| Reference | Study population | Age (years) | Participation rate | Exposure measured by: | Diagnostic criteria | Adjusted for/matched by: | Comparisons | Results, OR (95% CI) | Design | Quality of the study (+, ++, +++); strength/weakness |
| Typpö, 198524 | Cases: 224 with hip OA (incl 90 farmers) Controls: 255 without hip OA (incl 70 farmers) All radiologically examined by veno-, angio-, uro-, colo-, or cystography, hips and abdomen | 16–86, mean: 57 | – Selected among 416 males and 503 females | Questionnaire Present occupation (job title) | Retrospective radiological hip OA mild, moderate, severe | – | Farmers vs office workers | Case- control | + Strength: none. Weakness: participation rate missing. No statistical testing. No adjustment for potential con-founders. Only data for a part of the study population (n = 505) | |
| Mild/moderate hip OA | 1.8 (0.97 to 3.34) | |||||||||
| Severe hip OA | 1.98 (1.01 to 3.87) | |||||||||
| Jacobsson 198720 | 85 males waiting for hip replacement 262 males who have had urography of whom 106 had hip OA | 70–76 | – | Questionnaire “working as a farmer” | Joint-space <3 mm or waiting for THR | Age, height, weight | Farm work vs others | Case-control | + Strength: none. Weakness: participation rate not described. Results only sparsely described. No statistical analyses | |
| THR | 1.8 (1.12 to 3.02) | |||||||||
| JSN | 2.9 (1.2 to 7.37) | |||||||||
| Thelin 199034 | Cases: 98 males who have had THR at two hospitals in Sweden Controls: 201 random sample of Swedish males | 55–70, mean 65 | 91% | Questionnaire Occupational history from age 15 | THR | Age, injuries, tobacco, hospital | Farming vs no farming (<1 year) | Case-control | ++ Strength: high participation rate. Specified exposure. Weakness: relatively few cases (but narrow confidence intervals around the point estimates indicate adequate statistical power). Case definition (THR, risk of selection bias). Exposure definition (job title, risk of recall bias) | |
| 1–10 years | 2.1 (1.1 to 4.3) | |||||||||
| >10 years | 3.2 (1.8 to 5.5) | |||||||||
| Drove tractor regularly | 2.2 (1.3 to 3.9) | |||||||||
| Milking regularly | 2.2 (1.3 to 3.7) | |||||||||
| Vingård 199127 | Male farmers: 35981 (479 OA) Female farmers: 1739 (12 OA) Controls: male: 91057 (320 OA); female: 24145 (112 OA) | Born 1905–45 | Register-based | Physical demands classified by two experienced occupational health physicians Same occupation in 1960 and 1970 | Hospitalised in 1981–3 ICD-8 713.00 (hip OA) | Age, county | Farmers vs workers with low exposure to physical work demands | Cohort | ++ Strength: design and material well described. High number of participants. Weakness: exposure (job title + light, medium, heavy exposure, risk of misclassification). Case definition (hospitalised because of hipOA, risk of selection bias) | |
| Males | 3.78 (2.1 to 4.4) | |||||||||
| Females | 1.47 (0.86 to 2.9) | |||||||||
| Croft 199216 | Cases: 245 males with hip OA farmers: all degree: 52 OA severe: 19 OA Controls: 294 without hip OA farmers: 65 examined by urography | 60–75 | 68% | Blinded interview Occupational history Job title classified by experienced persons | Joint-space All <2.5 mm Severe <1.5 mm | Age, sport, BMI | All OA cases (<2.5 mm) | 0.9 (0.6 to 1.4) | Case-control | +++ Strength: use of intravenous urograms (may avoid selection bias). JSN definition. Results and analysis well described. Weakness: exposure measurement (lift >25.4 kg, frequency not further explained, risk of misclassification) |
| Severe cases (<1.5 mm) | 1.6 (0.8 to 3.1) | |||||||||
| Farming vs <1 year | ||||||||||
| 1–9 years | 1.0 (0.3 to 3.1) | |||||||||
| ⩾10 years | 2.0 (0.9 to 4.4) | |||||||||
