5. Renal anatomy—tests routinely performed in all potential donors (n = 131).

GFR: In contrast to the changes in evaluating renal anatomy, little has changed in estimating glomerular filtration rate (Figure 6A). Almost all programs (90%) obtain a 24-h urine for creatinine clearance, with the remaining programs using a radioactive isotope or iodinated tracer. The biggest change regarding acceptable creatinine clearance for donation is that few programs now have no specific cutoff, and no programs use 40 or 60 mL/min/1.73 m² as cutoffs (Figure 6B). Most programs now use 80 mL/min/1.73 m² as their cutoff, although 25% of programs now require the creatinine clearance be within two standard deviations of the mean creatinine clearance for the donor's age. Of those responding that they have a different cutoff for creatinine clearance, the most common comment was that they require at least 90 mL/min/1.73 m² to avoid taking donors who could be classified as chronic kidney disease (CKD) Stage II (defined as 60–89 mL/min/1.73m²).

Figure 6. (A) Kidney function – tests routinely performed in all potential donors (n = 131). (B) Exclusion criteria by creatinine clearance (in mL/min/1.73 m²).

Proteinuria: The evaluation of donors for proteinuria has undergone minor changes between 1995 and 2007. Most programs (76%) use a 24-h urine collection for protein (Figure 7A). Some programs rely on a spot urine protein to creatinine ratio, and almost one-half of programs now use urinary albumin as a screen, while only 5% checked albumin in 1995. The most common exclusion criterion is 300 mg/day proteinuria, as it was in 1995, but almost as many programs now use 150 mg/day as a cutoff, unless the proteinuria is postural (Figure 7B). Since the 1995 survey did not ask about using a cutoff of 150 mg/day, the number of programs doing so then cannot be assessed.

Figure 7. (A) Proteinuria—tests routinely performed in all potential donors (n = 131). (B) Exclusion criteria by level of proteinuria (n = 121).

Hematuria: Essentially all programs (93% of 122 respondents) still perform a urinalysis as part of a donor evaluation, but only 68% routinely do a urine culture. The most common practice now (43%) is to only accept donors with hematuria, defined as greater than 10 red blood cells per high powered field (rbc/hpf), if urological evaluation and renal biopsy are negative. Twenty-one percent of programs automatically exclude potential donors with greater than 10 rbc/hpf, regardless of work-up. In 32% of programs, hematuria was defined as greater than 3 rbc/hpf (an option not offered in the 1995 survey), and programs would accept donors with this amount of hematuria if the urological and renal evaluation were negative.

Kidney stones: Most commonly, programs accept a potential donor with a history of nephrolithiasis if no stones are present and metabolic studies are normal (Figure 8). Compared to 1995, more programs now accept the donor as long as stones are no longer present, and fewer programs automatically exclude donors with a history of stones. Five percent of programs reported no policy regarding stone history.

Figure 8. Policy toward history of kidney stones (n = 120).

Polycystic kidney disease (PKD): Ninety percent of (n = 120) respondents currently accept a donor who is related to a recipient with PDK as long as the CT scan or ultrasound is normal, and the donor is over 30 years old. Only 7.5% of programs will accept such a donor with negative imaging regardless of donor age, in contrast to the greater than 23% of programs in 1995 that had no age requirement. Genetic testing for PKD was not available in 1995, but 25% of programs currently use such testing to clear donors.

Factors influencing practice patterns

Flexibility in applying selection criteria: To obtain information about how flexibly programs might apply selection criteria, we asked a series of questions about decision-making flexibility in the context of several donor and recipient factors (Figure 9). When asked about the impact of younger donor age on how flexibly to apply criteria, programs tended to be not at all flexible (Figure 9A). In contrast, programs were much more flexible in the event of older donor age (Figure 9B). Programs were not at all flexible in applying the selection criteria based on the severity of the recipient's medical condition (Figure 9C), but were more flexible based on the preexisting emotional relationship between donor and recipient (Figure 9D).

Figure 9. How flexible is your application of the donor selection criteria, based on the following factors? (A) Younger donor age (n = 110). (B) Older donor age (n = 111). (C) Severity of recipient's medical condition (n = 109). (D) Pre-existing emotional relationship between donor and recipient (n = 110).

We performed a number of analyses to determine if program size was associated with more or less restrictive exclusion criteria for donors. Unlike the 1995 survey, which found a number of associations between larger programs and less restrictive policies, we found no relationship between program size and any exclusion criteria. In addition, practice patterns did not significantly differ by geography, as assessed by UNOS region.

