- Email: [email protected]
The lumbrical provocation test in subjects with median inclusive paresthesia
935
The Lumbrical Provocation Test in Subjects With Median Inclusive Paresthesia Andrea I. Karl, BA, Maureen L. Carney, MBA, PA-C, Matthew P. Kaul, MD ABSTRACT. Karl AI, Carney ML, Kaul MP. The Lumbrical Provocation Test and discriminating among subjects with median inclusive paresthesia. Arch Phys Med Rehabil 2001;82:935-7. Objective: To investigate the value of the Lumbrical Provocation Test (LPT) in predicting carpal tunnel syndrome (CTS) among patients with symptoms suspicious for CTS. Design: Prospective unigroup technique with blinded comparison of a clinical diagnostic test with 2 commonly used methods of diagnosing CTS: electrodiagnosis and a hand diagram. Setting: Outpatient veterans referred by a heterogeneous group of specialists and generalists to a Veterans Affairs medical center electrodiagnostic laboratory. Patients: Ninety-six consecutive patients who were referred to the electrodiagnostic laboratory with median inclusive paresthesia were evaluated. Interventions: LPT: hold hand as fist for 1 minute (to evaluate changes in paresthesia); electrodiagnostic evaluations: median and ulnar mixed nerve, antidromic sensory, and motor latencies; and hand symptom diagram to describe pain. Main Outcome Measures: Evaluation of symptoms of paresthesia (with or without pain) inclusive of the median nerve distribution distal to the wrist. Prevalence sensitivity, specificity, and positive and negative predictive value of LPT, and electrodiagnosis, and hand diagram tests. Results: Compared with the findings obtained with electrodiagnosis, the sensitivity of the LPT was .37; specificity, .71; positive predictive value, .59; and negative predictive value, .50. Compared with the findings obtained with the use of the hand diagram as a clinical measure, the sensitivity of the LPT was .43, specificity, .71, positive predictive value, .59, and negative predictive value, .56. Conclusion: The LPT has minimal use in predicting CTS in patients with median inclusive paresthesia compared with 2 commonly used methods of diagnosing CTS. Key Words: Carpal tunnel syndrome; Electrodiagnosis; Rehabilitation. © 2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation ARPAL TUNNEL SYNDROME (CTS) is a common diagnosis in patients in the primary care setting, particuC larly in workers engaging in highly repetitive movements of the wrist.1 Many have attempted to develop a clinical diagnostic
From the Oregon Health Sciences University School of Medicine (Karl, Carney); and Veterans Administration Medical Center (Kaul), Portland, OR. Accepted in revised form August 29, 2000. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Matthew P. Kaul, MD, EMG Laboratory, Portland Veterans Affairs Medical Center, PO Box 1035, Portland, OR 97201-1035. 0003-9993/01/8207-6310$35.00/0 doi:10.1053/apmr.2001.23898
test capable of evaluating patients with median inclusive paresthesia. It would be particularly valuable to confirm that noninvasive clinical findings can identify a subgroup of patients at low or high risk for CTS who could be managed without electromyography studies. Such a test would offer many advantages in the clinical setting, including low cost, universal accessibility, increased convenience, and potential for avoiding patient discomfort. To date, clinical diagnostic tests for use in CTS have had wide variability in sensitivity and specificity depending on the spectrum of disease in the study group and the configuration of the control group.2 Standard clinical diagnostic tests, such as Phalen’s test, were poor predictors of CTS in several welldesigned studies.3-8 In addition, a test should help to differentiate between a particular disease, such as CTS, and other disorders with similar and commonly confused presentations. Toward this end, several provocation tests for CTS have recently been developed.9 Of these tests, the Lumbrical Provocation Test (LPT) developed by Yii and Elliot is the least studied, with only 