Abstract
Background and objectives: Clinical outcomes after kidney transplant have improved considerably in the United States over the past several decades. However, the degree to which this has occurred uniformly across the country is unknown.
Design, setting, participants, & measurements: Regional variations in graft failure after kidney transplant during three different time periods were examined. These time periods were chosen to coincide with major shifts in immunosuppressant usage: Era 1, cyclosporine usage, 1988 through 1989; Era 2, introduction of tacrolimus and mycophenolate mofetil, 1994 through 1995; and Era 3, widespread use of tacrolimus and mycophenolate mofetil, 1998 through 1999. Patient data were obtained from the United States Renal Data System database. For each period, regional differences in time from transplant to graft failure (organ removal, death, or return to dialysis) were examined. For each region, differences in graft failure over time were examined.
Results: One-year graft survival rates ranged from 76% to 83% between regions in Era 1 (n = 13,669), from 84% to 89% in Era 2 (n = 17,456), and from 87.5% to 92% in Era 3 (n = 20,375). Three-year graft survival ranged from 65% to 75% between regions in Era 1, from 84% to 89% in Era 2, and from 77% to 86% in Era 3. Adjusted models for donor and recipient characteristics showed improvements in graft survival over time in all United Network for Organ Sharing regions with minimal variation across regions.
Conclusions: Regional differences in graft survival after kidney transplant are minimal, particularly when compared with the dramatic improvements in graft survival that have occurred over time.
Regional variations in health outcomes have been described in a wide variety of health care settings and are often attributed to variations in practice patterns (1–11). For kidney transplant patients, there are known regional differences in wait-list mortality, access to deceased-donor kidneys, and time to transplant (2,12). However, relatively little is known about regional variation in outcomes after kidney transplant and whether known improvements in graft survival that have occurred over time have occurred uniformly across all regions of the United States.
Graft failure after kidney transplant across all United Network for Organ Sharing (UNOS) regions and within each region over time were examined. It was hypothesized that graft failure would vary by region and that these variations would have become less pronounced over time.
Materials and Methods
UNOS Regions and Time Periods
To facilitate organ procurement, allocation, and transplantation, the United States is divided into 11 UNOS regions, approximately corresponding to the U.S. Census regions (Figure 1). Kidney allograft failure across the 11 UNOS regions (based on the patient's region of residence) during three different 2-yr time periods of kidney transplantation (eras) were compared; they were chosen to correspond with major shifts in immunosuppressant use over time: 1988 and 1989 (Era 1: predominantly cyclosporine era), 1994 and 1995 (Era 2: introduction of tacrolimus and mycophenolate mofetil), and 1998 and 1999 (Era 3: widespread use of tacrolimus and mycophenolate mofetil). These eras were chosen to allow for at least 3 yr of post-transplant follow-up time for all patients.
Figure 1.
Map of United Network for Organ Sharing (UNOS) regions. Reprinted from reference 17, with permission.
Data Sources and Patients
The study was conducted using data from the United States Renal Data System (USRDS), a comprehensive national registry for patients with end-stage renal disease. Specifically, the PATIENTS and TXUNOS files were used to ascertain information on recipient and donor characteristics for this study. From the registry, all patients aged 18 years and older who underwent their first kidney transplant during one of the three time periods selected for analysis were identified. Recipients who received another organ in addition to a kidney were excluded. Differences in 1-yr and 3-yr allograft failure across regions during each of the eras chosen were examined.
Predictor Variables
The primary predictor variable for all analyses was each recipient's UNOS region of residence at the time of transplant based on the residence ZIP code reported in the TXUNOS file. Outcomes by UNOS region during any given era were compared with the national average for that era. Multivariable analyses were adjusted for a wide range of recipient and donor characteristics (individual variables are listed in the footnote to Table 3). A secondary analysis that compared outcomes within each region during different time periods was performed. For this analysis, Era 1 served as the referent.
