Low back pain (LBP) is a major public health problem that affects all age groups in high-income, middle-income and low-income countries(Reference Hartvigsen, Hancock and Kongsted1). It has enormous economic impact on patients, governments and healthcare services(Reference Steffens, Maher and Pereira2). Lifetime prevalence of LBP is 50–80 % in adults(Reference Andersson3–Reference Hoy, Bain and Williams5), while the prevalence in children usually becomes similar to that in adults by the age of 15–18 years(Reference Kovacs, Gestoso and Gil del Real6–Reference MacDonald, Stuart and Rodenberg9). Additionally, LBP is a major contributor to years lived with disability(Reference Forouzanfar, Alexander and Anderson10). In adolescents, LBP is a major cause of school absenteeism and loss of educability(Reference O’Sullivan, Beales and Smith11–Reference Akbar, AlBesharah and Al-Baghli13). LBP in childhood and adolescence is also a significant risk factor for LBP in adulthood, which highlights the importance of prevention at an early age(Reference Hestbaek, Leboeuf-Yde and Kyvik14).
The underlying causes of non-specific LBP are poorly understood in all age groups, and studies on risk factors associated with LBP have reported controversial results(Reference Forouzanfar, Alexander and Anderson10). Having LBP during childhood or adolescence has been consistently reported to be a risk factor for LBP in adulthood(Reference Taylor, Goode and George15,Reference Hestbaek, Leboeuf-Yde and Kyvik16) , while all other factors show weak or inconsistent association(Reference Zadro, Shirley and Ferreira17–Reference Calvo-Munoz, Kovacs and Roque19). Recently, vitamin D deficiency has been proposed to be a predisposing factor for LBP, particularly in adults. This has attracted huge attention because vitamin D supplementation would potentially be an easy preventive intervention. Although there is no definitive mechanism to explain how vitamin D levels can influence LBP, there are several plausible mechanisms through which vitamin D deficiency may increase the risk of LBP. First, vitamin D deficiency may have proinflammatory effect, which may contribute to LBP(Reference DeLuca20). This is supported by several studies which demonstrated that vitamin D supplementation reduces inflammatory markers(Reference Grossmann, Zughaier and Liu21,Reference Hopkins, Owen and Ahearn22) . Second, vitamin D levels may modulate sensory neuron excitability(Reference Garcion, Wion-Barbot and Montero-Menei23,Reference Stumpf and O’Brien24) and influence muscle strength(Reference Rosendahl-Riise, Spielau and Ranhoff25,Reference Stockton, Mengersen and Paratz26) , which could both be related to LBP. Finally, low vitamin D levels decrease the uptake of Ca and reduce bone mineralisation (osteomalacia or osteoporosis), which may potentially lead to back pain(Reference Zadro, Shirley and Ferreira27).
Several epidemiological studies have attempted to establish the association between vitamin D and LBP and reported conflicting results. A recent literature review of nineteen previous studies (eleven cross-sectional studies and eight case–control studies) reported a significant association between vitamin D deficiency and LBP, pooled OR 1·6 (95 % CI 1·20, 2·12)(Reference Zadro, Shirley and Ferreira17). However, when the analysis was stratified by sex and age, the association was not significant in men and was stronger among women <60 years with no association in women ≥60 years. In fact, the authors reported that the association was confined to only studies from the Middle-Eastern/Mediterranean region attributing this to climatic and cultural factors. After this review, another nested case–control study showed no association between vitamin D status and the risk of LBP(Reference Heuch, Hagen and Mai28) and another review showed high prevalence of hypovitaminosis D in patients with LBP(Reference Bansal, Boya and Vatte29). A recent review of randomised controlled trials that aimed to investigate the impact of vitamin D supplement on non-specific LBP or LBP resulting from osteoporosis showed that vitamin D supplementation is no more effective than placebo(Reference Zadro, Shirley and Ferreira27) and a major Cochrane Database systematic review has precluded a large beneficial effect of vitamin D in the treatment of painful conditions(Reference Straube, Derry and Straube30). Despite these findings, a recent review showed that vitamin D supplementation is effective in reducing the pain score in chronic widespread pain(Reference Yong, Sanguankeo and Upala31).