2 publications evaluating its use.8,9 The initial study of the LPT used as a population base a group of patients referred for carpal tunnel release surgery.9 It was observed that the lumbrical muscles often moved into the carpal tunnel with finger flexion. The hypothesis was that compression during this proximal migration of the lumbricals might exacerbate CTS. A total of 29 consecutive patients scheduled for CTS surgery were studied by using the LPT, Phalen’s test, Tinel’s test, and the Carpal Compression Test. The results reported were that the LPT performed best among the 4 earlier-mentioned clinical tests, with a sensitivity of 97%. In the asymptomatic control group, specificity was 93%. Based on these findings, the report concluded that the proximal migration of the lumbricals might explain CTS paresthesia induced by sustained grip.9 The other early investigators of the LPT reported a similar high specificity (92%) for the LPT. Their sensitivity, however, was 62%.8 We undertook our study as an independent investigation into the clinical value of the LPT in predicting CTS compared with electrodiagnosis and hand diagram– diagnosed CTS, 2 commonly accepted methods of diagnosing CTS. METHODS Ninety-six consecutive patients referred to the Portland Veterans Affairs Medical Center Electrodiagnostic Laboratory with median inclusive paresthesias were assessed (table 1). Patients with symptoms of median paresthesia distal to the wrist were entered into the study. Patients were excluded if they had had prior carpal tunnel release, ipsilateral proximal median neuropathy, arteriovenous shunt–related median paresthesia, or paresthesia only induced by neck range of motion. The LPT was performed immediately before electrodiagnostic testing as part of the routine physical examination. In bilateral cases, the patient’s more symptomatic hand was assessed. If symptoms were symmetric, the dominant hand was assessed. When other provocative tests were performed during routine physical examination, complete return to baseline symptoms was ensured before performing the LPT. Arch Phys Med Rehabil Vol 82, July 2001
936
THE LUMBRICAL PROVOCATION TEST, Karl
Table 1: Clinical and Demographic Features of the LPT (n ⴝ 135) Subjects Feature
n
%
Women Men White Duration of symptoms (mo) ⬍3 3–12 ⬎12 Bilateral symptoms Unilateral symptoms Comorbidity Diabetes Thyroid disease Rheumatoid arthritis
6 90 93
6 94 97
6 34 56 59 37
6 35 58 61 39
17 6 0
18 6 0
The LPT was performed according to published techniques.9 Each patient was asked to make a fist for 1 minute. The test was considered positive when median inclusive paresthesia was induced or exacerbated. All clinical testing and electrodiagnostic testing was performed by 1 examiner (MPK). An electrodiagnostic grading score of mild, moderate, or severe CTS was given to each patient based on electrodiagnostic criteria as recommended by Stevens.10 This was used to evaluate for an adequate spectrum of CTS severity (table 2). Electrodiagnostic Criteria Of the following 3 standard electrodiagnostic evaluations, CTS was diagnosed based on the results of the first method (described later), unless it was unobtainable. In that case, the second or third method was used. Method 1. Median and ulnar mixed nerve latencies were obtained by placing a block electrode (reference more proximal) over the respective nerve at the wrist 3cm proximal to the distal wrist crease. There were 8cm between the stimulation sites and the recording sites. Supramaximal stimulation was delivered in the palm, between the second and third metacarpal bones for the median nerve and between the fourth and fifth metacarpals for the ulnar nerve (palm diff). Method 2. Median and ulnar antidromic sensory latencies were obtained by placing ring electrodes over digits 2 and 5, active electrode placed proximally, with a 4-cm interelectrode spacing at digits 2 and 5 whenever possible. Supramaximal stimulation was delivered at the wrist over the respective nerve 14cm proximal to the active ring electrode (D2–D5 diff).
