Evidence for Work-Related Musculoskeletal Disorders

To the Editor: Our thanks to the Journal for providing readers the opportunity to evaluate the scholarship presented by Dr Hadler¹ and the counterargument by Drs Silverstein, Silverstein, and Franklin² with regard to causation of carpal tunnel syndrome (CTS). Hadler indicts the status quo of the workers' compensation system. The Silversteins and Franklin target physical risk factors requiring ergonomic intervention. Hadler and the Silverstein group each present viewpoints based on extensive literature review.

Typical, sound epidemiology has been characterized as a systematic refutation of hypotheses, starting with the null hypothesis and continuing to hypotheses concerned with confounding factors, biases, and modifiers.³ Although much has been published, how solidly does that literature specifically refute the sequential hypotheses concerning lack of a causal relationship between CTS and workplace ergonomic stressors?

One Silverstein reference (Stock)⁴ concluded that the body of literature was about evenly divided over whether there was sufficient evidence to establish causal connection between occupational factors and various disorders; the author followed this with a de novo review identifying only a single study^5-7 representing strong evidence of a causal link between forceful and repetitive hand activity and CTS. A subsequent review by Vender et al⁸ referenced by Hadler and utilizing the same epidemiologic criteria as the former review by Silverstein et al was unable to validate any studies because of serious flaws detected in the assessments of control groups and the measurements of exposures and outcomes. A recent analysis of the literature by Gerr et al⁹ regarding the relationship between video display terminal use and upper-extremity musculoskeletal disorders concluded, "Epidemiologic studies have been limited by poor ascertainment of both exposure and health outcome. Many have failed to control for any potential confounding. An almost exclusive reliance on cross-sectional study designs has resulted in possible bias from selective survival, exposure-effect reversal, and poor estimates of exposures occurring prior to development of the disorder. Given the inconsistency of the literature and the growing controversy surrounding this issue, prospective study of this question using objective methods for assessment of exposure and health outcome is recommended."

In May 1996, the Council for the American Society for Surgery of the Hand¹⁰ issued a statement of concern over the trend for clinicians to assign inappropriate diagnoses (cumulative trauma disorders and repetitive strain injuries) and specific diagnoses lacking objective clinical indications, and to attribute occupational causation despite "a lack of epidemiologic evidence." This strong stand from such a highly respected organization of hand surgeons makes Silverstein's characterization of such evidence as "impressive", "credible," and "convincing"² seem extreme, particularly considering her inability to detect efficacy from ergonomic changes in a 3-year follow-up study.¹¹

Despite the uphill battle for those who believe in a material causal relationship between CTS and occupational factors, this theory remains only a possibility and would have to be established by studies with representative control groups and well-defined, objective diagnoses. Our review of the literature fails to support the interpretations and citations that the Silverstein group has used to make their points about causation, in the JOEM counterargument. For example, like Gerr et al,⁹ we have found that most studies about dose effects of keyboard utilize subjective symptoms rather than a medically valid diagnosis or case definition as the out-come measure. Even so, a statistically significant, consistent dose effect, whether based on symptoms or symptoms confirmed by median nerve abnormalities, is not observed. In a review of the literature on work-related musculoskeletal disorders, purportedly including CTS, in computer keyboard operation, Punnett¹² states, regarding relative risks of shoulder, arm, and hand disorders, "No evidence was found regarding the risk with fewer than 4 hours/day." Punnett cites one of our articles¹³ as suggesting that CTS is elevated in keyboard operators relative to administrative/clerical workers, but this was not our finding or stated conclusion. The article¹³ was about median nerve pathology, not the clinical condition of CTS, and keyboard use predicted less, not more, median-nerve slowing in that worker population.

Like others,^8,9 we have observed the case definition for CTS to vary greatly from one study to the next.¹⁴ Many studies that analyze keyboard use discuss subjective symptoms but not the median-nerve pathology.^8,9,12,14-17 Nerve-conduction studies are the gold standard for confirming the diagnosis of CTS.^8,14,18-20 Silverstein et al^2,5-7 fail to include nerve studies in their diagnosis of CTS and base their conclusions about etiology on frequency of symptoms and nonspecific clinical signs.^2,5-7,9,14 Published studies that do not consider median-nerve pathology^2,4-8,15,16 or that include incomplete, unreliable, or misinterpreted nerve study data^17,21-23 are of little value in determining the causes of CTS.^4,8,9,14

As suggested by Gerr et al,⁹ we have done prospective 5- and 11-year follow-up studies of the effects of personal and work factors on the development of CTS in industrial workers, using an objective case definition of CTS.²⁴ Our case definition of CTS is confirmed by median-nerve conduction abnormalities specifically involving the carpal tunnel.²⁰ What we have found is the opposite of what the cumulative trauma disorder hypothesis of Silverstein et al^2,5-7 would predict. More cumulative occupational keyboard use and more cumulative avocational keyboard time both predicted less, not more, median-nerve slowing and CTS. Longer cumulative duration of employment, after adjustment for age, also predicted less median-nerve slowing and CTS. We also observed a beneficial novice worker effect.²⁴ Our findings are supported by research which indicates that full-time, continuous employment is associated with a beneficial healthy worker effect.^25,26

Through careful scientific investigations and public expressions by authors such as Hadler, what we term personal factors are surfacing for consideration by officials involved in public policy decisions regarding causation of CTS. Unlike Hadler, the Silversteins and Franklin are in government positions that involve setting and administering public policy. This demands that they extend their perspective beyond as-yet unproven ergonomic solutions to consider factors such as heredity, body mass index (obesity), lack of exercise, wrist dimensions, and health habits when considering government-mandated workplace modifications and when setting policy for adjudication of claims for CTS.

Placing the burden of proof on those who would allege a particular causal association rather than those who would deny it may at first seem unwarranted and unfair, yet provides time-honored, fundamental, and indispensable safeguards to scientific credibility.

Peter A. Nathan, MD

Portland Hand Surgery and Rehabilitation Center; Portland, OR