Pain medication use for musculoskeletal pain among children and adolescents: a systematic review

Literature selection

A literature search strategy was developed using a combination of medical subject headings (MeSH) and text words related to the child/adolescent (e.g. child, children, adolescent, teenagers, schoolchildren, school), MSK pain (e.g. back pain, neck pain, knee pain, chronic widespread pain), study design (e.g. prospective study, retrospective study, cross-sectional study) and pain medications (e.g. paracetamol, aspirin, opioids, NSAID, ibuprofen, codein) categories. The search strategy is outlined in Supplementary Material. Key words were identified by the senior investigator (AA) after consulting systematic reviews on the topic of MSK pain and pain medications and discussions with the other reviewers and research librarians at Aalborg University.

All databases were searched from inception to the 5th of May 2021. The EMBASE, CINAHL and PsychINFO scientific databases were searched. The search terms included in the strategy were tailored to each specific scientific database.

Eligibility criteria

Eligible studies had to report data on the outcomes of this review. The primary outcome was the prevalence of the use of pain medication for MSK pain in children and adolescents. The secondary outcomes were the factors and side effects associated with the use of pain medication for MSK pain in children and adolescents. Studies designed to investigate the 1) prevalence 2) use (e.g. doses, frequency, duration) 3) factors and side effects associated with the use of pain medication for MSK pain were included. Studies where information on the use of pain medication for MSK pain was reported in the abstract were also included. Pain medications such as paracetamol, non-steroidal antinflammatory drugs and opioids were considered. MSK pain was considered as pain arising from the components of the MSK system (muscles, joints, bones, ligaments and tendons) [31, 32]. Studies reporting on adolescents with either MSK pain of a duration <3 months or MSK pain of a duration >3 months, as defined by the International Association for the Study of Pain [33], were considered. Studies had to include children and adolescents aged 6–19 years old. According to the World Health Organization (WHO) consolidated guidelines [34], June 2013, “child” is a person 19 years or younger unless national law defines a person to be an adult at an earlier age and “adolescent” a person into the 10 to 19 age category. In addition, the age range 6–19 is chosen as the age of 6 is considered as the starting age for a child to be able to reliably self-report pain [35, 36]. Peer-reviewed studies published in English, Scandinavian languages, Spanish, Italian were eligible for inclusion. Studies published in any other language (e.g. Portoguese) that could be translated by the authors were also considered for inclusion. Studies with a prospective, retrospective or cross-sectional study design were included. Randomized controlled trials (RCTs) were excluded as an intervention was delivered and the required selection criteria would likely compromise generalizability. Studies were excluded if: data were reported on adults only, if separate information for children and adults in case of a mixed sample were not available or if children under the mean age of 6 were included in the sample. Studies were further excluded if pain medication was not used for MSK pain or if focusing on populations with specific diseases or conditions (e.g. cancer pain) where pain is assessed and reported but is a result of the disease or underlying condition. Studies reporting treatment in the acute setting (e.g. treatment for acute pain related to injuries) were excluded. Studies were also excluded if full-text was not available or if translation was not possible. An attempt was made to obtain the full-text of the article through the interlibrary loan service when the full-text was not available.

Selection process

The process of literature selection was carried out by four authors (NAJ, AA, CG and CLS) and is summarized in the following stages: 1) Initial search of articles, upload of all references on the reference manager software (Covidence) and elimination of duplicates. This stage was performed by one author (AA) with the support of the librarians at Aalborg University Library. 2) Screening of titles and abstracts to identify articles potentially suitable for inclusion. This stage was performed by two authors (NAJ and AA). 3) Full text reading and selection of the articles that met the inclusion criteria. Four reviewers (NAJ, AA, CG and CLS, two authors for each paper) who assessed eligibility of the studies undertook this stage.

Data extraction

Data was extracted by AA, NAJ, CG and CLS. Data was extracted independently by a minimum of two authors for each study and then cross checked in order to ensure consistency in the data extraction process. In cases of disagreement, the article was discussed between two authors until agreement was reached. In case disagreement remained, a third author (MSR) would decide the outcome of the disagreement. A standardized Excel data extraction form was used during the data extraction process (Supplementary Material).

