On your marks. Get set. Evidence!

What an inspiration the recent Olympic Games and Paralympics has been for us all! Across the world, people from every nation have sat for hours, watching TV, and thinking about how physical activity is good. I myself walked a whole two feet across the lounge to change the channel manually rather than use the remote control. It took a lot of energy bars to recover that day.

We all know the health benefits of exercise in terms of increasing muscle strength and improving cardiovascular and respiratory fitness. We’re constantly hearing of the virtues of maintaining a healthy body weight and of being able to drone on mindlessly about how you did some moving at a gym. Increasingly there is talk of the benefits of exercise for mental and emotional health. However, talk and anaerobic respiration may just give you a stitch if you’re not careful. What you really need is properly collated evidence to demonstrate that an intervention might be doing something beneficial without causing harm.

Depression in young people

There have been a number of studies investigating the effect of physical activity and/or exercise on the symptoms of depression in adults, but not quite as many in children and adolescents.

• Depression has been estimated to affect between 2% and 8% of young people, with the peak incidence of depression generally thought to be around puberty
• Depression in this age group is typically limited in duration, except about 40% of affected children experience a recurrent attack and around 30% of affected children will self-harm.

If something as seemingly simple as exercise or physical activity could alleviate or remove the symptoms of depression, then much suffering could potentially be avoided. Except there are more than a few hurdles to get over first.

Previously, two systematic reviews reported that exercise had small to moderate beneficial effects in reducing the symptoms of depression. However, both of these reviews noted that most studies included small sample sizes and were of low methodological quality, limiting the value of any potential conclusions about the benefits of exercise for mental health in this age group. But research into exercise is bound to involve a few repetitions, so another systematic review was conducted involving much of the same evidence (Carter et al, 2016). In addition, an attempt to combine the existing evidence in a meta-analysis was completed in order to explore the potential effect of exercise interventions on depressive symptoms for adolescents (aged 13 to 17 years).

Weight training (methods)

• The usual suspect databases were searched for studies in adolescents looking at the impact of exercise or physical activity-based interventions on various mental health conditions. Exercise was defined as planned, structured, repetitive, and purposeful physical activity aiming to improve or maintain at least one component of physical fitness. Which is great, because you can argue that this includes darts.
• For a study to be included, its participants had to be aged between 13 and 17 years old and investigate the impact of an intervention promoting exercise or physical activity (of specified duration and time period) using a continuous, validated outcome measure of depressive symptoms. Basically, symptoms of depression had to be measured in a way that had been shown to work before and that resulted in some sort of score.
• Only randomised controlled trials were included.
• Studies with participants with physical health problems, intellectual disabilities or eating disorders were excluded.
• The primary outcome measure was depressive symptoms measured using a continuous validated scale.
• The Delphi list was used to determine the methodological quality of studies included. Well, most of the Delphi list. The item that requires “blinding” of the therapist was removed as it was deemed difficult for them to not know if they were getting people to exercise or not.

Cardio (results)

• Out of 543 potentially relevant studies, 11 studies (1,449 participants in total) met the inclusion criteria; 8 of which were suitable to be included in the overall meta-analysis. The rest took a running jump.
• The mean age of participants ranged from 14.7 to 17 years, i.e. adolescents “shut up, why do you want to know? I’m going to my room!”
• All of the trials used self-reported outcome measures of symptoms of depression.
• Based on the baseline depression scores, the severity of participants’ depression was interpreted as severe in 2 trials, moderate to severe in 1 trials, moderate in 2 trials, mild in 2 trials, below mild in 3 trials and unable to be classified in 1 trial. As such, the majority of trials included adolescents with below mild to moderate symptoms of depression.
• In all trials, exercise included some form of aerobic and/or resistance/strength training. The majority of trials did not provide a measure of exercise intensity. This obviously seemed like too much hard work.
• The duration of the exercise interventions studied ranged from 6 to 40 weeks, with the majority of interventions involving exercise 3 times per week. The analysis observed a moderate level of heterogeneity across the studies.
• 4 trials scores 5/8 on the Delphi list with 2 scoring 4/8 and below.
• In the 8 studies included in the meta-analysis, exercise showed an overall moderate (statistically significant) effect on depressive symptoms (Standardised mean difference [SMD] = -0.48; 95% confidence intervals [CI] = -0.87 to -0.10; P=0.01).
• If only trials with a Delphi score of 5 or higher were included, exercise was shown to have a moderate, non-statistically significant effect on depressive symptoms (SMD = -0.41, 95% CI = -0.86 to 0.05; P=0.08).
• Including only trials that investigated the effect of exercise in clinical populations led to a moderate statistically significant effect of exercise on depressive symptoms, while the effect in general population samples was not statistically significant.

