Introduction

Osteoarthritis (OA) is a structural and functional degenerative joint disease involving the cartilage and the surrounding tissue [1]. OA is the most common joint disease in the world, with a prevalence of 10–15% of adults, and causes chronic pain and disability [2,3]. In a study in Korea, OA prevalence was 3.3% in males and 16.0% in females aged 50 years and older [4]. Because the average life span is gradually increasing, OA is also likely to increase with the size of the aged population. OA has considerable effects on social life and quality of life. OA has become an especially conspicuous problem in the recent era, in which people tend to actively work into old age.

Knees and hips are clinically frequent sites for developing OA [5,6]. Previous studies have shown that OA is associated with certain occupational risk factors, such as kneeling, squatting, climbing, heavy lifting, and vibration [7]. However, no study has examined whether OA is more or less likely to develop among people working in different occupational clusters. Generally, examination for OA has not been included in standard checkups, and OA has gone largely unobserved. Moreover, symptoms of OA are not correlated with radiologic findings [8,9], and severe radiological OA with pathological progression is often seen in patients who are asymptomatic or have mild symptoms.

OA is a slowly progressing disease, and the damaged cartilage is not able to recover normal status [5]; thus, prevention should be the focus. Early examination and diagnosis are necessary for the health of individuals and communities. Further, a finding that certain occupational clusters have higher prevalence rates of OA would indicate that greater awareness and caution regarding OA should be employed by these groups. We investigated OA prevalence by occupational cluster using data from the 2010–2013 Korean National Health and Nutrition Examination Survey (KNHANES).

Materials and Methods

Results

Table 1 provides the characteristic of study. The total number of weighted study subject were approximately 5.8million in males, approximately 6.1million in females. In males, 27.4% were white collar, 14.2% were pink collar, 41.3% were blue collar, and 17.1% were green collar. In females, 14.1% were white collar, 32.2% were pink collar, 34.2% were blue collar, and 19.5% were green collar. In males, 5.3% had OA, in spite of, in females, 18.4% had OA.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristic of study subject.

https://doi.org/10.1371/journal.pone.0170229.t001

Table 2 shows the demographics of unweighted study samples according to OA. In both male and female, occupation cluster, age, alcohol consumption, HTN, and income level had significant differences between OA group and non-OA group, while smoking status had no significant differences. BMI and DM had significant differences between OA group and non-OA group in female but not in male.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Demographics of unweighted study samples according to OA. (n = 9905).

https://doi.org/10.1371/journal.pone.0170229.t002

Compared to the white collar group, the ORs of the pink, blue, and green collar groups on OA were 1.27 (95% CI 0.67–2.38), 1.93 (95% CI 1.27–2.94), and 3.86 (95% CI 2.51–5.94), respectively, in males, and 3.41 (95% CI 2.32–5.02), 4.27 (95% CI 2.93–6.23), and 8.17 (95% CI 5.58–11.95), respectively, in females. In the multiple logistic regression model, adjusted for age and BMI (Model I) the ORs for pink, blue, and green collar were 1.34 (95% CI 0.71–2.53), 2.07 (95% CI 1.36–3.15), and 3.14 (95% CI 2.05–4.82), respectively, in males, and 2.66 (95% CI 1.81–3.93), 3.05 (95% CI 2.08–4.47), and 4.12 (95% CI 2.77–6.13), respectively, in females. In the multiple logistic regression model, adjusted for age, alcohol intake, smoking, BMI, hypertension, DM, and income level (Model II) the ORs for pink, blue, and green collar were 1.23 (95% CI 0.64–2.36), 1.85 (95% CI 1.18–2.88), and 2.91 (95% CI 1.86–4.54), respectively, in males, and 2.53 (95% CI 1.71–3.73), 2.86 (95% CI 1.94–4.21), and 3.90 (95% CI 2.60–5.83), respectively, in females. The p-value for the Cochran-Armitage trend test about OA according to occupation in order white, pink, blue, green was 0.001 or less in both males and females. (Table 3, Fig 1)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Osteoarthritis according to occupation clusters.

https://doi.org/10.1371/journal.pone.0170229.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Odds ratios and 95% confidence intervals of occupation on osteoarthritis among males and females.

https://doi.org/10.1371/journal.pone.0170229.t003

Discussion

We classified the occupations into four occupation-cluster groups by physical loading. The amount of manual labor, in order from highest to lowest, was green, blue, pink, and white collar group. Actually, our study showed that the green-collar group showed the highest prevalence rate of OA, followed by the blue-collar group, and then the pink-collar group. The white-collar group showed the lowest prevalence in both male and female. In the Cochran-Armitage trend test, about OA according to occupation in order white, pink, blue, and green, also statistically significant. On the basis of these results, we argued that physical labor can induce and develop OA. Particularly in the blue and green-collar groups, occupation was more strongly associated with OA than any other risk factor with the exception of age, including alcohol consumption, smoking status, obesity, hypertension, and DM (Table 2). Thus, occupation type was the second mostly important risk factor among the covariates tested in our study.

