Thoracic Spine Pain in Adolescents and Its Association With Electronic Devices and Other Factors

Purpose: To determine the prevalence of thoracic spine pain (TSP) in adolescents and to analyze its association with sociodemographic characteristics, use of electronic devices, habitual practice of physical activity and mental health problems. Methods: Cross-sectional study with 1628 students of public schools from the city of Bauru/SP. Data collection: 1. Characterization of the participants, use of electronic devices (TV, computer, tablet and phone), psychosocial factors and lifestyle; 2. Strengths and diculties questionnaire (SDQ); 3. Baecke Questionnaire; 4. Nordic questionnaire. Descriptive analyzes, bivariate and multivariate logistic regression were used. Results: The overall prevalence of TSP was 51.5% (CI 49.1 to 53.9). Women showed a prevalence of 59.6% (CI 56.2 to 62.9) and men of 43.1% (CI 39.7 to 46.5), with signicant difference. The variables associated with TSP were: female gender (PR= 1.96), use of computer for more than 3 hours per day (PR= 1.29), use of computer for more than 3 times per week (PR= 1.35), distance from eye to the computer screen (PR= 1.29), use of cell phone in semi-lying position (PR= 1.37), time of use of cell phone for more than 3 hours per day (PR= 1.44), use of tablet in sitting position (PR= 1.47) and presence of mental health problems (PR= 2.62) Conclusion: There is a high prevalence of TSP in adolescents and a marked association with female sex, the use of electronic devices, the presence of mental health problems. The level of physical activity was a protective factor.


Introduction
Thoracic spine pain (TSP) is relatively common in children and adolescents. In a review study, TSP in adolescents in the last twelve months and lifetime varied from 4.2-9.7%, and 15.6-19.5%, respectively, being comparable lumbar pain (7.0-50.8% and 7.0-72.0%) and cervical pain (7.6-13.0% and 3.0-28.0%) [1]. There is also evidence that the presence of TSP should not be underestimated in adolescence. In contrast to adulthood, when TSP incidence and prevalence is considerably lower than that of LBP and NP, TSP incidence and prevalence are similar to that of LBP and NP in children and adolescents [1].
A review article found few studies regarding the risk factors of TSP, re ecting the general observation that the thoracic spine receives noticeably less attention in the literature compared to NP and LBP [1,2]. The literature relates TSP to female sex [1,2,3], back pack weight and chair height at school [4], as well as parental smoking [2,4], increased body weight [1,2], and mental health problems [5]. To the best of our knowledge there are no studies that have assessed tablet and phone use, and there is no national or international data regarding on their relationship with TSP until now. Since we already know that the use of phones and tablets is associated with NP and LBP [4], and the thoracic spine articulates with both the cervical and lumbar vertebrae, the use of these devices and their association with TSP cannot be ignored, as it may be that they represent some risk for that region of the spine.
Preliminary evidence suggests that TSP in youth is as important as LBP, and therefore it is a cause for concern. Epidemiological studies have been characterized as important predictors of TSP later in life. In children, it can cause marked discomfort and impairment in daily life, causing long term problems.
Studies suggest that children and adolescents who report TSP have increased use of health services, absenteeism, impairment in school, restrictions on physical activity and often experience the coexistence of other health, physical or mental complaints [6].
The present study will contribute to other epidemiological investigations and collaborate for the knowledge of the national estimates about the prevalence of TSP in adolescents and its risk factors. It will also contribute to systematic reviews and meta-analysis, once the available data regarding the association among TSP and sociodemographic variables, the use of electronic devices, the practice of physical activity and mental health problems in adolescents from Brazil and other countries are scarce.
This study may also aid at the understanding of its global cause and contribute to public policies aimed to control this problem, based on preventive and/or therapeutic interventions.
Considering these points, the objective of the present study was to determine the prevalence of TSP and to analyze its association with the sociodemographic characteristics, the use of electronic devices, the practice of physical activity and the presence of mental health problems.

Design and population
This is a cross-sectional study, performed with the project data "Back pain and associated factors in students of high school: a longitudinal study" (Financed by Fapesp, Process: 2016/182837), with of 14to 18-year-olds of both sexes attending the rst and second years of high school in the morning in the urban area of Bauru, SP, Brazil. This study was approved by the institution Research Ethics Committee.

