Open Access is an initiative that aims to make scientific research freely available to all. To date our community has made over 100 million downloads. It’s based on principles of collaboration, unobstructed discovery, and, most importantly, scientific progression. As PhD students, we found it difficult to access the research we needed, so we decided to create a new Open Access publisher that levels the playing field for scientists across the world. How? By making research easy to access, and puts the academic needs of the researchers before the business interests of publishers.
We are a community of more than 103,000 authors and editors from 3,291 institutions spanning 160 countries, including Nobel Prize winners and some of the world’s most-cited researchers. Publishing on IntechOpen allows authors to earn citations and find new collaborators, meaning more people see your work not only from your own field of study, but from other related fields too.
To purchase hard copies of this book, please contact the representative in India:
CBS Publishers & Distributors Pvt. Ltd.
www.cbspd.com
|
customercare@cbspd.com
Physical activity is believed to promote mental health. However, research has not yet reached a consensus on whether physical activity declines panic and anxiety symptoms in children, adolescents, and early adulthoods. The current chapter carried out a meta-analysis to investigate the association between physical activity and panic/anxiety based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. Search is conducted on 22nd April 2022, which follow databases: MEDLINE (Ovid), EMBASE, Web of Science, Cochrane Central Register of Controlled Trials, and SPORTDiscus. Fifteen articles (N = 994) were identified and included, where four studies reported measurement in panic symptoms and fourteen studies reported measurement in anxiety symptoms. The meta-analysis among the pooled effect sizes demonstrated a small significant effect of physical activity intervention reducing panic disorder (d = −. 45, SE = .12, Z = −3.65, p < .001) and a middle effect reducing anxiety (d = −.51, SE = .15, Z = −3.38, p < .001) in children, adolescents and early adulthoods. Age or gender ratio was not found to be significant in predicting the effect sizes. More evidence is required to produce a solid conclusion.
Sports Department, Guilin University of Electronic Technology, China
Sport and Health Science Department, University of Exeter, UK
Yihao Liu
Psychology Department, University of Exeter, UK
*Address all correspondence to: lw679@exeter.ac.uk
1. Introduction
Physical activity (PA) is one of the most accessible interventions for anxiety disorders for the public, whereas few systematic research has been done to examine its efficacy. PA is defined as a movement that causes an increase in energy expenditure in the movement of people [1], including walking, running, doing housework, jogging, or other sports that are defined as PA behaviour [2, 3]. Dollman et al. (2015) have classified PA intensity with metabolic equivalent (MET) as light PA (MET: 0–2.99), moderate-to-vigorous PA (MET:3–5.99), and vigorous PA (MET: ≥6) through a scale of energy expenditure [4, 5, 6]. PA intervention aims to promote the health of people through exercise training, sports and habit of PA [7, 8]. A PA intervention study should specify the processing target (e.g., children, adolescents, or some clinic population), PA type (e.g., aerobic exercise, resistance exercise, or yoga), PA intensity (e.g., light, moderator, moderator-to-vigorous), PA frequency (e.g., three sessions per week, 60 minutes per session), PA duration (e.g., three months, 1 years), and outcome measurement [9]. The PA intervention research is easier to quantify and more accessible to the generalised population. It is also found to benefit mental health. Therefore, research is also focused on the effect of PA intervention on panic disorder.
For example, Ensari, Petruzzello [10] conducted an RCT experiment to deliver a 40-minute yoga programme for eighteen participants with high-anxious. The result suggested a significant main effect of the task on panic and respiratory measures (p < .05). When collapsed over inhibition task and condition, there was a small reduction in cognitive anxiety from baseline to immediately post and 1-h post-condition (p < .05) [10]. Similarly, Naderi, Naderi [11] investigated the effect of physical exercise on anxiety among victims of child abuse and reported a significant reduction in anxiety (p < .001). Accordingly, the author argued that such improvement is comparable to empirically supported treatments for panic and GAD. However, such effects of PA intervention on panic disorder used a variety of PA types, intensity levels, or frequency and, therefore, produced isolated effects. The effect of PA could vary regarding individual differences, such as age and gender [12]. A most recent systematic review in 2022 tended to examine the effect of regular exercise interventions on the panic disorder in adults. They only retrieved eight studies in this field and demonstrated no clear evidence suggesting whether regular exercise programs (at least two 20-minute sessions per week for at least six weeks) reduce panic-related symptoms. The study argued for more RCT studies to support more robust and clear evidence for better understanding [11].
Moreover, there were even fewer studies focused on the effect of PA intervention on the panic disorder or general anxiety disorder (GAD) in children and adolescents, respectively. This may be because it is difficult to categorise different anxiety disorders among children. Furthermore, there is a lack of tools to measure child panic in laboratory settings. For example, there is the anxiety scale for children with autism spectrum disorder, revised children’s manifest anxiety scale and social anxiety scale for children to measure anxiety disorder in children and adolescents [9, 13, 14], whereas the panic disorder is measured based on the sensitivity of anxiety scales in children and adolescents [15]. These unclear concepts may increase the research difficulty, which supports the evidence for PA intervention on panic disorder in children and adolescents. Therefore, this chapter aims to carry out a quantitative review to clarify and explore the evidence in this area. Moreover, the secondary objective is to investigate the effect of PA intervention on anxiety disorder in children and adolescents.