Table 3.
| Region
|
Odds ratio (95% CI)
|
||
|---|---|---|---|
| Era 1
|
Era 2
|
Era 3
|
|
| All regions | Referent 1.00 | Referent 1.00 | Referent 1.00 |
| 1 | 0.86 (0.68–1.09) | 0.98 (0.84–1.15) | 0.91 (0.67–1.22) |
| 2 | 0.91 (0.78–1.05) | 1.04 (0.87–1.24) | 1.17 (1.01–1.38) |
| 3 | 0.93 (0.81–1.07) | 0.97 (0.81–1.16) | 0.83 (0.65–1.05) |
| 4 | 1.18 (0.94–1.48) | 1.05 (0.91–1.22) | 1.18 (0.95–1.46) |
| 5 | 1.26 (1.11–1.43) | 0.92 (0.80–1.05) | 0.90 (0.75–1.05) |
| 6 | 1.01 (0.78–1.30) | 0.82 (0.68–0.99) | 0.89 (0.59–1.35) |
| 7 | 0.84 (0.76–1.16) | 0.88 (0.75–1.04) | 0.93 (0.73–1.18) |
| 8 | 0.87 (0.72–1.06) | 1.14 (0.92–1.39) | 1.01 (0.80–1.28) |
| 9 | 1.26 (1.03–1.55) | 1.24 (1.06–1.45) | 1.08 (0.91–1.28) |
| 10 | 0.96 (0.82–1.14) | 1.07 (0.91–1.25) | 1.13 (0.93–1.36) |
| 11 | 0.88 (0.72–1.07) | 0.98 (0.83–1.15) | 1.03 (0.87–1.22) |
CI, confidence interval; UNOS, United Network for Organ Sharing.
Era 1, 1988 and 1989; Era 2, 1994 and 1995; Era 3, 1998 and 1999.
Multivariate analysis was adjusted for many factors. The recipient factors included age and time on dialysis (vintage) before transplant; race (African-American versus other); Hispanic ethnicity versus non-Hispanic ethnicity; wait time: time on wait list (1-yr increments, with 0–1 the referent); sex; body surface area: <1.6, 1.6–1.8, >1.8–2.0 (referent), >2.0–2.2, and >2.2 m2; dialysis modality before transplant (peritoneal versus hemodialysis); transplant before versus after initiation of dialysis; comorbid conditions, including diabetes mellitus, chronic obstructive pulmonary disease, hypertension, cerebrovascular disease, peripheral vascular disease, and cardiovascular disease; recipient insurance status at transplant (private, preferred provider organization, or health maintenance organization versus other); functional status (independent in activities of daily living versus not); employment status at transplant (employed versus unemployed); educational level (college-educated versus not); 3-yr average median household income [ascertained by state from 2002 U.S. Census Bureau Income data (18) and categorized into four roughly equal groups: $40,000 (referent), $40,001–$44,000, $44,001–$48,000, and >$48,000]; transplant center; and most recent and peak panel reactive antibody status (0%, 1–50%, or >50%). Donor characteristics included age, sex, race [categorized as white (referent), African-American, or other], donor type (living versus deceased), donor UNOS region (similar to recipient UNOS region versus not), number of HLA antigen mismatches (0–6), cold ischemia time [categorized as <1 h (referent), 1–12 h, 13–24 h, or ≥25 h], and cause of donor death [vascular (cerebrovascular and cardiovascular) versus nonvascular].
Outcome Measures
Outcome measures were allograft failure within 1 and 3 yr of first kidney transplant, respectively. The standard UNOS “composite” definition of allograft failure as “organ removal, death, or return to dialysis” (13) was adopted. Date of graft failure was ascertained from the TXUNOS file.
Statistical Analyses
Logistic regression analysis was used to examine the unadjusted and adjusted associations of UNOS region of residence with both 1-yr and 3-yr allograft failure. These associations within each region over time were also examined using the earliest era as a referent. Multivariable models of regional effects were adjusted for all donor and recipient characteristics that were significantly associated with the outcome on univariate analysis (P < 0.05). All multivariable models were adjusted for a random effect for transplant center. Clustering at the level of the transplant center was accommodated for by using the robust variance method of Lin and Wei (14). The odds ratio (OR) of each region versus the overall national average for that time period was calculated as the approximately weighted linear combination of the model's coefficient.