Sample calculation
Adolescents enrolled in the rst and second grades of secondary education in public schools in the city of Bauru/SP were enrolled. According to data provided by the State Department of Education, 9,000 students were enrolled in 2017 [4].
Age groups and sex were de ned called sampling areas for which minimum numbers in the sample were guaranteed that would allow subsequent analyses. The sampling areas were students of both sexes in the 1st and 2nd years of secondary education [4].
To determine the sample size, the formula to calculate samples for nite populations was used with the following parameters: con dence level, 95%; prevalence, 50%, unknown percentage complement (100-p); population size, 9,000 high school students from state public schools; and maximum permissible error, 3%. Thus, the sample size for the group was at least 990 individuals, to which we added a 20% expected loss and 15% for association studies, reaching a total of 1,366 adolescents [4].
The sample size calculation considered a plan with cluster sampling in two stages, where the primary sampling units (UPA) are the schools and Secondary Sampling Units (USA) are the classes concerning the three years of secondary education of the selected schools. The sample of school children was formed by all the students of USA classes selected in the sample of UPA schools [4].
The UPA schools were initially strati ed by their geographical location in accordance with the division of the city into eight sectors. Schools with classes of 1st and 2nd years of secondary education were counted to reach the required number in each sector. In three consecutive sectors between the west and east of the city (counterclockwise), there were no public state high schools; therefore, these three sectors were rejected, and only the ve that concentrate the 30 schools between the east and west of the city (counterclockwise) were considered [4].
In each geographic stratum, the sample was obtained in two stages. First, schools were selected using a method of selection with probabilities proportional to the size. Size considered in the selection of the schools was the total number of students in the two years of secondary education in each school, and that the percentage of each year in relation to the total number of students (9000): 36.9% and 33.6%, respectively. These percentages were applied to the sample (1,366 students). Second, the total number of students to be interviewed per sector and per year of high school was determined. To reach the total determined for each sector, the schools were randomly selected, as were the classes of the schools [4].

Inclusion and exclusion criteria
Included in this study were students who were part of the study in 2017 and who answered the questionnaire alone, were aged 14-18 years, whose parents provided informed consent, and who did not report low back pain in 2017. Students who changed school, or who changed school in the morning to night school, were excluded from study.

Data collection procedure
The State Department of Education authorized the research. After parental/guardian consent was provided, baseline data were collected from March to June 2017 by undergraduate and graduate students trained [4].
After parental/guardian permission was received, the researchers moved from room to room applying questionnaires. For the data collection, the following procedure was performed: rst, in each classroom, the researcher explained the research objectives, and the students were informed about the voluntary nature of their participation, right to leave the study at any time, and right to con dentiality of their data. Subsequently, the questionnaires were distributed with instructions and recommendations for completion, although no deadline was established. During the process, any questions expressed were promptly answered by the interviewer collecting the data. While completing the questionnaire, students did not communicate to minimize possible undesirable interference with their responses [4].
For each school, three extra visits were made that aimed to collect the data from the students who were absent from class related to previous collections. Students who were not found after three visits were considered lost. Students who refused to answer the questionnaire by personal choice were considered refusals [4].

Variable description
The TSP represented the response variable of interest in this study. The variable TSP was observed using the Nordic questionnaire, which was validated and adapted to the Brazilian culture [7]. The question used to de ne the outcome of TSP was: "In the past twelve months, have you had any pain or discomfort in the thoracic spine?" In the present study, TSP pain was de ned as a non-speci c pain experienced in the spinal cord extending from the cervicothoracic joint (C7-T1) to the thoracolumbar junction (T12-L1) and the corresponding posterior aspect of the trunk, and it can occasionally cause radiating pain into the anterior chest wall [8]. In addition to the verbal questionnaire, an image of the spinal regions in different colors was also presented, so the interviewees could better specify the low back region where the pain was [7].
Sociodemographic aspects, variables related to electronic devices, the usual practice of physical activity and mental health were considered independent. Sociodemographic factors include sex, age, marital status (single, married, and widowed/separated), ethnicity (white, black, brown, yellow and indigenous) and academic year [4]. Table 1 shows the questions asked to the participants regarding electronic devices use [4,9] (TV, computer, tablet or mobile phone). To estimate the level of habitual physical activity practice, we used the Baecke Questionnaire of Habitual Physical Activity validated in Brazil [10]. Through the application of this instrument, it was possible to determine the score of each domain of physical activity. The sum of the scores of each section comprised a value of total dimensionless, that is, habitual physical activity. For the classi cation of habitual physical activity, we used the formula proposed by Baecke et al [10]. Students were subdivided into quartiles according to the individual total score provided by the instrument, which resulted in the following physical activity groups: sedentary (1st quartile); moderately active (2nd and 3rd quartiles); and active (4th quartile) [11].
Mental health was evaluated by the Strengths and Di culties Questionnarie (SDQ) validated in Brazil by Fleitlich-Bilyk [12]. The possibilities of results indicated by the instrument for all ve subscales and the total number of di culties are three: "Normal" (healthy): indicates that there are no di culties regarding what is being assessed; "Borderline": indicates that the child or adolescent already presents some di culty which, if not properly cared for, can deteriorate and jeopardize their development; "Abnormal" (clinical): indicates that there are major di culties relating to what is being assessed, requiring specialized intervention [12,13].

