Article Text
PDF
Abstract
Objective It is postulated that increased load from pinch and grasp in occupational tasks increases the risk of thumb carpometacarpal (CMC1) osteoarthritis (OA). We sought to characterise the relationship between doctor-diagnosed CMC1 OA and occupation in a large working population.
Methods We performed a matched case–control study using a Swedish healthcare register. We identified residents aged 30–65 years in 2013 with physician-diagnosed CMC1 OA from 1998 to 2013. We matched four controls per person with CMC1 OA by age, sex, education and postcode. Swedish Standard Classification of Occupations was used to assign occupation. Occupation was categorised as light, light–moderate, moderate and heavy labour. We used conditional logistic regression to estimate ORs with 95% CIs.
Results We identified 3462 patients with CMC1 OA and matched 13 211 controls. The mean age of the CMC1 OA group was 63 (SD 7) years, with 81% women. The ORs for CMC1 OA in men were 1.31 (95% CI 0.96 to 1.79) for light–moderate, 1.76 (95% CI 1.29 to 2.40) for moderate and 2.00 (95% CI 1.59 to 2.51) for heavy compared with light work. Women had ORs for CMC1 OA of 1.46 (95% CI 1.32 to 1.61) for light–moderate, 1.27 (95% CI 1.10 to 1.46) for moderate and 1.31 (95% CI 1.07 to 1.59) for heavy compared with light work.
Conclusions The association between increased manual load in occupation and risk of CMC1 OA is more pronounced in men than in women, likely due to higher workload in the heavy labour category.
- occupation
- thumb carpometacarpal joint
- osteoarthritis
Statistics from Altmetric.com
Key messages
What is already known about this subject?
There are minimal data in the literature, limited to small case series and low-volume cohort studies, about the impact of occupation on the development of thumb base osteoarthritis (OA), which is common and can be disabling.
What are the new findings?
We suggest a dose–response relationship between heavier occupational load and the development of thumb base OA in men. Occupations with high load on thumb joints are associated with higher odds of thumb base OA in both men and women.
How might this impact on policy or clinical practice in the foreseeable future?
These findings may allow physicians to recommend joint protective exercises in those at high risk of developing thumb base OA. These data could also be useful in developing protocols to decrease stress on these hand joints to be implemented in the workforce.
Introduction
The development of osteoarthritis (OA) of the thumb carpometacarpal (CMC1) joint is typically characterised by degeneration and narrowing of the CMC1 joint on plain radiographs and often by pain with grip and pinch.1 The CMC1 joint has been shown to bear the highest load among the joints of the thumb,2 with biomechanical studies demonstrating that a simple pinch placed loads of 12 kg across the CMC1 joint, compared with 3.0 and 5.4 kg at the interphalangeal and metacarpophalangeal joints. Strong grasp causes loads up to 120 kg across the CMC1 joint.
While increased risk of joint degeneration and OA due to the impact of repetitive and heavy loading has been relatively consistently reported for the knee and spine,3 4 the same evidence has not been shown for CMC1 OA. Limited case–control studies have shown elevated ORs of hand OA in those engaged in work requiring dexterity or repetitive work5 or with vibrational tools.6 There are also a few small case series of thumb OA in ironers7 and cashiers.8
The impact of occupational load on the subsequent development of CMC1 OA is not well characterised. Thus, our goal was to learn whether occupations with higher physical demand place workers at higher risk of CMC1 OA. We therefore cross-referenced a population health register with individual-level data with occupation registration using a national database. We hypothesised that heavier occupational demands would be associated with higher occurrence of doctor-diagnosed CMC1 OA in a working-age population.
Material and methods
We report this study according to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines. For this matched case–control study, we used data from the Skåne Healthcare Register (SHR). This register includes all healthcare contacts by any resident of the Skåne region in southern Sweden, capturing visit data and diagnosis coding from all public healthcare providers (both outpatient and inpatient) in a region with 1.2 million inhabitants. The treating physician assigns diagnoses using International Classification of Diseases-10 (ICD-10) codes. We identified cases, specifically all men and women between 30 and 65 years living in the Skåne region in 2013, who were assigned the ICD-10 code of M18 (the specific code for CMC1 OA) during two or more visits to a physician during the time period from 1998 to 2013. Requirement of two coded diagnoses entered by physicians provides a more stringent presumption of a diagnosis made on the basis of examination and/or radiographs. We matched up to four randomly sampled control subjects from the general population in year 2013, who had at least one healthcare contact during the specified time period, for each identified case. We further matched by year of birth, sex, education level and postcode (geographical area). The year 2013 was chosen because this was the most recent year with complete occupational data reporting.
