| 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 |
| Kellgren and Lawrence, 19529 | 84 miners, 45 manual workers, 42 office workers | 40–50 | 73 | Occupation | Radiological grade 2–4 TFJAP and lateral view of the right knee | – | Miners vs office workersMiners vs manual workersManual workers vs office workers | 2.77, 1.2 to 6.33.03, 1.36 to 6.790.91, 0.34 to 2.48 | Cross-sectional | ++Strength: Well-described design and materials. Blinded evaluation of the radiographsWeakness: Not-weight-bearing radiographs. Exposure: Job title (risk of misclassification) |
| Wickström et al, 198310 | 252 reinforcement workers, 231 painters | 20–64 | 85 | Job titleVideo-recordings | Radiological grade1–4 and ⩾3 | Age | Reinforcement workers vs paintersAllSevere | 1.1, 0.72 to 1.741.1, 0.31 to 4.33 | Cross-sectional | ++Strength: Exposure assessment done by video-taping. Weakness: Relative few subjects in the age group >50 years. Only active workers included (healthy workers effect) |
| Enderlein and Kasch, 19895 | 173 welders77 shipbuilders107 pipe-fitters116 controls | 18–65 | – | Job title, keratoses on the kneesWelders 75%Shipbuilders 30%Pipe-fitters 10%Controls 0% | Radiological and clinical kneeOA grade 1–3 | –Population divided in age groups and by earlier traumas | WeldersShipbuildersPipe-fitters | 3.42.52.4 | Cross-sectional | ++Strength: Inclusion of subjects with high, moderate and low exposure for kneeling. Separate subjects with and without earlier trauma. Weakness: Poor description of design and materials, no adjustment for confounders |
| Vingård et al, 199114 | Male construction workers: 38 095 (114 OA) Controls: low exposure: Male: 91 057 (200 OA) Same occupation in 1960 and 1970 | Born 1905–1945 | Register-based | Classified by two experienced occupational health physicians | Hospitalised in 1981–83ICD-8: 713.01ICD-10: M17 | Age, county | Construction workers vs low exposure | RR 1.36, 1.13 to 1.79 | Cohort | ++Strength: High number of participantsWeakness: Exposure classification: job title+ light, medium, heavy (risk of misclassification) Case-definition = hospitalised because of kneeOA (risk of selection bias) |
| Vingård, 199217 | Cases: 181 men with disability pension due to kneeOA (34 construction workers; 13 painters and carpet layers) Controls: 298 from the general Swedish population | Born 1915–1934 | – | InterviewOccupational history | Primary diagnosis kneeOA made by a physician | Age | Construction workers vs never exposed to any of 20 most exposed occupationsPainters and carpet-layers | RR 5.1, 2.6 to 10.6RR 23.1, 3.0 to 178.3 | Case–control | ++Strength: Description of material and designWeakness: Diagnosis from physicians certificates (risk of misclassification) Exposure = high workload on knees, not especially related to heavy lifting (risk of misclassification) Subjects in physical demanding occupations may have an increased risk of getting disability pension (risk of selection bias) |
| Kivimäki, 199219 | 168 floor-layers146 painters (controls) | 25–45 | 72 | Job titleVideo-recording for kneeling activities | TFJ + PFJOsteophytes and JSNWeight-bearing radiographs | Age, occupation, smoking, BMI, knee injury | Floor-layers vs paintersTFJKnee osteophytesPFJ (caudal) PFJ (cranial) | 0.87, 0.17 to 4.361.96, 1.25 to 3.062.85, 1.85 to 4.41.3, 0.98 to 1.94 | Cross-sectional | ++Strength: Well-described design and materials, video-taped exposure measurements, weight-bearing radiographsWeakness: Exposure assessment: only for kneeling working activities. Only subjects up to the age of 49 years included |
| Jensen et al, 199421 | Total: 2 664 192 men and women(7588 OA) 10 223 construction workers (35 OA) 13 447 carpenters (40 OA) | 20–59In 1981 | Register-based | Occupation 1981 | Hospitalised 1981–90ICD-8: 713.01ICD-10: M17 | Age | MenConstruction workersCarpenters | SHR 144, 101 to 201SHR 159, 117 to 217 | Cohort | ++Strength: High number of participantsLongitudinal design (avoid information bias) Weakness: Use of job title (risk of misclassification). Main job has been registered in 5-year periods (risk of misclassification). Case definition = diagnostic code (risk of misclassification + risk of selection bias) |