We also examined associations between more restrictive policies in multiple areas of the medical evaluation. We found surprisingly few associations among policies regarding acceptance of older donors, higher BMI, more elevated glucose and more severe hypertension. Only programs with more strict policies regarding hypertension tended to have more strict criteria regarding elevated glucose in donors (p = 0.003).

Risk counseling

The long-term risk of renal failure related to kidney donation is usually described to donors as ‘small’ rather than ‘nil’, an almost complete reversal of the terms used in 1995 (Figure 10). The long-term risk of hypertension related to kidney donation is also usually described as ‘small’ (82%) rather than ‘nil’ (8%), with 10% saying the risk is unknown. Similarly, the long-term risk of proteinuria is usually described as ‘small’ (79%) rather than ‘nil’ (7%), with 14% saying the risk is unknown. In 1995, 60–65% of programs informed potential donors that there is a ‘small’ risk of developing proteinuria or hypertension related to donation.

Figure 10. Donors are told that the long-term risks of renal failure related to donation are (n = 115):

Discussion

Our findings represent practices as they are reported, and we cannot confirm how accurately they reflect actual practice, a limitation of any survey data. In addition, we can not exclude the possibility of a systematic bias among programs that chose to respond compared to programs that did not. The mean number of transplants performed was greater among responders than nonresponders, but a wide range of transplant volumes was represented among responders, and we did not find any differences in practices based on program size. Therefore, we do not believe that the practices of responding programs systematically differ from those that did not respond. In addition, the sum of 2005 transplants reported by survey respondents is 10 965. This figure represents 67% of the 16 481 kidney transplants performed in the United States that year, but likely underestimates the collective transplant experience of the sample because 15 (11%) centers did not report volume statistics. Of note, we received responses from 71% of the programs that received our email.

The expanding use of living donors has been well documented (1), and our findings about the use of living unrelated donors are consistent with this trend. Between 1986 (11) and 1995 (6), the percentage of US transplant programs performing transplants using living unrelated donors increased from 16% to 31%, while 91% of programs perform such transplants now, and for the first time, no programs said they would refuse to consider an unrelated living donor. Thus, over two decades, opposition to living donation, especially living unrelated donation, has gone from strong to weak, to essentially nonexistent.

Consistent with the growth in living donation, programs have made a number of medical exclusion criteria less strict. For example, older donors, donors taking antihypertensive medications and donors with a history of kidney stones are now more likely to be accepted than in 1995. Forty-one percent of programs consider donors with well-controlled hypertension on one medication, and 8% will consider donors on two medications. One possible explanation for shifting practices regarding hypertension is that ‘borderline’ hypertension is now more likely treated with medications (12). In addition, most adults with hypertension 12 years ago were treated using a ‘step’ approach, whereby high doses of a first agent were used before adding a second. However, current approaches to hypertension are more likely to start off using two agents at low dose, so that the use of two agents is no longer necessarily reflective of more severe hypertension (12). Untreated hypertension is clearly associated with elevated cardiovascular risk (13) and kidney disease (14), but these risks are significantly reduced with good control (15). In the absence of data on long-term risks after donation by well-controlled hypertensives, these donors should be evaluated cautiously.

In some areas of the medical evaluation, criteria have not become notably less or more restrictive, but more widespread or detailed testing is being applied. This may be due to the willingness of programs to consider higher risk donors in recent years. For example, cardiac testing and GTT appears to be applied more broadly. In some areas, entirely new tests have become available, such as the use of CT angiography to evaluate renal anatomy and genetic testing for donors who have family members with PKD.

Previously published practice guidelines suggest that diabetics, defined as FBG ≥ 126 mg/dL or 2-h OGTT ≥ 200, should be excluded as kidney donors (8). However, we find that the current practices are quite variable with respect to potential donors with glucose intolerance, defined by the American Diabetes Association (ADA) as FBG between 100 and 126 mg/dL or 2-h OGTT between 140 and 200 mg/dL (16). Only 37% of programs use FBG over 100 mg/dL to exclude donors, while 61% exclude donors based on 2-h OGTT over 140 mg/dL. Although ADA recommendations for population screening for diabetes and glucose intolerance may not be directly applicable to the evaluation of potential donors, donors meeting criteria for glucose intolerance have an increased risk of developing diabetes, and should be evaluated cautiously.

Despite the expansion of some criteria for donation, other criteria have become more restrictive. For example, programs no longer accept donors under 18 years of age. Stricter criteria are also applied now with respect to obese donors, with most programs excluding potential donors with a BMI over 35. Closer attention to obesity may be related to significant increases over the last few decades in the prevalence of obesity in the United States, and the increased recognition of associated morbidity and mortality (17). The negative impact of obesity on renal outcomes in the general population has been extensively described (18,19). In addition, obesity has been shown to worsen outcomes after uninephrectomy. Obese patients, defined as BMI above 30, almost universally developed proteinuria after nephrectomy, and had a much higher rate of renal failure compared to those with BMI below 30 (9). These findings suggest that the increased attention to obesity in donors is appropriate, and that caution is warranted in evaluating donors with BMI over 30, especially if the BMI is over 35.