Method 3. Median and ulnar motor latencies were obtained by placing active recording surface electrodes over the midpoint of the abductor pollicis brevis and abductor digiti minimi. Reference electrodes were placed 4cm distally. Supramaximal stimulation was delivered at the wrist over the respective nerves 8cm proximal to the active surface electrode (motor diff). CTS was diagnosed when (1) palm diff peak latency difference was greater than 0.3ms,10 or (2) D2–D5 diff peak latency difference was greater than 0.5ms,11 or (3) motor diff onset latency difference was greater than 1.7ms.10 Midpalm was our first choice for the electrodiagnostic method of diagnosing CTS. If midpalm was unobtainable, then D2–D5 diff and/or motor diff were used to diagnose CTS. All electrodiagnostic testing was performed with a Nicolet Viking IVD.a For sensory responses, we used a bandpass of 20Hz to 2kHz and sensitivity between 5 and 50V/div appropriate for the size of the response. Palm temperatures were assessed immediately before electrodiagnostic testing. The extremity was warmed to greater than 31.5°C whenever a temperature less than this was recorded. Clinical Standard In consideration of the controversy surrounding the electrodiagnostic diagnosis of CTS, we also assessed patients with clinical criteria. Brief hand symptom diagram instructions were given by 1 examiner (MPK). The hand diagrams were scored blindly by the same examiner according to criteria described by Katz et al.3 We used their method of grouping results into possible or unlikely versus classic or probable groups. For our purposes, there were no unlikely category patients because our inclusion criteria required median inclusive symptoms. We considered that subjects with a classic or probable diagram rating qualified for the diagnosis of CTS. A classic designation is given when a diagram shows tingling, numbness, or decreased sensation with or without pain in at least 2 of digits 1, 2, or 3; symptoms in the palm and dorsum of the hand excluded; and fifth finger symptoms, wrist pain, or radiation proximal to the wrist allowed. A probable rating is defined as classic symptoms plus symptoms in the median aspect of the palm (table 2). Analysis The data were analyzed with 2 ⫻ 2 tables for both the electrodiagnostic and clinical (hand diagram) criteria to determine prevalence, sensitivity, specificity, positive predictive value, and negative predictive value (table 3).
Table 3: 2 ⴛ 2 Tables of the LPT vs Electrodiagnostic and vs Clinical Electrodiagnosis and Hand Diagram Standard Table 2: Internal Spectrum Categorized by Electrodiagnosis and Hand Symptom Diagram
Electrodiagnosis
Results of LPT Symptom Spectrum
Level (n ⫽ 80) Mild Moderate Severe Category (n ⫽ 96) Classic Probable Possible
n
%
23 38 19
29 47 24
19 26 51
20 27 53
Arch Phys Med Rehabil Vol 82, July 2001
LPT (⫹) LPT (⫺) Total
CTS(⫹)
CTS(⫺)
Total
19 32 51
13 32 45
32 64 96
Clinical Hand Diagram
LPT (⫹) LPT (⫺) Total
CTS(⫹)
CTS(⫺)
Total
20 27 47
14 35 49
34 62 96
THE LUMBRICAL PROVOCATION TEST, Karl
RESULTS A total of 96 consecutive patients with median inclusive paresthesia were assessed. The mean age ⫾ standard deviation was 52.6 ⫾ 13.8 years. Other baseline demographic and clinical features of the patients in our study are summarized in table 1. Electrodiagnostic Results In 18 cases, palm diff could not be used for CTS diagnosis. In these subjects, 8 cases were diagnosed with CTS by D2–D5 diff and 10 cases were diagnosed with motor diff. In all cases in which CTS was excluded, the midpalm assessment was used. Fifty-one patients (53%) were diagnosed by electrodiagnostic criteria as having CTS in at least 1 upper extremity. The sensitivity (proportion of patients with CTS who had a positive LPT) was .37. The specificity (proportion of those without CTS who had a negative LPT) was .71. The positive predictive value (proportion of those with a positive LPT who had CTS) was .59. The negative predictive value (proportion of those with a negative LPT who did not have CTS) was .50. These values were calculated from the 2 ⫻ 2 tables listed in table 3. Clinical Hand Diagram Results The sensitivity of the LPT compared with the clinical hand diagram was .43, the specificity was .71, the positive predictive value was .59, and the negative predictive value was .56. These values were calculated from the 2 ⫻ 2 tables listed in table 3. DISCUSSION The prevalence of CTS in our sample population was .53. For a diagnostic test to be valuable, the predictive value of a test must be substantially greater than the sample prevalence. The positive predictive value of the LPT was .59, compared with both the electrodiagnostic and hand diagram results. Thus, the positive predictive value of the LPT is essentially no different than the prevalence of CTS in our sample, and, therefore, the test is of no diagnostic value in distinguishing outcomes among subjects with a positive test. For a negative test to be of value, the negative predictive value should be substantially greater than the prevalence of subjects without the CTS. Our negative predictive value was .50 with electromyography gold standard, and .56 with the hand symptom diagram. This added no value in predicting outcomes among subjects with negative test results. Our findings differ substantially from those reported in other evaluations of the LPT.8,9 We think that this is in large measure caused by differences in methodology. In contrast to previous research on the LPT, our study compared outcomes among patients with a broad range of symptom severity and a broad spectrum of clinical presentation. Our approach limited the sensitivity bias that occurs when a diagnostic test is performed on patients with severe or classic disease, and then the test is assumed to be equally effective among patients with less classic disease. Sensitivity may be substantially lower in the more equivocal population. To control for sensitivity bias, the disease-positive group must have a broad spectrum of clinical presentations and a broad range of severity. We documented both of these for our disease-positive group. Previous investigators studying the LPT acknowledge their study’s limitation in this regard.9 We think this accounts for the substantially lower sensitivity of the LPT found in our study (.37 with electromyography, .43 with the hand diagram).