For all analyses, if >10% of continuous data and >5% of categorical values were missing, a “missing” category was created, so that the explanatory value of the covariate could still be captured without making unnecessary assumptions. Additionally, a small percentage of the study cohort (0.35%) were missing wait list dates. In these patients, the date of first service for end-stage renal disease treatment was used as the beginning date of their wait time. All statistical analyses were conducted using STATA statistical software, version 8.0 (StataCorp LP, College Station, TX).
Results
Study Cohort
A search of the USRDS identified a total of 13,669 adult kidney transplant patients in Era 1, 17,456 in Era 2, and 20,375 in Era 3. The mean (SD) age of patients increased from 41.9 (13.2) years in Era 1 to 47.1 (13.2) years in Era 3 (Table 1). The percentage of Hispanic patients increased from 8.4% in Era 1 to 12% in Era 3. The burden of comorbid conditions among allograft recipients also appears to have increased over time. For example, from Era 1 to Era 3, the percentage of recipients with diabetes increased from 16.6% to 27.5%, and the percentage with cardiovascular disease increased from 6.1% to 10.5% (Table 1). Socioeconomic indicators improved from Era 1 to Era 3, with a higher percentage of patients being college-educated (20% and 36%, respectively) and employed (7.9% and 15.7%, respectively). Median time on dialysis before transplant (vintage) also increased from 13.6 to 20.3 mo, and the percentage of transplants performed preemptively increased from 0.1% to 13.1%.
Table 1.
Characteristics of kidney transplant recipients by eraa
| Variable | % missing, if any | Era 1 (n = 13,669) | Era 2 (n = 17,456) | Era 3 (n = 20,375) |
|---|---|---|---|---|
| Mean age, yr (SD) | 41.9 (13.2) | 44.8 (13.0) | 47.1 (13.2) | |
| Female, % | 39.5 | 39.3 | 40.5 | |
| African-American, % | 21.7 | 24.6 | 23.9 | |
| Hispanic, % | 8.4 | 11.0 | 12 | |
| Mean BMI (SD) | 7 | 24.5 (4.5) | 25.5 (5.0) | 26.3 (5.3) |
| Employed, % | 11 | 7.9 | 11.2 | 15.7 |
| College education, % | 48 | 20 | 32 | 36 |
| Functional status,b % | 16 | 65.5 | 77.0 | 83.1 |
| Median income,c $ | 43,786 (5,359) | 43,774 (5,560) | 44,038 (5,601) | |
| Recipient primary insurance status,d % | 35 | 0.2 | 28.1 | 39.1 |
| Peak PRA ≥50, % | 4 | 10.4 | 6.9 | 6.5 |
| Most recent PRA ≥50, % | 4 | 5.8 | 3.2 | 2.8 |
| Median vintage, mo | 13.6 | 18.2 | 20.3 | |
| Prior dialysis,e % | 40 | 8.3 | 21.1 | 17.3 |
| Preemptive transplant, % | 0.1 | 9.7 | 13.1 | |
| Comorbid conditions, % | ||||
| Treated COPD | 44 | 0.0 | 0.8 | 0.9 |
| HTN | 33 | 83.3 | 79.6 | 81.7 |
| CBVD | 45 | 1.1 | 2.0 | 2.0 |
| CVD | 44 | 6.1 | 5.6 | 10.5 |
| PVD | 45 | 1.1 | 4.6 | 4.0 |
| DM | 32 | 16.6 | 27.0 | 27.5 |
BMI, body mass index; CBVD, cerebrovascular disease; COPD, chronic obstructive pulmonary disease; CVD, cardiovascular disease; DM, diabetes mellitus; HTN, hypertension; PRA, panel reactive antibody; PVD, peripheral vascular disease.
Era 1, 1988 and 1989; Era 2, 1994 and 1995; Era 3, 1998 and 1999.
Independent in activities of daily living.
Three-year average.
At transplant (private, health maintenance organization, or preferred provider organization versus other).
Peritoneal dialysis versus hemodialysis.
