Open access Original research Childhood overweight and obesity and back pain risk: a cohort study of 466 997 children Antony J Palmer,1 Jose Luis Poveda,2,3,4 Daniel Martinez-­Laguna,2,3,4 Carlen Reyes,2,3,4 Jeroen de Bont,5 Alan Silman,1,6 Andrew J Carr,1,6 Talita Duarte-­Salles ‍ ‍,5 Daniel Prieto-­Alhambra ‍ ‍1,6 To cite: Palmer AJ, Poveda JL, ABSTRACT Martinez-­Laguna D, et al. Strengths and limitations of this study Objectives To assess the association between age, sex, Childhood overweight and socioeconomic group, weight status and back pain risk in obesity and back pain ►► Large dataset of nearly half a million children. a large general population cohort of children. risk: a cohort study of 466 ►► Generalisable outcome of back pain: when parent Design and setting A dynamic cohort of children aged 4 997 children. BMJ Open believes pain is sufficiently severe to seek input 2020;10:e036023. doi:10.1136/ years in the Information System for Research in Primary from healthcare professionals. bmjopen-2019-036023 Care (SIDIAP) electronic primary care records data in ►► Study does not address psychometric factors that Catalonia. Multivariable Cox models were fitted to explore ►► Prepublication history and play important role in the aetiology of back pain. the association between back pain and weight status additional material for this ►► Long-­term follow-­up into adulthood not available. categories according to the WHO 2007 growth reference paper are available online. To groups (body mass index for age z-­score). Models view these files, please visit the journal online (http://​dx.​doi.​ were adjusted for age, sex, socioeconomic status and org/​10.​1136/​bmjopen-​2019-​ nationality. reported in the cervical, thoracic and lumbar 036023). Participants Children seen at age 4 years at paediatric spine,5 and prevalence increases with age,6 primary care clinics between 1 January 2006 and 31 An abstract entitled ‘Childhood reaching levels comparable with adult popu- obesity is associated with back December 2013 and followed up until 31 December 2016 or age 15 years. lations by 18 years.7 Most studies assess self-­ pain: a cohort study of 466,997 individuals’ was presented Outcome measures Incident back pain registered by reported pain, and the number of children at the International Society paediatricians at primary care using the International and adolescents consulting healthcare profes- for Pharmacoeconomics and Statistical Classification of Diseases and Health Related sionals with back pain is not known. Outcomes Research Europe Problems, 10th Edition code M54. There is very limited evidence for primary 2018 conference in Barcelona on the 10–14 November 2019 Results 466 997 children were followed for a median prevention of back pain.8 In children, only as a poster presentation. The 5.0 years (IQR 5.1). In multivariable models, overweight education and ergonomic school furni- abstract was published in Value and obesity increased back pain risk, with adjusted HRs of ture have been studied but without robust in Health, October 2018, Volume 1.18 (95% CI 1.09 to 1.27) and 1.34 (95%CI 1.19 to 1.51) evidence of benefit.9 Although interventions 21, Supplement 3, Page S290, for overweight and obesity, respectively. Females were at DOI: https://​doi.​org/​10.​1016/​j.​ that incorporate exercise are effective in greater risk of back pain than males with adjusted HR 1.40 jval.​2018.​09.​1729. (95%CI 1.35 to 1.46). Adjusted HR was 1.43 (95%CI 1.33 adults, these have not been investigated in to 1.55) for back pain in children from the most deprived children.8 Interventions are most likely to be Received 27 November 2019 Revised 15 May 2020 socioeconomic groups compared with the least deprived successful when targeting specific risk factors Accepted 13 July 2020 socioeconomic groups. in populations predisposed to developing Conclusions Maintaining a healthy weight from an back pain; however, predictive factors are early age may reduce the prevalence of back pain in currently not well characterised.10 both children and adults. Overweight female children The prevalence of childhood obesity and from deprived socioeconomic groups are at greatest back pain are both increasing.7 11 12 Crucially, risk of back pain and represent a target population for back pain during youth predicts pain in intervention. © Author(s) (or their adulthood,13 14 and obesity during childhood employer(s)) 2020. Re-­use predicts obesity in adolescence.15 Although permitted under CC BY-­NC. No commercial re-­use. See rights INTRODUCTION obesity is associated with chronic back pain and permissions. Published by Back pain is the greatest cause of disability in adults,1 there is only limited evidence of BMJ. worldwide, and this burden continues to an association during childhood.16–18 In addi- For numbered affiliations see increase in magnitude.1 It is common among tion, the increased prevalence of back pain end of article. children and adolescents and may affect up in females from lower socioeconomic groups Correspondence to to three quarters of individuals before adult- in adults1 has not been consistently repro- Dr Talita Duarte-­Salles; hood.2 Childhood back pain results in activity duced in adolescent cohorts.7 Childhood ​tduarte@​idiapjgol.