Consensus Criteria for the Classification of Carpal Tunnel Syndrome in Epidemiologic Studies

Bradley Evanoff, David Rempel,
Peter Amadio, Marc de Krom,
Gary Franklin, Al Franzblau,
Ron Gray, Fred Gerr,
Mats Hagberg, Tom Hales,
Jeffrey N. Katz, Glenn Pransky

INTRODUCTION

Current controversies regarding occupational risk factors for carpal tunnel syndrome (CTS) stem, in part, from debate over what constitutes an acceptable case definition. We attempted to create consensus case definiton criteria for CTS which would be (1) used for epidemiologic studies in various settings (not for clinical case diagnosis), (2) simple and practical, (3) recognizable by and defensible to practicing clinicians, and (4) evidence based. Conference participants rank-ordered different case definitions by estimated test performance, with the realization that no single ìepidemiologic case definitionî can be created, as different case definitions are appropriate for different study settings.

METHODS

The participants in the process included all the authors of this abstract. Participants were selected from relevant medical specialties, including epidemiology, neurology, occupational medicine, orthopedics, and rheumatology; all had experience conducting epidemiologic, clinical, or outcomes studies of carpal tunnel syndrome. The group met twice in 1996. At the meetings a modified nominal group process was adopted to achieve the best possible consensus while acknowledging major differences of opinion.

RESULTS

Based on published and clinical experience, the group reached agreement on the following conceptual issues:

There is no perfect gold standard for CTS. Although EDS is considered the most accurate single test, false negatives and positives are well documented. The use of EDS alone as a case definition for CTS probably results in substantial misclassification.

The combination of EDS and symptom characteristics provide the most accurate data on CTS diagnosis. Physical examination findings add little diagnostic value if EDS and symptom characteristics are available.

In the absence of EDS, combinations of symptom characteristics and physical examination provide the greatest diagnostic information.

With this background, two sets of case definitions were established. The first set (Table 1) requires both the assessment of symptoms and an appropriate electrodiagnostic study (EDS). The second set (Table 2) is used when the EDS is not available, and require the assessment of symptoms and/or a physical examination. Both sets of case definitions require a symptom questionnaire instrument capable of classifying symptoms as ìclassic/probable,î ìpossible,î and ìunlikely,î as shown in Table 3. (Katz 1990, Franzblau 1993). Case definitions that include EDS are believed to have better specificity than case definitions that do not include EDS, although the use of EDS alone may be problematic in epidemiologic studies because of the difficulty in classifying individuals who are asymptomatic but have an abnormal EDS.

Only a qualitative ranking can be provided for Table I, since empirically determined sensitivities and specificities for these case definitions are unknown. For case definitions that do not include EDS (Table 2), empirically determined sensitivities and specificities are provided based on the gold standard case definition used by the investigator (usually EDS). As noted above, the use of EDS alone as the ìgold standardî for testing other case definitions is problematic, and may result in a lowering of the apparent test performances shown in Table 2.

The likelihood of carpal tunnel syndrome was judged greatest for ìclassicî or ìprobableî symptoms in combination with positive EDS results (Table 1). The likelihood of CTS was judged to be low when symptoms were absent in the median nerve distribution, regardless of the results of the EDS. Consensus could not be reached for subjects with classic or probable symptoms in combination with a negative EDS.

Among the case definitions that did not include EDS (Table 2), the greatest PPV was observed for the combination of classic/probable symptoms, a physical examination abnormality compatible with CTS, and nocturnal symptoms (PPV=0.44 when prevalence is 10%). The PPV was poorest for a positive physical examination result alone (PPV=0.16 when prevalence is 10%). Note that sensitivity is sacrificed at the highest levels of specificity (e.g., addition of nocturnal symptoms).

Table 1. Estimated likelihood of carpal tunnel syndrome for case definitions of carpal tunnel syndrome (CTS) that include electrodiagnostic studies (EDS). The criteria also require symptom classification (Table 3).

¹ No consensus achieved on whether this should be - or +.

Table 2. Sensitivity, specificity, and positive predictive value (PPV) for case definitions of CTS that do not include EDS. See Table 3 for symptom classification. PPV calculated assuming disease prevalence of 0.10.

1. Unpublished data from 822 workers, median to ulnar latency difference 0.5 ms (Franzblau 1997). pe=physical examination positive.
2. Study of 144 workers. Gold standard used EDS and symptoms (Gerr 1995).
3. Study of 408 workers (Franzblau 1994).
4. Study of 130 workers. Physical examination: (Tinelís test, Phalenís test, or two-point (4 mm) discrimination in dominant hand (Franzblau 1993)).

Table 3. Classification of symptom quality and location for use with hand diagrams or focused questions. Modified from Katz (1990) and Franzblau (1993, 1994).

Characterization of Symptoms: Different combinations of symptom location and qualities have been assessed in different studies and no single best symptom-based classification scheme has emerged. A recommended classification scheme, modified from Katz (1990) and Franzblau (1993, 1994) is presented in Table 3. It requires documentation of the symptom location and symptom character (numbness, tingling, burning, or pain). Little information is available on the predictive value of symptom duration or symptom frequency. Franzblau, et al. (1993) required symptoms on at least three separate episodes or at least one episode lasting greater than one week within the past year. De Krom, et al. (1990a) required symptom frequency of at least twice per week.
Electrodiagnostic testing: There is currently no consensus regarding which EDS technique is best for detecting CTS (Ross and Kimura 1995). We recommend the selection of one or more simple and acceptable measures in order to allow cross-study comparisons (e.g., Katz 1990, de Krom 1990a, Franzblau 1994, Gerr 1995). Studies should be performed according to the current and future guidelines prepared by the American Academy of Neurology, the American Association of Electrodiagnostic Medicine, and the American Academy of Physical Medicine and Rehabilitation (AAN 1993a and b, AAEMQAC 1993). Investigators must monitor and control for skin temperature and should also control for age, height, and other potential covariates when interpreting the results of EDS or constructing models with EDS as dependent variables (Stetson 1992, Letz 1994). If population norms are used, investigators should determine whether the range of ìnormalî used in their testing procedures were derived from a control population appropriate to the population studied.