| Croft 199231 | 167 male farmers (28 OA) 71 (83) sedentary workers (20 OA) from general practice of 1231 males | 60–76 | 60% | Questionnaire Interview Years as farmer at least 1 year | Radiological hip OA Joint space <1.5 mm or hip replacement | Age, height, weight, Heberden’s nodes | Farmers | Cross-sectional | +++ Strength: selected from general practice, outcome measurement = JSN <1.5 mm. Weakness: low participation rate (60%). Over-representation of symptomatic farmers among the participants. No exclusion criteria described | |
| 1–9 years | 5.8 (1.1 to 31.5) | |||||||||
| ⩾10 years | 10.1 (2.2 to 45.9) | |||||||||
| Vingaard 199226 | Cases: 140 males, disability pension due to hip OA (17 farmers) Controls: 298 males from the general population | Born 1915–34 | 73% | Interview Occupational history Job title classified by experienced persons | Primary diagnosis of hip OA made by a physician | Age | Farmers and forest workers vs others | Relative risk 13.8 (4.0 to 48.1) | Case-control | ++ Strength: description of design and material. Weakness: diagnosis from physician certificates (risk of misclassification). Subjects in heavy occupations may have an increased risk of getting disability pension (risk of selection bias). Exposure “high workload on hips” may not be especially related to heavy lifting (risk of misclassification) |
| Axmacher 199330 | 565 male farmers (45 OA) chosen among 16250 active farmers in the population in Malmö county 1250 from general population in another study used as controls (10 OA) | 40–64 | 83% | Questionnaire Working as a farmer | Retrospective review of colon- and urography + non-weight-bearing joint space <4 mm | Age, gender | Farmers vs urban controls | Cohort | + Strength: high number of farmers included. Weakness: results are compared with an earlier population study (use as controls). No statistical analysis. Non-weight-bearing radiographs (risk of non-differential misclassification). Case definition: JSN <4 mm (high risk of misclassification) | |
| Males | 12.0 (6.7 to 21.4) | |||||||||
| Females | 2.3 (0.33 to 12.3) | |||||||||
| Jensen 199432 | Male farmers: total 63990 (1131 hip OA) Total: 125159 males (9674 hip OA) | 20–59 in 1981 | Register-based | Job title in 1981 | Hospitalised 1981–90 ICD-8 713.00 hip OA | Age | Farmers vs other occupations | SHR 273 258–7290 | Cohort | ++ Strength: high number of participants. Longitudinal design (may avoid information bias). Weakness: use of job title (risk of misclassification). Main job registered during 5-year periods (risk of misclassification of workers who only have worked a few years in their main job). Case definition: diagnostic code (risk of misclassification + risk of selection bias). Danish report, not peer-reviewed |
| Thelin 199733 | Cases 216 with radiological hip OA (136 farmers) Controls: 479 randomly selected from a local population register. (185 farmers) | <70 | 86% | Questionnaire Worked as farmer at least 1 year Specified physical activity as a farmer | Retrospective readings of radiological hip OA joint space <3 mm | Age, gender, residence-matched investigation of an association between hip OA sports, and smoking | Farming vs no farming | Case-control | +++ Strength: well described design and material. High number of participants, high participation rate. Case definition: radiological findings in previously taken x rays may avoid some risk of selection bias. Specified exposure in farming. Weakness: farmers with heavy physical demands may seek hospital (and have x ray taken) more often than other subjects (risk of selection bias) | |
| 1–10 years | 1.58 (0.59 to 4.23) | |||||||||
| 11–20 years | 2.81 (1.31 to 6.03) | |||||||||
| 21–30 years | 7.35 (2.87 to 18.8) | |||||||||
| >30 years | 3.82 (2.41 to 6.06) | |||||||||
| Farm worker | ||||||||||
| 1–10 years | 1.88 (1.23 to 2.86) | |||||||||
| 11–20 years | 2.53 (1.36 to 4.72) | |||||||||
| 21–30 years | 4.41 (1.31 to 14.8) | |||||||||