Quality assessment

The quality assessment was carried out by using a modified version of the Epidemiological Appraisal Instrument (EAI). The modified version of the EAI consisted of 31 items and is displayed in Figure 1. For each item, a response was given among the following options: Yes (2 points given), Partial (1 point given), No (0 points given), Unable to determine (0 points given) and Not applicable (0 points given). The quality score range for each study was therefore 0–62. The EAI has been appraised and shown as being an excellent tool for the assessment of observational studies, cross-sectional studies and prospective studies [37]. During the quality assessment process, each article was assessed by a minimum of two authors. In cases of disagreements the items in question were discussed between two authors. A third author (MSR) would decide on the final verdict if an agreement could not be reached.

Figure 1:

Modified version of the Epidemiological Appraisal Instrument (EAI).

Data synthesis

Data retrieved from studies identified were synthetized, as reported in the “results” section and discussed to reflect on the credibility and thoroughness of the results. Due to heterogeneity in terms of design and measures used in the included studies, it was not possible to conduct a meta-analysis. Therefore, a narrative synthesis was considered as the most suitable method to report the results of the review. To summarize the characteristics and outline the findings of the included studies, information was presented in the text in the following sections: “Study selection process”, “Quality and characteristics of the studies”, “Prevalence of the use of pain medication”, “Factors associated with the use of pain medication”, “Side effects”, “Duration of the use of pain medication”. Information was also presented in tables.

Study selection process

The electronic search initially provided 20,135 studies, of which 17,694 remained after duplicate deletion. After title and abstract screening 334 studies were screened for eligibility, of which 314 were excluded because they did not meet the eligibility criteria. Finally, 20 studies met the eligibility criteria and were included. The process of selection of studies in the review is outlined following the PRISMA model, Figure 2.

Figure 2:

Flowchart showing the process of selection of the studies in the review.

Quality and characteristics of the studies

Table 1 outlines the characteristics of the included studies, which were all published in English language. The included studies were from 11 countries (seven from the USA, three from Finland, two from Greece and one from Italy, Germany, Thailand, Norway, Iceland, Spain, The Netherlands and Australia each), reporting data on a total of 629,424 participants (range 80–268,228). More than half of the studies (n=11; 55%) had a sample size >1,000 participants. The majority of studies (n=9) were carried out in the school setting, while one study was from a sport club (a youth swimming team), one study from a mixed sample of sport clubs (including several youth sports teams) and schools, two from pediatric departments, two from birth cohorts, two from pain clinics and three studies were registry-based. The registry-based studies included data from four different registries: the Tennessee Medicaid program [40], and three registries of privately insured populations: the MarketScan commercial claims database [38], the HealthCore Integrated Research Database and the Arkansas Medicaid [51]. Thirteen studies were cross-sectional, six were retrospective studies and one had a prospective study design. The majority of the studies reported on adolescents aged 15–16 years old. Further, in most studies the sample had an equal distribution of boys and girls while four studies had a higher proportion of girls than boys [38, 42, 43, 57]. One study did not report on the gender distribution [46]. Five studies (25%) reported information on the ethnicity of participants, and 10 studies (50%) reported on the participation rate (for three studies this was not applicable as they were registry-based). The studies were published between year 2000 and 2020, with the majority (15/22) published after 2010.

Based on the EAI the average quality score of studies was 36.0 (potential maximum score: 62). By applying a categorization of studies in tertiles, the scores for the studies of higher quality ranged from 39 to 47, from 31 to 38 for studies of medium quality and from 24 to 30 for studies of lower quality according to our criteria. The quality assessment for each study is outlined in Supplementary Material.