DISCUSSION
Case definitions of CTS which combine a positive EDS and ìclassicî or ìprobableî symptoms will have the best predictive value. Case definitions using combinations of symptom characteristics and physical examination findings alone are useful in some study settings, but are likely to result in more misclassification of disease status than definitions using symptoms and EDS. The selection of an appropriate case definition will depend on the purpose and context of the epidemiologic study. For example, studies carried out for surveillance would likely use case definitions with high sensitivity, and therefore, lower specificity.

A number of studies use EDS as the gold standard for CTS because it is an ìobjectiveî test, and because it presumably measures the underlying pathophysiologic process of CTS. However, there are several problems with the use of EDS as the sole classification criterion for CTS in epidemiological studies. CTS is, by definition, a clinical syndrome with a characteristic symptom complex and, in severe cases, clear physical examination findings. Case definitions based on EDS alone ignore these additional data which are likely to improve the accuracy of classification. Indeed, the specificity of EDS alone is very low (12-22%) when compared to a gold standard of abnormal EDS plus symptoms (Nathan et al. 1994, Franzblau et al. 1997).

In conclusion, we have developed preliminary criteria for the classification of CTS in epidemiologic studies. The criteria must be regarded as preliminary, and may be changed by the results of future studies. We encourage other investigators to participate in the validation process and to use these criteria so that epidemiologic studies of CTS can share common case definitions

REFERENCES:

American Academy of Neurology, the American Association of Electrodiagnostic Medicine, and the American Academy of Physical Medicine and Rehabilitation. Practice parameter for electrodiagnostic studies in carpal tunnel syndrome. Neurology 1993 (a); 43:2404-5.

American Academy of Neurology, Quality Standards Subcommittee. Practice parameter for carpal tunnel syndrome. Neurology 1993 (b); 43:2406-9.

American Association of Electrodiagnostic Medicine Quality Assurance Committee. Literature review of the usefulness of nerve conduction studies and electromyography for the evaluation of patients with carpal tunnel syndrome. Muscle Nerve 1993; 16:1392-1414.

De Krom M, Kester A, Knipschild P, Spaans F. Risk factors for carpal tunnel syndrome. Am J Epi 1990 (a), 132:1102-1110.
De Krom MCTFM, Knipschild PG, Kester ADM, Spaans F. Efficacy of provocative tests for diagnosis of carpal tunnel syndrome. Lancet 1990 (b); 335:393-395.

Franzblau A, Werner R, Valle J, Johnston E. Workplace surveillance for carpal tunnel syndrome: a comparison of methods. J Occup Rehab 1993; 3:1-14.

Franzblau A, Werner R, Albers JW, et al. Workplace surveillance for carpal tunnel syndrome using hand diagrams. J Occup Rehab 1994; 4:185-98

Franzblau A et. al. Test-retest reliability of an upper extremity discomfort questionnaire in an industrial population. Scand J Work Enivron Health 1997 (submitted).

Gerr F, Letz R, Harris-Abbot D, Hopkins L. Sensitivity and Specificity of vibrometry for detection of carpal tunnel syndrome. JOEM 1995; 37:1108-1115.

Katz JN, Larson MG, Sabra A, Krarup C, Stirrat CR, Sethi R, Eaton HM, Fossel AH, Liang MH. The carpal tunnel syndrome: Diagnostic utility of the history and physical examination findings. Ann Int Med 1990; 112:321-327.

Katz JN, Stirrat CR, Larson MG, Fossel AN, Eaton HM, Liang MH. A self-administered hand symptom diagram for the diagnosis and epidemiologic study of carpal tunnel syndrome. J Rheumatol 1990; 3:1-14.

Katz JN, Punnett L, Simmons BP, Fossel AH, Mooney N, Keller RB. Workersí compensation recipients with carpal tunnel syndrome: The validity of self-reported health measures. Am J Pub Health 1996;86:52-56.

Letz R and Gerr F. Covariates of human peripheral nerve function: I. Nerve conduction velocity and amplitude. Neurotox and Terat 1994; 16:95-104.

Nathan PA, Keniston RC, Myers LD, Meadows KD. Obesity as a risk factor for slowing of sensory conduction of the median nerve in industry. J Occup Med 1992; 34:379-383.

Nathan PA, Takigawa K, Keniston RC, Meadows KD, Lockwood RS. Slowing of sensory conduction of the median nerve and carpal tunnel syndrome in Japanese and American industrial workers. Journal of Hand Surgery - British Volume. 19(1):30-4, 1994 Feb.

Ross MA, Kimura J. AAEM case report #2 The carpal tunnel syndrome, Muscle Nerve 1995, 18:567-573.

Stetson DS, Albers JW, Silverstein BA, Wolfe RA. Effects of age, sex, and anthropometric factors on nerve conduction measures. Muscle Nerve 1992; 15:1095-104.

Werner RA, Franzblau A, Albers JW, Buchele H, Armstrong TJ. Use of screening nerve conduction studies for predicting future carpal tunnel syndrome. Occup and Environ Med 1997; 54:96-100.