| Greinemann, 199736 | Cases: 500 miners working at least 25 years undergroundControls: 500 without knee demanding work | 50 | – | Occupation | Radiological and clinical examination | – | Miners vs non-minersTFJ OAPFJ OA | 14.8, 7.3 to 30.13.83, 2.21 to 6.7 | Cross-sectional | +Strength: High number of participants (miners and non-miners). Weakness: Missing description of study design, inclusion and exclusion criteria, participation rate, case definition, and statistical analysis |
| Jensen et al, 200024 | 50 floor-layers, 50 carpenters, 49 controls without knee demands selected from responders in a questionnaire study of workers in Copenhagen area | 26–72 | 78In questionnaire study | Job titleVideo-recordings for kneeling | Kellgren and Lawrence grade 2–4 TFJNot weight bearing | Stratified by age, knee complaints | Radiographic knee OA (%) Floor-layersCarpentersGraphic designersSymptomatic kneeOA >50 years of age (%)Floor-layersCarpentersGraphic designers | 148664226 | Cross-sectional | ++Strength: Description of material and designs, video-taped exposure measurements. Blinded assessment of the radiographs. Clinical investigationWeakness: Few participants (underestimation). Selection of participants for the radiological study: stratified sample– only restrictive statistical analysis. Radiographs non-weight-bearing (non-differential risk of misclassification) Exposure to lifting not video-recorded |
| Sandmark et al, 200027 | Cases: 325 men, 300 women with TKRControls: 264 men, 284 womenFrom 14 counties in Sweden | 55–70 | 80 | Interviews questionnaireOccupational historyLifts (kg) | TKR between 1991–1993 | BMI, smoking sportsAge-matched | Men: ConstructionFarmersForestry workers Women: Farmers | 3.1, 1.5 to 6.43.2, 2.0 to 5.22.1, 1.0 to 4.62.4, 1.4 to 4.1 | Case–control | +++Strength: Well-described design and material, high number of participantsWeakness: Case definition: TKR (risk of selection bias) Exposure assessment retrospective (risk of recall bias). Use of job titles (risk of misclassification) |
| Yoshimura et al, 200431 | KneeOA cases: 101 women from six hospitals in three cities of JapanControls: 101 women, random sample from the local population | ⩾45 | 84% cases59% controls | QuestionnairePhysical demandsJob title | Radiological grade3–4 kneeOA | Knee injury | Construction workers among cases and controlsFirst jobMain job | 2.62, 1.37 to 5.01.30, 0.69 to 2.5 | Case–control | ++Strength. Same design as the Coggon study25Weakness: Differences in participation rate among cases and controls. 50% of cases and 14% of controls had a previous knee injury. Few exposed to heavy lifting. Cases defined as TKR (risk of selection bias) Exposure assessment: retrospective (risk of recall bias) |
| Holmberg et al, 200432 | Cases: 778 with kneeOA (338 men 440 women) Controls: 695 (293 men 402 women) From three counties of Sweden57 construction workers (only men) | Mean: 63 | 89% | QuestionnaireOccupation since the age of 15 years | Retrospectivex-ray 1999–2000TFJ OA | HeredityWeightSmokingCivil statusSelf-employment, knee injury, meniscal injury, sports | MenConstruction workers1–10 years11–30 years>30 years Farmers1–10 years11–30 years>30 years WomenFarmers1–10 years11–30 years>30 years | 1.5, 0.4 to 4.52.5, 1.0 to 6.01.6, 0.6 to 4.61.3, 0.6 to 2.10.8 0.3 to 2.11.7, 0.7 to 4.00.8, 0.4 to 1.92.1, 1.0 to 4.52.0, 0.7 to 5.5 | Case–control | ++Strength: High frequency of participantsWeakness: Cases identified via x-ray departmentsExposure: Retrospective (risk of recall bias) Use of job title (risk of non-differential misclassification) |
AP, anterior-posterior; JSN, joint space narrowing; PFJ, patello-femoral joint; SHR, standardised hospitalisation ratio; TFJ, tibio-femoral joint.
+, Poor quality score 1–5; ++, medium quality score 6–10; +++, high quality score >10.