Prevalence of the use of pain medication

The prevalence of pain medication use in children and adolescents with MSK pain is reported in Table 2. Prevalence estimates varied depending on the study setting. The prevalence in an American study [57] (n=80) from a pain clinic (75%) and an American [43] (n=102) and Finnish [52] (n=1,637) study from sport clubs (62–72%) was higher compared to estimates from studies carried out in school settings (8–42%). Studies from school settings (n ranged from 366 to 11,018) were from Finland [56], Iceland [41], Thailand [49], Italy [45], Germany [53], Norway [55], Spain [44], The Netherlands [50] and the U.S.A. [54]. Among school-based studies, a low prevalence (10.4%) of pain medication use was reported in one study [45] while the remaining studies [41, 44, 49, 50, 53], [54], [55], [56] reported prevalence estimates ranging from 15 to 42%. Studies from the pediatric department (n=532) [48] and (n=692) [47] reported a prevalence estimate between 17.5 and 23.5%. Registry-based studies (all from the U.S.A.) provided slightly different information. One study (n=201,940) [40] reported that among all the opioid prescription for several indication (i.e. dental, medical or surgical procedure, trauma, infection, headache or other pain), 5.5% of hydrocodone, 4.7% of codeine, 3.1% of oxycodone and 18.3% of tramadol prescriptions were for back or other MSK pain. The second registry-based study (2 registries included, n up to 64,354) [51], reported data on opioid prescription for chronic non-cancer pain conditions among privately insured adolescents. In 2001, 21.5–37.8% of all the prescription were for joint pain, and 21.2–39.0% for back/neck pain. In 2005, these figures increased to 21.6–41.0% (joint pain), and 21.4–44.5% (back/neck pain) [51]. Only three studies [42, 51, 57] included children and adolescents with MSK pain of a duration >3 months or with a ICD-9-CM code for MSK pain and showed the highest prevalence of use of pain medication (from 20 up to 75%). The third registry based-study [38] (n=268,228) showed that 13% of children and adolescents with low back pain were prescribed opioids within 12 months from their diagnosis.

Eleven studies reported on the type of medication used. These were mainly paracetamol and NSAIDs in studies from school settings [41, 44, 55], sport club [52] or pediatric department [47, 48]. Opioids were reported only in four American studies, one from a pain clinic [42] and three registry-based studies [38, 40, 51].

Five studies (4 from school settings and 1 from a sport club) reported on the frequency of use of pain medication. Depending on the categorization (bullied vs. non bullied), Icelandic adolescents reported using “sometimes” the following medication when experiencing pain: 40–50% paracetamol, 39–44% ibuprofen, 3–7% aspirin and 3–8% diclofenac [41]. In another study, the frequency of use of pain medication in Italian school adolescents with mild back pain was 2 times/month, while it was 5 times/month among those with moderate to severe back pain [45]. In a study among German school children who reported experiencing pain in the last three months, 39.1% reported using pain medication for their pain sometimes and 12.4% often or always [53]. In a study including Norwegian school-adolescents, 10% reported that they were daily users of over-the-counter analgesic, while 35% of them used over-the-counter analgesics less than weekly and 39% did not use them in the past previous 4 weeks [55]. In one study among 102 adolescent American swimmers, 72% reported using pain medications to manage shoulder pain, of whom 20% used them <1 time per month, 33% 1–3 times per month and 47% ≥ 1 times per week [43].

Factors associated with the use of pain medication

Fourteen studies (70%) reported information on the factors associated with the use of pain medication for MSK pain in children and adolescents (outlined in Table 3). Mixed-evidence was found regarding the association between the use of pain medication and increasing age, with four studies [38, 41, 51, 53] reporting a positive association and two studies reporting no evidence of association [44, 50]. Regarding female gender, six studies [38, 39, 44, 46, 52, 53] reported higher use in girls compared to boys and two studies [50, 55] reported a non-statistically significant association. However, it should be noted that for both age and female gender, the studies reporting no evidence of association also included the use of pain medication for non-MSK pain conditions (e.g. headache, abdominal pain) in the analysis.

Several pain characteristics were associated with the use of pain medication as assessed with a statistical test, or as shown by a marked increase in the percentage of pain medication use. These pain characteristics were: pain severity [44, 45], pain intensity [42, 46, 50, 53], pain duration [42, 44, 53], disability levels [42, 46] and the presence of ≥2 pain conditions or pain in various body sites [50, 51, 55]. Conflicting results were reported for pain frequency, which was associated with the use of pain medication in two studies [41, 56], of which one study [41] considered pain frequency as weekly symptoms of either headache, stomachache, backache, and/or neck and shoulder pain. Conversely, other studies reported that healthcare utilization [53] (e.g. visit to the doctor and/or using pain medication) or pain medication use [50] were not predicted by pain frequency. However, in these latter studies the use of pain medication might also include use for other conditions such as headache and abdominal pain, as separate analysis for each pain condition was not carried out.