| >30 years | 6.43 (1.83 to 22.5) | |||||||||
| In agriculture Y/N | 2.70 (1.94 to 3.77) | |||||||||
| Drive tractor Y/N | 2.05 (1.45 to 2.88) | |||||||||
| Milk full-time Y/N | 2.98 (2.07 to 4.28) | |||||||||
| Coggon 199815 | Cases: 210 males 401 females (19 farmers) waiting for hip replacement in three English districts Controls: 210 males, 401 females (8 farmers), random sample from general practice in same area | 45–91, mean: 70 | 68% 84% of cases, 58% of controls | Interview Occupation held for >1 year from school age Specified physical activity | Cases waiting for surgery | – | Farmers vs others | 2.5 (1.10 to 5.70) | Case-control | + Strength: the same study also investigated heavy lifting (table 2). Weakness: only the numbers of farmers are mentioned. No analysis was made. Case definition: waiting list for THR (risk of selection bias). Participation rate relatively low Retrospective exposure measurement (risk of recall bias). No adjustment for confounders |
| Yoshimura 200029 | Cases: (103 females 11 males) waiting for hip replacement in 2 districts in Japan (19 farmers or fishermen) Controls: 114 from the local population (17 farmers or fishermen) | >45, mean: 64 | 91% | Questionnaire: Occupation since leaving school; physical activity in their first and main job | Waiting for hip replacement | Matched by age, gender, residence | Farmers or fishermen vs non farmers or fishermen | 1.14 (0.57 to 2.33) | Case-control | + Strength: same design as Coggon et al, 1998. The study also investigated the exposure heavy lifting. Weakness: few participants, especially few males (including farmers). Farmers and fishermen included in same group. Case definition (THR, risk of selection bias). Retrospective exposure data (risk of recall bias) |
| Tüchsen 200336 | Total between 34068 (355 OA) and 63990 (458 OA) All actively working males in Denmark in 1981, 1986, 1991, 1994 | 20–59 | Register-based | Occupation classified by occupation (job title) 1980, 1985, 1990, 1993 | Hospitalised with hip OA (ICD-8 713.00 or ICD-10 M16) 1981–5 1986–90 1991–3 1994–9 | Farmers vs others | SHR | Cohort | +++ Strength: all hospital admissions due to hipOA during 20 years are included. Longitudinal design. Relatively young subjects (20–59 years), although the results showed increased risk of hipOA. Weakness: case definition (diagnostic code, risk of misclassification). No adjustment for age, BMI, traumas, or sports activities. Subjects not active in their trade were not included (risk of healthy worker effect) | |
| 1981–5 | 281 (259 to 304) | |||||||||
| 1986–90 | 283 (269 to 298) | |||||||||
| 1991–3 | 285 (268 to 302) | |||||||||
| 1994–9 | 286 (262 to 313) | |||||||||
| Farm worker vs others | ||||||||||
| 1981–5 | 114 (89 to 147) | |||||||||
| 1986–90 | 138 (118 to 161) | |||||||||
| 1991–3 | 160 (140 to 183) | |||||||||
| 1994–9 | 189 (158 to 227) | |||||||||
| Thelin 200435 | Cases: 369 farmers with hip OA Controls: 389 farmers without hip pain From a Norwegian farmers’ cooperative (30000 persons) | 40–71 | 86% | Interview Work tasks as a farmer | Retrospective and new radiological hip OA <3 mm | Age, sex, residence-matched | Working >5 h/day in livestock housing | 13.3 (1.2 to 145) | Case-control | ++ Strength: high number of farmer participants. Weakness: no control group of non-farmers. The study cannot conclude if there is a higher risk among farmers than non-farmers, but can investigate if there are risks within specific work tasks |
| Milking >40 cows daily | 4.5 (1.9 to 11.0) | |||||||||
| Working at large farms >100 ha | 0.14 (0.05 to 0.4) |
THR, total hip replacement; SHR, standardised hospitalisation ratio; BMI, body mass index (weight/height2); OA, osteoarthritis.
+, Poor quality score 1–5; ++, medium quality score 6–10; +++, high quality score >10.