Most of the studies included in the literature are confined to adult population, and there is paucity of data on adolescents’ population. Adolescence is a period of growth; hence, LBP in this age group may have a different risk profile and pathways from that in adulthood. Particular covariates have to be taken into account at this age group, and therefore, results from studies on adults cannot be directly extrapolated to adolescents. As an example, the association between LBP and the weight of school bag has been under a continuous debate(Reference Calvo-Munoz, Kovacs and Roque32) but hardly considered in the previous studies. Heavy weight carried in school bags can distort the natural spinal curvature in the middle and lower back causing muscle strain and irritation to the spine joints and the rib cage(Reference Chow, Ou and Wang33). Furthermore, carrying school bag with centre of gravity positioned at the thoracic vertebrae region as T7 causes significant forward displacement in body posture(Reference Grimmer, Dansie and Milanese34), while carrying school bag with centre of gravity at the lumbar region usually leads to an increase in spinal flexion as well as reductions in pelvic anteversion and rectus abdominis muscle activity(Reference Devroey, Jonkers and de Becker35). Previously, from the same data set, we have demonstrated that vitamin D deficiency is very common in adolescents in Kuwait(Reference Al-Taiar, Rahman and Al-Sabah36). In this study, we aimed to investigate the association between vitamin D level and LBP among adolescents in public middle schools in Kuwait while adjusting for obesity, physical activity and the weight of school bags.
Methods
Participants
Data were collected within a project that aimed to estimate the prevalence of vitamin D deficiency among middle schoolchildren (11–16 years). The details of the project have been described previously(Reference Al-Taiar, Rahman and Al-Sabah36). In brief, a nationally representative sample of students was selected from public middle schools in Kuwait using multistage cluster sampling. Each school was given probability to be selected proportionate to its relative size as judged by the number of students in that school compared with the total number of students in the province (probability proportional to size sampling). The participants were selected from twelve public middle schools. A team of investigators drew blood samples and collected data on socio-demographic factors from the parents of selected adolescents through a self-administered questionnaire. Data were also collected on factors related to lifestyle such as physical activity and smoking through face-to-face interviews with the adolescents. Another team of researchers at a later time (minimum of 12 months from the time of blood collection) was formulated to interview the students and collect data on back pain and its related factors. This was conducted on a subgroup of students (n 762). This study including the original project was approved by The Ethics Committee at the Health Sciences Centre, Kuwait University (no. DR/EC/2338) and The Ethics Committee at The Ministry of Health in Kuwait (no. 2015/248). Data collection was initiated only after written informed consent was obtained from the parents and verbal assents from the adolescents.
Collection of blood samples and laboratory methods
In the original project, 5 ml of venous blood were drawn by a trained nurse. On the same day, the samples were centrifuged and the serum was transferred to Eppendorf tubes and stored at −80°C for analysis. We measured serum 25-hydroxyvitamin D (25(OH)D) concentration, which is known to be the best marker of vitamin D status that reflect both the amount of vitamin D produced in the skin after sun exposure and the amount consumed in foods(Reference Holick37). We used liquid chromatography-tandem MS to measure 25(OH)D, which is the ‘gold standard’ method of vitamin D assessment(Reference Yetley, Pfeiffer and Schleicher38,Reference de la Hunty, Wallace and Gibson39 ). This assay has a detection limit of 2 ng/ml for 25(OH)D3, with the intra-assay CV of 3·7 % and the inter-assay CV of 5·3–6·0 %. Complete blood count, parathyroid hormone, vitamin B12, Fe, ferritin, transferrin and transferrin saturation were all measured in an accredited clinical biochemistry laboratory in a teaching hospital, where these tests are routinely conducted under strict quality control. Serum intact parathyroid hormone was assessed using the Access Intact PTH chemiluminescent immunoassay with the Unicel DxI 800 Beckman Coulter analyzer using commercial kit (catalogue no. A16972). Serum vitamin B12 and erythrocytes folate in haemolysed whole blood were analysed with Roche commercial kits (catalogue no. 04745736 190) and (catalogue no. 03253678 122), respectively.