​org limitation and school absenteeism.3 4 Pain is obesity may represent a modifiable risk factor Palmer AJ, et al. BMJ Open 2020;10:e036023. doi:10.1136/bmjopen-2019-036023 1 Open access for back pain, and weight optimisation confers benefits Population beyond musculoskeletal health. The cohort included all children assessed at age 4 years The aims of this study were: (1) to determine the inci- (±6 months) by a paediatric doctor or nurse in any of the dence of community paediatrician consultations due to contributing primary care centres between 1 January 2006 back pain and (w) to explore the association between and 31 December 2016 with at least one valid height and back pain and age, sex, socioeconomic group and weight weight measurement recorded in the SIDIAP database status in a large general population cohort of children. (figure 1). These assessments represent a routine paedi- atric health surveillance programme within the region. Study design METHODS A dynamic cohort study was conducted with longitu- Data source dinal data extraction for all participants from their index Data were obtained from anonymised electronic medical appointment (age 4 years old±6 months) until the earliest records from the Sistema d’Informació pel Desenvolupa- of: the date they reached 15 years of age (when individ- uals transition to adult care), migration away from a ment de l’Investigació en Atenció Primària (SIDIAP; www.​ Catalan Institute of Health centre, death or end of study sidiap.​org) (Information System for Research in Primary (31 December 2016). Care). The database includes longitudinal data on patient demographics, clinical visits to primary care centres and Study exposures and outcomes International Statistical Classification of Diseases and Recorded exposures included patient demographics Health Related Problems, 10th Edition (ICD-10) coded measured at the index appointment of a screening health problems from 296 primary care centres in Cata- programme: age (months), sex, height (metres), weight lonia. Data are available for greater than 80% of the 7.5 (kg), socioeconomic group and nationality. Age and sex million population of Catalonia, totalling more than 1.1 referenced BMI (body mass index for age z-­score (BMIz)) million children under the age of 15 years. was calculated according to the WHO 2007 guidance, Figure 1 Study population flow chart. BMI, body mass index; SIDIAP the Information System for Research in Primary Care. 2 Palmer AJ, et al. BMJ Open 2020;10:e036023. doi:10.1136/bmjopen-2019-036023 Open access and children were categorised as ‘underweight’, ‘normal RESULTS weight’, ‘overweight’ or ‘obese’.19 These weight status cate- A total of 466 997 children were included in the study gories represent the number of SD away from the mean: with median follow-­up 5.03 years (IQR 5.11). A total of 21 underweight <−2, normal weight −2 to +2, overweight >+2 638 and 150 children were transferred out or died during and obesity >+3. Biologically implausible values of BMIz follow-­up, respectively. Participant demographics at the were excluded according to WHO Expert Committee 1995 initial assessment, stratified by baseline WHO weight cate- guidance, and implausible values in growth trajectory data gories (BMIz) are summarised in table 1. up to the age of 10 years were identified using a conditional The 11-­year cumulative incidence of individuals diag- growth percentile model.20 Socioeconomic group was nosed with back pain at any site by a community paedi- measured using the MEDEA (Mortalidad en áreas pequeñas atrician increased with age and reached 8.93% (95% CI Españolas y Desigualdades Socioeconómicas y Ambientales) 8.61% to 9.25%) by age 15 years (figure 2 and online deprivation index calculated using five census-­based socio- supplemental figure 1). economic indicators: unemployment, manual occupation, Back pain was most frequently experienced in the temporary employment, illiterate adults and school leavers lumbar region and the cumulative incidence during before 16 years of age. Group 1 represents the least depri- childhood (age 4–15 years old) at different sites was 