We then cross-referenced the medical information on an individual level with data from the Swedish Occupational Register, maintained by Statistics Sweden. This register uses the Swedish Standard Classification of Occupations (SSYK) from 2012, based on the International Standard Classification of Occupations-08 (ISCO) from 2008. The Swedish Occupational Register contains occupation data on all persons aged 16 or older, registered in Sweden, whose employers report their occupation yearly, using the various SSYK categories. The SSYK classification divides occupations into 28 categories (online supplementary appendix A).
Supplemental material
To define exposure, we assigned each person to the profession in which they were registered in Statistics Sweden during 2013. We created a categorisation of occupations by load, based on the work of Holtermann et al,9 and divided occupations into four levels according to the physical demands of work: light, light–moderate, moderate and heavy labour.9
We also examined a subgroup of occupations categorised as repetitive, defined as that involving repeated cycles of hand/wrist use, and compared this to the light and heavy occupational groups (online supplementary appendix B). This group included occupations of restaurant staff, nurses, dental hygienists and computer workers.
Supplemental material
Statistical analysis
Descriptive data from the study sample are presented as means and SD or frequencies and percentages, as appropriate. For analysis of the association between occupation exposure and CMC1 OA, we used conditional logistic regression analysis. The estimates are presented as ORs with 95% CIs.
Results
We identified 3813 persons with physician-diagnosed CMC1 OA in the Skåne region in 2013. We excluded the 32 persons in the military (due to low numbers), 41 persons with missing data and those who were not assigned to any occupational group in 2013 (7% of OA cases and 8% of eligible controls). We matched four controls per case in 89% of cases, while 38 cases had no available control and were excluded. The resulting study sample consisted of 16 673 persons in total, with 3462 persons with CMC1 OA and their 13 211 matched controls.
The case sample with CMC1 OA had a mean age of 63 (SD 7) years in 2013, with 2770 women (81%) and 692 men. The majority of workers with CMC1 OA and controls were categorised in the light or light–moderate occupational groups (n=12 651/16 718, 75.7%). Men dominated the heavy occupation group, despite the fact that the CMC1 OA sample was more than three-fourths female (table 1).
Occupational exposure for men and women CMC1 OA cases and controls
The overall estimated ORs for CMC1 OA were 1.46 (95% CI 1.33 to 1.60) for light–moderate manual work, 1.34 (95% CI 1.18 to 1.53) for moderate manual work and 1.58 (95% CI 1.38 to 1.82) for heavy manual work as compared with the light work category as the referent. When stratified by sex, the analysis showed a dose–response pattern in men, with heavier occupational demand associated with higher OR for doctor-diagnosed CMC1 OA (table 2). We did not find a similar association for women, although all three occupational categories were associated with higher odds of CMC1 OA compared with the referent light population.
Occupational level and OR for CMC1 OA diagnosis
Because the work classification was based on loading stress of occupations, we also separately analysed a subgroup of occupation descriptions as involving repetitive use of the hands. We analysed the repetitive group for odds of CMC1 OA compared with light and heavy occupational groups. We identified a total of 2393 persons (including cases and controls) categorised into occupations defined as repetitive and noted that repetitive occupations conferred a higher OR of a diagnosis of CMC1 OA. The odds of a CMC1 OA diagnosis were comparable between repetitive and heavy occupations in women, with a higher risk in men based on heavy work as compared with repetitive work (table 3).
Repetitive occupations and OR for CMC1 OA diagnosis
Discussion
While the impact of occupational load on joint biomechanical forces makes intuitive sense, there are limited data to support a relationship between occupation and CMC1 OA. This is an important inquiry with regard to occupational productivity and economic impact. A recent analysis evaluated healthcare contacts and productivity costs in an employed cohort treated for CMC1 OA with either injection or surgery, and estimated total healthcare and productivity costs for surgical patients at €5770 and €5548, respectively.10 These data indicate that CMC1 OA has a meaningful impact on occupational function.