Factors pertaining to the psychological domain (presence of one or more mental health condition, being bullied and low-self-esteem) were also associated with the use or prescription of pain medication, as reported in three different studies [41, 51, 55]. Consulting a greater number of doctors was associated with the use of pain medication in one study [42]. Regarding physical activity, being a member of a sport club was associated with the use of pain medication (especially in boys) in one study [52]. However, no differences between high and low-frequency pain medication users in relation to participation in sports were observed in another study [55]. Having an opioid prescription was associated with a future opioid prescription in one study [38] Conflicting results were found for an association between use of pain medication and educational level [50, 55]. Finally one study [55] found statistically significant associations between the use of pain medication and other factors such as sleeping <7 h per night, daily consumption of caffeinated drinks, binge drinking, performing spare-time paid work and absence from school three days or more.

Side effects

Only one study, registry based [40], reported on side effects (n=275) observed. Side effects related to the use of opioids included neurologic or respiratory symptoms (49% of cases), of which 25.8% was central nervous system depression, 3.3% respiratory depression and 30.2% neuropsychiatric symptoms. Other individuals had gastrointestinal (20.7%), dermatologic (18.5%), allergic (12.7%) or other (1.5%) symptoms [40]. However, the risk of experiencing the side effects was different depending on the type of opioid used, with tramadol and oxycodone users who had increased risk for all opioid‐related side events compared to codeine and hydrocodone users. There were no studies reporting on side effects related to the use of non-opioid pain medication (paracetamol or NSAIDs).

Duration of the use of pain medication

Only two American studies (10%) [40, 51] reported on the duration of use of pain medication. In one study [40], opioids were prescribed for a mean of 4.4 days, while in another study [51] opioids were prescribed for a mean of 11.5 days in 2001 (Arkansas Medicaid registry) and 7.5 days (HealthCore registry) and 13 days (Arkansas Medicaid registry) in 2005. However, these estimates might also include the use of opioids for other non-MSK pain conditions.

Summary

We identified 20 studies, which reported that 8–75% of children and adolescents use pain medication for MSK pain. The lowest prevalence was found in studies from school-based samples (8–42%) while 67–75% of adolescents in sports clubs or pain clinics used pain medication for their MSK pain. The most consistent factors associated with the use of pain medications were pain characteristics and psychological factors. Mixed evidence was found for increasing age and female gender, although most studies reported that these factors were associated with use of pain medication, and the studies reporting no evidence of association also included the use of pain medication for other pain conditions in the analysis.

Only two studies reported on the duration of the use of pain medication and only one study on side effects related to the use of pain medication, which assessed only opioids, however. The lack of data regarding side effects (especially related to the use of paracetamol and NSAIDs) and duration of pain medication use requires further investigation.

Comparison with previous studies and interpretation of results

We found a higher use of pain medication among children and adolescents with more severe or long-lasting MSK pain. This aligns with previous studies reporting that pain medication is the first line of treatment for chronic pediatric MSK pain [17]. Previous reviews on the use of medication in children and adolescents focused mainly on other pain conditions or symptoms (e.g. headache, allergies, fever, cold and cough, insomnia, skin problems and menstruation) [22, 23, 58], which limits comparison. These reviews report that 2–92% of children and adolescents self-medicate for these conditions [22, 58]. Importantly, these studies found that factors such as female gender, older age, maternal education and familial practices were associated with self-medication suggesting that children and adolescents with certain traits are more likely to self-medicate across different conditions [22]. This finding could be interpreted as social learning involved in self-medication as previous research have shown that mothers’ use of medication was associated with their children’s use [23, 59]. In our review we found mixed evidence for increasing age and female gender, although most studies reported that these factors were associated with use of pain medication, and future studies are needed to confirm this association. If evidence of association will be found, the higher use of pain medication in girls might be explained by use due to the management of menstrual pain [60]. In addition, older adolescents might tend to use pain medication independently because they want to engage in social activities despite their pain [53, 61] and, depending on the regulations on sales of pain medication of specific countries, might be able to buy pain medication themselves without the parental control [62, 63].

The higher prevalence of pain medication use found in populations regularly engaging in sport activities is in agreement with previous findings in adult athletes [64]. This could be explained by cultural factors shared by sport participants (e.g. pain medication are used to win the competition or not to lose the role in the team) [64]. This is supported by the fact that 84% of swimmers in one study [43] agreed with the statement: “Taking time off of swimming due to an injury is not a practical option if I want to succeed at a high level.” [43]. Self-medication may be one of the strong drivers of the high use. The severity of self-medication is supported by Iacovelli et al. [45] who showed that 63–80% of adolescents used analgesic for their back pain without a doctor’s prescription.