4.33% (95% CI 4.08% to 4.58%) for lumbar pain, 2.46% vation and group 5 the most deprivation. MEDEA catego- (95% CI 2.29% to 2.63%) for thoracic pain and 2.86% ries also include rural communities and unclassified urban (95% CI 2.71% to 3.01%) for cervical pain (table 2). communities. Cumulative risk was higher in females (10.31% (95% CI Outcomes were ascertained from diagnoses recorded 9.82% to 10.80%)]) than males (7.62% (95% CI 7.20% in primary care paediatric records during visits to to 8.03%)) and increased with weight category: under- contributing health centres between 1 January 2006 and weight 4.08% (95% CI 1.20% to 6.95%), normal weight 31 December 2016 in the Catalan universal healthcare 8.78% (95% CI 8.45% to 9.12%), overweight 10.14% system. Prespecified ICD-10 codes were used to identify (95% CI 8.89% to 11.39%) and obese 11.83% (95% CI first episodes of back pain after 4 years of age: lumbar 9.43% to 14.24%). spine pain (M54.1, M54.3, M54.4 or M54.5), thoracic Individuals classified as overweight or obese had an back pain (M54, M54.6, M54.8 or M54.9) or cervical increased risk of developing back pain at any site. This spine pain (M54.2). association remains present when adjusting for age, sex, socioeconomic group and nationality. In the multivariable Statistical analysis cox regression model, low socioeconomic group, female Population demographics were expressed as frequencies sex and non-­Spanish nationality all increased the risk of and percentages for categorical data and means and SD back pain (table 3). There was no interaction between sex or medians and IQR for continuous data where appro- and weight categories (p=0.86). priate. Cumulative incidence was calculated using actu- The association between weight category and back pain arial lifetables for first time event of a recorded code of was maintained when stratifying for lumbar, thoracic and interest (lumbar, thoracic or cervical spine pain) and cervical spine pain (online supplemental tables 1-3). stratified for weight category. Results were tabulated and Obese children had an increased risk of developing pain expressed as Kaplan-­Meier plots. in all areas of the back compared with individuals with Multivariable Cox models were used to explore the asso- a normal weight status. In each of the lumbar, thoracic ciation between back pain and weight categories (BMIz), and lumbar regions, females remained at greatest risk of adjusted for age, sex, nationality (Spanish or other) and back pain, as did individuals in the most deprived socio- socioeconomic group (MEDEA). Weight categories and economic groups compared with the least deprived socio- socioeconomic status groups were treated as factor vari- economic groups. ables. Age was treated as a continuous variable. MEDEA Birth weight was available for 310 772 (66.55%) of the participants, and inclusion as a covariable in the cox scores were categorised into quintiles to minimise risk of regression model did not modify the observed associa- reidentification. tions (data not shown). Birth weight was not associated A prespecified interaction between weight category and with back pain other than in the cervical region, with sex was tested by introducing a multiplicative term in the adjusted HR of 1.10 (95% CI 1.02 to 1.20) per kg increase Cox equation. Sex-­stratified results are reported where in weight at birth (online supplemental table 4). the p value for interaction appeared borderline signifi- cant (p<0.1). A sensitivity analysis was conducted making further adjustment for birth weight as a marker of genetic DISCUSSION and/or in utero conditioning of adiposity. Almost 1 in 10 children develop at least one episode of back pain during childhood, and lower back (lumbar) Patient and public involvement pain is the most common site. This study demonstrates There was no patient and public involvement in this that children who are overweight or obese at an early age study. have an increased risk of developing back pain during Palmer AJ, et al. BMJ Open 2020;10:e036023. doi:10.1136/bmjopen-2019-036023 3 4 Table 1 Baseline characteristics of the included and excluded study population Excluded Included (no BMI available at age 48±6 months) Normal Normal Total weight Overweight Obese Underweight Total weight Overweight Obese Underweight Open access Participants  N (%) 466 997 430 681 26 526 9 250 540 336 924 278 030 37 052 19 546 2 296 (92.2) (5.7) (2.0) (0.1) (82.5) (11.0) (5.8) (0.7) BMIz*  Mean (SD) 0.46 0.80 2.40 3.66 −3.51 0.43 0.11 1.72 2.91 −2.71 (1.09) (0.88) (0.28) (0.51) (0.47) (1.17) (0.87) (0.50) (0.73) (0.64) Age in months  Mean (SD) 49.13 49.13 49.13 49.15 49.06 54.30 49,00 76,69 83.53 85.72 (2.00) (1.99) (2.01) (2.11) (2.50) (36.61) (34.58) (36,85) (31.80) (39,24) Age in months  Female N (%) 226 868 210 274 12 455 3891 248 162 842 136 061 17 981 7771 1029 (48.58) (48.8) (47.0) (42.1) (45.9) (48.3) (48.9) (39.8) (48.5) (44.8) Socioeconomic status  Rural N (%) 94 825 88 571 4 666 1 488 100 59 418 49 898 6110 3006 404 (20.3) (93.4) (4.9) (1.6) (0.1) (17.6) (84.0) (10.3) (5.1) (0.7)  Urban quintile 1 (least 57 439 53 877 2 750 761 51 50 398 42 208 5557 2284 349 deprived area) (12.3) (93.8) (4.8) (1.3) (0.1) (15.0) (83.7) (11.0) (4.5) (0.7)  Urban quintile 2 68 720 63 799 3 715 1 126 80 47 392 38 886 5471 2734 301 (14.7) (92.8) (5.4) (1.6) (0.1) (14.1) (82.1) (11.5) (5.8) (0.6)  Urban quintile 3 71 416 65 660 4 170 1 507 79 47 313 38 274 5564 3145 330 (15.3) (91.9) (5.8) (2.1) (0.1) (14.0) (80.9) (11.8) (6.6) (0.7)  Urban quintile 4 72 801 66 407 4 569 1 746 79 49 127 39 808 5755 3239 325 (15.6) (91.2) (6.3) (2.4) (0.1) (14.6) (81.0) (11.7) (6.6) (0.7)  Urban quintile 5 77 129 69 581 5 259 2 172 117 55 468 44 928 6134 3962 444  (most deprived area) (16.5) (90.2) (6.8) (2.8) (0.2) (16.5) (81.0) (11.1) (7.1) (0.8)  Missing address 24 667 22 786 1 397 450 34 27 808 24 028 2461 1176 143 (5.3) (92.4) (5.7) (1.8) (0.1) (8.3) (86.4) (8.8) (4.2) (0.5) Nationality  Spanish N (%) 415 829 383 739 23 568 8083 439 272 974 225 915 29 833 15 669 1557 (89.04) (89.10) (88.80) (87.40) (81.30) (81.0) (81.3) (80.5) (80.2) (67.8)  Other N (%) 51 168 46 942 2958 1167 101 63 950 52 115 7219 3877 739 (10.96) (10.90) (11.20) (12.6) (18.70) (19.0) (18.7) (19.5) (19.8) (32.2) *BMIz measured age 48±6 months for included participants and measurement closest to age 4 (but <3.5 and >4.5) years available for those included. BMIz, body mass index for age z-­score. Palmer AJ, et al. BMJ Open 2020;10:e036023. doi:10.1136/bmjopen-2019-036023 Open access lifetime prevalence of 40% (95% CI 34% to 54%) in 30 studies.7 The large variation in reported prevalence most likely reflects the different tools used to diagnose back pain, which in most studies is self-­reported, as well as the population characteristics. Studies to date have recruited individuals from educational or community institutions. In our study, back pain was diagnosed by community paediatricians and only includes children with pain suffi- ciently severe to seek input from healthcare services. Existing literature reports inconsistent associations between childhood back pain and BMI.21 22 A systematic review of studies including 5567 participants categorised into normal weight or overweight, concluded that over- weight children are at increased risk of lower back pain with a risk ratio of 1.42 (95% CI 1.03 to 1.97); however, the included studies offer only low-­quality evidence.17 Our results demonstrate that children who are overweight or Figure 2 Age-­specific cumulative incidence (95% CI) of obese have an increased risk of developing back pain in back pain at any site. the lumbar, thoracic and cervical spine, after adjustment for confounding factors. Furthermore, there is a sugges- tion of a dose–response relationship between back pain childhood. The cumulative incidence of back pain diag- and weight categories. nosed by community paediatricians increased with age As with weight categories, there is conflicting evidence and was almost 9% in children with a normal weight, regarding the association between sex and childhood compared with almost 12% for obese children. Adjusting back pain. Some studies report no difference between for age, sex, socioeconomic status and nationality, a child sexes22–24 and others report a higher prevalence in with obesity has a 34% excess risk of developing back pain females.25–27 A meta-­analysis of 27 studies revealed no before the age 15 years. Female children are 40% more statistically significant relationship between sex as life- likely to develop back pain than males, and children from time prevalence.7 Our study demonstrates a higher prev- most deprived socioeconomic groups are at a more than alence of back pain in females, as demonstrated in adult 40% higher risk of developing back pain than those of the cohorts,28 and was the case for the lumbar, thoracic and least deprived socioeconomic groups. cervical spine. The incidence of back pain in this study is lower than in Adults with chronic back pain are more likely to other cohorts, where reported lifetime prevalence during have received less education and have lower house- adolescence ranges from 4.7% to 74.4%2 with a mean hold income.29 Children of mothers with no educa- tional qualifications are more likely to report frequent Table 2 Cumulative incidence of back pain stratified by complaints of headache, stomach ache or sickness