Purpose: Understanding center-level differences in usage of marginal deceased-donor kidneys (suboptimal kidneys, SOK) can help centers increase their deceased-donor transplant rate (DDKT), and to direct SOK offers to centers willing to accept them.

Methods: We studied 5,646,748 offers of 8,539 SOK donors from 1/1/2018-12/31/2019 using SRTR data. Pediatric centers, center annual volume<1, or multiple match-runs were excluded. We defined SOK offer as: age>60, cold ischemia time (CIT)>24 hours, hepatitis C (HCV), serum creatinine (SCr)>2.0 mg/dL, donated after cardiac death (DCD), KDPI>85, or infectious risk donors (IRD). We used multilevel logistic regression to calculate the median odds ratio (MOR), a measure of center-level variation, for each subtype of SOK.

Results: There were 18,471 (64%) SOK among 28,676 DDKT performed, Among 197 centers, the percentage of SOK acceptance ranged from 0 to 44.7% with the median (IQR) of 10.6% (6.9%, 14.6%). The SOK subtype with the least center-level variance was DCD (acceptance range: 0-95%, median [IQR]: 10.2% [6.4%, 15.9%], MOR=2.11) and IRD (acceptance range: 0-63.6%, median [IQR]: 16.0% [9.2%, 22.5%], MOR=2.11). The SOK subtype with the most center-level variance was CIT>24 (acceptance range: 0-55.1%, median [IQR]: 2.7% [0.8%, 6.6%], MOR=3.15). For comparison, the MOR of non-SOK kidneys was 1.90.

Conclusions: There was substantial center-level variation in acceptances of SOK offers for each of the SOK subtype categories. Informing centers of donor phenotypes for which that center's acceptance rate is lower than the national average may motivate centers to accept more marginal kidneys, improving access to DDKT for patients at those centers.