Data Analysis
The data analysis was performed using the Statistical Package for the Social Sciences version 18.0. An independent student who did not participate in the study introduced the data. The accuracy of the data entry was tested in 10% of randomly chosen questionnaires. An error was detected and corrected. Another set of 5% of the randomly chosen quizzes was tested, and no error was found.
In the descriptive analysis, the prevalence and con dence intervals of all variables included in the study was calculated. Statistical association methods were used using bivariate and multivariate logistic regression analysis between the prevalence of TSP (dependent variable) and all independent variables studied with determination of signi cance level and estimated relative risk of the 95% of con dence interval.
Bivariate and logistic regression analyses were performed according to the theoretical-conceptual hierarchical model. For all variables, a reference category was established that was considered the lowest risk for the occurrence of the outcome. To construct the hierarchical model, the variables were organized into four levels according to the temporal and causal relationship for TSP. The adjustment of the rst level was performed by all variables belonging to it. The second was adjusted by variables of the previous level that presented values of p < 0.10 after adjustment and by those belonging to it. The third was adjusted by variables of the rst and second levels with values of p < 0.10 after adjustment besides those belonging to it. Finally, the fourth level variable was controlled for the previous three levels. To select the variables that will remain in the regression model, the backward selection process was used, leaving in the nal model all variables with values of p < 0.05 [14,15].

Results
A total of 1,628 students were studied, already deducted from the nal percentage of 2.05% refusals. The socio-demographic characteristics of the sample can be seen in Table 2. The variables related to the use of electronic devices in high school adolescents are presented in Table 3.  Table 4 shows that there was a signi cant association among TSP and female sex and the presence of mental health problems. TSP was associated to type of PC, daily time of cell phone use of more than 3 hours and to the use of tablet presented in Table 5.

PR = Prevalence ratio/ CI = Con dence interval
In multiple factor analysis, after adjustment by logistic regression according to the hierarchical model, at the rst level, female sex remained associated with pain and was maintained as an adjustment variable for the next level. At the second level, clinical mental health problem acted as a predictor of pain and remained as an adjustment variable for the next level along with sex. At the third level, use of computer for more than 3 hours per day, use of computer for more than 3 times per week, use of cell phone in semilying position, time of use of cell phone for more than 3 hours per day, use of tablet in sitting position remained associated with pain remained as an adjustment variable for the next level, together with the previous ones. On the fourth level, only the activity level physics acted as a pain predictor after adjustments to the same level and previous variables (Table 6).

