Work-Related Disorders
of the Upper Limb

Keith T Palmer

Honorary Reader and MRC Clinical Scientist

Cyrus Cooper

Professor of Rheumatology
MRC Epidemiology Resource Centre, University of Southampton

Reports on the Rheumatic Diseases Series 5 : Topical Reviews

• Upper limb disorders are a heterogeneous group of conditions – some specific, some non-specific and some controversial
• Many of these disorders are suspected of being caused by physical work activities (e.g. repetitive jobs and those involving heavy physical work and awkward postures)
• Evidence on work association comes mainly from cross-sectional surveys, with rather few investigations of strong research design (cohort and intervention studies)
• Within limits these suggest that posture may exacerbate neck–shoulder problems and that a combination of repetition, force and posture may cause or aggravate elbow and wrist disorders

Introduction

Work-related musculoskeletal disorders (WMSDs) are an important source of occupational morbidity. Estimates of the size of the problem depend on case definition and the source of the statistics, but by most measures they represent one of the two major categories of work-related illness in Britain today (the other being mental illness). The Health and Safety Executive (HSE) estimates1 that in 2004/5 WMSDs affected a million people and cost society £5.7 billion in direct and indirect costs. Although back pain was the most common complaint (450,000 prevalent cases), disorders of the neck and upper limb were almost as common (375,000 cases).

Work-related upper limb disorders (WRULDs), the topic of this review, comprise a heterogeneous group of conditions – 165 in all according to a report by the US National Institute for Occupational Safety and Health (NIOSH). The majority of these are ill-defined with natural histories that are rather loosely characterised. In some instances there is dispute as to whether a recognised clinical entity exists at all; in others it is the epithet 'work-related' that generates strong opinions,2 reflected in several contradictory legal rulings.

Why has the apparently simple question: 'Which musculoskeletal problems are known to be work-related?' proved so hard to answer? To understand, we need first to state the nature of the problem and the clinical end points more precisely – for example, to differentiate if we can between exposures that cause and those that aggravate or accelerate a condition; to distinguish discomfort or disability resulting from accepted rheumatological conditions from ill-defined regional pain disorders; to clarify the background level of the disorder and its natural history in unexposed populations, and the personal and non-occupational risk factors (e.g. age, sex, anthropometrics, physical strength, medical history, mental health, threshold for complaint, leisure time pursuits) that could confound associations; and finally to explore associations with workplace factors that are themselves hard to measure (e.g. work postures, repetitiveness, the torque forces at joints, workplace psychosocial demands). We also need to recognise that although a large amount of research has been conducted, much of it is cross-sectional and subject perhaps to healthy worker selection bias (in which those worst affected may have selected themselves out of exposed employment); and in general, case definitions have not been applied in a uniform or standardised way.3,4

The challenges are such that views about work causation are polarised. However, there is rather more consensus in the research than in the clinical community. We focus here on a selection of soft tissue disorders of the upper limb. For each in turn we briefly review the evidence for an apparent association with work activities and present what we regard to be the current consensus among researchers in the field. We conclude by mentioning the current path on which research inquiries are headed.

Shoulder disorders

Fraying of the rotator cuff tendons, thickening of the bursae and proliferative changes of the synovium are common age-related phenomena. They are often asymptomatic in life, but may be discovered at post-mortem. However, shoulder pain is also common in working-aged people from the general community. In one Swedish study set in Malmö,5 some 14% of middle-aged subjects reported shoulder pain lasting ≥1 day in the past month, and 3% had taken sick leave because of it in the past year. In a survey of 40–74 year-old residents of Stockholm6 the point prevalence of shoulder pain was 20%, while Dimberg et al7 found that 13% of workers in a Swedish aeroengineering factory had current shoulder pain.

Injury in the occupational setting appears to arise through recurrent impingement of a relatively hypovascular area of the cuff against the acromioclavicular arch, an event that can be aggravated by superior migration of the humeral head in abduction and elevation, or by an acromial spur or degenerative acromioclavicular joint. Studies of blood flow in the supraspinatus muscle and muscle fatigue confirm the relevance of postural risk factors to the development of shoulder disorders. Occupational studies have therefore focused in the main on workers who elevate the shoulder repeatedly or for sustained periods, such as assembly line, crop and other production workers. The outcome studied has usually been rotator cuff tendinitis, defined operationally as localised shoulder pain and tenderness over the humeral head.8 Although other shoulder problems may arise, co-exist with or become confused with tendinitis, or contribute to work incapacity, their relation to work activity has rarely been examined in detail.

