Histology and Not Donor Specific Antibody Helps Predict 5-Year Kidney Transplant Failure Using 1 Year Post-Transplant Data

M. Moreno Gonzales,¹ A. Bentall,^2,3 M. Stegall,¹ R. Borrows.²

¹Department of Transplant Surgery, Mayo Clinic, Rochester, MN
²Department of Nephrology, Queen Elizabeth Hospital Birmingham, Birmingham, United Kingdom
³School of Immunity and Infection, University of Birmingham, Birmingham, United Kingdom.

Meeting: 2015 American Transplant Congress

Abstract number: 34

Keywords: Graft failure, Kidney transplantation

Introduction: The objective of this study was to validate the RISK score (Am J Kid D 2014; 63[4]:643) for renal allograft failure using a different cohort and to assess if protocol biopsy data and anti-donor HLA antibody improved the model's ability to predict 5-year graft failure using 12 month data.

Methods: The validation cohort included solitary kidney transplant (Tx) recipients at Mayo Clinic, Rochester, MN between January 1999 and December 2008. Negative crossmatch Tx recipients (n=2159) were assessed for death-censored failure 5 years post-Tx. Risk scores were evaluated for prognostic utility (discrimination, calibration, and risk reclassification). Weighted regression coefficients for baseline, 12-month demographic and clinical predictor characteristics were used.

Results: Multivariate analysis identified glomerulitis and interstitial fibrosis and Class II DSA as associated with graft failure.

Table 1 – Multivariate Analysis

Group Analysis	Number	Variable	Hazard Ratio	p-value
Histology Alone Group	981	g	2.53 (1.87, 3.42)	<0.001
		ci	1.66 (1.35, 2.06)	<0.001
DSA Data Alone	622	Class II MFI<800	1.17 (0.34, 3.97)	0.001
		Class II MFI>800	4.34 (1.98, 9.52)	0.001
Combined Histology and DSA	556	g	2.74 (1.77, 4.24)	<0.001
		ci	1.90 (1.27, 2.85)	<0.001
		Class II MFI<800	0.73 (0.16, 3.39)	0.002
		Class II MFI>800	4.57 (1.89, 11.1)	0.003

Histology (g and ci scores) improved discrimination for death censored Tx failure (C statistic= 0.90 vs 0.84 without histology). Class II DSA did not improve predictability (C-statistic=0.83). Both scores demonstrated good calibration (Hosmer-Lemeshow P< 0.05 in with both histology and DSA variables). Adding histology to the RISK score resulted in statistically significant risk reclassification for death-censored Tx failure (net reclassification improvement [NRI], 29.0%, p<0.001 vs. NRI, 1.2%, p=0.90). Class II DSA had no effect on reclassification.

Conclusions: The RISK score was validated in a large cohort accurately identifying at-risk transplants and the addition of 1 year protocol biopsy data improves the assessment of risk, improving the classification in 30% of patients into high or low risk populations.

« Back to 2015 American Transplant Congress