Prevalence of chronic low back pain and its associated factors in the general population of South Korea: a cross-sectional study using the National Health and Nutrition Examination Surveys

Data from versions IV-1, -2, and -3 of the Korea National Health and Nutrition Examination Survey (KNHANES) performed in 2007, 2008, and 2009, respectively, were analyzed. This survey has been conducted annually since 1998 by the Korea Centers for Disease Control (KCDC). To evaluate the health and nutritional status of the general South Korean population, a nationwide sampling method (clustered, multistage, stratified, and randomized) is used for proportional distribution according to geographic area, sex, and age. The survey participants are different every year and are not serially monitored, resulting in a random sampling every year. The KNHANES evaluates data from three sources: health questionnaires, health and physical examinations, and nutrition questionnaires that are administered by experienced interviewers, registered nurses, and laboratory technicians [26]. The KNHANES IV-1 (2007), IV-2 (2008), and IV-3 (2009) examinations and health surveys were completed by 4,594, 9,744, and 10,533 participants (total: 24,871 participants), respectively. The present analysis was confined to 17,038 respondents aged 10–100 years who answered the CLBP examination survey and had no missing data regarding the demographics and health questionnaires (Fig. 1).

Participants in the survey who answered “yes” to all three of the following questions were considered to have LBP: (1) “Have you ever had LBP?” (2) “Do you currently have LBP?” and (3) “Have you complained of LBP for more than 90 days during the past year?” [26, 27]

We analyzed the participants’ demographics and socioeconomic status, comorbidities, and lifestyle habits through health interviews and examinations. All participants were asked whether they had received diagnoses of major comorbidities by physicians, such as hypertension, diabetes mellitus, dyslipidemia, ischemic heart disease (e.g., myocardial infarction, angina), stroke, liver cirrhosis, asthma, chronic obstructive pulmonary disease (COPD), arthritis, or chronic kidney disease. Major cancers, including lung, stomach, liver, colon, breast, prostate, uterine, and cervical cancers, were surveyed in all participants.

Age was divided into age groups. Body mass index (BMI) was calculated as bodyweight (kilograms) divided by height (meters) squared and categorized as underweight (< 18.5 kg/m²), normal weight (18.5–24.9 kg/m²), and obese (≥ 25.0 kg/m²) [28]. Patients were categorized into non/ex-smokers or current smokers based on their present smoking status. Alcohol consumption was categorized as none or ≥ 1 drink/month. Occupations were divided into five groups: unemployed (e.g., students, housewives), office workers (e.g., managers, professionals), sales and services, machine fitting and simple labor (e.g., technicians, device and machine operators, low-level laborers), and agriculture, forestry and fishery [29]. Household income was divided into four quartiles. Education level was categorized into four groups according to the highest level of education completed: elementary school (≤ 6 years), middle school (7–9 years), high school (10–12 years), and university or college (≥ 13 years). Physical activity was divided into three categories: “walking” was defined as walking for ≥ 30 min, 5 or more days per week; “moderate physical activity” was defined as mid-intensity physical activity for ≥ 30 min, 5 or more days per week; and “high physical activity” was defined as high-intensity physical activity for ≥ 20 min, 3 or more days per week [12]. Participants in this survey who felt sad or had a depressive symptom for 2 consecutive weeks during the past year were considered to have “depressive symptoms” [10]. Serum vitamin D levels were categorized into two groups: insufficiency (< 20 ng/ml) and normal (≥ 20 ng/ml).

General demographics and co-variables were compared between the participants with and without CLBP. Student’s t-test was used to compare continuous variables, and the chi-square test was used for categorical variables. Univariate logistic regression analysis was used to identify individual factors associated with CLBP. Co-variables with a p value < 0.05 in the univariate analysis were subsequently evaluated by multiple logistic regression analysis using backward stepwise selection with a 0.05 significance level. Categorical variables were expressed as numbers and percentages, whereas continuous variables were expressed as means and standard deviations. Statistical significance was denoted by a two-tailed p value < 0.05. Odds ratios (ORs) with corresponding 95% confidence interval (CIs) were calculated accordingly. Sampling weights were applied to the study population to represent the South Korean population without bias. Statistics were performed using Stata/MP 15.1 (StataCorp., 2017, Stata Statistical Software: Release 15; College Station, TX, USA; StataCorp LP).