The results showed no differences in OA prevalence between white collar and pink-collar workers in males, but did see a difference in females. These results suggest that males in the pink collar group worked at similar levels to males in the white collar group, whereas females in the pink collar group had more physical activity, similar to females in the blue collar group.

Occupational classification in ISCO-08 was based on 10 major groups. Major Group 1: managers; Major Group 2: professionals; Major Group 3: technicians and associate professionals; Major Group 4: clerks; Major Group 5: service and sales workers; Major Group 6: skilled agricultural and fishery workers; Major Group 7: craft and related trade workers; Major Group 8: plant and machine operators and assemblers; Major Group 9: elementary occupations; and Major Group 0: armed forces occupations [16]. Occupation was re-classified into four groups according to the social and culture environment of Korea and skill level. White collar workers included Major Group 1, 2, and 3, pink included Major Group 4 and 5, green included Major Group 6, and blue included Major Group 7, 8, and 9 [17]. In Korea, green collar workers have much more physical activities than blue.

Females showed a higher prevalence rate of OA than males in this study. Compare to white collar group, odds ratios of pink, blue and green collar groups in females were 1.5−2 times higher than those in males. Srikanth et al.’s meta-analysis of gender differences in prevalence of OA similarly showed that females had a significantly higher incidence of knee and hip OA than males [18]. This finding suggests that hormonal factors may affect the development and progression of OA, and structural differences, such as knee–ankle angle and cartilage volume, may also influence OA [19,20]. Women have higher femoral lateral bowing in knee OA and increased rates of cartilage loss and progression of cartilage defects at the knee than men.

Additionally, in the current study, age was associated with OA prevalence in both males and females. Age and obesity are well known risk factors of OA [21,22]. OA is a degenerative disease where chondrocytes can perform only a limited number of replications. Because chondrocytes are the main factor associated with the development and progression of OA, OA is an inevitable consequence of aging [23]. A previous study using data from KNHANES 2010–2012 found that smoking was a risk factor of OA in Korean males [4]. However, we include data of KNHANES 2013 and found that smoking was not associated with OA. Furthermore, inverse correlation was reported between smoking and knee OA [24,25]. Bae et al. showed that HTN was also associated with OA using KNHANES 2007–2009 data [26]. However, the findings of these earlier studies using KNHANES 2007–2009 data were not consistent with those of the current study.

Alcohol eventually has a pathological effect on joints and increases OA development; both animal and case-control studies have verified this relationship between alcohol consumption and OA [27,28]. However, previous study showed that alcohol consumption was not associated with OA. DM was not associated with OA in either gender in this study. Overweight status has been shown to be strongly associated with OA, and physical activity tends to lower BMI [29,30]. Therefore, the fact that the green and blue collar occupations involved higher levels of physical activity may have influenced the results of this study.

Our study has some strengths. First, this study was based on authoritative nationwide data (KNHANES 2010−2013). Thus, our study can show recent Korean status and tendencies of relationship between OA and occupational clusters. In addition, we defined OA as a combination of clinical symptoms and radiologic findings. Radiologic OA considers only pathophysiological joint signs present in radiographic images, whereas symptomatic OA considers joint pain or crepitus [31,32]. These two components are both important in the diagnosis of OA. We also examined the association of OA with occupation type according to gender stratification. As mentioned earlier, gender is an important factor of OA development and progression. Finally, this study establishes relationship between occupational clusters and the prevalence of OA, which was the first study, to the best of our knowledge.

Some limitations are present in this study. First, the cross-sectional design precludes conclusions about causal relationships, so further prospective studies and intervention trials should be undertaken to establish a causal association between occupational cluster and OA prevalence. However, the relationship of OA to occupation is important in that the progression of OA makes physical activity difficult, limiting OA patients’ ability to engage in occupation-related physical activities. Second, the occupational clusters were based on self-reports. Consequently, these data may have been influenced by systematic errors in individuals’ consideration of their occupations, which may have led to non-differential misclassification or underestimation. However, the questionnaire subdivided occupation into four groups and presented examples of each occupational cluster, fostering participants’ selections of the correct occupational types.

The present study found that the prevalence rate of OA is associated with occupational cluster. The prevalence of OA, in order from highest to lowest, was observed in the green, blue, pink, and white collar occupation clusters. Therefore, workers in occupational clusters at high-risk for OA should use caution and preventative measures for OA. Additionally, early diagnosis and prevention systems for OA should be prepared to improve public health care.

Acknowledgments

No grants or other support were received for this study.