The previous Machado, Telles [11] review obtained a similar outcome measurement to examine the effect of PA on anxiety disorder. Their target sample did not limit the population age, which means there is no evidence that physical activities affect panic and anxiety disorder in children and adolescents. Consequently, the current analysis aims 1) to summarise and explain the potential impact of PA interventions on panic symptoms and critically comment on the research that exists and 2) to analyse the effect of PA intervention on panic and anxiety symptoms in children, adolescents and early adulthood, through meta-analysis, which provides reference evidence for further research.
2.1 Materials and methods protocol and registration
The study is based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. A protocol for this review was registered with PROSPERO (CRD42022334054).
Figure 1.
The forest plot of the effect size of physical activity on panic in children, adolescents and early adulthood.
2.2 Search strategy and databases
Search is conducted on 22nd April 2022, which follow databases: MEDLINE (Ovid), EMBASE, Web of Science, Cochrane Central Register of Controlled Trials, and SPORTDiscus. Search terms based on the PICO format (participants/population, intervention/exposure, comparison, outcome) were divided, and adjusted according to the respective databases’ Thesaurus and Medical Subject Headings (MeSH) terms (Bramer et al., 2018) through Ovid. Articles must be available in English, there will be no restriction on the publication period. The full list of search terms is provided in Supplementary Material 1.
2.3 Eligibility criteria
The inclusion criteria of the current analysis were: 1) Population: studies included participants who primarily exhibited panic or anxiety symptoms or were diagnosed with panic or anxiety disorders in children, adolescents (5–19 years) and early adulthoods (19–22 years). Participants may present secondary comorbid other illnesses, such as diabetes; 2) Intervention: studies applied regular PA intervention (i.e. walk, jog, aerobic, strength, or multimodal training), which is prescribed to reduce panic disorder or GAD, social anxiety disorder, obsessive–compulsive disorder, post-traumatic stress disorder, and agoraphobia. PA interventions can be combined with other treatment procedures were also included; 3) Comparators: studies included a control group as a comparator, which is not received the PA intervention delivery; 4) Outcomes: studies that took panic symptoms or panic disorder as the primary outcome. And the second outcome was anxiety symptoms, GAD, social anxiety disorder, obsessive–compulsive disorder, post-traumatic stress disorder, or agoraphobia; 5) Study Design: studies carried out in randomised trials (RCT) comparing an intervention(s) encompassing PA with a group(s) without PA intervention or encompassing PA at the baseline with post PA intervention.
The exclusion criteria were: 1) studies reported in non-English language; 2) studies reported insufficient information to estimate effect size or other essential data.
2.4 Article selection and data extraction
Study selection: Data will be formatted in RIS format and will be managed using the endnote software. PRISMA 2020 guidelines will be applied for reporting the screening process [16]. Two researchers will undertake the removal process through an independent screen. Firstly, all duplicate articles will be removed before reviewing titles and abstracts. Those not fitting the inclusion criteria will be removed. Then, full-text versions will be collected when the articles meet the screening criteria. Discussions with a third reviewer will resolve discrepancies between the two independent reviewers.
Data extraction: Two independent reviewers will extract the following four data dimensions. Including fundamental characteristics (e.g., author, public year, country, population, sample, age, sex, weight status), intervention characteristics (e.g., PA intensity/frequency/duration, intervention program), methodology (e.g., data analysis method), and effect size.
2.5 Risk-of-bias (quality) assessment
Two reviewers will independently score the studies according to the National Institutes of Health study quality assessment tool for Quality Assessment of Controlled Intervention Studies from the National Heart, Lung, and Blood Institute (https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools). The quality assessment focused on the specification of eligibility criteria, generalisability, intervention description, outcome assessment and incomplete data [17]. There were 14 items assessed in the quality assessment, as shown in Appendix S1. The tool does not assign numeric values or definite judgements of the quality of the studies, although it has good, fair, and poor results. Good quality studies have less bias risk and are more valid. A fair study is prone to some bias but insufficient to invalidate its findings. A poor study has a high-risk of bias and is considered invalid [18]. The results of quality assessment are presented in Table 1.