Donor characteristics have also changed over time. Mean age of donors increased from 30.6 (16.1) years in Era 1 to 37.0 (17.9) years in Era 3, and the percentage of female donors increased from 40.0% to 48.3% (Table 2). The percentage of Asian and African-American donors has also increased, as has the percentage of living donation, from 20.9% to 36.4%. The percentage of 6-antigen mismatched kidneys has increased slightly, and mean (SD) cold ischemia time has decreased from 19.2 (14) hours in Era 1 to 13.1 (11) hours in Era 3 (Table 2). The percentage of kidneys from donors who died of vascular disease increased from 30.2% to 41.9% over the course of the study.
Table 2.
Characteristics of kidney transplant donors by eraa
| Variable | % missing, if any | Era 1 | Era 2 | Era 3 |
|---|---|---|---|---|
| Mean (SD) age, yr | 1 | 30.6 (16.1) | 33.2 (17.3) | 37.0 (17.9) |
| Female, % | 3 | 40.0 | 44.4 | 48.3 |
| Race, % | 3 | |||
| African-American | 9.1 | 12.5 | 12.1 | |
| Asian | 1.1 | 2.0 | 2.4 | |
| White | 89.1 | 84.5 | 84.4 | |
| Donor type, % living | 20.9 | 29.4 | 36.4 | |
| Same UNOS region as recipient, % | 78.2 | 75.1 | 74.2 | |
| Mean (SD) CIT, h | 4 | 19.2 (14) | 15.4 (12) | 13.1 (11) |
| HLA antigen mismatch, % | 17 | |||
| 0 | 10.3 | 12.8 | 13.6 | |
| 1 | 5.5 | 5.7 | 5.3 | |
| 2 | 13.9 | 15.3 | 13.8 | |
| 3 | 25.2 | 26.6 | 24.9 | |
| 4 | 23.6 | 20.9 | 18.5 | |
| 5 | 15.8 | 13.8 | 15.9 | |
| 6 | 5.6 | 5.0 | 7.9 | |
| Vascular cause of death,b % | 30.2 | 37.4 | 41.9 |
CBVD, cerebrovascular disease; CIT, cold ischemia time; CVD, cardiovascular disease; UNOS, United Network for Organ Sharing.
Era 1, 1988 and 1989; Era 2, 1994 and 1995; Era 3, 1998 and 1999.
CBVD and CVD.
Despite the older age and less favorable comorbidity profile of allograft recipients and the older age and higher prevalence of vascular disease among donors, 1-yr allograft survival improved by ≈10% from 80.5% to 90.6% from Era 1 to Era 3 and 3-yr allograft survival improved by ≈12% from 69.4% to 81.8%. Improvements in 1-yr and 3-yr allograft survival occurred with considerable uniformity across regions. Although outcomes varied slightly by region—with 1-yr graft survival ranging from 76% to 83% in Era 1 versus 87.5% to 92% in Era 3, and 3-yr graft survival ranging from 65% to 75% in Era 1 versus 77% to 86% in Era 3—these differences were much less striking than differences within regions over time (Figure 2).
Figure 2.
Absolute rates of kidney allograft survival across all UNOS regions in Era 1 (1988–1989), Era 2 (1994–1995), and Era 3 (1998–1999). (A) One-year graft survival. (B) Three-year graft survival.
Regional Differences in Allograft Outcomes
After adjustment for differences among regions in donor and recipient characteristics, regional variation in 1-yr allograft failure was minimal in all three eras (Table 3). In Era 1, 1-yr allograft failure differed significantly from the national average in only two regions: region 5, OR [95% confidence interval (CI)] 1.26 (1.11–1.43), and region 9, 1.26 (1.03–1.55). In Era 3, 1-yr allograft failure differed significantly from the average in one region [region 2, OR (95% CI) 1.17 (1.01–1.38)]. However, confidence intervals for regions performing better and worse than the national average overlapped with those of other regions performing at the national average. Over time, no region had consistently better or worse 1-yr survival compared with the national average.