937
Just as important as the spectrum of disease in the diseasepositive group is the selection of subjects for the diseasenegative group. Specificity bias has been shown in many different studies in which asymptomatic comparison subjects were used to determine the specificity of diagnostic or clinical tests. This false elevation of specificity can be substantial. For example, the specificity of the Phalen’s test in diagnosing CTS was found to be .97 when the study was performed with asymptomatic controls.2 However, when the same study was performed with symptomatic non-CTS controls, the specificity dropped to .61. We controlled for specificity bias by including a symptomatic non-CTS control group. Because it is among patients suspected of having CTS that the LPT would be used, it is critical to determine whether the test performs differently among different control populations. One prior investigator of the LPT considered this. However, the symptomatic non-CTS control sample was too small (12 subjects) to draw valuable conclusions.8 We think our results are generalizable to the predominately male veteran population based on a number of lines of reasoning. Our hospital does not have a built-in disincentive to ordering electrodiagnostic testing. Surgeons in our hospital generally require preoperative electrodiagnostic confirmation of CTS in all surgical candidates, even in patients with classic symptoms. CONCLUSION The LPT does not have use in predicting CTS in veterans with symptoms suspicious for CTS compared with electrodiagnosis and hand diagrams, 2 commonly accepted methods of diagnosing CTS. References 1. Dawson DM. Current concepts: entrapment neuropathies of the upper extremities. N Engl J Med 1993;27:2013-8. 2. Gerr F, Letz R. The sensitivity and specificity of tests for carpal tunnel syndrome vary with the comparison subjects. J Hand Surg [Br] 1998;23:151-5. 3. Katz JN, Larson MG, Sabra A, Karup C, Stirrat CR, Sethi R, et al. The carpal tunnel syndrome: diagnostic utility of the history and physical examination findings. Ann Intern Med 1990;112:321-7. 4. de Krom MC, Knipschild PG, Kester AD, Spaans F. Efficacy of provocative tests for diagnosis of carpal tunnel syndrome. Lancet 1990;335:393-5. 5. Buch-Jaeger N, Foucher G. Correlation of clinical signs with nerve conduction test in the diagnosis of carpal tunnel syndrome. J Hand Surg [Br] 1994;19:720-4. 6. Golding DN, Rose DM, Selvarajah K. Clinical tests for carpal tunnel syndrome: an evaluation. Br J Rheumatol 1986;25:388-90. 7. Heller L, Ring H, Costeff H, Solzi P. Evaluation of Tinel’s and Phalen’s signs in diagnosis of the carpal tunnel syndrome. Eur Neurol 1986;25:40-2. 8. De Smet L, Steenwerckx A, Van den Bogaert G, Cnudde P, Fabry G. Value of clinical provocative tests in carpal tunnel syndrome. Acta Orthop Belg 1995;61:177-82. 9. Yii NW, Elliot D. A study of the dynamic relationship of the lumbrical muscles and the carpal tunnel. J Hand Surg [Br] 1994; 19:439-43. 10. Stevens JC. AAEM Minimonograph #26. The electrodiagnosis of carpal tunnel syndrome. Muscle Nerve 1997;20:1477-86. 11. Stevens JC. AAEM Minimonograph #26. The electrodiagnosis of carpal tunnel syndrome. Muscle Nerve 1987;10:99-113. Supplier a. Nicolet Biomedical Inc, 5225 Verona Rd, Bldg 2, Madison, WI 53711-4495.