than age, sex and site of pain children of mothers a higher degree (risk ratio of 2.06, 95% CI 1.64 to 2.59).30 Studies have not found an associ- Cumulative incidence (%) 95% CI ation between socioeconomic status and the prevalence of childhood back pain,31–34 although there is evidence All participants 8.93 8.62 to 9.25 that childhood socioeconomic status may affect the Sex risk of developing back pain as an adult.35 Our cohort  Female 10.31 9.82 to 10.80 demonstrated a clear association between ecological  Male 7.62 7.20 to 8.03 socioeconomic status and childhood back pain with a monotonic association, and studies suggest this trend BMIz categories continues into adulthood.35  Underweight 4.08 1.20 to 6.95 Study strengths include the large dataset of nearly half  Normal weight 8.78 8.45 to 9.12 a million children, which to our knowledge is the largest  Overweight 10.14 8.89 to 11.39 study exploring risk factors for childhood back pain. The  Obese 11.83 9.43 to 14.24 outcome is a standardised measure of back pain, which is when a child or their parent believes the back pain is Site of pain sufficiently severe to seek input from healthcare services.  Cervical 2.86 2.71 to 3.01 Diagnosis is then confirmed by a community paediatri-  Thoracic 2.46 2.29 to 2.63 cian. The diagnosis of back pain in this cohort is there-  Lumbar 4.33 4.08 to 4.58 fore likely to represent a severe episode of pain, although there are no data on the duration, relief or aggravation BMIz, body mass index for age z-­score. of symptoms. Palmer AJ, et al. BMJ Open 2020;10:e036023. doi:10.1136/bmjopen-2019-036023 5 Open access Table 3 Crude and multivariable Cox regression models with HRs for development of back pain at any site Univariate Multivariable* HR 95% CI HR 95% CI Age  Months 1.02 1.01 to 1.03 1.02 1.01 to 1.03 Sex  Male Ref – – –  Female 1.40 1.35 to 1.46 1.40 1.35 to 1.46 BMIz categories  Underweight 0.78 0.41 to 1.50 0.78 0.41 to 1.51  Normal range Ref – – –  Overweight 1.18 1.10 to 1.27 1.18 1.09 to 1.27  Obese 1.34 1.20 to 1.51 1.34 1.19 to 1.51 Socioeconomic status  Rural area 1.29 1.20 to 1.38 1.29 1.20 to 1.39  Urban quintile 1 Ref – – –  Urban quintile 2 1.16 1.07 to 1.25 1.16 1.07 to 1.25  Urban quintile 3 1.28 1.18 to 1.38 1.27 1.18 to 1.38  Urban quintile 4 1.32 1.23 to 1.43 1.33 1.23 to 1.44  Urban quintile 5 1.41 1.31 to 1.52 1.43 1.33 to 1.55  Urban not classified 1.32 1.19 to 1.47 1.32 1.18 to 1.47 Nationality  Spanish Ref – – –  Other 0.75 0.69 to 0.82 0.73 0.67 to 0.79 *Adjusted for all the variables reported on in this same table. BMIz, body mass index for age z-­score. Most epidemiological studies use patient reported obesity in Europe.37 There were an insufficient number of outcome measures to diagnose back pain; however, this underweight participants to draw conclusions regarding does not evaluate access to healthcare services, which is back pain in this population group, where there is likely key from a health economics perspective. The compre- to be a higher proportion of patients with systemic hensive SIDIAP database makes it possible to adjust for medical conditions. Also, a higher proportion of partic- salient confounding variables, identified as age, sex, ipants were overweight or obese in participants without socioeconomic status and nationality, which confirmed a valid BMIz at 4-­year follow-­up assessment. An area not overweight or obese weight categories are indepen- addressed by this study is whether children with back pain dent risk factors for back pain. Despite adjustment for at one site are more likely to also present with back pain confounding factors, there may be other important factors at a different site. There may also be variation in how that have not been considered. Psychometric factors clinicians code back pain, particularly when anatomically play an important role in the aetiology of back pain and differentiating the lumbar, thoracic and cervical regions. have not been evaluated in this study,21 nor have physical Obesity is associated with back pain in adults,38 39 and activity levels or ergonomics of child’s environment. In our study demonstrates that this association is already addition, weight categories were established based on age evident during childhood. Critically, childhood back pain and sex referenced BMI using WHO guidance19 but may has been shown to predict back pain in adults,13 14 poten- not accurately reflect cardiometabolic health.36 tially after the initiation of pathology or persistence of Long-­term follow-­up of this cohort into adulthood is risk factors.40 Our study does not quantify the duration of not currently available. While this study demonstrates an symptoms but chronic pain may