Discussion
In the present study, the prevalence of TSP was 51.5%, higher than that reported in other countries, such as Australia (20.0%) [1], Portugal (13.2%) [3]; in the region of Southern Denmark (36%) [16]. It was also higher than previously Brazilian studies carried out in a South State of Brazil, that showed a prevalence of 26.2% in the city of Rio Grande [17], 36,9% in the city of Pelotas [14]. On the other hand, the prevalence was similar to that reported in Canada [18] of 59.0%. These prevalence variations may be a result of sociocultural, demographic, economic and occupational differences of the investigated regions.
Then, considering these high and still underestimated prevalence of TSP in adolescents, it is important to recognize its risk factors. In the present study, it was observed that TSP was associated with female sex, computer use for more than 3 hours a day, computer use for more than 3 times a week, short distance from eye to computer screen, use of the cell phone in semi-lying position, cell phone use for more than 3 hours a day, use of tablet in sitting position and presence of mental health problems.
Women was approximately 1.96 times more likely to develop pain in comparison to men, a result similar to other studies with adolescents [1,2,3,6,19]. Female sex presents structural and psychosocial differences in relation to males, such as musculoskeletal differences, which make women predisposed to overload in the vertebral column. Additionally, among adolescents women, the long periods sitting in school, the hormonal changes resulting from puberty and the low levels of physical activity predisposes them to report pain with greater frequency [20].
The use of computer and cell phone for more than 3 hours a day and more than 3 times a week was associated with TSP, corroborating with other investigations [3,19]. In fact, a great number of studies suggest the association between spinal pain and electronic devices use, specially cell phone and computer [3,4,19,21]. The semi-lying and sitting postures to use the cell phone and tablet, respectively, were factors associated with TSP in the present study, corroborating with the results from Briggs et al. [1].
Long periods in inadequate postures using a portable device, usually cause the exion of the cervical and thoracic spine, exerting great pressure on muscles, tendons and ligaments that support the head and maintain good posture. Then, these stressed soft tissues cannot adequately support the spine, which suffers structural deterioration, affecting musculoskeletal structures. When the weight of the head and upper body is moved forward and out of its central axis, the stress in all soft tissues and the spine are exponentially increased. In addition, the exion and internalization of the cervical increases the stress in the thoracic and low back regions, since all the dorsal muscles and structures of the spine work together to provide stability and movement to the trunk and limbs [22].
Additionally, inadequate postures during activities of daily living are worthy of concern, as these postures can increase the compression strength in the intervertebral discs, leading to disc malnutrition, compromising the integrity of the musculoskeletal system, predisposing to fatigue and higher levels of pain [23]. Frequent or sustained neck exion and shoulder protrusion resulting in scapular dysfunction during upper limb movement, activities with the use of electronic equipment, may result in increased thoracic kyphosis and, consequently, this associated with a higher risk of thoracic spine pain [24].
The clinical category related to mental health problems was associated with TSP in adolescents, corroborating with previous studies [25,26]. The emergence of emotional symptoms is common among students, which can be triggered by the moment experienced, pressure in the school environment, economic di culties and relationship problems.
There seems to be a relationship between emotional symptoms and physical manifestations, such as increased secretion of the hormone cortisol and changes in the hormonal regulation of the adrenal glands, which generates inhibitory effects on the immune response, digestion and symptoms of excessive body wear, tiredness, fatigue, muscle pain, joint disorders and reduced physical capacity.
Psychosomatic symptoms are common among children and adolescents and can manifest in different ways, for example, headache, abdominal pain and musculoskeletal problems, including pain in the spine.
On the other hand, the pain itself can also lead to some degree of social or emotional impact, resulting in low scores on stress and well-being parameters [26,27]. Further studies are needed to determine the emotional relationship with TSP since the design of the present study does not allow us to determine what would be the cause of this association, but rather contributes to new questions.
In present study physical activity was associated with lower TSP. Some studies have shown that physically active adolescents were less likely to have back pain [28,29], other authors have identi ed that physical activity can contribute to pain [30,31], while other investigations have not noted an association between the variables [32].
The mechanisms of association between physical activity and spinal pain are not clear. One of the hypotheses about the association between physical activity and lower spinal pain prevalence is that it can lead to increased muscle strength and exibility [31,32].
The results of this study should be interpreted in their own context of limitations, i.e., the data were based entirely on interviews, so that an inaccurate estimate of the data and memory bias is inevitable, no data were collected regarding the use of other technologies, and another limitation is that students come from public schools, which limits the generalization of data to those from private schools. The strength of this study is that it is the rst Brazilian study to examine the role of factors in the onset of TSP in young people, the use of validated questionnaires to evaluate the results and the number of people interviewed.
Considering that the available data from both national and international literature are scarce, this study will contribute as a reference to other epidemiological investigations and shall cooperate with the national evaluation of prevalence and risk factors, systematic reviews and meta-analyzes.

Conclusion
The TSP had high prevalence and marked association with female sex, use of computer for more than 3 hours a day and 3 times a week, use of the cell phone for more than 3 hours a day, use of cell phone in semi-lying position, use of tablet in sitting position, presence of mental health problems and the level of physical activity was a protective factor.

Declarations
Funding: This study was supported by funding from the FAPESP (São Paulo Research Foundation).
Con icts of interest: The authors declare no con ict of interest. Availability of data and material: The data that support the ndings of this study are available from the corresponding author, AV, upon reasonable request.
Code availability: Not applicable.
Authors' contributions: AV participated in the design and coordination of the study, analysis and interpretation of data and preparation of the manuscript. LDC participated in the analysis and interpretation of data and the preparation of the manuscript. TPFB conducted data collection, analysis and interpretation of data and preparation of the manuscript. LAF participated in analysis and interpretation of data and preparation of the manuscript. WLR participated in the analysis and interpretation of data and preparation of the manuscript. JAAF participated in the analysis and interpretation of data and preparation of the manuscript. NMM participated in the analysis and interpretation of data and preparation of the manuscript. PDOP participated in the analysis and interpretation of data and preparation of the manuscript. All authors read and approved the nal manuscript.