The IV-1, IV-2, and IV-3 versions of the KNHANES were approved by the KCDC Institutional Review Board (approval no. 2007-02CON-04-P, 2008-04EXP-01-C, 2009-01CON-03-2C). Informed consent was obtained from all participants when the surveys were conducted.

The baseline characteristics of all participants are shown in Table 1. A total of 17,038 participants were included in the final analysis, including 2,693 with CLBP and 14,345 without. The prevalence of CLBP was 15.8% in South Korean subjects, with a prevalence of 11.8% in men and 24.5% in women. Prevalence by age group was as follows: 2.25% in teens, 6.85% in 20 s, 7.88% in 30 s, 9.22% in 40 s, 15.34% in 50 s, 25.51% in 60 s, 33.91% in 70 s, 36.93% in 80 s, and 29.41% in 90 s (Fig. 2).

Table 2 shows the results of the univariate logistic regression analysis. Underweight BMI (< 18.5), walking activity, high-intensity physical activity, and liver cirrhosis were found to be statistically insignificant.

Table 3 shows the results from the multiple logistic regression analysis. The included associated factors were age, gender, alcohol consumption, household income, education level, mid-intensity physical activity, depressive symptoms, vitamin D level, and comorbidities (e.g., stroke, ischemic heart disease, knee osteoarthritis, asthma, COPD, and cancer history). Among these factors, advanced age, female gender, mid-intensity physical activity, depressive symptoms, stroke, ischemic heart disease, knee arthritis, asthma, COPD, and cancer history were positively associated with LBP. In contrast, alcohol consumption ≥ 1 drink per month, increased household income, higher education level, and vitamin D insufficiency were negatively associated with CLBP.

In a review of the worldwide prevalence of LBP that included 54 countries, the mean prevalence of CLBP was estimated to be 20.6% (95% CIs 19.4–21.9%) [8]. We found that the prevalence of CLBP in the general South Korean population was 15.8% (11.8% in men and 24.5% in women), which is slightly lower than that previously reported.

Old age and female gender are well-established risk factors for CLBP [6, 8]. At 70 years of age, the OR increased to 8.996. In a previous study, the CLBP prevalence was highest in those aged 45–64 years [30]. However, in our study, the positive association between age and CLBP increased up to the 8th decade of life. The increased risks in women are similar to those in a previous study [21]; as age increases, the prevalence of CLBP increases, and advanced age is regarded as a risk factor [31, 32]. On the other hand, it has also been reported that age is not associated with CLBP or radiating to the lower extremity pain [33]. In our study, the ORs increased with increasing age, reaching a peak in the 70 s, and decreased in the 90 s. Female gender is known to be a risk factor and had an OR of approximately 1.54 in a previous study [33]. In our study, the OR was 1.763, which is similar to that previously reported.

In previous studies, smoking and alcohol were reportedly associated with CLBP [31‐34]. In particular, smoking varies according to the study, but it has been consistently observed as a risk factor for CLBP. Alcohol consumption was found to be a significant risk factor in some studies but was not reported by a systematic review [32]. However, the results of our study suggest that smoking is not a relevant factor for CLBP, and alcohol consumption is rather negatively associated with CLBP. However, alcohol consumption cannot be considered a protective factor due to the inability of this study to determine a causal relationship. Future studies are needed.

Socioeconomic factors, such as education level, household income, and occupation, are widely accepted as factors that are associated with multiple health outcomes. In previous studies, these factors were also found to be associated with CLBP [33, 35, 36]. Similar to these studies, we found that education level was strongly negatively associated with CLBP [33, 35]. Because education level is associated with an understanding of health and treatment, similar results are likely to be seen in several studies. Household income and occupation were significantly associated with LBP in the univariate analysis, but not in multivariate analysis. In our study, education level was the only socioeconomic factor consistently found to be associated with CLBP.

