Objective: To investigate the 10-year biopsy-proven recurrence rates and risk factors for immunoglobulin A nephropathy recurrence in kidney transplant recipients. Methods: We included 299 kidney transplant recipients from 1995 to 2015, who had biopsy-proven underlying immunoglobulin A nephropathy and underwent zero-hour biopsy. The primary end-point was recurrence of immunoglobulin A nephropathy. We compared clinical, treatment and graft failure among those with and without recurrent immunoglobulin A nephropathy. A time-to-recurrence analysis was carried out using the competing risk analysis and time-dependent Cox model. Results: Of 299 recipients, 80 had recurrent immunoglobulin A nephropathy (66.3% with clinical biopsy and 33.7% with protocol biopsy, post-transplant biopsy rate: 90.6%). The 10-year recurrence rate was 34.3% (95% confidence interval 27.6-41.1). Relateddonor transplantation (hazard ratio 2.28, P = 0.009) and post-transplant increased proteinuria (hazard ratio 1.59, P < 0.001) were identified as potential risk factors for immunoglobulin A nephropathy recurrence. The 10-year rates were 41.5% in related donor recipients and 16.3% in unrelated donor recipients. There was no conclusive evidence that the calcineurin inhibitor, antimetabolites, basiliximab and rituximab reduce immunoglobulin A nephropathy recurrence. Immunoglobulin A nephropathy recurrence was associated with an increased risk of death-censored graft failure (hazard ratio 5.29, 95% confidence interval 1.39-20.17, P = 0.015). However, related donor itself was not associated with an increased risk of graft failure. Conclusions: The present results have clinical implications in that the signs of recurrent immunoglobulin A nephropathy should be evaluated carefully in recipients receiving related-donor transplants. There is a need for further studies related to genetic and/or familial interactions in kidney transplant recipients with immunoglobulin A nephropathy and related donors. †Data from repeated determinations were analyzed by time-dependent Cox model. ‡Median [interquartile range].