Regional differences in 3-yr graft failure were slightly more pronounced than for 1-yr allograft failure. In Era 1, graft failure rates deviated significantly from the national average in five regions (Table 4). Three regions performed better than the national average, and two performed worse than the national average, with ORs (95% CIs) ranging from 0.78 (0.64–0.96) to 1.26 (1.04–1.54) (Table 4). In Era 3, the odds of graft failure deviated significantly from the national average in two regions: region 4, OR (95% CI) 1.17 (1.02–1.34) and region 9, 1.20 (1.05 to 1.39). Again, confidence intervals for these regions performing above and below average overlapped considerably with those for other regions.
Table 4.
| Region
|
Odds ratio (95% CI)
|
||
|---|---|---|---|
| Era 1
|
Era 2
|
Era 3
|
|
| All regions | Referent 1.00 | Referent 1.00 | Referent 1.00 |
| 1 | 0.90 (0.76–1.07) | 0.95 (0.78–1.16) | 0.93 (0.76–1.13) |
| 2 | 0.93 (0.82–1.05) | 1.09 (0.95–1.23) | 1.13 (1.00–1.30) |
| 3 | 0.90 (0.80–1.00) | 0.91 (0.81–1.04) | 0.87 (0.73–1.02) |
| 4 | 1.26 (1.04–1.54) | 1.20 (1.08–1.34) | 1.17 (1.02–1.34) |
| 5 | 1.26 (1.10–1.44) | 0.96 (0.87–1.07) | 0.90 (0.77–1.04) |
| 6 | 1.03 (0.85–1.25) | 0.87 (0.72–1.05) | 0.77 (0.54–1.08) |
| 7 | 0.94 (0.76–1.17) | 0.85 (0.74–0.98) | 0.95 (0.80–1.13) |
| 8 | 0.78 (0.64–0.96) | 1.11 (0.90–1.26) | 1.10 (0.91–1.32) |
| 9 | 1.09 (0.92–1.30) | 1.12 (1.00–1.24) | 1.20 (1.05–1.39) |
| 10 | 0.99 (0.87–1.14) | 1.06 (0.92–1.21) | 1.03 (0.85–1.25) |
| 11 | 0.84 (0.71–1.00) | 0.97 (0.85–1.10) | 1.02 (0.90–1.15) |
CI, confidence interval; UNOS, United Network for Organ Sharing.
Era 1, 1988 and 1989; Era 2, 1994 and 1995; Era 3, 1998 and 1999.
Multivariate analysis was adjusted for many factors. The recipient factors included age and time on dialysis (vintage) before transplant; race (African-American versus other); Hispanic ethnicity versus non-Hispanic ethnicity; wait time: time on wait list (1-yr increments, with 0–1 the referent); sex; body surface area: <1.6, 1.6–1.8, >1.8–2.0 (referent), >2.0–2.2, and >2.2 m2; dialysis modality before transplant (peritoneal versus hemodialysis); transplant before versus after initiation of dialysis; comorbid conditions, including diabetes mellitus, chronic obstructive pulmonary disease, hypertension, cerebrovascular disease, peripheral vascular disease, and cardiovascular disease; recipient insurance status at transplant (private, preferred provider organization, or health maintenance organization versus other); functional status (independent in activities of daily living versus not); employment status at transplant (employed versus unemployed); educational level (college-educated versus not); 3-yr average median household income [ascertained by state from 2002 U.S. Census Bureau Income data (18) and categorized into four roughly equal groups: $40,000 (referent), $40,001–$44,000, $44,001–$48,000, and >$48,000]; transplant center; and most recent and peak panel reactive antibody status (0%, 1–50%, or >50%). Donor characteristics included age, sex, race [categorized as white (referent), African-American, or other], donor type (living versus deceased), donor UNOS region (similar to recipient UNOS region versus not), number of HLA antigen mismatches (0–6), cold ischemia time [categorized as <1 h (referent), 1–12 h, 13–24 h, or ≥25 h], and cause of donor death [vascular (cerebrovascular and cardiovascular) versus nonvascular].
By comparison, odds of graft failure at 1 yr and 3 yr improved markedly over time in all but one region (Table 5). The magnitude of differences in graft survival over time was far more pronounced than those between regions during the same time period.
Table 5.
