Objective: Self-injury (SI), self-injurious behaviour (SIB), including suicidal or non-suicidal self-injury (SSI, NSSI) represent an increasing problem among teenagers amounting to a 6–17% prevalence rate in adolescence, yet little data exists on detailed characteristics and associated factors of SI reaching clinical severity. There is also a scarcity of data distinguishing between suicidal and non-suicidal subsamples of self-injuring patients, i.e. showing which predictors contribute to develop self-injurious behaviour with a previous suicidal history (SSI). Method: Clinical diagnoses and characteristics of risk behaviour were examined in a crosssectional design in suicidal and non-suicidal subgroups of Hungarian adolescent outpatients practising self-injurious behaviour. From the total pool of 708 new patients consecutively referred with various psychiatric problems in five regional child psychiatric centres in Western-Hungary over an 18-month period, 105 adolescent outpatients suffering from self-injurious behaviour participated in the study (28 males and 77 females aged from 14 to 18 years, mean age 15.97, SD 1.05). The Ottawa/Queen’s self-injury questionnaire (OSI) was used to measure the characteristics of risk behaviour, while the comorbid clinical diagnoses were confirmed by the M.I.N.I. Plus International Neuropsychiatric Interview. Descriptive statistics presented the frequencies of the characteristics of SI, bilateral comparisons were used to reveal relevant items to differentiate between sex, duration of practice and SSI versus NSSI and logistic regression was performed to identify significant predictors of suicidal subtype of self-injuring practice. Results: A total of 60% of the clinical SI population experienced a present or past episode of major depression. The motivation of patients to resist impulses and to discontinue malpractice was low. Cutting and scratching was the most common self-injuring methods. Two-thirds of the sample practised the impulsive type of SI, while 30% practised premeditated SI having an incubation time from 30 min to days and weeks before carrying out SI. Although duration of SI did not distinguish the sample in important aspects, girls and boys differed in several aspects of SI practice. SSI adolescents differed from their NSSI peers in a number of important characteristics including the frequency of actions, injured areas, methods, specific stresses and motivations. SSI adolescents were more likely to favour cutting of the lower leg and drug overdose as modes of SI. SSI adolescents were more likely to report addictive features than their peers with no suicidal motivation. From the aspect of self-injurious practice, logistic regression analysis found only two significant predictors for the combined pathology