Arch Phys Med Rehabil Vol 82, July 2001
The Lumbrical Provocation Test in Subjects With Median Inclusive Paresthesia Andrea I. Karl, BA, Maureen L. Carney, MBA, PA-C, Matthew P. Kaul, MD ABSTRACT. Karl AI, Carney ML, Kaul MP. The Lumbrical Provocation Test and discriminating among subjects with median inclusive paresthesia. Arch Phys Med Rehabil 2001;82:935-7. Objective: To investigate the value of the Lumbrical Provocation Test (LPT) in predicting carpal tunnel syndrome (CTS) among patients with symptoms suspicious for CTS. Design: Prospective unigroup technique with blinded comparison of a clinical diagnostic test with 2 commonly used methods of diagnosing CTS: electrodiagnosis and a hand diagram. Setting: Outpatient veterans referred by a heterogeneous group of specialists and generalists to a Veterans Affairs medical center electrodiagnostic laboratory. Patients: Ninety-six consecutive patients who were referred to the electrodiagnostic laboratory with median inclusive paresthesia were evaluated. Interventions: LPT: hold hand as fist for 1 minute (to evaluate changes in paresthesia); electrodiagnostic evaluations: median and ulnar mixed nerve, antidromic sensory, and motor latencies; and hand symptom diagram to describe pain. Main Outcome Measures: Evaluation of symptoms of paresthesia (with or without pain) inclusive of the median nerve distribution distal to the wrist. Prevalence sensitivity, specificity, and positive and negative predictive value of LPT, and electrodiagnosis, and hand diagram tests. Results: Compared with the findings obtained with electrodiagnosis, the sensitivity of the LPT was .37; specificity, .71; positive predictive value, .59; and negative predictive value, .50. Compared with the findings obtained with the use of the hand diagram as a clinical measure, the sensitivity of the LPT was .43, specificity, .71, positive predictive value, .59, and negative predictive value, .56. Conclusion: The LPT has minimal use in predicting CTS in patients with median inclusive paresthesia compared with 2 commonly used methods of diagnosing CTS. Key Words: Carpal tunnel syndrome; Electrodiagnosis; Rehabilitation. © 2001 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation ARPAL TUNNEL SYNDROME (CTS) is a common diagnosis in patients in the primary care setting, particuC larly in workers engaging in highly repetitive movements of the wrist.1 Many have attempted to develop a clinical diagnostic
From the Oregon Health Sciences University School of Medicine (Karl, Carney); and Veterans Administration Medical Center (Kaul), Portland, OR. Accepted in revised form August 29, 2000. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Matthew P. Kaul, MD, EMG Laboratory, Portland Veterans Affairs Medical Center, PO Box 1035, Portland, OR 97201-1035. 0003-9993/01/8207-6310$35.00/0 doi:10.1053/apmr.2001.23898
test capable of evaluating patients with median inclusive paresthesia. It would be particularly valuable to confirm that noninvasive clinical findings can identify a subgroup of patients at low or high risk for CTS who could be managed without electromyography studies. Such a test would offer many advantages in the clinical setting, including low cost, universal accessibility, increased convenience, and potential for avoiding patient discomfort. To date, clinical diagnostic tests for use in CTS have had wide variability in sensitivity and specificity depending on the spectrum of disease in the study group and the configuration of the control group.2 Standard clinical diagnostic tests, such as Phalen’s test, were poor predictors of CTS in several welldesigned studies.3-8 In addition, a test should help to differentiate between a particular disease, such as CTS, and other disorders with similar and commonly confused presentations. Toward this end, several provocation tests for CTS have recently been developed.9 Of these tests, the Lumbrical Provocation Test (LPT) developed by Yii and Elliot is the least studied, with only 2 publications evaluating its use.8,9 The initial study of the LPT used as a population base a group of patients referred for carpal tunnel release surgery.9 It was observed that the lumbrical muscles often moved into the carpal tunnel with finger flexion. The