Reviewed research

Authors Duke, SA, Balzer BWR, Steinbeck, KS

Review Date September 2014

Citation Journal of Adolescent Health 2014; 55: 315-322


A recent Australian longitudinal study followed 158 males and 119 females from ~11-12 years of age (Tanner stage 3) for 3 years, with regular measurement of urine testosterone and estradiol, and measures of mood, behaviour and sleep.

The only associations with naturally increasing testosterone levels in boys were a small increase in rule-breaking and a reduction in sleep time. Measures of social problems, thought problems, attention problems, aggressive behaviours, self-harm, suicidal ideation and impulsivity/aggression were unrelated to testosterone or estradiol levels in boys or girls.

The study concludes that “[p]uberty is only one part of the complex developmental stage of adolescence that includes psychosocial transitions and neurocognitive maturation. If mood and behavioural changes are attributed solely to hormones in the second decade many young people will be deprived of health care interventions that would go far towards improving their health and wellbeing.”

This Australian study provides a response to the summary of the article reviewed below, “that there are insufficient longitudinal data of high methodological quality to currently confirm that the changing testosterone levels during puberty significantly affect male adolescent behaviors and mood”: now there are sufficient data, and they do not show significant affects on male adolescent behaviours and mood.


During puberty there is an approximate 30-fold increase in testosterone production in males. Many assumptions are often made around the nonphysical effects of this large increase, most commonly that it causes changes in mood and behaviour (e.g., increase in aggressive behaviour and risk taking). It is important to ensure that such changes are not falsely attributed to puberty hormones and the true diagnoses ignored.



To determine the available evidence for the effects of endogenous testosterone on behaviour and mood in young adolescent men.



Seven databases (Medline, Pre-Medline, Education Resources Information Centre, PsycINFO, Embase, Scopus and Web of Science) were searched. Selection criteria included the use of human participants, a community sample comprised of healthy male adolescents aged 9-18 years and not taking exogenous testosterone, and a validated mood and/or behaviour assessment with a timed testosterone measurement. Outcomes included those behaviour and affect disorders thought to have a clear increased incidence and prevalence in the second decade of life and as a result of puberty hormones: aggressive/disruptive/conduct disordered behaviour, substance abuse, social competency and social interactions, depression and anxiety, and self-image concerns. Due to the heterogeneity of the measures and outcomes across papers, data are presented as descriptive only.



Twenty-seven studies met inclusion criteria. The studies’ findings were grouped together for clarity based on common themes:

Aggressive and/or disruptive behaviour: One longitudinal study found no evidence of a systematic increase in aggression during the three years of the study, while most of the cross-sectional studies reported ambiguous and conflicting findings on the relationship between aggression and testosterone levels; significant associations, when described, were often only for a subset of measures.

Substance use, self-image, social interaction and competency, depression, and anxiety: There were very few studies looking at behaviours, mood, or affect beside aggressive and/or disruptive behaviour. One study found a positive association between testosterone and drug use; however, this increased when family, peer, neighbour, and school contexts became more harmful suggesting testosterone does not act independently of sociocultural context. No studies found an association between testosterone and self-image concerns or competence. One study demonstrated an inverse relationship between testosterone level and anxiety and depression; however, an additional five studies found no such relationship.



The authors conclude that there is insufficient evidence to support the assumptions often made by clinicians and the general population around endogenous testosterone and nonphysical effects such as aggressive behaviour and risk taking. These assumptions need to be challenged due to the lack of evidence and the likelihood that the aetiology for such behaviours are likely be multifactorial with strong psychosocial interactions. Unlike normal pubertal testosterone rise, which is not modifiable, these predictors can be targeted for intervention.


Points to Note
  1. It is often assumed that testosterone changes in puberty are associated with nonphysical changes such as increased aggressive behaviour and risk taking.
  2. This is the first systematic review to examine the effect of endogenous testosterone on behaviour, mood, or affect in male adolescents.
  3. All but one study was cross-sectional and causality cannot be inferred. Further, the validity of the testosterone assays used was often poor, particularly in older studies.
  4. None of the included studies assessed neurocognitive change and its association with testosterone, a topic that is currently under investigation by many research groups.
  5. It is possible that the speed of hormone change may be more important than the absolute hormone levels. Future research should consider if and how the onset of testosterone rise and the rate of testosterone change may influence behaviour.
  6. Assumptions around testosterone during puberty and nonphysical changes need to be challenged; there is no evidence to support them and it is likely that such changes evolve due to many factors including the psychosocial context of the individual.


Website: http://www.ncbi.nlm.nih.gov/pubmed/25151053

Reviewed by
Updated on