To What Extent Is Severe Osteoarthritis Preventable?

Occupational and Non-Occupational Risk Factors for Knee and Hip Osteoarthritis

Tea Kontio; Markku Heliövaara; Eira Viikari-Juntura; Svetlana Solovieva


Rheumatology. 2020;59(12):3869-3877. 

In This Article


Study Population

A nationally representative random sample of 8028 individuals aged 30 years or over who lived in mainland Finland and were alive on 1 July 2000 was drawn by the Social Insurance Institution of Finland using a two-stage sampling design. A large national Health Examination Survey (the Health 2000 Study), which comprised questionnaires and interviews, as well as a health examination, is described in detail elsewhere.[20] The Health 2000 Study population, 5254 individuals aged 30–59 years, comprised our study base (Figure 1). We excluded persons who had a prior hospitalization due to knee or hip OA (n = 80), received a disability retirement benefit due to knee or hip OA prior to participation in the Health 2000 Study (n = 11), or had missing information on weight (n = 463) or physical work-related factors (n = 58). Our cohort consisted of 4642 individuals (2247 men and 2395 women) who were followed from the day of their participation in the Health 2000 Study (baseline) until the first hospitalization due to knee or hip OA, death or end of follow-up (31 December 2015), whichever came first. Date of death from the Population Information System was provided to the study by the Social Insurance Institution of Finland.

Figure 1.

Flow chart of sample formation

The Section for Epidemiology and Public Health of the Ethics Committee of the Hospital District of Helsinki and Uusimaa approved the study design of the Health 2000 Study 31 May 2000. Participation in the study was voluntary and the subjects were asked to sign an informed consent form. The consent contained a request for permission to use participants' data in studies on health, various diseases and their determinants, including linkage to the registers.

Hospitalizations Due to Knee or Hip Osteoarthritis

The National Hospital Discharge Register provides information on the start and end dates as well as primary and secondary diagnoses for all hospital admissions. This register is held by The National Institute for Health and Welfare, and is a previously known reliable source of information.[21] The diagnoses were classified according to the International Classification of Diseases (ICD). In Finland, the 8th edition of the ICD was used before year 1986, the 9th edition in 1986–95 and the 10th edition from 1996 onward. The diagnoses of primary interest of this study were knee OA (ICD-8: 713.01; ICD-9: 175.16; and ICD-10: M17) and hip OA (ICD-8: 713.00; ICD-9: 175.15; and ICD-10: M16).

Potential Risk Factors

Age at baseline was divided into 10-year categories. Height and weight were measured in light indoor clothing during the health examination. BMI was calculated by dividing weight (kg) by the square of height (m2). Overweight and obesity were defined based on BMI using the World Health Organization recommendation of BMI <25 kg/m2 (normal), 25–29.9 kg/m2 (overweight) and ≥30 kg/m2 (obese). Data on length of education (in years), occupation/occupations, seniority (years in each occupation), physical workload factors in each occupation and smoking were collected during the home interview at baseline. Level of education was defined as (i) high (>12 years), (ii) middle (9–12 years) or (iii) low (<9 years).

Occupational History and Physical Workload Factors

Exposure to the following physical workload factors in each occupation was assessed with the home interview: physically heavy work (work involving, e.g., lifting and carrying heavy loads, excavating, shovelling or hammering), manual handling of heavy loads (lifting, carrying or pushing items heavier than 20 kg at least 10 times per day), kneeling or squatting (work demanding kneeling or squatting for 1 h per day or longer), and standing or walking (work demanding standing or walking for ≥5 h per day). Cumulative exposure for each physical workload factor was calculated as the total number of years having been exposed during the entire work career. First, we calculated the cumulative exposure for each workload factor prior to the baseline using self-reported information on the number of years a person had worked in occupations involving the workload in question. Second, we calculated cumulative exposure during the follow-up using information regarding employment and unemployment periods as well as earnings-related pensions from the Finnish Centre for Pensions to estimate the years a person had worked in the most recent occupation reported at baseline. We assumed that persons did not change their occupation during follow-up.

Furthermore, we used K-means cluster analysis to empirically identify three latent clusters of composite cumulative workload to capture the duration and pattern of co-occurrence of exposures during working life more thoroughly. The clusters were referred to as low (n = 2324, 50%) intermediate (n = 1855, 40%) and high composite cumulative workload (n = 463, 10%), and are described in Supplementary Table S1, available at Rheumatology online.

Smoking was assessed with the following question: 'Have you ever smoked in your life', with two possible responses 'No' and 'Yes'. Smoking data were missing for 182 persons. This dichotomous question was chosen because the other questions including frequency and amount of smoking had more missing values.

Information on leisure-time physical activity (LTPA) was collected via self-administered questionnaire and dichotomized into no or irregular (two to three times a month or less) and regular (at least once a week). Information on LTPA was missing for 53 persons.

Major injury of the knee or hip prior to baseline and during follow-up was obtained from the National Hospital Discharge Register. The ICD codes used to identify these injuries are listed in the Supplementary Material, available at Rheumatology online.

Statistical Analyses

Competing risk regression models (stcrreg, STATA version 14, StataCorp 2015, College Station, TX, USA) were used to examine associations of possible risk factors with hospitalization due to knee or hip OA, respectively. We accounted for the impact of death (as competing risk) on the outcome. Adjusted hazard ratios with 95% CIs were estimated. First, we explored the effect of age and gender on the outcome. Second, we estimated the association between each potential risk factor (education, prior injury, BMI, LTPA, smoking, cumulative physical work load factors and composite cumulative workload) and hospitalization due to knee or hip OA controlling for age and gender. Finally, we ran the fully adjusted model including the following risk factors: age, gender, prior injury, BMI and composite cumulative workload. In addition, we repeated the analyses separately for those ≤60 or >60 years of age at the end of follow-up.

We calculated the PAF to estimate the proportion of new cases of hospitalizations due to knee or hip OA that could be attributed to prior injury, BMI and cumulative workload. The formulas are presented in the Supplementary Material, available at Rheumatology online.

A sensitivity analysis was carried out (described in Supplementary Material, available at Rheumatology online).