Adjusted allograft failure over time within each UNOS regiona
|
A: One-year adjusted allograft failure
| |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Era
|
Region 1
|
Region 2
|
Region 3
|
Region 4
|
Region 5
|
Region 6
|
Region 7
|
Region 8
|
Region 9
|
Region 10
|
Region 11
|
| Odds ratio (95% CI) | |||||||||||
| Era 1 | Referent | Referent | Referent | Referent | Referent | Referent | Referent | Referent | Referent | Referent | Referent |
| Era 2 | 0.58 | 0.71 | 0.61 | 0.61 | 0.40 | 0.30 | 0.63 | 1.78 | 0.60 | 0.55 | 0.51 |
| (0.25–1.34) | (0.42–1.20) | (0.38–0.98) | (0.36–1.01) | (0.26–0.62) | (0.12–0.76) | (0.38–1.05) | (0.84–3.79) | (0.32–1.12) | (0.31–0.97) | (0.32–0.79) | |
| Era 3 | 0.41 | 0.61 | 0.32 | 0.54 | 0.26 | 0.18 | 0.52 | 1.21 | 0.37 | 0.39 | 0.32 |
| (0.13–1.22) | (0.33–1.11) | (0.18–0.56) | (0.31–0.91) | (0.16–0.43) | (0.07–0.47) | (0.29–0.92) | (0.55–2.60) | (0.21–0.65) | (0.21–0.71) | (0.19–0.52) | |
| B: Three-year adjusted allograft failure | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Era
|
Region 1
|
Region 2
|
Region 3
|
Region 4
|
Region 5
|
Region 6
|
Region 7
|
Region 8
|
Region 9
|
Region 10
|
Region 11
|
| Odds ratio (95% CI) | |||||||||||
| Era 1 | Referent | Referent | Referent | Referent | Referent | Referent | Referent | Referent | Referent | Referent | Referent |
| Era 2 | 0.62 | 0.73 | 0.55 | 0.63 | 0.46 | 0.29 | 0.59 | 2.21 | 0.72 | 0.56 | 0.57 |
| (0.30–1.30) | (0.50–1.07) | (0.39–0.77) | (0.42–0.97) | (0.32–0.66) | (0.12–0.67) | (0.44–0.79) | (1.31–3.71) | (0.39–1.32) | (0.35–0.87) | (0.39–0.81) | |
| Era 3 | 0.54 | 0.59 | 0.35 | 0.53 | 0.29 | 0.18 | 0.54 | 1.72 | 0.57 | 0.40 | 0.39 |
| (0.22–1.30) | (0.39–0.89) | (0.22–0.55) | (0.31–0.91) | (0.19–0.45) | (0.07–0.43) | (0.36–0.80) | (0.92–3.21) | (0.30–1.06) | (0.23–0.68) | (0.27–0.58) | |
Multivariate analysis was adjusted for many factors. The recipient factors included age and time on dialysis (vintage) before transplant; race (African-American versus other); Hispanic ethnicity versus non-Hispanic ethnicity; wait time: time on wait list (1-yr increments, with 0–1 the referent); sex; body surface area: <1.6, 1.6–1.8, >1.8–2.0 (referent), >2.0–2.2, and >2.2 m2; dialysis modality before transplant (peritoneal versus hemodialysis); transplant before versus after initiation of dialysis; comorbid conditions, including diabetes mellitus, chronic obstructive pulmonary disease, hypertension, cerebrovascular disease, peripheral vascular disease, and cardiovascular disease; recipient insurance status at transplant (private, preferred provider organization, or health maintenance organization versus other); functional status (independent in activities of daily living versus not); employment status at transplant (employed versus unemployed); educational level (college-educated versus not); 3-yr average median household income [ascertained by state from 2002 U.S. Census Bureau Income data (18) and categorized into four roughly equal groups: $40,000 (referent), $40,001–$44,000, $44,001–$48,000, and >$48,000]; transplant center; and most recent and peak panel reactive antibody status (0%, 1–50%, or >50%). Donor characteristics included age, sex, race [categorized as white (referent), African-American, or other], donor type (living versus deceased), donor UNOS region (similar to recipient UNOS region versus not), number of HLA antigen mismatches (0–6), cold ischemia time [categorized as <1 h (referent), 1–12 h, 13–24 h, or ≥25 h], and cause of donor death [vascular (cerebrovascular and cardiovascular) versus nonvascular].