Study
1
2
3
4
5
6
7
8
9
10
11
12
13
14
overall
Kenis-Coskun et al., 2022
Y
Y
Y
NR
NR
Y
Y
Y
Y
Y
Y
Y
Y
Y
Good
Tanksale et al., 2021
Y
Y
N
N
N
Y
Y
Y
N
Y
Y
NR
NA
Y
Fair
Nazari et al., 2020
Y
Y
N
N
N
Y
Y
Y
NR
Y
Y
Y
Y
Y
Fair
Romero-Pérez et al., 2020
Y
Y
N
N
N
Y
Y
Y
Y
Y
Y
NR
Y
Y
Fair
Yu et al., 2020
Y
Y
N
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Fair
Mucke et al., 2020
Y
Y
N
N
N
Y
Y
Y
Y
Y
Y
Y
NA
Y
Good
Akko et al., 2020
Y
Y
N
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
Good
Luna et al., 2019
Y
Y
N
N
N
Y
Y
Y
Y
Y
Y
NR
Y
Y
Fair
Naderi et al., 2019
Y
Y
N
N
N
Y
Y
Y
Y
Y
Y
NR
Y
Y
Fair
Ensari et al., 2019
Y
N
N
NR
NR
Y
Y
Y
Y
Y
Y
NR
Y
Y
Fair
Polis et al., 2017
Y
Y
N
N
N
Y
Y
Y
NR
Y
Y
NR
NA
Y
Fair
Smits et al., 2009
Y
N
NR
NR
NR
Y
Y
Y
Y
Y
Y
NR
Y
Y
Fair
Broman-Fulks & Storey, 2008
Y
N
NR
NR
NR
Y
Y
Y
Y
Y
Y
NR
Y
Y
Fair
Lindwall et al., 2005
Y
Y
N
N
N
Y
N
N
N
NA
Y
NR
Y
Y
Poor
Crew et al., 2004
Y
Y
N
N
N
Y
Y
Y
Y
N
Y
NR
NA
Y
Fair
Table 1.
Summary of quality of included studies.
2.6 Effect size estimation
The studies we aimed to include in the current analysis were RCTs with both between-subject, within-subject and mixed designs. Consequently, a combined effect size of standard mean difference (Cohen’s d) was calculated for each study to produce a pooled effect size based on the improved method provided in Morris [19]. Firstly, the effect size of studies reported only median and interquartile range was estimated based on the improved formula [20, 21], where d refers to Cohen’s d; q1 refers to the first quartile; q3 refers to the third quartile:
d=q3−q11.35
Moreover, studies reported pre-calculated Cohen’s d without providing the mean or standard deviation was estimated standard error using the 95% confidence interval for meta-analysis weighting [22], where SE refers to standard error, CLup and CLlow refer to the upper and the lower bound of the confidence interval, respectively:
SE=CLup−CLlow3.92
The effect sizes were from between-subject design, within-subject design and mixed design studies were extracted, converted and matched to Cohen’s d with an online converter tool ‘Psychometrica’ (https://www.psychometrica.de/effect_size.html), which follows the method proposed in Morris [19] and Lakens [23].
2.7 Statistical data analysis
The statistical analysis was carried out in STATA v.17. The estimated effect sizes of adolescent panic and anxiety were pooled in a quantitative meta-analysis in a random effect model and Cohen’s d to determine the overall effect among the studies. A significant result of this analysis indicates that there was a significant effect of physical activities/exercise on adolescent panic or anxiety across the studies. The effect size is considered small when SMD is between 0.2 and 0.5, medium when it is between 0.5 and 0.8, and large when it is above 0.8 [24].
Then, a heterogeneity test of the available data was carried out with the random effect model. A non-significant result in the heterogeneity test would mean that the effect sizes were homogenous and measured a consistent effect on the same side. The heterogeneity I2 is considered moderate when I2 > 50%, and it is considered high when I2 > 75% [21]. Egger’s regression-based tests were then used to assess the publication bias. A significant result in this test would suggest potential publication bias in the current analysis.
Finally, exploratory meta-regression analyses were performed on the available characteristic data, including age, gender and intervention duration, to determine potential predictors of the current effect. The research method would be entered into subgroup analysis to determine whether the study design made a difference in the overall effect.
As shown in Appendix 2, keywords searching in MEDLINE (Ovid), EMBASE, Web of Science, Cochrane Central Register of Controlled Trials, and SportDiscus resulted in 1691 articles. After reviewing, based on inclusion criteria, twelve studies were identified, and four studies were excluded based on exclusion criteria, leaving eight studies for the meta-analysis. The included 11 studies were all RCTs in different designs (within-subject, between-subject and mixed design).
3.2 Characteristics of included literature
As shown in Table 2, 15 RCT studies with sufficient data were included. Among these studies, 14 obtained anxiety measurements, including social anxiety disorder [14, 25, 26] and GAD [10, 13, 27, 28, 29, 30, 31, 32], and 4 studies obtained panic measurements [10, 31, 32, 33]. There were 994 participants included in total, and 478 and 354 participants were allocated in the intervention and control groups, respectively. The mean age of the participants was 13.21 ± 4.16 years old, ranging from 8 to 24 years old. Physical activity intervention duration was from a minimum of 2 weeks to 8 months, and frequency was 2 or 3 sessions per week, with two studies performing a one-session intervention [10, 34]. Two studies also provided discussion and reflection sessions after the physical activity sessions [34, 35] and two studies were conducted on the school campus [14, 28].
The lead researcher delivered the Yoga program face-to-face to all participants on the university campus within a group of less than 5 parent–child dyads.
