|dc.description.abstract||Background: Low back pain (LBP) is a common complaint in adolescents with prevalence rates in the teenage years reported to be as high as in adults and a wide range of lifestyle, socioeconomic, psychological and physical factors have been associated with LBP in the adolescent population. However, in New Zealand there has been only one study published to date, which has looked at the prevalence rates of LBP in the adolescent population (11-14 years). The Zealand study examining LBP in adolescents showed that psychological, social and emotional factors may play a stronger role than the physical factors. In yet another study on LBP in adolescents, investigators concluded that certain aspects of diet may influence on LBP in adolescents. Other investigators have linked physical factors to LBP in adolescents by demonstrating the association between increased physical activity and strong back flexor muscles with LBP. Overall the current literature suggests possible factors associated with LBP in the adolescent population are wide ranging and the issue is complex.
Aims: The primary aim of the current study was to determine the current and period prevalence’s of LBP when categorized according to reporting period (LBP lifetime, LBP recurrent and LBP location confirmed (LC)) specifically in adolescent schoolgirls in Otago, aged between 13-18 years. The secondary aim was to examine the relationship between the three reporting periods of LBP and key lifestyle factors of physical activity, smoking habits, food and drink consumption levels along with anthropometric measurements and back extensor endurance (BEE) estimates.
Methods: This was a cross-sectional observational study. Consenting adolescent females (n=322) from six Otago schools completed a self-report questionnaire, the Otago Back Pain & Lifestyle Study Questionnaire (OBPLSQ). The questionnaire was designed and customized for the current study and comprised 48 items. The items were based on previously validated questionnaires and covered demographics, physical activity levels (PALs), LBP, smoking habits and food and drink consumption levels. The questionnaire was made available on-line to the participants in their respective school’s computer suite. The anthropometric measurements of height (cm) and midpoint waist circumference (cm) were taken directly from the participants on the same day they completed the questionnaire. Bio-electrical impedance analysis was undertaken to gather information regarding the participants’ body fat percentage, fat mass (kg), fat-free/lean mass (kg), body mass index and body fat percentages and BE endurance score (seconds).
Data analysis: Descriptive statistics (mean, standard deviation and range) were used to describe the participants’ characteristics. Uni-variate and multinomial logistic regression (MLR) analysis were undertaken on the three dependent variables of LBP self-report categories (LBP lifetime, LBP recurrent, LBP LC) to identify any significant lifestyle factors explaining the risk of LBP. The sixteen predictor variables used in this analysis were age, ethnicity, waist to height ratio (WtHR), body mass index standard deviation (BMI z score), BEE, fat percentage, PAL’s (New Zealand physical activity questionnaire (NZPAQ), metabolic equivalents (METs), health behaviour in school children (HBSC), current smoking, fruit and vegetable intake along with food indices (fruit and vegetable (FV), fibre, calcium, variety, treat) derived from food and drink consumption section). Predictor variables demonstrating p ≤ 0.2 at the uni-variate level analysis were entered into the MLR models for further analysis.
Results: Two hundred and ninety seven participants (92%) completed the entire questionnaire and had their physical measurements taken. The mean (SD) age of the participants in the current study was 14.3 (SD 1.2) years. Prevalence levels of LBP were LBP lifetime (57.6%), LBP recurrent (26.6%) and LBP LC (24.2%). From the uni-variate analysis the likelihood of LBP LC was found to be almost three times (OR=2.9 95% (confidence interval) CI 2.56, 3.01, p=0.04) greater in those participants who were current smokers. The predictor variables of WtHR, BMI z score, fat percentage, BEE and age were also found to be associated with different categories of LBP at various levels of significance. Five predictor variables (WtHR) (odds ratio) (OR=72.17, 95% CI 55.34, 93.79 p=0.05), (BMI z) score (OR=1.34, 95%CI 0.98, 3.33 p=0.01), fat percentage (OR=1.04, 95%CI 0.56, 1.87 p=0.00), BEE (OR=0.99, 95%CI 0.32, 1.24 p=0.01) and variety index (OR=1.11, 95%CI 1.01, 1.55 p=0.09) met the threshold criteria to be included in MLR models using the dependent variable of Lifetime LBP. Seven predictor variables: age (OR=1.28, 95%CI 1.03,1.71p=0.01), BMI z score (OR=1.24, 95%CI 0.64, 3.52 p=0.08), fat %age (OR=1.03, 95%CI 0.43 p=0.02), BEE (OR=0.99, 95%CI 0.46, 1.98 p=0.00), PAL’s (OR=1.25, 95%CI 0.09,1.66 p=0.19), current smokers (OR=2.5, 95%CI 1.19 p=0.08) and variety index (OR=1.13, 95%CI 0.77, 2.67p=0.11) were analysed with the dependent variable of recurrent LBP. Four predictor variables: age (OR=1.26, 95%CI 1.01, 1.60 p=0.02), BEE (OR=0.99, 95%CI 0.77, 2.25p=0.01), PAL’s (OR=1.33, 95%CI 1.11, 1.43 p=0.10) and current smokers (OR=2.9, 95%CI 2.56, 3.01 p=0.04) were included in the MLR analysis with the predictor variable of LBP (LC). No significant relationships were identified from any of the MLR analyses.
Conclusion: The LBP prevalence levels reported in the current study are in accordance with available literature. Of all the variables examined the anthropometric measurements showed the strongest associations with LBP when compared to the self-report lifestyle variables. In the preliminary analysis no consistent pattern emerged between the three self-report categories of LBP as although the factors of age, WtHR, BMI z score, fat percentage, BEE, current smoking were significantly related to LBP the level of risk varied between the three back pain categories. However, these significance levels were lost when these same factors were examined in the more robust models (MLR) and current smoking was found to be the most significant predictor of LBP and this was specifically evident in the adolescents categorized under the LC category. The emergence of only LC category LBP as most significant in relationship to smoking emphasizes the importance of using the body chart in studies of such kind and being able to relate the pain to specific body part which helps to improve the reliability of the responses. Finally, the results demonstrate that the reporting period is very important when factoring in risk factors associated with LBP in female adolescence, as it is clearly observed that there is difference in the significance levels of same risk factor for example smoking, WtHR within different categories of LBP in the same set of population.||