Data collection on low back pain and covariate assessment
Using a structured questionnaire, three trained data collectors gathered data on LBP through face-to-face interview. The questions on LBP were developed after an extensive literature review of studies that investigated LBP among adolescents. The details of the questionnaire, which has been used in previous studies on LBP in high school students in Kuwait, were published previously(Reference Akbar, AlBesharah and Al-Baghli13). In brief, we collected data on the period prevalence of LBP, including the lifetime, 6-month and 1-month prevalence. A photo card was used to show each participant the exact location of LBP. Lifetime prevalence was evaluated by the question ‘Have you ever felt low back pain that lasted a day or longer? (Yes, No, I don’t remember)’, while 6-month period prevalence was gauged by the question ‘Have you felt low back pain that lasted a day or longer in the last six months? (Yes, No, I don’t remember)’. Questions about the frequency of LBP, its impact on daily life activities, treatment needed, absenteeism from school due to LBP, previous back injuries and relatedness of LBP to menstrual cycle (amongst females) were included. A 0–10 Numeric Pain Rating Scale was used to rate the intensity of LBP in past month. Data on depressive symptoms were collected using the Children’s Depression Inventory(Reference Kovacs40), which has been translated and adopted in our setting(Reference Al-Balhan41).
Although the association between weight of school bags and LBP is under debate, we measured the weight (kg) of every student’s bag using a portable digital luggage scale (SAFEWAYR) that was calibrated before each use. We also collected data on the perceived heaviness of school bags by asking the students the question ‘How would you describe the weight of your school bag? (Light, Normal, Heavy, Too heavy)’. Data were also collected on the number of school bags a student carries and the use of lockers to reduce the weight of their bags. The weight of each adolescent was measured to the nearest 0·1 kg and the height to the nearest 0·1 cm in a standardised manner using a digital scale after removing shoes and any heavy clothing.
Statistical analysis
BMI was calculated as weight (kg) divided by height squared (m2), and BMI-for-age z-scores were calculated using WHO growth charts. 25(OH)D status was defined according to the Endocrine Society(Reference Holick, Binkley and Bischoff-Ferrari42) and the Society for Adolescent Health and Medicine(43) as deficiency <50 nmol/l; insufficiency 50–75 nmol/l; sufficiency ≥75 nmol/l. The association between 25(OH)D and LBP was assessed using unconditional logistic regression with adjustment for potential confounders. Separate analyses were performed with 25(OH)D fitted as a continuous variable and as a categorical variable. We categorised 25(OH)D using acceptable cut-off points(Reference Holick, Binkley and Bischoff-Ferrari42) or quartiles. First, crude OR were calculated, and then variables with P < 0·2 were introduced sequentially to the model while noting the impact of this on the association between 25(OH)D and LBP. Some studies have suggested that the association between 25(OH)D and LBP is dependent on age and sex(Reference Zadro, Shirley and Ferreira17,Reference Heuch, Hagen and Mai28 ); we, therefore, performed separate tests for the interaction between 25(OH)D and sex as well as age. As a sensitivity analysis, we used stepwise logistic regression to explore if the conclusion on the association between 25(OH)D and LBP would be different with stepwise variable selection. Data analysis was conducted using Stata (StataCorp. 2011, release 12), and factors that showed P < 0·05 were deemed to be statistically significant.