Table 1 summarises some older findings, based upon a comparison of occupational titles. Although the risk estimates vary, most reports have concluded that there is fair evidence for an association between overhead work and shoulder problems. In more recent investigations exposure contrasts have been defined in terms of work activity, rather than the crude measure of occupational title. A number of exposures (e.g. repetitive movements, vibration, high psychological demands) have thus been studied. Van der Windt et al,16 in their systematic review, comment that in studies of good methodological quality associations were 'generally not strong'; while another review by NIOSH17 talked about 'evidence' (meaning some degree of support) for associations with repeated or sustained shoulder postures with greater than 60º of flexion or abduction, but 'insufficient evidence' in relation to vibration and forceful movements. Only recently have prospective studies been undertaken. But in an Arthritis Research Campaign (arc) cohort study of new employees, lifting, carrying, pushing/pulling, monotonous employment and work with the hands above shoulder height were all found roughly to double risks of the relatively softer outcome of new onset shoulder pain during a 2-year follow-up.18

TABLE 1.  Some occupational studies of shoulder tendinitis. Cases Controls Study design RR Study Male shipyard workers and plate workers Office workers C 11–13 Herberts et al 19849 Female packers Female shop assistants C 2.6 Luopajarvi et al 197910 Male and female packers Knitters C 2.2 McCormack et al 199011 Sewers Knitters C 2.4 McCormack et al 199011 Board manufacturers Knitters C 2.1 McCormack et al 199011 Garment workers Health service workers C 2.2 Punnett et al 198512 Assembly workers General population C 3.4 Ohlsson et al 198913 Female data entry workers Other female office workers C 0.54 Kuorinka et al 197914 Male industrial workers with shoulder tendinitis Age and workshop matched non-cases CC 11* Bjelle et al 197915 * RR for work with hands at or above shoulder level C cross-sectional study; CC case-control study; RR relative risk

Elbow disorders

Epicondylitis is a pattern of pain at the origins of the extensors of the fingers and wrists on the lateral epicondyle, or at the origin of the flexors on the medial epicondyle of the humerus. Lateral epicondylitis is substantially the commoner of the two. The point prevalence of elbow pain has been reported at 11–13% in aeroengineering workers7 and textile workers,11 but clinically verified epicondylitis is less prevalent and among the textile workers only 2% had tender as well as painful elbows.7

Lateral epicondylitis is assumed to arise from overexertion of the finger and wrist extensors in repeated hand dorsiflexion or in alternating forearm pronation and supination. In those cases coming to clinical attention there is often a history of unaccustomed, forceful, repetitive use. However, the pathological lesion remains a point of debate, with at least 25 suggested causes.19 The most widely shared theory is that macroscopic or microscopic tears occur between the common extensor tendon and the periosteum of the lateral humeral epicondyle, although such tears have not been identified in surgical procedures consistently enough to resolve the argument.

Over 20 epidemiological studies have examined workplace physical risk factors and their relation to epicondylitis. A major review by NIOSH17 found 'insufficient' evidence for associations with repetition and postural factors when considered individually, and only weak evidence in relation to forceful work; but the evidence was considered 'strong' when these factors existed in combination. In fact, in 1997 when the review was written just one cohort study had been conducted.20 This showed an incidence rate 7–10 times higher in meat cutters and sausage makers (designated as jobs with high force and repetition) relative to other workers. A sports and biomechanical literature and some cross-sectional studies added plausibility to the finding. In 2006 the evidence base is scarcely more extensive, although one case-control and one cohort study point in the same direction. Haar et al21 compared work exposures in 267 cases and 388 referents recruited from Danish general practices and found associations with various activities performed for three-quarters or more of the time vs. never/almost never – namely, arms lifted in front of the body (odds ratio (OR) 4.0 for women); hands bent or twisted (OR 7.4 for women, 3.2 for men); same movements of the arm (OR 3.7 for women); and work requiring precise movements (OR 5.2 for men). In France, where workers have routine statutory medicals, Leclerc et al22 reported that jobs that involved repetitive turning and screwing doubled the risks of incident epicondylitis after allowing for other factors.