Participants were randomly allocated to a waiting list without receiving intervention
1 s/w (6 weeks)
ASC-ASD
Between-subject
NA
Nazari et al., 2020
40
20 (20)
20 (20)
11.11 (2.29)
NR
20 mins of Pilates exercises & 20 mins of body weight-bearing exercise & 20 mins aerobic exercises (V-forward, V-back & march)
NR
3 s/w (16 weeks)
RCMAS
Mixed
Heart rate
Romero-Pérez et al., 2020
105
54 (54)
51 (51)
9.88 (0.83)
60
5-min warm-up, a 40-min aerobic exercise
The CG participants continued with their usual activities at the end of the classes
2 s/w (20 weeks)
RCMAS
Mixed
NA
Yu et al., 2020
188
106 (99)
82 (72)
9.8 (0.7)
35
20-min class recess in the morning and one extra gym class (40 min) after school in the afternoon, including Jogging 20 min in the morning break every weekday; Rope skipping 40 min on Monday and Thursday; Playing badminton 40 min on Wednesday and Friday; 200-m relay race 40 min on Tuesday.
Children in the control school followed their usual practice with no extra intervention
(weekday daily) 5 s/w (8 months)
SAQ-C24
Mixed
NA
Mucke et al., 2020
60
30 (NA)
30 (NA)
17.9 (1.24)
0
The exercise group performed an exercise session at moderate intensity on a bicycle ergometer (30 mins)
During the next 30 mins, the control group read an article from a magazine of their choice
One session
STAI
Between-subject
Heart Rate
Akko et al., 2020
44
23 (23)
21 (21)
9.35 (0.6)
39
45 min after school running and running-based games of moderate-intensity
preadolescent children participated in assisted homework sessions to prevent attention bias and to control for retest effects
3 s/w (10 weeks)
STAI
Mixed
NA
Luna et al., 2019
113
69 (69)
44 (44)
13.82 (0.79)
49
55-minute sessions volleyball sessions, including warming-up, training and friendly matches and regular stage competition; Meetings for comprehension and reflection with the intervention of the responsibility roles
Traditional collective sport (basketball) with a conventional teaching style in which the teachers directed all tasks without students’ participation
2–3 s/w (6 weeks)
SAS-A
Mixed
NA
Naderi et al., 2019
22
11 (11)
11 (11)
8 to 11
22
Aerobic dancing: Warming up (10–15 min); Basic movements (35–40 min); Cooling down (10 min)
Not receiving any intervention
3 s/w(8 weeks)
STAI
Mixed
NA
Ensari et al., 2019
18
9 (9)
9 (9)
22.1 (5.0)
18
40 min Yoga session, which was designed based on published recommendations under the guidance of an instructor with the specific order of poses
Stretching exercises
One session
API & SAI
Mixed
Heart Rate
Polis et al., 2017
23
12 (5)
11 (8)
11 to 18
NR
Evening yoga session led by yoga instructor
NR
2 s/w (6 weeks)
STAI
Mixed
NA
Smits et al., 2009
92
48 (48)
44 (44)
19.43 (1.31)
51
20 min treadmill exercise on a computer-controlled treadmill, maintaining 70% of max heart rate
20-min resting period
One session
API
Mixed
Heart Rate
Broman-Fulks & Storey, 2008
24
12 (12)
12 (12)
19.04 (1.90)
19
20 min aerobic exercise: treadmill running
Report to the lab at the same time but no exercise
3 s/w (2 weeks)
ASI-R
Mixed
Heart Rate
Lindwall et al., 2005
110
56 (27)
54 (35)
16.35 (1.56)
110
Preferred multiple aerobics exercises (45 min) and discussion (15 min), including water aerobics, step-up, badminton, kickboxing, spinning dancing, climbing, bowling, karate, jujitsu, yoga and different ballgames
Control group were put on a waiting list with no forms of alternative activities organised
2 s/w (24 weeks)
SPAS
Mixed
NA
Crew et al., 2004
66
66 (66)
NA
NR (Grade 4)
33
The aerobic group exercised by means of stationary cycling, track running, and jumping on a minitrampoline. The physical activity group engaged in a variety of physical activities such as shooting baskets for skill improvement, playing a common children’s game called foursquare, and walking
NA
3 s/w (6 weeks)
STAI
Within-subject
Heart Rate
Overall
994
561 (488)
433 (371)
13.21 (4.16)
478 (51.3%)
Table 2.
Characteristics of included studies.
IG = intervention group; CG = control group; Pre-I: pre-intervention; Post-I: post-intervention; s/w = sessions per week; NR = not reported; NA = not applicable; ASC-ASD: Anxiety Scale for Children–Autism Spectrum Disorder RCMAS = Revised Children’s Manifest Anxiety Scale; SAQ-C24 = Social Anxiety Scale for Children; SAS-A = Social Anxiety Scale for Adolescents; STAI = State–Trait Anxiety Inventory; SAI = State Anxiety Inventory; SPAS = Social Physique Anxiety Scale; RCADS = Anxiety and Depression Scale in Children-Revised; API = Acute Panic Inventory; ASI-R = Anxiety Sensitivity Index-Revised. The sports type are labelled in bold.
NA: not applicable, NR: not reported, CD: cannot determine, overall of Good: more than 10Y (NA = Y), Fair 8 to 10Y, Poor: Below 8.