Discussion
In a large national cohort of kidney transplant recipients, 1-yr and 3-yr graft survival rates after initial kidney transplant improved substantially from 1988 through 1999. These improvements occurred with considerable uniformity across the country. Regional differences in graft survival were minimal in any era and were far less pronounced than the overall improvements in graft survival that occurred during this time period.
Although surprising, the finding of minimal change in regional variation in graft survival after kidney transplant is consistent with literature showing an inconsistent and relatively weak effect of transplant center on graft survival. Gjertson (15) observed superior 1-yr graft survival among patients at large versus small transplant centers (≥400 versus <100 living donor kidneys) but found that 5-yr adjusted allograft survival did not vary by size of transplant center. Terasaki and Cecka (16) showed no significant differences in overall short-term (1-yr) graft survival in larger versus smaller centers. Large centers had slightly improved long-term (10-yr) outcomes among high-risk recipients (older patients and those with diabetes mellitus).
Changes that might explain the substantial improvements in graft survival that occurred during the 1990s include greater standardization of transplant recipient selection criteria, evolution of immunosuppressive strategies, development of methods for dissemination of knowledge via professional societies and large-scale professional meetings, development of an organized effort to improve practices (e.g., decrease in cold ischemia time, increase in the living donor pool, development of transplant nephrology as a subspecialty within nephrology), and emphasis by UNOS on maintaining a comprehensive team approach to improve long-term outcomes.
In contrast with what was found for graft failure, there appear to be relatively large regional differences in kidney transplant rates and wait list times. For example, Ellison et al. (2) reported significant regional variation in rates of transplant within 4 mo of wait listing, ranging from 5.6% (Region 1) to 19.8% (Region 3) of patients undergoing transplant within 4 mo of being wait-listed. Data for 2006–2007 from the Organ Procurement and Transplantation Network indicate significant regional variation in wait time (13). Based on the wait-list registrations as of December 31, 2007, the current median active wait times vary from 385 d (Region 1) to 670 d (Region 5). It is possible that lack of regional variation in transplant outcomes compared with transplant rates and wait times reflects greater uniformity across regions in the characteristics of transplant recipients versus the wider population of ESRD patients or even wait-listed patients. Lack of regional differences in transplant outcomes may also reflect relatively rapid dissemination across regions of key care practices after transplant.
Limitations
Despite the completeness and the large sample size of the United States Renal Data System data, the results may reflect residual confounding by various domains of clinical practice that were not available. Unique center-related variables were adjusted for by using the transplant center code as a surrogate marker. Finally, the presence of missing values is an inherent limitation of registry database analyses. All selected data elements were present in a majority of transplant recipients. The authors did not wish to exclude too many patients from the analysis, to have sufficient power to compare results across regions. However, they did not wish to bias the results with imputation if a high fraction of patients had missing values for a particular covariate; thus the missing variable was created as described previously. This approach was consistent with the overall conservative analytic approach, that is, to explain as much variation as possible by known factors, and to diminish the association that could be attributed to regional variation.
Conclusions
One-year and 3-yr graft survival after initial kidney transplant varied minimally across UNOS regions regardless of the time period examined, particularly when compared with the marked improvements that have taken place over time in graft survival within most regions.
Disclosures
None.
Acknowledgments
The data reported herein were supplied by the United States Renal Data System. The interpretation and reporting of these data are the responsibility of the authors and in no way reflect the official policy or interpretation of the U.S. Government.
This 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 of the Department of Health and Human Services.
Published online ahead of print. Publication date available at www.cjasn.org.