Results
Of 1416 adolescents with 25(OH)D measurement, 762 were interviewed to collect data on LBP and its related factors. The mean age was 12·26 (sd 0·81) years, and 386 (sd 50·66 %) were males. The lifetime prevalence of LBP was 32·28 % (95 % CI 28·97, 35·73 %), which was significantly higher among females compared with males (36·44 v. 28·24 %; P = 0·016). Also, the 6-month prevalence of LBP was 21·26 % (95 % CI 18·40, 24·33 %), which was 24·27 and 18·44 % in females and males, respectively (P = 0·050). The 1-month prevalence was 13·25 % (95 % CI 10·93, 15·87 %) and was also significantly higher among females compared with males (16·27 v. 10·39 %; P = 0·017). Among those who reported LBP during the past month, 66·99 % sought no treatment, 19·42 % used home remedies (e.g. hot packs), 5·83 % used analgesics over the counter and 7·77 % sought medical treatment from a public or a private clinic. In this group, the pain severity was assessed using 0–10 numeric pain rating scale, which showed a mean of 5·52 (sd 1·63) score. The geometric mean of 25(OH)D was 30·26 (sd 1·77) nmol/l, which was significantly lower among females compared with males, 22·22 (sd 1·69) nmol/l v. 40·87 (sd 1·55) nmol/l; P < 0·001. The prevalence of vitamin D deficiency was 93·88 % among females and 66·84 % among males (P < 0·001).
The association between socio-demographic factors and LBP in the past 6 months is shown in Table 1. In this analysis, we excluded two participants who answered ‘I don’t know’ when asked about the LBP during the past 6 months. Except for sex in which females tended to have higher prevalence of LBP, socio-demographic characteristics were not significantly associated with LBP in univariable analysis. Table 2 shows the association between LBP in the past 6 months and lifestyle factors such as consumption of food not prepared at home, sleeping hours during weekdays and weekends and physical activity. The association between LBP in the past 6 months and BMI categories as well as factors related to school bags in addition to depression score is depicted in Table 2. In this analysis, sleeping hours during the weekdays (but not weekends) (P = 0·024) and perceived weight of school bag (P < 0·001) were associated with LBP in the past 6 months. Neither the absolute weight of school bag (P = 0·528) nor the weight of school bag as a percentage of body weight (P = 0·059) was significantly associated with LBP in the past 6 months in univariable analysis. However, the weight of school bag as a percentage of body weight became significantly associated with LBP in multivariable analysis (Table 3). The other two factors significantly associated with LBP in univariable analysis were sitting posture (on bed, desk, floor and couch) and depression score. The association between various laboratory measurements and LBP in the past 6 months is demonstrated in Table 4. None of the laboratory results showed significant association with the LBP during the past 6 months in univariable analysis. This was evident whether these laboratory measurements were fitted as continuous or categorical variables.
Association between low back pain during the past 6 months and socio-demographic factors in 760 adolescents in univariable analysis
(Numbers and percentages; odds ratios and 95 % confidence intervals)
Ref., reference.
Association between low back pain during the past 6 months and lifestyle factors, physical activity, BMI and depression score in addition to factors related to school bag in univariable analysis
(Numbers and percentages; odds ratios and 95 % confidence intervals)
Ref., reference; CDI, Children’s Depression Inventory.
Factors associated with low back pain during the past 6 months in multivariable analysis
(Numbers and percentages; odds ratios and 95 % confidence intervals)
Ref., reference; CDI, Children’s Depression Inventory.
Association between low back pain during the past 6 months and parathyroid hormone, calcium, vitamin B12, anaemia, ferritin and folate in univariable analysis
(Numbers and percentages; odds ratios and 95 % confidence intervals)
Ref., reference.