Tendinitis

In the US National Health Interview, 20 per 1000 adults reported that a doctor had told them they had 'tendinitis';23 while in the UK Primary Care Study for 1981, the incidence of tenosynovitis, tendinitis, synovitis and bursitis combined was 10.9 per 1000 persons per year,24 being more common in women than men at all ages.

There have been several cross-sectional studies of hand–wrist tendinitis in the workplace, but rather few prospective studies. Occupations described as at higher risk include board manufacturers, sewers and packers, and assembly line workers (Table 2). In one study, jobs that combined high force and high repetition carried a much greater risk of tendinitis than those lacking such features.25 And a cohort study, which compared the incidence of hand–wrist tendinitis in packers and sausage makers (jobs deemed to entail work strenuous to the musculotendinous junction of the upper limb) with office workers and supervisors, found high risk ratios of 24–36.20

click to enlarge

Carpal tunnel syndrome

The relation between carpal tunnel syndrome (CTS) and work activities has been investigated repeatedly. For example, a review by Hagberg et al28 identified 15 cross-sectional studies and 6 case-control studies that met high-quality criteria for case ascertainment. In the cross-sectional studies, the prevalence of CTS varied between different occupational groups over a 10-fold range, with particularly high prevalences and odds ratios in grinders, grocery store workers, frozen food factory workers and platers. High-risk groups were assumed to share in common the risk factors of repetitive and forceful gripping. Silverstein et al29 classified the occupation of workers from 7 different industries according to the degree of force and repetition required, and found, compared with low-force low-repetition jobs, that both high force and high frequency increased the risk moderately (about 2–3 fold), but that combined high force and frequency generated an odds ratio exceeding 15.

The strongest associations in case-control studies have been with the use of vibratory tools and with activities that frequently flex or extend the wrist.30 The link with vibratory exposure has also been described in occupational surveys, notably among foresters, and is accepted for compensation purposes by the UK State Industrial Injuries Scheme.

It can be demonstrated experimentally that extreme flexion and extreme extension of the wrist increase the pressure in the carpal tunnel sufficiently to impair blood perfusion of the median nerve, so that epidemiological and physiological investigations provide a coherent view of causation.

Chronic upper limb pain

No account of the upper limb would be complete without consideration of non-specific upper limb pain and so-called 'repetitive strain injury' (RSI) or 'cumulative trauma disorder' (CTD). Terms like this are unhelpful and contentious in the sense that they are ambiguous in their coverage. Moreover, they assume a cause as well as a diagnosis (whereas some of the conditions casually lumped under these headings may also have well-recognised non-occupational causes).

The problem is reflected in the confused position of recent legal judgements. In 1981 an enquiry over arm complaints in an Inland Revenue office accepted that disorders such as lateral epicondylitis and tenosynovitis could sometimes arise from the use of a visual display unit (VDU). In 1991, the High Court ruled that two British Telecom keyboard operators had 'RSI' induced by their work (a high number of key strokes per hour, bonus for higher totals, long hours in constrained positions on defective seating). The complaints here were tenosynovitis and epicondylitis. In 1994 a legal secretary received an award for tenosynovitis provoked by periods of intense typing. By contrast, in 1993 the complaints of a Reuters journalist were dismissed and it was ruled that 'RSI did not exist'. Such a ruling plainly does not disprove the existence of lateral epicondylitis and tenosynovitis! However, whether these disorders arise from repetitive work, and whether there is a condition of non-specific arm pain which presently defies clinical diagnosis and which arises from repetitive work, are points at issue.

In this review 'RSI' refers to chronic upper arm pain for which no diagnosis can be made and which has been ascribed to occupational over-use.

Given the lack of a clear definition and validated and accepted diagnostic criteria, it may be imagined that interpretation of the literature requires more than the usual degree of caution. However, in Australia the Bureau of Statistics separately codes compensation awards for injuries without explicit diagnosis ascribed to repetitive movement. Sequential data over the period 1980–7 showed dramatic changes with time. The number of successful claims among women in 1984–5 was five times greater than in 1980–1, and that among men 50% greater. Rates were higher in blue-collar workers, especially men involved in the manufacture of textiles, clothing and footwear, and food and beverages; and in women from parts of the manufacturing sector. Around this time Hocking reported on the Australian RSI epidemic as it affected one large employer, the nationalised telephone operator Telecom Australia.31 Almost 4000 cases were reported in the company, with nearly half in telephonists.

