For patients with kidney failure, kidney transplantation prolongs survival, improves quality of life, and reduces costs compared with dialysis. However, according to Elizabeth M. Sonnenberg, MD, and colleagues, there is a large gap between the demand for transplantation and supply of organs available. Identification of centers that are associated with the best outcomes of transplantation and encouraging patients to utilize those centers is a potentially valuable strategy for maximizing the benefit of transplantation.
Previous studies have shown an association between improved outcomes and high-volume centers in a variety of surgical fields. However, there are few data revealing an association between kidney transplantation volume and survival. Dr. Sonnenberg et al. conducted a retrospective cohort study to examine whether a center volume-outcome relationship exists for contemporary kidney transplantation, specifically for recipients with diabetes, recipients ≥65 years of age, and recipients of high kidney donor profile index (KDPI ≥85) kidneys. The researchers sought to test the hypothesis that compared with low-volume centers, high-volume centers would have decreased graft failure and patient mortality. Results were reported in the American Journal of Kidney Diseases [2019; 74(4):441-451].
The researchers utilized data from the Organ Procurement and Transplantation Network to identify a cohort of adults ≥18 years of age who underwent kidney-only transplantation between January 1, 2009, and December 31, 2013. Transplantation centers were stratified into quartiles: Q1, low (annual range, 2-65); Q2, medium (annual range, 66-110); Q3, medium-high (annual range, 111-195); and Q4, high (annual range, 198-315). Centers performing <10 transplantations during the study period were excluded. The primary outcomes of interest were all-cause graft failure and mortality within 3 years of transplantation. Recipients of living and deceased donor organs were analyzed separately.
The final cohort included 79,581 kidney transplantations performed at 219 centers. Deceased donor transplantations constituted a lower proportion of transplantations at Q4 centers compared with Q1 centers (57.1% vs 67.1%; P<.001). Of their total deceased donor volume, Q4 centers used a greater proportion of high-KDPI kidneys compared with Q1 centers (12.5% vs 8.3%; P<.001).
There were slight variations in patient characteristics among the volume quartiles. Deceased donor grafts at Q4 centers had longer cold ischemia times: 10.8% of grafts at Q4 centers had ≤36 hours compared with 1.1% of grafts at Q1 centers. There were also variations in donor characteristics across volume quartiles. Deceased donor kidneys at Q4 centers had higher median KDPI score compared with those used at Q1 centers (53 vs 44; P<.001). Q1 centers used a larger proportion of donation after cardiac death donors compared with Q4 centers (16.4% vs 13.6%; P<.001).
There were significant differences in unadjusted 3-year all-cause graft failure and mortality rates across volume quartiles. There were also significant differences across volume quartiles in unadjusted Cox models. The differences were small in absolute terms. Unadjusted rates of all-cause graft failure were 14.9% in Q1, 15.3% in Q2, 14.9% in Q3, and 16.7% in Q4. Rates of mortality were 9.1% in Q1, 8.8% in Q2, 8.4% in Q3, and 9.8% in Q4.
In analyses of graft failure, center volume had a borderline statistically significant interaction with recipients with diabetes (interaction term P=.05) and high-KDPI kidney recipients (interaction term P=.05). There was no interaction between recipient age and center volume for either outcome.
A total of 5128 recipients with diabetes who received a deceased donor transplant (38.1% of all deceased donor transplants) received care at a Q1 center; 3973 recipients with diabetes who received a deceased donor transplant (35.3% of all deceased donor recipients) were treated at a Q4 center. While not clinically meaningful, there were significant differences across volume quartile in unadjusted 3-year all-cause graft failure and mortality rates. The highest unadjusted rates of all-cause graft failure were in Q4 centers (19.7% vs 18.4%, 18,1%, and 17.7% in Q1, Q2, and Q3 centers, respectively; P=.02); rates of mortality were also highest in Q4 centers (13.8% vs 13.2%, 12.6%, and 11.7% in Q1, Q2, and Q3 centers, respectively; P=.01).
Because low-volume centers transplanted a larger proportion of deceased donor organs, all quartiles used a substantial number of high-KDPI kidneys: 1117 at Q1 centers versus 1401 at Q4 centers. In unadjusted analysis of 3-year all-cause graft failure and mortality rates of high-KDPI kidneys, the rates were lowest for Q3 centers: all-cause graft failure in Q3 centers, 23.3% versus 26.5% for Q1 centers, 28.0% for Q2 centers, and 26.5% for Q4 centers; mortality in Q3 centers, 13.0% versus 16.0% at Q1 centers, 17.5% at Q2 centers, and 15.0% at Q4 centers. There was no significant effect of center volume on all-cause graft failure and mortality in multivariable Cox frailty models.
There were some limitations to the study, including unmeasured confounding from patient comorbid conditions and organ selection and the potential for measurement errors in registry data.
In conclusion, the researchers said, “This study found no evidence that increased center volume was associated with improved outcomes for kidney transplant recipients. Importantly, this finding remained consistent among increased-risk recipients and increased-risk donors. For nephrologists, who influence where patients seek a transplantation evaluation, these results would argue against referral to larger centers based on volume alone. Other patient-specific considerations, such as proximity to center, may matter more than center volume when selecting a center. Additional research is needed to understand qualities and practices of transplantation centers that generate superior outcomes for patients.”
- Researchers conducted a retrospective cohort study to test the hypothesis that high-volume kidney transplantation centers would generate superior outcomes compared with low-volume centers.
- In multivariable Cox regression models, there was no significant association between center volume and all-cause graft failure or mortality within 3 years post-transplantation.
- Transplant recipients with diabetes had slightly lower 3-year mortality rates at centers with medium-high volume compared with centers with low volume (adjusted hazard ratio, 0.85; 95% confidence interval, 0.73-0.99).