This was a cross-sectional study that used health checkup data. The check-up was conducted by the All Japan Labor Welfare Foundation, a health checkup center in Japan. Subjects in the present study were adults aged over 20 years who underwent the annual health checkup between April 2013 and March 2014. Of the 552,005 subjects, 551,871 subjects consented to participate in this study. Of these, we excluded 132,868 subjects with missing data on any variables used in the present study. Thus, 419,003 participants were included in the analysis. Written informed consent for the use of personal health checkup data in this study was obtained from each participant. This study protocol was approved by the Ethics Committee of the All Japan Labor Welfare Foundation (Approval No. 9-1-0007) and the medical ethics committee of Showa University School of Medicine (Approval No. 2407).
Data on age, sex, lifestyle behaviors, medical history, and current use of medications were collected using a self-administered questionnaire. Trained staff measured height to the nearest 0.1 cm using a stadiometer and weight to the nearest 0.1 kg using a scale. Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m). Blood pressure was measured in the sitting position using an automated sphygmomanometer (HEM-907, Omron, Kyoto, Japan). LBP was defined as self-reported LBP under treatment (i.e., a “yes” answer to the question “Do you have LBP under treatment including follow-up?”) .
Blood samples were collected and stored in a cooler at 4°C for transporting to an external laboratory (SRL, Tokyo, Japan). Triglyceride levels were measured using an enzymatic method (AU5400; Beckman Coulter, Tokyo, Japan), while low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C) were measured using a direct method (AU5400; Beckman Coulter, Tokyo, Japan). Hemoglobin A1c (HbA1c) level was determined using latex agglutination turbidimetry (JCA-BM9130, JEOL, Tokyo, Japan).
Hypertension was defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, or medication use for hypertension . Diabetes was defined as HbA1c (National Glycohemoglobin Standardization Program) ≥ 6.5% or medication use for diabetes . Dyslipidemia was defined as triglyceride ≥ 150 mg/dL, HDL-C < 40 mg/dL, LDL-C ≥ 140 mg/dL, or medication use for dyslipidemia .
Assessment of lifestyle risk score
Questionnaire items on lifestyle behaviors were based on the standardized questionnaire used for the National Health Promotion Program [21, 22], which started in Japan in 2008 and aimed to prevent lifestyle-related diseases (e.g., metabolic syndrome and cardiovascular disease). The following 12 items related to lifestyle behaviors were used in the present study: 1) smoking habits (current, former, none), 2) alcohol intake (everyday, sometimes, none), 3) exercise ≥ 30 min/day, ≥ twice a week, and ≥ 1 year (yes, no), 4) physical activity equal to walking ≥ 60 min/day (yes, no), 5) walking faster than others in the same generation (yes, no), 6) weight gain ≥ 10 kg since age 20 years (yes, no), 7) body weight change ≥ 3 kg during the preceding 1 year (yes, no), 8) eating speed (fast, normal, slow), 9) eating dinner within 2 hours before bed ≥ three times per week (yes, no), 10) having a snack after dinner ≥ three times per week (yes, no), 11) skipping breakfast ≥ three times per week (yes, no), and 12) adequate sleeping (yes, no).
For each lifestyle factor, we created a binary variable; 1 represented health risk (unhealthy) and 0 showed no health risk. The criteria of health risk were determined with reference to the recommended guideline or current health-related studies [7, 23-29]. Specifically, we assigned a score of 1 for each item as follows: 1) current smoking, 2) drinking alcohol every day, 3-5) a response of no, 6-7) yes, 8) eating fast, 9-11) yes, and 12) no response. A lifestyle risk score was calculated by combining the scores of the 12 lifestyle factors (range: 0-12)  and was categorized into the following three groups by tertile of the total score; low (score: 0-3), moderate (4-5), and high risk (6-12).
Data on the participants’ characteristics are presented as n (%) or median (25th, 75th percentile). Characteristics of the study participants with and without LBP were compared using chi-squared test for categorical variables or Wilcoxon rank-sum test for continuous variables.
To evaluate the association between LBP and lifestyle risk levels, a logistic regression analysis was performed to calculate the odds ratio (OR) and 95% confidence interval (CI) for LBP. Model 1 was a crude model, and model 2 was adjusted for age and BMI. Model 3 was further adjusted for hypertension, diabetes, and dyslipidemia. We then examined the association between LBP and lifestyle risk score stratified by age (20-39, 40-59, and ≥ 60 years) and BMI (underweight: < 18.5, normal weight: 18.5-24.9, and overweight/obesity: ≥ 25 kg/m2) categories in model 3. Test for trend was conducted with the lifestyle risk score groups considered as continuous variables. All statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA). A p value < 0.05 was considered statistically significant.