There was no difference in the geometric mean of 25(OH)D between those with and without LBP in the past 6 months, 28·50 nmol/l and 30·82 nmol/l, respectively (P = 0·122). The association between 25(OH)D and LBP during the past 6 months before and after adjusting for potential confounders is shown in Table 5. There was no association between 25(OH)D and LBP in univariable or multivariable analysis whether 25(OH)D was fitted as a continuous, or as a categorised variable using quartiles or acceptable cut-off points. In this analysis, the interaction between sex and 25(OH)D was not significant in all crude (P ≥ 0·157) or adjusted (P ≥ 0·651) analyses. We also used stepwise logistic regression method, in which 25(OH)D was not selected neither in forward nor backward selection. When 25(OH)D was forced in this analysis, it was not significant in any model. Vitamin D binding protein was measured in a subsample of 404 participants; hence, we were able to calculate free vitamin D levels as described by Tsuprykov et al. (Reference Tsuprykov, Chen and Hocher45). There was no significant association between free 25(OH)D and LBP in univariable (P = 0·603) or multivariable (P = 0·457) analysis. We also repeated the above analysis with LBP in the past month or lifetime LBP and found no significant association between 25(OH)D and any of these outcomes. Finally, factors that were significantly associated with LBP in the past 6 months in multivariable analysis were governorate (province) in which the school was allocated (P = 0·039), perceived weight of school bag (P = 0·001), weight of school bag as a percentage of the body weight (P = 0·017) and sleeping hours during weekdays (P = 0·037) (Table 3). Using backward and forward stepwise selection, perceived weight of school bag, age of adolescents and sleeping hours during weekdays were the only significant predictors for LBP in the past 6 months.
Association between plasma 25-hydroxyvitamin D (25(OH)D) and low back pain in the past 6 months before and after adjusting for potential confounders
(Numbers and percentages; odds ratios and 95 % confidence intervals)
Q1–Q3, quartile 1 to quartile 4. Ref., reference.
* Adjusted for all variables with P < 0·2 including sex, age, school governorate (province), sleeping hours during weekdays, perceived weight of school bag, the way of sitting during homework or reading, weight of school bag as percentage of the body weight, parathyroid hormone and depression score.
Discussion
In this study, we estimated the prevalence of LBP among adolescents in the middle schools in Kuwait and investigated the association between LBP and vitamin D levels. We found the lifetime prevalence of LBP to be 32 %, while the 6-month prevalence to be 21·26 %. There was no association between 25(OH)D levels and LBP in the past 6 months neither in univariable nor in multivariable analysis. This was evident whether 25(OH)D was fitted as a continuous variable or a categorical variable using acceptable cut-off points or quartiles.
The lifetime and the 6-month prevalence of LBP in middle school students were lower than that reported among high school students in Kuwait(Reference Akbar, AlBesharah and Al-Baghli13). This can be explained by the notion that the prevalence of LBP increases with age and reaches that of the adult population at the age of 15–18 years(Reference Calvo-Munoz, Gomez-Conesa and Sanchez-Meca8,Reference MacDonald, Stuart and Rodenberg9 ). The prevalence was significantly higher among females compared with males, which is consistent with several other studies(Reference MacDonald, Stuart and Rodenberg9,Reference Sundell, Bergstrom and Larsen46 –Reference Alghadir, Gabr and Al-Eisa48). Prevalence of LBP in adolescence varies considerably between different studies with lifetime prevalence ranges between 11·60 and 85·56 % as reviewed by Calvo-Munoz et al. (Reference Calvo-Munoz, Gomez-Conesa and Sanchez-Meca8). Some of these differences could be attributed to methodological differences, including the variation in LBP definition and the age of the targeted groups.