3.3 Risk-of-bias (quality) assessment
Overall, three reports showed good quality, whereas most studies were fair. We can find that in terms of random methods, most of the studies do not report the process of random sampling and there are basically no studies that apply computer random sampling methods. In addition, the information reported on the sample power is relatively lacking.
3.4 Effect of PA intervention on panic
Main effect. As shown in Figure 1, the pooled effect size revealed a significant overall effect of physical activity on panic symptoms from four studies (Random effects; d = −.45, SE = .12, Z = −3.65, p < .001). Estimation of the homogeneity suggested that the chance of inconsistent distribution of the effect sizes was not significant, Q(3) = 1.36, p = .715. Sensitivity analysis revealed small-to-no heterogeneity across the effect sizes of the studies, I2 = 0.00%. These results mean that the effect sizes across four studies suggested a significant small effect of physical activity in reducing panic symptoms among children, adolescents and early adulthood compared to the controls.
Publication bias. As shown in Figure 2, Egger’s regression-based tests suggested no estimated publication bias, β = .57, SE = .1,06 t = .54, p = .644. One study with small sample size (N = 24) reported high standard error [32].
Figure 2.
Funnel plot of the effect size estimates of physical activity on the panic symptoms in children and adolescents.
Exploratory meta-regression. Age and gender ratios were entered into meta-regression analysis to determine potential predictors for the combined effect sizes. As summarised in Table 3, meta-regression analysis suggested no significant predictor for the combined effect sizes of current studies, R2 = .00, Qw(2) = 1.35, p = .508. This means that age or gender ratios were not significant predictors of the current pooled effect sizes.
Predictor
Coefficient B
SE
95% CI
z
p
Age
−.03
.03
[−.11 .04]
−.99
.325
Gender Ratio (Female)
−.06
.66
[−1.35 1.23]
0.09
.927
Table 3.
Meta-regression analysis on age and gender predicting pooled effect sizes of PA intervention reducing panic disorder.
3.5 Effect of PA intervention on anxiety
Main effect. As shown in Figure 3, the pooled effect size revealed a significant overall effect of physical activity on anxiety from 14 studies (Random effects; d = −.51, SE = .15, Z = −3.38, p < .001). Estimation of the homogeneity suggested that the chance of inconsistent distribution of the effect sizes was not significant, Q(13) = 20.83, p = .076. Sensitivity analysis revealed moderate-to-small heterogeneity across the effect sizes of the studies, I2 = 47.52%. These results mean that the effect sizes across 14 studies suggested a significant medium effect of physical activity in reducing anxiety among children and adolescents compared to the controls.
Figure 3.
The forest plot of the effect size of physical activity on anxiety disorder in children and adolescents.
Publication bias. As shown in Figure 4, Egger’s regression-based tests suggested no estimated publication bias, β = −.58, SE = .63, t = −.92, p = .377. One study was outside of the funnel [10], of which the recorded main effect was a three-way interaction (pre vs. post & PA vs. Control & inhalation task 1 vs. task 2 vs. task 3). Such a small sample size (N = 18) and physiological measurement design (inhalation task) reported a greater effect size and small standard error comparing to other self-report questionnaires. It only counted 8.34% of the weight, which was not considered producing bias to the overall result. And another study reported high standard error because the sample size was very small (N = 23) [30].
Figure 4.
Funnel plot of the effect size estimates of physical activity on anxiety disorder in children and adolescents.
Exploratory meta-regression and subgroup analysis. Age and gender ratios were entered into meta-regression analysis to determine potential predictors for the combined effect sizes. Meta-regression analysis suggested no significant predictor for the combined effect sizes of current studies, R2 = .00, Qw(2) = .29, p = .865. This means that age or gender ratios were not significant predictors of the current pooled effect sizes.
The research method (mixed-design vs. between-subject & within-subject design) was entered into subgroup analysis to determine whether the main effect changed between the two interactive groups. No significant group difference in homogeneity was identified in either research method, Qb(1) = 1.26, p = .26.
The current chapter carried out a meta-analysis on 15 reports to investigate the pooled effect sizes of PA intervention on panic and anxiety symptoms among children, adolescents and early adulthood. Compared to the controls, we reported a small effect of PA on reducing panic symptoms and a middle effect on reducing anxiety symptoms. To our knowledge, this is the first study to summarise the effect of PA on panic and anxiety symptoms in children, adolescents and early adulthood with meta-analysis.
The primary finding of the current meta-analysis is that the pooled effect sizes from 4 reports yielded a significant small effect (d = −.45) of PA intervention in reducing panic symptoms. This means PA obtains such potential to be used in therapeutic intervention for panic in children, adolescents and early adulthoods. The mechanism of PA affecting panic symptoms was suggested to be related to the metabolism of CO2 and lactate. Accordingly, CO2 and lactate hydrolysed into HCO3- were found to moderate the blood pH level and influence brain acidosis [35]. Specifically, such change in pH level and brain acidosis is related to the activation of the amygdala and, furthermore, regulates the reaction of fear [36, 37]. Here, we would not discuss the biochemical pathology further but highlight the importance of CO2 and lactate. Empirical evidence suggested that patients with panic disorder were found to have chronically low end-tidal CO2 [38] and, vice versa, CO2 inhalation reduces panic symptoms [39]. Similarly, early literature reported an increased lactate level among panic patients [40, 41]. As a result, regulating the brain acidosis condition is believed to attenuate panic symptoms. Apart from direct CO2 inhalation, appropriate physical activity also regulates CO2 inhalation and lactate levels, eventually balancing the blood pH and brain acidosis, which is potentially the mechanism of how PA intervention works.
