Relapsed Acute Lymphoblastic Leukemia (ALL) is among the most common causes of cancer-associated deaths in children. However, little is known about the implications of deviations from ALL treatment protocols on survival rates. The present study elucidates the various characteristics of treatment deviations in children with relapsed ALL included in the ALL-REZ BFM 2002 (i.e., Relapse Berlin-Frankfurt- Münster) trial and determines their prognostic relevance for relapse and death rates. Among 687 patients, 100 were identified with treatment deviations, further classified, and examined by occurrence time, cause and type. Protocol deviation was considered a time-dependent variable and its impact on Disease Free Survival (DFS) and Overall Survival (OS) was examined using the time-dependent model Mantel Byar. Five years after the relapse diagnosis, deviations were significantly related to both inferior DFS (38%) and OS (57%) rates compared to protocol conformed treatment (DFS = 61%; OS = 70%, P < 0.001). Based on multivariate analyses, protocol deviation proved to be an independent adverse prognostic factor of DFS. Moreover, deviations triggered by chemotherapy-induced toxicity were associated with a higher relapse rate compared to deviations due to insufficient response. Therefore, to avoid impairment of results by deviations, future clinical trials, and treatment strategies should focus on less toxic treatments and stricter protocol compliance.