Although there are several plausible pathways in which low vitamin D levels may contribute to LBP, the findings of epidemiological studies on this issue remain controversial. A literature review(Reference Zadro, Shirley and Ferreira27) showed that most studies were on adult populations or included participants with a wide age range (adolescents and adults) without sufficient numbers to conduct an age-stratified analysis. In our study, we found no significant association between 25(OH)D levels or 25(OH)D status and LBP in adolescents. Our results are different from the findings of another study that investigated this issue among school-aged children(Reference Alghadir, Gabr and Al-Eisa48). In this study, the authors reported a lower mean of 25(OH)D among school-aged children with severe LBP (n 45) compared with those with moderate LBP (n 85) and no or minimal LBP (n 120). In their results, however, arithmetic mean may not represent the centre of the 25(OH)D distribution (e.g. mean = 9·5 ng/ml, while sd = 4·7 ng/ml in the severe LBP group). Our findings match the results of a pilot study among adolescent male ballet dancers(Reference Ducher, Kukuljan and Hill49) as well as a recent nested case–control study among adults (adults 19–55 years)(Reference Heuch, Hagen and Mai28). Both studies showed no significant association between vitamin D status and the risk of LBP. Furthermore, a recent review of studies (mainly on adult population) concluded that there is an association between vitamin D deficiency and LBP overall(Reference Zadro, Shirley and Ferreira17). However, age- and sex-stratified analysis showed the association existed only in women <60 years with no association in men or in women >60 years. In fact, the authors found the association to be confined to the Middle East/Mediterranean region, with no association between vitamin D and LBP in the studies outside this region. In another recent review of studies on adult population (30–66 years), authors found higher prevalence of hypovitaminosis D in patients with LBP compared with those without LBP(Reference Bansal, Boya and Vatte29). One of the possible explanations for these mixed results of observational epidemiological studies is the difference in the measurement of vitamin D as well as the differences in the definition of LBP. Not only are the results of observational studies conflicting but also the randomised controlled trials that aimed to investigate the impact of vitamin D supplementation on LBP or musculoskeletal pain in general showed mixed results. A recent review of randomised controlled trials showed that vitamin D supplementation is no more effective than placebo in treatment of non-specific LBP or LBP resulting from osteoporosis(Reference Zadro, Shirley and Ferreira27). A major Cochrane Database systematic review has also precluded a large beneficial effect of vitamin D supplementation in the treatment of painful conditions(Reference Straube, Derry and Straube30). Some recent reviews showed no clear benefit of vitamin D supplementation in treatment of chronic non-specific musculoskeletal pain(Reference Gaikwad, Vanlint and Mittinity50), while other reviews showed that vitamin D supplementation is effective in reducing the pain score in chronic widespread pain(Reference Yong, Sanguankeo and Upala31, Reference Wu, Malihi and Stewart51, Reference Helde-Frankling and Bjorkhem-Bergman52). Currently, vitamin D supplementation cannot be considered as an independent treatment for chronic pain(Reference Martin and Reid53).
While the lack of association between 25(OH)D and LBP could be genuine in our study, it could be due to the fact that our adolescents were mostly vitamin D deficient (80·2 % of the study group were vitamin D deficient and only 4·2 % had sufficient vitamin D level). In spite of large study size, the small number of adolescents with sufficient vitamin D level lowered the power of the study to detect small differences in the prevalence of LBP. Power calculation showed that our study has 73 % power to detect a difference of 10 % in the prevalence of LBP between those with and without vitamin D deficiency at 5 % level of significance (two-sided) given that the prevalence of LBP we found in the study was 21 %. However, the study will have more than 90 % power to detect 15 % difference in the prevalence of LBP between those with and without vitamin D deficiency, while all other assumptions remain the same. Furthermore, in children or adolescents even if the association between 25(OH)D and LBP vitamin D deficiency exists, low vitamin D level would have less time to exert its influence on the risk of LBP compared with adults. Finally, the proposed mechanism for low vitamin D to increase the risk of LBP through reducing bone mineralisation (osteomalacia or osteoporosis) seems less likely during adolescence compared with adulthood. This makes the association between vitamin D levels and LBP more elusive in children and adolescents compared with that in adults.
With respect to other risk factors, a recent literature review on risk factors for LBP in children and adolescents showed inconsistent association between all presumed risk factors and LBP(Reference Calvo-Munoz, Kovacs and Roque19). Recently, the association between the weight of school bag and LBP among schoolchildren has attracted huge attention. We have previously reported the importance of the perceived weight of school bag compared with the absolute weight of school bag (or the bag weight as a percentage of body weight) for practical and logical reasons among high school students(Reference Akbar, AlBesharah and Al-Baghli13). In our present study, the perceived weight of school bag was significantly related to LBP in univariable and multivariable analyses. We reiterate our previous recommendation that parents should be encouraged not to allow their children to carry school bags, which their children describe as heavy. Such recommendation would be easy to implement because it does not require weighing the child or the school bag. This does not mean that the weight of school bag as a percentage of body weight is not important. In our previous study, the weight of school bag as a percentage of the body weight was significantly associated with LBP in univariable analysis but not in multivariable analysis among high school students. In the present study, while the weight of school bag as a percentage of body weight was not significantly associated with LBP in univariable analysis (P = 0·059), it was significant in multivariable analysis (P = 0·017). It is worth noting that the perceived weight of school bag was not correlated with the weight of school bag as a percentage of body weight but was correlated with the absolute weight of school bag. A recent literature review showed the association between backpack weight and LBP to be inconsistent across studies(Reference Calvo-Munoz, Kovacs and Roque19) and that the available evidence does not support that school bags weighing >10 % of the body weight are associated with higher prevalence of LBP among schoolchildren(Reference Calvo-Munoz, Kovacs and Roque32).