Walking and regular physical activity is an excellent method of pain reduction for patients with CLBP [16, 37]. Previous studies have reported that walking reduces CLBP [12], and that physical activity significantly reduces LBP in those who sit for extended periods [9]. Furthermore, physical activity is especially important in patients with high BMIs [38, 39]. In this study, obesity, walking, and high-intensity physical activity were not independently associated with CLBP. In addition, mid-intensity physical activity was found to be positively associated with CLBP. Previous studies have shown that strenuous exercise is associated with CLBP [40]; however, no association has been reported between mid-intensity exercise and LBP. It is unclear due to the cross-sectional study design, and determining a causal relationship may be difficult in this circumstance because patients with CLBP can only perform mid-intensity exercises.

Psychiatric problems, including depression, are associated with CLBP and are considered to be risk factors for increasing CLBP [41]. Among them, depression was reported to be closely associated with CLBP, radiation pain, and poor surgical outcomes and was reported relatively consistently in many papers [33, 41, 42]. In this study, the results were in line with previous studies and showed a high OR.

Numerous studies have demonstrated the association between vitamin D deficiency and CLBP [43, 44]. This association is supported by several theories, but the exact mechanism is not yet clear [43]. Furthermore, some have refuted the association between vitamin D levels and CLBP [45], and others have determined that vitamin D supplementation is not effective for controlling CLBP [46]. However, our study showed the opposite result; vitamin D deficiency was negatively associated with CLBP. Rather than interpreting this negative association as a result of our study, it is better to conclude that vitamin D is not associated with CLBP.

The association between comorbidities and CLBP is well established [19, 47]. In theory, comorbidities may affect allostatic loads and cause pain through the dysregulation of physiological mechanisms, but these mechanisms have not been precisely identified [47]. In a previous study, CLBP increased relatively consistently with the number of comorbidities, with an OR of 5.05 when there were four or more comorbidities. In this study, hypertension, dyslipidemia, diabetes, chronic kidney disease, and liver cirrhosis were not associated with CLBP, but cerebrovascular events, cardiovascular disease, pulmonary disease, knee arthritis, and cancer history were found to be positively associated with CLBP.

To the best of our knowledge, this is the first study investigating personal factors associated with CLBP to this extent in a representative sample of the general South Korean population. The greatest strength of our study was the increased external validity of our findings due to the KNHANES data. The KNHANES has the advantage of obtaining a large amount of data from the general population nationwide [26]. With the use of nationwide sample data, the results can be generalized to the greater community.

There are some limitations to our study. First, this study was conducted through a national health and nutrition examination survey, which was designed to be cross-sectional. Therefore, we cannot determine causal relationships between the identified associated factors and CLBP. However, as mentioned above, this dataset was extracted from the South Korean population to minimize sampling errors, and the results can be considered very representative. Second, the KNHANES was designed to minimize sampling errors by utilizing a clustered, multi-stage, random sampling method. However, selection bias may exist due to missing data. Participants were selected from our raw data to minimize selection bias, but missing data inevitably led to bias. Unlike other studies, such as cohort studies and clinical trials, the imputation of missing values is impossible in our dataset. Therefore, we excluded participants with missing data, which was necessary for analysis. Third, the simple CLBP survey used in this study did not evaluate the severity, source, or duration of LBP, which would require instruments for measuring pain on a scale (e.g., the visual analog scale pain score). Fourth, this study could not analyze some prognostic factors for CLBP, such as previous episodes of CLBP, the severity of pain, and disability. However, this study concomitantly analyzed many other associated factors that were not included in other studies. Finally, our study produced slightly different results from those of previous studies. These previous studies determined associations by analyzing relatively few factors, but our study analyzed nearly all the individual, socioeconomic, lifestyle, and mental factors and comorbidities associated with CLBP. To control for confounding factors, we used multiple logistic regression analysis to assess meaningful associations. We believe this method is an excellent statistical technique for identifying associated factors [48].