continue into adulthood. association between weight status and back pain during A longitudinal study of the British Birth Cohort suggests childhood, it does not provide evidence of causation. back pain is unlikely to be a cause of obesity given the Furthermore, it is not known whether the findings in absence of a temporal association.24 Reasons for the this cohort are representative of other populations, such increased risk of back pain in overweight or obese chil- as children from USA, where there are higher rates of dren are not known, but the aetiology may be mechan- obesity; however, Spain has among the highest rates of ical, biological or psychological.1 The absence of an 6 Palmer AJ, et al. BMJ Open 2020;10:e036023. doi:10.1136/bmjopen-2019-036023 Open access association between birth weight and back pain suggests Contributors JLP, DM-­L, JdB, TD-­S and DP-­A designed the study. JLP and DP-­A it is weight gain at a young age that increases the risk of performed the statistical analysis. AJP and DP-­A drafted the manuscript. All authors contributed to the data interpretation, revised manuscript drafts, approved the back pain. Crucially, strategies to reduce the prevalence final version and contributed intellectually important content. DP-­A attests that all of back pain in adulthood may be most effective when authors meet authorship requirement and that no others meeting the criteria have commenced during childhood. been omitted. DP-­A is the guarantor of the paper and takes responsibility for the There is very limited evidence for primary prevention of integrity of the work from inception to published article. back pain, particularly in children.8 In adolescent popu- Funding Funding was received from La Marató de TV3 Foundation (Grant Number: 2 01 621–30). TD-­S is funded by the Department of Health of the Generalitat lations, only education and ergonomic school furniture de Catalunya (Strategic Plan for Research and Innovation in Health (PERIS) have been studied. There is moderate quality evidence 2016–2020 Modality Incorporation of Scientists and Technologists. Grant Number: that education is not effective and very low quality SLT002/16/00308). DP-­A is funded by a National Institute for Health Research evidence that ergonomically designed furniture could Clinician Scientist award (CS-2013-13-012). AJP is funded by a National Institute for Health Research Academic Clinical Lectureship. The research was supported prevent back pain compared with conventional furni- by the National Institute for Health Research (NIHR) Oxford Biomedical Research ture.9 Exercise, particularly when combined with educa- Centre. tion, can prevent back pain in adulthood and warrants Disclaimer The views expressed are those of the author(s) and not necessarily investigation in child populations.8 those of the NHS, the NIHR or the Department of Health. The highest risk population for developing back pain Competing interests All authors have completed the ICMJE uniform disclosure is obese females from deprived socioeconomic groups. form at www.​icmje.​org/​coi_​disclosure.​pdf. No conflicts of interest are declared These individuals should be the priority target for inter- in relation to the submitted work. Outside of the submitted work, DM-­L receives personal fees from Amgen, Lilly, Novartis, Ferrer and Rubió. DP-­A received research vention. Tacking childhood obesity may reduce the grants from Les Laboratoires Servier, Amgen and UCB Biopharma SRL. DP-­A also prevalence of back pain both in children and adults, in received advisory, speaker and consultancy fees paid to his department from UCB addition to conferring other health benefits. The role for and Amgen. early intervention is further supported by a recent study Patient consent for publication Not required. demonstrating that the most rapid gain in weight during Ethics approval Scientific approval was obtained from the SIDIAP Scientific childhood is observed between 2 and 6 years of age.15 An Committee and ethical approval was granted by CEIC Idiap Jordi Gol (reference improved understanding of how socioeconomic factors number P16/179). influence back pain and obesity may facilitate the devel- Provenance and peer review Not commissioned; externally peer reviewed. opment of novel preventative strategies. In addition to Data availability statement Data may be obtained from a third party and are not physical activity levels, diet and environmental exposures, publicly available. In accordance with current European and national law, the data cultural beliefs and differing thresholds for consulting used in this study are only available for the researchers participating in this study. Thus, we are not allowed to distribute or make publicly available the data to other healthcare practitioners by socioeconomic group may parties. However, researchers from public institutions can request data from SIDIAP play a role.41 The additional observation that a higher and other sources (eg, Cancer Registries) if they comply with certain requirements. proportion of underweight individuals are of non-­Spanish Further information is available online (https://www.