Soluble CD30 correlates with clinical but not subclinical renal allograft rejection

Soluble CD30 (sCD30) has been proposed as a promising noninvasive biomarker for clinical renal allograft rejection, but its diagnostic characteristics regarding detection of subclinical rejection have not been assessed. We investigated sCD30 in 146 consecutive kidney allograft recipients under tacrolimus–mycophenolate‐based immunosuppression having 250 surveillance biopsies at 3 and 6 months as well as 52 indication biopsies within the first year post‐transplant. Allograft histology results were classified as (i) acute Banff score zero or interstitial infiltrates only, (ii) tubulitis t1, (iii) tubulitis t2‐3 and (iv) isolated vascular compartment inflammation. sCD30 correlated well with the extent of clinical (P < 0.0001), but not subclinical tubulointerstitial rejection (P = 0.06). To determine diagnostic characteristics of sCD30, histological groups were assigned to two categories: no relevant inflammation (i.e. acute Banff score zero and interstitial infiltrates only) versus all other pathologies (tubulitis t1‐3 and isolated vascular compartment inflammation). For clinical allograft inflammation, AUC was 0.87 (sensitivity 89%, specificity 79%; P = 0.0006); however, for subclinical inflammation, AUC was only 0.59 (sensitivity 50%, specificity 69%; P = 0.47). In conclusion, sCD30 correlated with clinical, but not subclinical renal allograft rejection limiting its clinical utility as a noninvasive rejection screening biomarker in patients with stable allograft function receiving tacrolimus–mycophenolate‐based immunosuppression.