Time trials (strengths and limitations)

• The authors noted that all except 3 of the included trials lacked an intention-to-treat analysis, which could be argued to be the preferred type of analysis if you’re trying to demonstrate the benefit of an intervention on a set of symptoms. I’d explain why mathematically, but I’ve got a note from my mum excusing me from doing so.
• The authors also noted that all of the trials included used self-reported outcome measures of depression with all of the associated problems. For example, participants may have been reporting improvement in their symptoms of depression because they knew the purpose of the study they were enrolled in (demand characteristics) or because they thought this response would be viewed favourably by others (social desirability bias/being a teenager).
• The population included wasn’t very clear. While the age group was well-defined, the clinical characteristics (or otherwise) were fairly broad, including adolescents with diagnosed mental health conditions and volunteers from the general population. It is unlikely that exercise will have the same effect across these different populations, or that symptoms of depression even represent the same phenomenon across these groups. Having one or two symptoms of depression would seem to be wildly different from having the full range and meeting the diagnostic criteria for a depressive disorder. Presumably, the general level of symptoms of depressions would be relatively low in the general population. This may be part of the reason why the effect of exercise only significantly improved symptoms of depression in the clinical population compared to the non-significant effect in the general population.
• It was also noted as a limitation that the frequency and intensity of exercise varied widely over the included studies (note the moderate heterogeneity), meaning no recommendations for the optimum frequency and intensity of exercise treatment can be made from this study. As such, even if exercise had been convincingly shown to benefit symptoms of depression, this study couldn’t describe how much exercise to do and how often.
• Problematically, it seems that the majority of studies included seemed to be of low methodological quality, with 4 or more trials scoring 5 or less out of 8) on the Delphi list. This was noted by previous systematic reviews exploring exercise as a treatment, with no particular reason for why the studies that already existed would suddenly improve now. This instantly throws any conclusion that can be made about the benefits of exercise for symptoms of depression into doubt. Even if a positive benefit was shown, we can’t be sure that it wasn’t because the studies were badly designed.

More promisingly perhaps was the fact that the studies in the clinical populations tended to score more highly for methodological quality, meaning we can be more confident of the statistically significant effect of exercise on symptoms of depression in adolescents with clinically recognised mood disorders. So possibly a point, but we might need to rely on Hawkeye. Not that one. The tennis one. Never that one.

In general, it should probably also be noted that just because someone describes an intervention as useful, it doesn’t mean that they are suggesting it’s easy. If you have severe symptoms of depression, completing even a minimal amount of physical activity is going to be extremely difficult. Regular exercise is likely to be more so. This may lead to problems with adherence in this particular intervention. The extent of this effect was not reported in the analysis. Ironically.

Manager’s interview (conclusions)

The authors concluded that the meta-analysis found a statistically significant treatment effect of exercise on depressive symptoms in adolescents. They noted the significant heterogeneity of studies, although argued that this may have primarily been due to 1 outlier study. Without this outlier, the magnitude of the effect of exercise was reduced to “small”.

The authors’ also concluded that exercise may be more beneficial for depressive symptoms in the clinical population than in the general population.

It was generally concluded (as it always is), that larger, high quality clinical trials in clinical practice with adequate descriptions of exercise interventions are needed for any firm conclusions to be drawn.

Match of the Day (summary)

This systematic review and meta-analysis attempted to explore the impact of exercise as an intervention for treating symptoms of depression in adolescents. The authors reported that exercise had a moderate, statistically significant impact, although the studies included were of generally low quality.

It was held that exercise had more benefit in clinical populations of adolescents with symptoms of depression compared with the general population, and that the studies in the clinical population were generally of higher quality.

Overall, as usual, it would seem that no real conclusion can be drawn about the benefits of exercise for the mental health of adolescents until a large, high quality study has been conducted. At least while they’re doing it they’ll get physically fitter I suppose. They think it’s all over. Well, it isn’t quite yet.

Links

Primary paper

Carter T, et al. (2016) The Effect of Exercise on Depressive Symptoms in Adolescents: A Systematic Review and Meta-Analysis. J Am Acad Child Adolesc Psychiatry 2016;55(7):580‒90. [PubMed abstract]

Other references

Hazell P. (2009) Depression in children and adolescents. BMJ Clin Evid 2009;2009:1008. [PubMed abstract]

Larun L, et al. (2006) Exercise in prevention and treatment of anxiety and depression among children and young people. Cochrane Database Syst Rev 2006;(3):CD004691. [PubMed abstract]

Brown HE, et al. (2013) Physical activity interventions and depression in children and adolescents : a systematic review and meta-analysis. Sports Med 2013;43(3):195‒206. [PubMed abstract]

Photo credits

• Kevin Dooley CC BY 2.0
• Quinn Dombrowski CC BY 2.0
• Jakob Davies CC BY 2.0
• Atom Ray CC BY 2.0