By the 1990s the compensation rate in Australia had declined to a more normal level, perhaps reflecting an earlier bubble fuelled by local concerns, awareness and adverse media publicity. Similar transient epidemics have been seen in other countries and other time periods: in Japan an epidemic of 'occupational cervico-brachial disorders' in 1958 and 1982, a crop of upper limb complaints in the Inland Revenue Department of the British Civil Service in 1981, and an epidemic of writer's cramp among clerks from the British Civil Service in 1830. The Japanese outbreak resulted in its Ministry of Labour introducing guidelines that restricted working time at the keyboard and imposed a maximum number of keystrokes per day, and a fall in the frequency of complaints followed.

The time variation in these outbreaks suggests that psychosocial variables may play an important part in presentation and recognition, if not also in development.

Large clinical series of cases referred with suspected RSI tend to show common features of gradual-onset diffuse pain, paraesthesiae, subjective swelling, and complaints of anxiety, irritation, mood change, fatigue and sleep disturbance.32 Clinical signs seem generally to be absent.

In this respect RSI has some similarities with fibromyalgia and this has led to the suggestion that RSI is a fibromyalgia variant based on regional rather than generalised pain. The relative contribution of personality, emotional state, health beliefs and psyche remain to be assessed in formal epidemiological studies, however, as most investigations have failed to distinguish non-specific from discrete rheumatological disorders.

Dupuytren's contracture

Dupuytren's contracture is a common condition that affects some 10% of older men in northern Europe. In general the condition appears to be idiopathic or familial, although some research evidence links the disorder with smoking, alcohol consumption, diabetes and epilepsy.

Case reports also exist of contracture arising (or perhaps being noticed) soon after an acute injury to the hand, such as a penetrating wound, crush injury or fracture. A review by Liss and Stock identified a limited but relevant literature on Dupuytren's contracture and cumulative occupational trauma.33 In one study a standardised morbidity ratio of nearly 2 was observed among bagging and packing plant workers after allowing for age and sex, and an odds ratio of 5.5 compared with other local workers who did not undertake these activities; while in another survey the odds were three times higher for heavy vs. light manual workers. The literature relating to the use of vibrating tools is slightly more extensive. Dupuytren's contracture appears to be about twice as common in men claiming vibration-induced white finger as in hospital controls admitted for elective surgery; while in two Italian studies, risks were increased 2–2.5-fold in quarry drillers and stone carvers compared with non-vibration exposed manual controls. Thus, there is a degree of evidence that Dupuytren's contracture may arise from occupational activities, this being strongest for exposure to hand-transmitted vibration.

Interventional studies?

Table 3 represents a summary by NIOSH of the evidence on work-relatedness of upper limb disorders at various anatomical sites;17 the strength of evidence varies from 'strong' for combinations of exposure at the elbow and wrist to 'insufficient' for a number of risk factors considered individually. This table fairly summarises the current state of knowledge, its conclusions being based on a mixed bag of observational epidemiology.

TABLE 3.  Summary of evidence of associations between mechanical factors and upper limb disorders. (Adapted from ref. 17.)
Anatomic site	Strong evidence of effect	Some evidence of effect	Insufficient evidence of effect
Neck and neck–shoulder
– Repetition		+
– Force		+
– Posture	+
– Vibration			+
Shoulder
–  Repetition		+
–  Force			+
–  Posture		+
–  Vibration			+
Elbow
–  Repetition			+
–  Force		+
–  Posture			+
–  Combination of these	+
Tendinitis of the hand/wrist
–  Repetition		+
–  Force		+
–  Posture		+
–  Combination of these	+
Carpal tunnel syndrome
–  Repetition		+
–  Force		+
–  Posture			+
–  Vibration		+
–  Combination of these	+

A better gold standard might be a high-quality blinded randomised control trial, to show that removal of a given exposure reduced the incidence of a given upper limb disorder in a target population. However, in practice some serious methodological challenges arise in mounting such trials – in blinding (where it may be possible to disguise the invention from the clinician assessing outcome, but not from the participant), in randomisation (which is most conveniently allocated at the factory, workshop or plant level), in standardised approach to diagnosis (a common omission), and in implementation (fear of litigation limits the number of employers willing to experiment – there being a natural bias towards solutions proposed by ergonomists on the basis of ergonomic theory and enforced by regulatory authorities). Not surprisingly, the evidence that has accrued so far is rather sparse.