In our study, physical activity was not significantly associated with LBP neither in univariable nor multivariable analyses. Although we collected data on physical activity by a questionnaire that has been previously validated using accelerometers in high school students (Spearman correlation 0·92; P < 0·001 for total steps count, not published), self-reported data on physical activity usually carry large amount of non-differential misclassifications. This could bias the association between physical activity and LBP towards the null, which may explain our findings. Previous studies have suggested U-shaped association between physical activity and LBP in adolescents, where both low and high levels of physical activity are associated with higher risk of LBP(Reference MacDonald, Stuart and Rodenberg9). Nevertheless, a recent review of the previous studies showed that the association between sport activities and LBP is controversial, with some studies showing positive association or inverse association or no association(Reference Calvo-Munoz, Kovacs and Roque19). Similarly, BMI was not associated with LBP neither in univariable or multivariable analysis in our study. This is consistent with the current literature review which showed that BMI is not a major determinant of LBP in children and adolescents(Reference Calvo-Munoz, Kovacs and Roque19).
To our knowledge, this is the largest study that investigated the association between vitamin D levels and LBP among adolescents in the Middle East. We have measured 25(OH)D using liquid chromatography-tandem MS, which is the ‘gold standard’ method(Reference Yetley, Pfeiffer and Schleicher38, Reference de la Hunty, Wallace and Gibson39) in an accredited laboratory. We collected data on all potential confounders in this age group including the measurement of the weight of school bags, physical activity, BMI and depression. In order to draw a robust conclusion, we analysed the association between 25(OH)D and LBP dealing with 25(OH)D as a continuous variable and as a categorical variable using quartiles or acceptable cut-off point. This is an important analytical aspect because the results of the previous studies on this issue were dependent on how the variables were analysed(Reference Zadro, Shirley and Ferreira27). We also calculated free vitamin D and investigated its association with LBP for the first time. Inevitably, data collection on LBP relied on recall, which may have caused non-differential misclassification in the outcome hence biased our findings towards the null. Finally, we did not collect data on other musculoskeletal conditions that may interfere with LBP such as hypermobility syndrome, connective tissue disorder and juvenile-onset scoliosis.
In conclusion, we have demonstrated that approximately one-third of children in middle schools in Kuwait had experienced LBP in their lifetime and one-fifth had LBP during the past 6 months. We found no significant association between vitamin D levels and LBP in adolescents in an area with high prevalence of vitamin D deficiency. Although it is important to have sufficient vitamin D level during childhood and adolescence for several health benefits, we concluded that vitamin D is not a major determinant for LBP among adolescents in our setting. We also concluded that the weight of school bag as a percentage of body weight is associated with LBP in adolescence.
Acknowledgements
We would like to thank Ms Nadien Rushdi for the project management and the team of data collectors.
This work was partially funded by Kuwait University, Research Sector (project no. WF 02/13).
Author contributions were: A. A.-T.: study design, data collection, data analysis and wrote the manuscript; A. R.: study design, data collection, data interpretation and review the paper with significant intellectual input; R. A.-S.: study design, data collection and review the paper with significant intellectual input; L. S.: study design, data collection and review the paper with significant intellectual input; A. H. B.: data collection and review the paper with significant intellectual input; S. B.: review the paper with significant intellectual input and M. A.-E.: review the paper with significant intellectual input.
The authors declare that there are no conflicts of interest.