References
- 1.Cooper GS, Yuan Z, Chak A, Rimm AA: Geographic and patient variation among Medicare beneficiaries in the use of follow-up testing after surgery for nonmetastatic colorectal carcinoma. Cancer 85(10): 2124–2131, 1999 [PubMed] [Google Scholar]
- 2.Ellison MD, Edwards LB, Edwards EB, Barker CF: Geographic differences in access to transplantation in the United States. Transplantation 76(9): 1389–1394, 2003 [DOI] [PubMed] [Google Scholar]
- 3.Fang J, Alderman MH: Is geography destiny for patients in New York with myocardial infarction? Am J Med 115(6): 448–453, 2003 [DOI] [PubMed] [Google Scholar]
- 4.Grumbach K, Anderson GM, Luft HS, Roos LL, Brook R: Regionalization of cardiac surgery in the United States and Canada. Geographic access, choice, and outcomes. JAMA 274(16): 1282–1288, 1995 [PubMed] [Google Scholar]
- 5.Hirth RA, Turenne MN, Woods JD, Young EW, Port FK, Pauly MV, Held PJ: Predictors of type of vascular access in hemodialysis patients. JAMA 276(16): 1303–1308, 1996 [PubMed] [Google Scholar]
- 6.Pilote L, Califf RM, Sapp S, Miller DP, Mark DB, Weaver WD, Gore JM, Armstrong PW, Ohman EM, Topol EJ: Regional variation across the United States in the management of acute myocardial infarction. GUSTO-1 Investigators. Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries. N Engl J Med 333(9): 565–572, 1995 [DOI] [PubMed] [Google Scholar]
- 7.Reddan DN, Frankenfield DL, Klassen PS, Coladonato JA, Szczech L, Johnson CA, Besarab A, Rocco M, McClellan W, Wish J, Owen WF Jr; Center for Medicare & Medicaid Services’ End-Stage Renal Disease Clinical Performances Measures Workgroup: Regional variability in anaemia management and haemoglobin in the US. Nephrol Dial Transplant 18(1): 147–152, 2003 [DOI] [PubMed] [Google Scholar]
- 8.Saran R, Dykstra DM, Pisoni RL, Akiba T, Akizawa T, Canaud B, Chen K, Piera L, Saito A, Young EW: Timing of first cannulation and vascular access failure in haemodialysis: an analysis of practice patterns at dialysis facilities in the DOPPS. Nephrol Dial Transplant 19(9): 2334–2340, 2004 [DOI] [PubMed] [Google Scholar]
- 9.Soucie JM, Coates RJ, McClellan W, Austin H, Thun M: Relation between geographic variability in kidney stones prevalence and risk factors for stones. Am J Epidemiol 143(5): 487–495, 1996 [DOI] [PubMed] [Google Scholar]
- 10.Stack AG: Determinants of modality selection among incident US dialysis patients: Results from a national study. J Am Soc Nephrol 13(5): 1279–1287, 2002 [DOI] [PubMed] [Google Scholar]
- 11.Vitale MG, Krant JJ, Gelijns AC, Heitjan DF, Arons RR, Bigliani LU, Flatow EL: Geographic variations in the rates of operative procedures involving the shoulder, including total shoulder replacement, humeral head replacement, and rotator cuff repair. J Bone Joint Surg Am 81(6): 763–772, 1999 [DOI] [PubMed] [Google Scholar]
- 12.U.S. Renal Data System, USRDS 2004 Annual Data Report: Atlas of end-stage renal disease in the United States: National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, 2004
- 13.United Network for Organ Sharing: Organ donation and transplantation. Available from: http://www.unos.org/policiesandbylaws/policies.asp. Accessed 2007.
- 14.Lin DY, Wei LJ: The robust inference for the Cox proportional hazards model. J Am Stat Assoc 84(408): 1074–1078, 1989 [Google Scholar]
- 15.Gjertson DW: Center and other factor effects in recipients of living-donor kidney transplants. Clin Transpl 209–221, 2001 [PubMed]
- 16.Terasaki PI, Cecka JM: The center effect: Is bigger better? Clin Transpl 317–324, 1999 [PubMed]
- 17.United Network for Organ Sharing: Organ donation and transplantation: Who we are: Regions [database on the internet]. Richmond, VA (cited 2006 Jan 11). Available from: http://www.unos.org/whoWeAre/regions.asp, 2005
- 18.U.S. Census Bureau. Three-year-average median household income by state: 2000–2002 [database on the Internet]. Washington, DC (updated 2004 May 13; cited 2005 Dec 8). Available from: http://www.census.gov/hhes/income/income02/statemhi.html, 2002