However, it is very important to bear in mind that this result is very primal and exploratory because it was summarised from four reports with limited sample sizes, of which one study targeted children (N = 28) [33], two studies targeted adolescents (N = 92; N = 24, respectively) [31, 32] and one study targeted early adulthoods (N = 18; age mean = 22.1) [10]. There is, to date, indeed a shortage of evidence in this field. A most recent 2022 systematic review only identified eight studies testing PA intervention’s effect on panic disorder among adults [11] and argued for more evidence. On the one hand, it is difficult to recruit children with panic symptoms or diagnosed with panic disorder and deliver interventions for them. Unlike general anxiety symptoms, panic disorder is often clinically diagnosed and needs to be treated with extra caution. On the other hand, the tool to measure panic symptoms are limited among children. The experiments included in the current analysis used measurements, including API, RCAD and ASI-R [10, 31, 32, 33]. Among these scales, RCAD and ASI-R are developed for anxiety and only address panic symptoms in their subscales, which were not specifically developed for panic disorder. API was developed for panic symptoms but not aimed at children. It is important to evaluate the assessment tool because the panic among children and adolescents could differ between ages [42]. In comparison, another scale developed after 2014 to measure panic symptoms, namely the Panic Disorder Severity Scale for Children, adapted for adolescents from 11 to 17 years old [43], and used in some studies testing the effect of CBT on panic disorder. Future studies could consider testing the effect of PA on children’s panic with suitable assessments.
The secondary finding of the current analysis is that PA intervention had a middle effect (d = −.51) in reducing anxiety among children and adolescents. The mechanism of PA reducing anxiety symptoms could be similar to panic symptoms, as mentioned earlier. However, it is noticeable that only four isolated effects were found significant among the included reports in which Kenis-Coskun, Aksoy [33] applied repeated rehabilitation exercises, Nazari, Shabani [13] applied continuous aerobic & resistance exercise intervention, Mücke, Ludyga [34] applied only one session of aerobic exercise at moderate intensity on the bicycle ergometer., and Ensari, Petruzzello [10] applied one Yoga session. A possible reason is that most of these experiments were conducted using a mixed design, with comparisons between experiment and control groups and within the experiment groups. Studies may have reported isolated pronounced effects between the experiment and control group or within the experiment groups, whereas the interaction was not significant [28] or not reported [9, 13, 25, 26, 29, 30]. Only one study reported no effect at all [27]. The data extracted from these studies reflected more combined effects than isolated between-group or within-group effects [19]. Subgroup analysis was also carried out to determine the potential difference between mixed-design and non-mixed design studies (only between-subject or within-subject designs), and no difference was identified between the effect sizes. Consequently, this result should be considered to reflect the actual effects of PA intervention on reducing anxiety.
Knowing that PA intervention reduces panic and anxiety symptoms across the empirical evidence, the next question is what type of exercise produces an optimal effect. The current meta-analysis could not perform the proper subgroup analysis to determine the best PA type due to insufficient data inclusion. Among the 15 studies, three studies applied resistance training, including push up [33], weight-lifting [13] and treadmill exercises [31], whereas others applied aerobic exercise in groups or individually. Among which, the push up and weight-lifting seems to produced the large and significant effect sizes reducing anxiety symptoms (d = −1.72, p < .001; d = −1.07, p = .002, respectively) [13, 33]. The effect of resistance training seemed somewhat contradictive to the mechanism in which the high lactate level could induce panic and anxiety symptoms when such anaerobic exercise produces lactate. There are two possible reasons to explain this. One possible reason is that only lactate in the brain induces the change in brain acidosis that leads to panic and anxiety symptoms [44], whereas muscle lactate produced by anaerobic exercise is independent of that in the brain [45]. It is the blood lactate which influences brain acidosis. The second reason is that the participants were measured almost immediately after the PA intervention cooled down, where the blood lactate level remained still. Hiscock, Dawson [46] tested 4 different types of weight-lifting technic to investigate the difference in muscle activation and blood lactate. The blood lactate was measured immediately after the exercise and no change in blood lactate was detected with a significant difference in muscle activation. Consequently, blood lactate was not influenced by different exercise intensities if measured immdiately after the intervention. Hypothetically, it is the regulation of CO2 during PA that potentially attenuates panic and anxiety symptoms. Future studies could compare, first, the effect of breathing practice from CBT and PA intervention with controlled blood lactate levels. Second, future studies could investigate the long-term effect of resistance training on blood lactate and panic & anxiety symptoms.