Three American studies reported on opioids use in the management of MSK pain. In one study 21–44% of adolescents with joint or back/neck pain were given an opioid prescription [51], in another study 13% of adolescents with low back pain were prescribed opioids [38], while in another study 4.7–18.3% of opioid prescriptions were for back or other musculoskeletal pain [40]. This is a significant concern as opioids are currently not recommended for management of chronic MSK pain and opioid use is associated with significant risks for the patients [65, 66]. Since both studies were performed in the U.S.A., these findings may be due to cultural and healthcare system differences between the U.S.A. and most of the OECD countries (with safer opioid regulations in the European Union [67]), along with the opioid epidemic in the U.S.A. [66, 68], [69], [70].

Strengths and limitations of this study

We took several steps to ensure the quality of the present systematic review. Firstly, three different databases covering the majority of the published health-science literature were systematically searched without any time restriction and including the most common languages for publication. In addition, although it could be argued that the Pubmed or Medline databases were not searched, the EMBASE database provides the same coverage with an additional focus on drugs and pharmacology. Secondly, we focused on study designs which were compatible with the provision of information on the prevalence of use of pain medication and excluded designs which could have provided a biased estimate (i.e. RCT). Thirdly, no restriction was applied to the study settings. This enabled us to observe the prevalence of pain medication use in a broad setting and among populations with different grades of MSK pain severity (e.g. schools vs. pain clinics) or which might be more likely to experience MSK pain due to overuse (e.g. sport clubs). Fourthly, the inclusion of studies from different countries and continents allowed us to obtain an understanding on differences in use of pain medication (and type) for MSK pain. The use of pain medication may be influenced by cultural factors (pain perception and reporting as well as the prescription of pain medication varies across countries and ethnicities [71], [72], [73]) and legislations applied in the specific countries (especially regarding the availability of over-the-counter and opioid pain medication) [67]. Lastly, the majority of studies included had a high number of participants or a response rate >70%, which supports generalizability of results.

This review also presents some limitations. Different methods were used for the assessment of MSK pain (e.g. intervals considered for the prevalence, frequency, duration of pain), of pain medication (e.g. prevalence, type, frequency) and the recall period considered between studies. Therefore, recall bias cannot be excluded [74]. This may also limit comparability of the results between studies. Due to the variability in how MSK pain was measured and lack of information on pain medication we could not perform a meta-analysis to create a pooled estimate. The variability in data collection in the original studies also prevented us from subgroup analysis to explore the association between duration of pain and use of pain medication. This review focused on studies reporting on the prevalence of pain medication use among children and adolescents with MSK pain as a primary aim of the study or that reported this information in the abstract. Therefore, it is possible that other studies that did not focus on pain medication as a primary outcome but reported such information in the text or tables were not included. However, it is likely that such studies would report a lower quality information on pain medication use as this was not the main objective of those studies.

Implication for future research

This review showed that pain medication is commonly used in children and adolescents to manage MSK pain (not related to injuries, as studies reporting on acute pain resulting from injuries were not included). However, we found a lack of information on the duration of use and the side effects of pain medication, especially regarding paracetamol and NSAIDs. Future studies should strive to collect detailed information on the use which will allow to differentiate between occasional use of pain medication for MSK pain of shorter duration, which might be safe, and regular and long-lasting use that might occur for MSK pain of a longer duration. Regular and long-lasting use of pain medication might result in side effects, as it has been shown in adults [75], but there is little information available from the best evidence in children and adolescents [17], [18], [19], [20], [21]. Therefore, future dependencies, overuse, misuse and potential health hazardous consequences of the use of pain medication in children and adolescents should be investigated and collected. Other information that should be collected in future studies include the use of medication for concurrent comorbidities (e.g. headache, menstrual pain) and the gateway to pain medication (e.g. sport coach, family, friends, general practitioner) [22, 76, 77]. Future studies should validate which factors are associated with the use of pain medication (e.g. bullying, being a member of a sport club, self-esteem). This might lead to the assessment and targeting of specific factors (e.g. psychological factors), which might be currently overlooked in daily practice.