​sidiap.​org/​index.​php/​menu-​ nationality may have implications for migrant health and solicitudes-​en/​application-​proccedure) or by contacting Anna Moleras (​amoleras@​ idiapjgol.​org). warrants further investigation. Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-­reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and CONCLUSION responsibility arising from any reliance placed on the content. Where the content Children who are overweight or obese are at a significantly includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, increased risk of developing back pain. The incidence of terminology, drug names and drug dosages), and is not responsible for any error back pain increases with age and is highest in overweight and/or omissions arising from translation and adaptation or otherwise. females from deprived socioeconomic groups. Given Open access This is an open access article distributed in accordance with the childhood back pain predicts back pain in adulthood, it is Creative Commons Attribution Non Commercial (CC BY-­NC 4.0) license, which recommended that preventative strategies commence at permits others to distribute, remix, adapt, build upon this work non-­commercially, an early age. Strategies should include ensuring children and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use maintain a normal healthy weight. is non-­commercial. See: http://​creativecommons.​org/​licenses/​by-​nc/​4.​0/. Author affiliations ORCID iDs 1 Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Talita Duarte-­Salles http://​orcid.​org/​0000-​0002-​8274-​0357 Sciences, University of Oxford, Oxford, UK Daniel Prieto-­Alhambra http://​orcid.​org/​0000-​0002-​3950-​6346 2 GREMPAL Research Group, Fundació Institut Universitari per a la recerca a l'Atenció Primària de Salut Jordi Gol i Gurina (IDIAPJGol) and CIBERFes, Barcelona, Spain REFERENCES 3 Universitat Autonoma de Barcelona, Barcelona, Spain 1 Hartvigsen J, Hancock MJ, Kongsted A, et al. What low back pain is 4 Instituto de Salud Carlos III, Barcelona, Spain and why we need to pay attention. Lancet 2018;391:2356–67. 5 Fundació Institut Universitari per a la recerca a l'Atenció Primària de Salut Jordi 2 Jeffries LJ, Milanese SF, Grimmer-­Somers KA. Epidemiology of Gol i Gurina (IDIAPJGol), Barcelona, Spain adolescent spinal pain: a systematic overview of the research 6 literature. Spine 2007;32:2630–7. NIHR Oxford Biomedical Research Centre, John Radcliffe Hospital, Oxford, UK 3 Skoffer B. Low back pain in 15- to 16-­year-­old children in relation to school furniture and carrying of the school bag. Spine Twitter Antony J Palmer @AJRPalmer and Talita Duarte-­Salles @TDuarte_Salles 2007;32:E713–7. Palmer AJ, et al. BMJ Open 2020;10:e036023. doi:10.1136/bmjopen-2019-036023 7 Open access 4 Bejia I, Abid N, Ben Salem K, et al. Low back pain in a cohort of 622 23 Petersen S, Bergström E, Brulin C. High prevalence of tiredness and Tunisian schoolchildren and adolescents: an epidemiological study. pain in young schoolchildren. Scand J Public Health 2003;31:367–74. Eur Spine J 2005;14:331–6. 24 Lake JK, Power C, Cole TJ. Back pain and obesity in the 5 Aartun E, Hartvigsen J, Wedderkopp N, et al. Spinal pain in 1958 British birth cohort. cause or effect? J Clin Epidemiol adolescents: prevalence, incidence, and course: a school-­based 2000;53:245–50. two-­year prospective cohort study in 1,300 Danes aged 11-13. BMC 25 Watson KD, Papageorgiou AC, Jones GT, et al. Low back Musculoskelet Disord 2014;15:187. pain in schoolchildren: occurrence and characteristics. Pain 6 MacDonald J, Stuart E, Rodenberg R. Musculoskeletal low back pain 2002;97:87–92. in school-­aged children: a review. JAMA Pediatr 2017;171:280–7. 26 Stanford EA, Chambers CT, Biesanz JC, et al. The frequency, 7 Calvo-­Muñoz I, Gómez-­Conesa A, Sánchez-­Meca J. Prevalence of trajectories and predictors of adolescent recurrent pain: a low back pain in children and adolescents: a meta-­analysis. BMC population-­based approach. Pain 2008;138:11–21. Pediatr 2013;13:14. 27 Kovacs FM, Gestoso M, Gil del Real MT, et al. Risk factors for 8 Foster NE, Anema JR, Cherkin D, et al. Prevention and treatment non-­specific low back pain in schoolchildren and their parents: a of low back pain: evidence, challenges, and promising directions. population based study. Pain 2003;103:259–68. Lancet 2018;391:2368–83. 