In this study, we evaluated a large regional population using cross-linked databases for occupation and healthcare. Our findings indicate a dose–response relationship between increasing occupational demand in men and a physician diagnosis of CMC1 OA, with the heavy occupations group at twice the odds of developing OA compared with controls. These findings are similar to a previous register-based study of hip and knee OA, which showed that while men and women both had increased cumulative risk dependent on occupation, men had overall higher incidence of OA related to occupation.11 A meta-analysis of hip OA reported a similar dose–response relationship in men, but not in women, between cumulative lifting of heavy loads and the development of OA.12 Based on these studies in weightbearing joints, as well as previous biomechanical studies showing that strong grasp activities confer the highest loads to the CMC1 joint,2 it can be postulated that occupational joint loading has a deleterious impact on articular cartilage leading to OA.
A recent meta-analysis of 19 studies evaluating finger and wrist OAs demonstrated a meaningful association between pinch grip work and CMC1 OA.13 Most studies of occupational load at the hand have focused on repetitive use activities. One study of chopstick use as a repetitive stress evaluated 1008 men and 1499 women with analysis limited to the hand joints (metacarpophalangeal, proximal and distal interphalangeal), and reported higher prevalence of OA in the interphalangeal joint of the thumb and the joints of the index and middle fingers.14 Two previous comparative studies focused on the impact of repetitive use occupations on CMC1 OA, with relatively small sample sizes, and showed increased risk of CMC1 OA with repetitive load.5 6 We found that women performing light–moderate occupation tasks were at highest risk of having a CMC1 OA diagnosis. Additionally, when repetitive occupations were analysed separately, those engaging in repetitive work were at higher risk of a CMC1 OA diagnosis. These findings are in agreement with previous literature and underscore the role of repetitive use as a mechanism for the development of CMC1 OA.
This study has limitations, mainly attributable to large database use. The SSYK occupational database has limited capture of self-employed persons in Sweden who can self-define their occupation and do not consistently register. The SHR uses codes input by physicians, and if patients had multiple codes at one visit, the CMC1 OA ICD-10 code may not have been captured accurately. We attempted to control for this limitation by requiring two separate M18 codes to increase specificity of our case definition. The SHR contains information about all visits to a physician, including primary, specialist and in-patient care. We did not control for type of healthcare contact or diagnosis at contact, but we did require the controls to have at least one doctor visit, as this was also required from the cases. If controls presented for heavy occupational load issues, we can suspect they did not seek care specifically for CMC1 OA. The random sampling of controls and the fact that the proportions of heavy occupation groups were similar in case and control groups support lower risk of dilution of the control group.
Risk factors for CMC1 OA, such as family history or history of trauma, as well as obesity, diabetes, parity and oestrogen use, were not controlled for in this study. Most of this information was not available in the register, which does not specifically collect information on family history, trauma history, body mass index, parity or medications. Finally, the occupational titles used in this study were primarily those defined by ISCO classification, but our subanalysis of repetitive occupations was self-defined and may have misclassified certain occupations.
In conclusion, this case–control study reveals a dose–response relationship between increasing level of occupational demand and diagnosis of CMC1 OA in men, as well as overall increased risk of CMC1 OA in the heavier labour groups. The impact of occupation on the development of CMC1 OA requires further study to demonstrate specific joint stress activities, as well as to investigate preventive measures.
Acknowledgments
We would like to acknowledge support from the Swedish Research Council, Greta and Johan Kock Foundation, Österlund Foundation, The Swedish Rheumatism Association and Governmental Funding of Clinical Research within National Health Service.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors JMW and IA planned the study. AT abstracted the data and performed the initial data analysis. JMW, IA, ME and AT performed substantive data evaluation and results review. JMW wrote the manuscript, which was edited extensively by IA, ME and AT.
Funding This study was funded by the Swedish Research Council, Greta and Johan Kock Foundation, Österlund Foundation, The Swedish Rheumatism Association and Governmental Funding of Clinical Research within National Health Service.
Competing interests JW receives salary for work as a deputy editor for the Journal of Hand Surgery.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information. These healthcare and occupational registers are maintained by state authorities in Sweden and are protected data per the ethics agreement. Thus the data are not made available for public or research review.