Criteria with respect to minimal creatinine clearance also have become more restrictive, in that programs no longer accept donors with creatinine clearances of 40 or 60 mL/min/1.73 m². This may be due to the increased awareness that such creatinine clearances would now be classified as Stage III CKD ‘moderate decrease in GFR’(20). The cutoff of 80 mL/min/1.73 m² that is still most commonly used to exclude potential donors is within Stage II CKD, ‘kidney damage with mild decrease in GFR’, defined as eGFR between 60 and 89 mL/min/1.73 m² (20). The continued use of this cutoff may be because 80 mL/min represents a reasonable minimum kidney function from a recipient's perspective (21), and also may reflect the influence of Bia et al.'s survey, which reported 80 mL/min as the most commonly used cutoff. However, many respondents commented that 90 mL/min/1.73 m² would be a better cutoff, because anything lower is Stage II CKD. In the absence of long-term follow-up data, it is impossible to argue strongly between the choice of 80 or 90, but the increased awareness of CKD stages may increase the use of 90 mL/min/1.73 m² in the future.

Another approach to avoiding the use of donors with impaired kidney function, reported in one-quarter of programs, is to make sure that the potential donor's creatinine clearance is within two standard deviations of the means for the donor's age. This approach focuses less on sufficient kidney function for the recipient, and more on excluding donors with lower than expected kidney function. However, ‘normal values’ by age differ among studies (22–24), and among different methods of measurement—for example, values based on creatinine clearance correlate poorly with direct measurements of glomerular filtration rate (25).

In addition, a number of technical issues make the interpretation of creatinine clearance data problematic. For example, our survey asked about exclusion criteria based on creatinine clearance normalized per 1.73 m², but we did not ask how commonly programs convert the 24-h creatinine clearance value to a normalized value—it is possible that many programs do not routinely do this. Not normalizing the value could potentially introduce substantial error, because not every donor is an ‘average’ person with a body surface area of 1.73 m². In addition, 24-h urine collections for creatinine clearance can be inaccurate due to many factors, including measurement variability between labs (26), patient collection error, and other factors that effect the production, filtration and tubular secretion of creatinine (20). Only about 10% of programs use a direct measure of GFR, based on a radioactive or iodinated marker, rather than an estimate based on creatinine. This proportion has not changed over the past 10 years, probably because of ongoing cost and availability barriers to using direct measures. However, given the expanding use of living donors, and the importance of accurate screening methods, the use of direct GFR measures may increase in the future.

In their subjective descriptions of donor risks, most programs now use the term ‘small’ to describe the rate of renal failure, hypertension and proteinuria related to kidney donation. The use of this term may reflect a belief that one should never say ‘never’, as well as medico-legal concerns about saying a risk is ‘nil’—some donors do eventually develop complications such as renal failure, and it may be impossible to determine whether the complication is related to kidney donation. However, existing data (2–5) suggest that, in young healthy donors, there is no increased risk of renal failure compared to nonnephrectomized controls. Therefore, we suspect that more programs compared to 1995 describe the risk of renal failure as ‘small’ rather than ‘nil’ because of the ongoing use of higher-risk donors.

Overall, programs continue to have differing approaches to the evaluation of living donors. In addition, we found no evidence that some programs are consistently strict in all their exclusion criteria, nor that other programs are consistently less strict: programs determine the selection criterion for each medical test individually. Unlike the 1995 study, we did not find associations between smaller program size and more strict application of exclusion criteria. This may reflect the more widespread dissemination of available data and experience from various programs.

In 1995, Bia et al. (6) concluded that the long-term follow-up of living kidney donors is essential, so that transplant programs can use outcome data to improve their evaluation strategies. However, there still exists little data on outcomes of living donation among patient subsets with conditions that might affect long-term kidney function, such as older age, hypertension and glucose intolerance. One result of program variability in selection criteria is that motivated donors can ‘shop around’ for a program that will accept them. Although we have no data on how common this is, our personal experience is that it is not uncommon. Our principal recommendation is for the establishment and widespread use of donor registries, to provide long-term data on outcomes among various categories of donors, and on the utility of different tests used in donor evaluations.

Acknowledgment

This work was supported by the Julie Henry Fund in the Transplant Center at Beth Israel Deaconess Medical Center.