However, three recent randomised trials in computer keyboard users attest that the challenge is now being taken up. Gerr et al34 monitored the incidence of neck–shoulder complaints in full-time workers randomised to receive an ergonomically adjusted workstation or the usual set-up; no important differences were found. By contrast, Aarås et al found a reduction in shoulder pain in parallel with a reduction in trapezius load in a small group of female data dialogue workers, after instituting a training programme and providing more ergonomic information.35 And most recently, Rempel et al tested the provision of ergonomic training, a wide forearm support surface and a trackball input device in call-centre workers using computers extensively.36 Armboard support was reported to halve the incidence of neck–shoulder disorders and reduce weekly pain scores in the intervention group. Whether these findings can be taken as evidence in support of work causation or not is, of course, a tricky matter of interpretation. But cohort studies of arm pain and work activity are being undertaken in increasing numbers across Europe and the US, including studies that include planned elements of intervention; so the evidence base on work causation is set to grow significantly in the near future.

What can the clinician do?

Each discrete disorder of the upper limb is managed medically along conventional lines. Options include: non-steroidal anti-inflammatory agents and analgesics; in many cases local corticosteroid injection, physiotherapy, local heat or pulsed ultrasound; splinting, and occasionally surgical decompression or release. The prognosis varies depending on diagnosis and management – from acute florid tenosynovitis (which settles quickly if thoroughly rested) through to epicondylitis (said to resolve in 8–12 months, but quite often lasting longer) and adhesive capsulitis (characteristically lasting 12–18 months).

An important work-related issue, but one which lacks a really strong evidence base, is the advice that should be given in relation to rest, rehabilitation and vocational alterations. Traditionally advised measures to support a planned return to work (as well as primary prevention) include: (i) job rotation, job enlargement, part-time working, or temporary job change – to provide respite from work involving repetitive monotonous use of the same muscles and tendons; (ii) 'adequate' rest breaks (often advised as an alternative to (i), although little information exists on the length that the break should be); (iii) task optimisation – that is, asking the employer and worker to think about the design and choice of equipment, tools, work layout and planning/throughput of tasks (often only a little thought is needed to reduce work effort and to avoid undesirable working postures. The HSE website provides useful guidelines – see www.hse.gov.uk/msd/hsemsd.htm#uld); (iv) a review of training – to ensure, likewise, that best working practices are being followed; (v) permanent redeployment – if cases prove recalcitrant or recurrent, and the link with unavoidable work activity seems very strong. Many parties also advise rest from all work activities as an immediate temporary measure while the job plans referred to above are laid. Such advice seems appropriate for discrete disorders of the upper limb, where pathology and acute inflammation are evident, but less certain in relation to non-specific complaints. For non-specific low-back pain, randomised controlled trial evidence has shown that patients tend to fare better if encouraged to remain active within the limits of pain, rather than strictly resting; and this begs the presently unanswered research question – whether the same might be true of those with diffuse non-specific arm pain.

Other Titles in this series

• June 2008 No 14
  Sports-Related Musculoskeletal Disorders
• October 2007 No 13
  Glucocorticoids in Rheumatic Diseases
• June 2007 No 12
  Immunisation of Immunosuppressed Patients with Rheumatic Diseases
• February 2007 No 11
  Rheumatological Manifestations of Malignant Haematological Conditions
• June 2006 No 9
  Qualitative Research: Its Relevance and Use in Musculoskeletal Medicine
• February 2006 No 8
  Rheumatic Diseases: Serological Aids to Early Diagnosis
• October 2005 No 7
  Malignancy and the Rheumatic Diseases
• June 2005 No 6
  Systemic Amyloidosis in the Rheumatic Diseases
• February 2005 No 5
  Heritable Collagen Disorders
• October 2004 No 4
  An Update on Gout
• June 2004 No 3
  Surgery for Arthritis: Total Hip and Knee Joint Replacement
• February 2004 No 2
  Idiopathic Inflammatory Myopathies
• September 2003 No 1
  Low Back Pain

download a pdf of this publication

6630.pdf
(618Kb)