Torture remains a severe human rights violation and a relevant medico-legal issue in asylum procedures. The Istanbul Protocol provides a standardized framework for documenting alleged torture and assessing the consistency between reported events and clinical findings. The objective of the study is to describe the demographic characteristics, reported contexts, injury patterns, availability of contemporaneous documentation, and Istanbul Protocol consistency ratings among asylum seekers referred for medico-legal evaluation at a single Italian center during 2009–2014 (a period overlapping with major geopolitical instability in several regions). We retrospectively reviewed 36 consecutive cases evaluated at the Unit of Legal Medicine of University of Bari, Southern Italy, between 2009 and 2014. Examinations followed the Istanbul Protocol and were conducted by a multidisciplinary team with cultural mediation. Data included demographics, country of origin, reported reasons for persecution, migration route, alleged torture methods categorized per Istanbul Protocol, injury distribution, time elapsed since the alleged events, presence of contemporaneous medical records, and consistency ratings. The cohort included 36 asylum seekers (32 men, 4 women; mean age 26 years), predominantly aged 21–30 years (58%). Countries of origin were heterogeneous (Pakistan 25%, Palestine 14%, Nigeria 11%, Turkey/Kurdistan 11%, other 39%). Most participants reported that the alleged torture occurred in the country of origin (92%). Detention in Libya during migration was reported by 31%. The most frequently documented categories were blunt trauma (69%), crushing injuries (47%), and psychological techniques (47%); burns were reported in 27%. Contemporaneous medical documentation was available in 17% of cases. No acute injuries were observed; in 58% the alleged events occurred 1–5 years prior to assessment and in 28% more than 5 years prior. Using the Istanbul Protocol consistency scale collapsed into three categories, full consistency was recorded in 22%, partial consistency in 75%, and inconsistency in 3%. In this retrospective single-center cohort, alleged torture reports were associated with diverse geopolitical contexts and were frequently assessed years after the events, with limited contemporaneous documentation. Despite these constraints, Istanbul Protocol–based evaluations yielded predominantly partial-to-full consistency ratings. These findings primarily inform the practical challenges of delayed medico-legal assessment in asylum seekers and support the need for earlier referral pathways and multicenter datasets; causal inferences regarding temporal or regional geopolitical events cannot be drawn from this design.