The last point to make is that the current meta-regression analysis did not suggest evidence of whether individual differences predicted the pooled effect sizes. Although it was argued in other literature that gender or age could moderate the effect of PA [12]. Previous literature reported gender differences in PA index as age grows [47], and more mature adolescents have more autonomy in PA than children [48], such that they can initiate behaviours that enhance the positive effects of PA on anxiety. As a result, the benefits of PA on depression and anxiety may increase with age. However, the current meta-regression analysis failed to demonstrate such effects on gender or age. One possible reason is that there were only 15 experiments identified in this study, which obtained limited power to detect any individual difference. Thirteen studies reported their gender ratio, and 12 reported the distribution of age. More studies are required to investigate individual differences’ role in the effect of PA intervention on anxiety and panic symptoms.
This meta-analysis has several limitations. Firstly, we used Cohen’s d rather than Hedges’ g to estimate the SMD, which the results might be biased in small sample studies. The reason to pick Cohen’s d was that some studies already reported pre-calculated Cohen’s d and did not provide sufficient information (N, mean, SD) to carry out transformation or bias correction. With limited numbers of the study identified, it would be impossible to exclude them from the analysis, and this could only be solved by including more studies with sufficient data. Secondly, the current study did not control the comorbidity among children and adolescents, which could affect the outcome of physical activities on panic and anxiety symptoms. As a result, the current results were very primal and exploratory, which should be considered cautiously.
The current meta-analysis reported a small effect of PA intervention on reducing panic symptoms and a middle effect on reducing anxiety symptoms in children, adolescents and early adulthood. Meta-regression analysis did not support age or gender predictors of the pooled effect size. More studies in this field are required to produce a more solid conclusion.
Appendix 2 Flow chart of studies retrieved and screened according to the PRISMA.
Appendix 3 Characteristics of included studies.docx.
Appendix 4 Summary of quality of included studies.
Appendix 5 Meta-analysis of raw data.
References
1.Caspersen CJ, Powell KE, Christenson GM. Physical activity, exercise, and physical fitness: Definitions and distinctions for health-related research. Public Health Reports. 1985;100(2):126
2.Kandola AA et al. Impact of replacing sedentary behaviour with other movement behaviours on depression and anxiety symptoms: A prospective cohort study in the UK biobank. BMC Medicine. 2021;19(1):133
3.Cescon M et al. Activity detection and classification from wristband accelerometer data collected on people with type 1 diabetes in free-living conditions. Computers in Biology and Medicine. 2021;135:104633
4.Hills AP, Mokhtar N, Byrne NM. Assessment of physical activity and energy expenditure: An overview of objective measures. Frontiers in Nutrition. 2014;1:5
5.Trost SG et al. Comparison of accelerometer cut points for predicting activity intensity in youth. Medicine and Science in Sports and Exercise. 2011;43(7):1360-1368
6.Howe CA et al. Comparison of accelerometer-based cut-points for Children's physical activity: Counts vs steps. Children (Basel). 2018;5(8):105
7.Bess HM et al. Efficacy of an individualized, motivationally-tailored physical activity intervention. The Society of Behavioral Medicine. 1998;20(3):174-180
8.Cavallo DN et al. A social media-based physical activity intervention: A randomized controlled trial. American Journal of Preventive Medicine. 2012;43(5):527-532
9.Tanksale R et al. Evaluating the effects of a yoga-based program integrated with third-wave cognitive behavioral therapy components on self-regulation in children on the autism spectrum: A pilot randomized controlled trial. Autism. 2021;25(4):995-1008
10.Ensari I, Petruzzello SJ, Motl RW. The effects of acute yoga on anxiety symptoms in response to a carbon dioxide inhalation task in women. Complementary Therapies in Medicine. 2019;47:102230
11.Machado S et al. Can regular physical exercise be a treatment for panic disorder? A systematic review. Expert Review of Neurotherapeutics. 2022;22(1):53-64
12.Bhui K, Fletcher A. Common mood and anxiety states: Gender differences in the protective effect of physical activity. Social Psychiatry and Psychiatric Epidemiology. 2000;35(1):28-35
13.Nazari M, Shabani R, Dalili S. The effect of concurrent resistance-aerobic training on serum cortisol level, anxiety, and quality of life in pediatric type 1 diabetes. Journal of pediatric endocrinology & metabolism : JPEM. 2020;33(5):599-604
14.Yu HJ et al. Improving the metabolic and mental health of children with obesity: A school-based nutrition education and physical activity intervention in Wuhan, China. Nutrients. 2020;12(1):194
15.Hayward C, Killen JD, Kraemer. T. H. C., C. B.Predictors of panic attacks in adolescents. Journal of the American Academy of Child & Adolescent Psychiatry. 2000;39(2):207-214
16.Page MJ et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. International Journal of Surgery. 2021;88:105906
