Increased Kidney Biopsy Scarring and Amyloid Deposition Predict Progression to Eskd in Patients with AL Amyloidosis

Background: AL amyloidosis frequently involves the kidney and causes significant morbidity and mortality. Many patients have a diagnostic kidney biopsy; however, there are no well-established techniques for deriving prognostic information from these biopsies. We hypothesized that the extent of amyloid and/or scarring in the kidney could be used to develop a pathologic scoring system to predict renal outcomes.

Methods: In a case control study, 39 consecutive patients treated for AL amyloidosis had kidney biopsies reviewed by two independent pathologists, including an expert renal pathologist at our facility. Kidney injury was assessed using novel scoring tools we developed for evaluating the severity of kidney AL amyloidosis. A Composite Scarring Injury Score (CSIS), a quantitative measure of scarring injury in glomeruli and tubulointerstitium, and an Amyloid Score (AS), a summation of assessments of extent of amyloid in the mesangium, glomerular capillary loops, interstitium, and vasculature, were generated. The scores were analyzed for correlation with baseline characteristics and the following outcomes: progression to end-stage kidney disease (ESKD), kidney response, kidney relapse in responders, and duration of kidney response.

Results: AS increased with the number of organs involved by AL amyloidosis (median AS: 5.0 for 1-2 organs involved, 7.25 for 3 organs, 9.75 for 4 organs, 10.75 for 5 organs, P=0.039). AS also correlated with initial troponin I (Spearman rho 0.432, P=0.024). There were no statistically significant differences in CSIS and AS between patients treated with autologous hematopoietic stem cell transplantation (n=23) or chemotherapy alone (n=16), those achieving (n=19) and not achieving (n=20) a hematological response, and in patients with (n=19) and without (n=20) cardiac involvement. CSIS and AS did not have a statistically significant correlation with baseline proteinuria, serum albumin, or brain natriuretic peptide.

Both CSIS (Spearman rho -0.359, P=0.025) and AS (Spearman rho -0.43, P=0.005) correlated with initial estimated glomerular filtration rate. Compared with patients who did not progress to ESKD (n=29), patients who progressed to ESKD (n=10) had significantly higher median CSIS (5.75 v. 3.30, P=0.048) and AS (9.75 v. 5.5, P=0.01). Receiver operator characteristic (ROC) curves associating progression to ESKD demonstrated that area under the curve was 71% for CSIS and 77% for AS, suggesting that CSIS and AS are useful indicators to predict progression to ESKD. CSIS and AS did not correlate with rates of kidney response (P=0.135 and 0.096), kidney relapse in responders (P=0.069 and 0.829), or duration of kidney response (P=0.06 and 0.06).

Conclusions: CSIS and AS correlated with progression to ESKD, demonstrating that a kidney biopsy, in addition to providing diagnostic information, can form the basis for a pathologic scoring system with prognostic significance in renal AL amyloidosis. A larger sample size is necessary to validate the predictive value of CSIS and AS, which should be considered in upcoming AL amyloidosis trials.