28 Hoy D, Bain C, Williams G, et al. A systematic review of the global 9 Michaleff ZA, Kamper SJ, Maher CG, et al. Low back pain in prevalence of low back pain. Arthritis Rheum 2012;64:2028–37. children and adolescents: a systematic review and meta-­analysis 29 Shmagel A, Foley R, Ibrahim H. Epidemiology of chronic low back evaluating the effectiveness of conservative interventions. Eur Spine pain in US adults: data from the 2009-2010 National health and J 2014;23:2046–58. nutrition examination survey. Arthritis Care Res 2016;68:1688–94. 10 Hill JJ, Keating JL. Risk factors for the first episode of low back pain 30 Fryer BA, Cleary G, Wickham SL, et al. Effect of socioeconomic in children are infrequently validated across samples and conditions: conditions on frequent complaints of pain in children: findings a systematic review. J Physiother 2010;56:237–44. from the UK millennium cohort study. BMJ Paediatr Open 11 Skinner AC, Ravanbakht SN, Skelton JA, et al. Prevalence of 2017;1:e000093. obesity and severe obesity in US children, 1999-2016. Pediatrics 31 Kristjánsdóttir G. Prevalence of self-­reported back pain in school 2018;141:e20173459. children: a study of sociodemographic differences. Eur J Pediatr 12 Cunningham SA, Kramer MR, Narayan KMV. Incidence of childhood 1996;155:984–6. obesity in the United States. N Engl J Med 2014;370:403–11. 32 Hestbaek L, Korsholm L, Leboeuf-­Yde C, et al. Does socioeconomic 13 Brattberg G. Do pain problems in young school children persist into status in adolescence predict low back pain in adulthood? A early adulthood? A 13-­year follow-­up. Eur J Pain 2004;8:187–99. repeated cross-­sectional study of 4,771 Danish adolescents. Eur 14 Hestbaek L, Leboeuf-­Yde C, Kyvik KO, et al. The course of low back Spine J 2008;17:1727–34. pain from adolescence to adulthood: eight-­year follow-­up of 9600 33 Jørgensen MB, Nabe-­Nielsen K, Clausen T, et al. Independent twins. Spine 2006;31:468–72. effect of physical workload and childhood socioeconomic status 15 Geserick M, Vogel M, Gausche R, et al. Acceleration of BMI on low back pain among health care workers in Denmark. Spine in early childhood and risk of sustained obesity. N Engl J Med 2018;379:1303–12. 2013;38:E359–66. 16 Krul M, van der Wouden JC, Schellevis FG, et al. Musculoskeletal 34 Power C, Frank J, Hertzman C, et al. Predictors of low back problems in overweight and obese children. Ann Fam Med pain onset in a prospective British study. Am J Public Health 2009;7:352–6. 2001;91:1671–8. 17 Paulis WD, Silva S, Koes BW, et al. Overweight and obesity are 35 Lallukka T, Viikari-­Juntura E, Raitakari OT, et al. Childhood and adult associated with musculoskeletal complaints as early as childhood: a socio-­economic position and social mobility as determinants of low systematic review. Obes Rev 2014;15:52–67. back pain outcomes. Eur J Pain 2014;18:128–38. 18 Hestbaek L, Leboeuf-­Yde C, Kyvik KO. Are lifestyle-­factors in 36 Tomiyama AJ, Hunger JM, Nguyen-­Cuu J, et al. Misclassification of adolescence predictors for adult low back pain? A cross-­sectional cardiometabolic health when using body mass index categories in and prospective study of young twins. BMC Musculoskelet Disord NHANES 2005-2012. Int J Obes 2016;40:883–6. 2006;7:27. 37 World Health Organisation Europe. Childhood Obsesity surveillance 19 de Onis M, Onyango AW, Borghi E, et al. Development of a who initiative, 2018. growth reference for school-­aged children and adolescents. Bull 38 Zhang T-­T, Liu Z, Liu Y-­L, et al. Obesity as a risk factor for low back World Health Organ 2007;85:660–7. pain: a meta-­analysis. Clin Spine Surg 2018;31:22–7. 20 Yang S, Hutcheon JA. Identifying outliers and implausible values in 39 Shiri R, Karppinen J, Leino-­Arjas P, et al. The association between growth trajectory data. Ann Epidemiol 2016;26:77–80. obesity and low back pain: a meta-­analysis. Am J Epidemiol 21 Jones GT, Watson KD, Silman AJ, et al. Predictors of low back pain 2010;171:135–54. in British schoolchildren: a population-­based prospective cohort 40 Dunn KM, Hestbaek L, Cassidy JD. Low back pain across the life study. Pediatrics 2003;111:822–8. course. Best Pract Res Clin Rheumatol 2013;27:591–600. 22 Mohseni-­Bandpei MA, Bagheri-­Nesami M, Shayesteh-­Azar M. 41 Vargas CM, Stines EM, Granado HS. Health-­equity issues related to Nonspecific low back pain in 5000 Iranian school-­age children. J childhood obesity: a scoping review. J Public Health Dent 2017;77 Pediatr Orthop 2007;27:126–9. Suppl 1:S32–42. 8 Palmer AJ, et al. BMJ Open 2020;10:e036023. doi:10.1136/bmjopen-2019-036023