17.Hale GE et al. Review: Physical activity interventions for the mental health and well-being of adolescents - a systematic review. Child Adolesc Ment Health. 2021;26(4):357-368
18.Mogre V et al. Adherence to self-care behaviours and associated barriers in type 2 diabetes patients of low-and middle-income countries: A systematic review protocol. Systematic Reviews. 2017;6(1):39
19.Morris SB. Estimating effect sizes from pretest-posttest-control group designs. Organizational research methods. 2008;11(2):364-386
20.Wan X et al. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Medical Research Methodology. 2014;14(1):1-13
21.Higgins JP et al. Cochrane handbook for systematic reviews of interventions. Chichester: John Wiley & Sons. 2019
22.Efron B. Nonparametric standard errors and confidence intervals. canadian Journal of Statistics. 1981;9(2):139-158
23.Lakens D. Calculating and reporting effect sizes to facilitate cumulative science: A practical primer for t-tests and ANOVAs. Frontiers in Psychology. 2013;4:863
24.Cohen J. Statistical power analysis for the behavioral sciences. New York: Academic press. 2013
25.Luna P, Guerrero J, Cejudo J. Improving adolescents subjective well-being, trait emotional intelligence and social anxiety through a programme based on the sport education model. International Journal of Environmental Research and Public Health. 2019;16(10):1821
26.Lindwall M, Lindgren EC. The effects of a 6-month exercise intervention programme on physical self-perceptions and social physique anxiety in non-physically active adolescent Swedish girls. Psychology of Sport and Exercise. 2005;6(6):643-658
27.Romero-Perez EM et al. Influence of a physical exercise program in the anxiety and depression in children with obesity. International Journal of Environmental Research and Public Health. 2020;17(13):4655
28.Akko DP et al. The effects of an exercise training on steroid hormones in preadolescent children - a moderator for enhanced cognition? Physiology and Behavior. 2020;227:113168
29.Naderi S et al. The effect of physical exercise on anxiety among the victims of child abuse. Sport Sciences for Health. 2019;15(3):519-525
30.Polis RL et al. Integration of yoga therapy into traditional residential treatment for at risk adolescent females: A community-based approach. Journal of Pediatric and Adolescent Gynecology. 2017;30(2):320-321
31.Smits JA et al. The effects of acute exercise on CO2 challenge reactivity. Journal of Psychiatric Research. 2009;43(4):446-454
32.Broman-Fulks JJ, Storey KM. Evaluation of a brief aerobic exercise intervention for high anxiety sensitivity. Anxiety, Stress, & Coping. 2008;21(2):117-128
33.Kenis-Coskun Ö et al. The effect of telerehabilitation on quality of life, anxiety, and depression in children with cystic fibrosis and caregivers: A single-blind randomized trial. Pediatric Pulmonology. 2022;57(5):1262-1271
34.Mücke M et al. The influence of an acute exercise bout on adolescents’ stress reactivity, interference control, and brain oxygenation under stress. Frontiers in Psychology. 2020;11:3091
35.Wemmie JA. Neurobiology of panic and pH chemosensation in the brain. Dialogues in Clinical Neuroscience. 2022;13:475-483
36.Coryell MW et al. Targeting ASIC1a reduces innate fear and alters neuronal activity in the fear circuit. Biological Psychiatry. 2007;62(10):1140-1148
37.Wemmie JA et al. Acid-sensing ion channel 1 is localized in brain regions with high synaptic density and contributes to fear conditioning. Journal of Neuroscience. 2003;23(13):5496-5502
38.Papp LA et al. Respiratory psychophysiology of panic disorder: Three respiratory challenges in 98 subjects. American Journal of Psychiatry. 1997;154(11):1557-1565
39.Griez E, van Den Hout MA. CO2 inhalation in the treatment of panic attacks. Behaviour Research and Therapy. 1986;24(2):145-150
40.Stewart PA. Independent and dependent variables of acid-base control. Respiration Physiology. 1978;33(1):9-26
41.Maddock RJ et al. Elevated brain lactate responses to neural activation in panic disorder: A dynamic 1H-MRS study. Molecular Psychiatry. 2009;14(5):537-545
42.Sheikh JI et al. Aging and panic disorder: Phenomenology, comorbidity, and risk factors. The American Journal of Geriatric Psychiatry. 2004;12(1):102-109
43.Elkins RM, Pincus DB, Comer JS. A psychometric evaluation of the panic disorder severity scale for children and adolescents. Psychological Assessment. 2014;26(2):609
44.Riske L et al. Lactate in the brain: An update on its relevance to brain energy, neurons, glia and panic disorder. Therapeutic advances in psychopharmacology. 2017;7(2):85-89
45.Martinsen EW et al. Tolerance to intensive exercise and high levels of lactate in panic disorder. Journal of Anxiety Disorders. 1998;12(4):333-342
46.Hiscock DJ et al. Muscle activation, blood lactate, and perceived exertion responses to changing resistance training programming variables. European Journal of Sport Science. 2016;16(5):536-544
47.Teixeira e Seabra AF et al. Age and sex differences in physical activity of Portuguese adolescents. Medicine and Science in Sports and Exercise. 2008;40(1):65-70
48.Courtney JB et al. Autonomous motivation and action planning are longitudinally associated with physical activity during adolescence and early adulthood. Psychology of Sport and Exercise. 2021;56:101974
Written By
Lin Wang and Yihao Liu
Submitted: 13 June 2022Reviewed: 23 June 2022Published: 16 July 2022
Open access peer-reviewed chapter