hypothesis was that compression during this proximal migration of the lumbricals might exacerbate CTS. A total of 29 consecutive patients scheduled for CTS surgery were studied by using the LPT, Phalen’s test, Tinel’s test, and the Carpal Compression Test. The results reported were that the LPT performed best among the 4 earlier-mentioned clinical tests, with a sensitivity of 97%. In the asymptomatic control group, specificity was 93%. Based on these findings, the report concluded that the proximal migration of the lumbricals might explain CTS paresthesia induced by sustained grip.9 The other early investigators of the LPT reported a similar high specificity (92%) for the LPT. Their sensitivity, however, was 62%.8 We undertook our study as an independent investigation into the clinical value of the LPT in predicting CTS compared with electrodiagnosis and hand diagram– diagnosed CTS, 2 commonly accepted methods of diagnosing CTS. METHODS Ninety-six consecutive patients referred to the Portland Veterans Affairs Medical Center Electrodiagnostic Laboratory with median inclusive paresthesias were assessed (table 1). Patients with symptoms of median paresthesia distal to the wrist were entered into the study. Patients were excluded if they had had prior carpal tunnel release, ipsilateral proximal median neuropathy, arteriovenous shunt–related median paresthesia, or paresthesia only induced by neck range of motion. The LPT was performed immediately before electrodiagnostic testing as part of the routine physical examination. In bilateral cases, the patient’s more symptomatic hand was assessed. If symptoms were symmetric, the dominant hand was assessed. When other provocative tests were performed during routine physical examination, complete return to baseline symptoms was ensured before performing the LPT. Arch Phys Med Rehabil Vol 82, July 2001
936
THE LUMBRICAL PROVOCATION TEST, Karl
Table 1: Clinical and Demographic Features of the LPT (n ⴝ 135) Subjects Feature
n
%
Women Men White Duration of symptoms (mo) ⬍3 3–12 ⬎12 Bilateral symptoms Unilateral symptoms Comorbidity Diabetes Thyroid disease Rheumatoid arthritis
6 90 93
6 94 97
6 34 56 59 37
6 35 58 61 39
17 6 0
18 6 0
The LPT was performed according to published techniques.9 Each patient was asked to make a fist for 1 minute. The test was considered positive when median inclusive paresthesia was induced or exacerbated. All clinical testing and electrodiagnostic testing was performed by 1 examiner (MPK). An electrodiagnostic grading score of mild, moderate, or severe CTS was given to each patient based on electrodiagnostic criteria as recommended by Stevens.10 This was used to evaluate for an adequate spectrum of CTS severity (table 2). Electrodiagnostic Criteria Of the following 3 standard electrodiagnostic evaluations, CTS was diagnosed based on the results of the first method (described later), unless it was unobtainable. In that case, the second or third method was used. Method 1. Median and ulnar mixed nerve latencies were obtained by placing a block electrode (reference more proximal) over the respective nerve at the wrist 3cm proximal to the distal wrist crease. There were 8cm between the stimulation sites and the recording sites. Supramaximal stimulation was delivered in the palm, between the second and third metacarpal bones for the median nerve and between the fourth and fifth metacarpals for the ulnar nerve (palm diff). Method 2. Median and ulnar antidromic sensory latencies were obtained by placing ring electrodes over digits 2 and 5, active electrode placed proximally, with a 4-cm interelectrode spacing at digits 2 and 5 whenever possible. Supramaximal stimulation was delivered at the wrist over the respective nerve 14cm proximal to the active ring electrode (D2–D5 diff).
Method 3. Median and ulnar motor latencies were obtained by placing active recording surface electrodes over the midpoint of the abductor pollicis brevis and abductor digiti minimi. Reference electrodes were placed 4cm distally. Supramaximal stimulation was delivered at the wrist over the respective nerves 8cm proximal to the active surface electrode (motor diff). CTS was diagnosed when (1) palm diff peak latency difference was greater than 0.3ms,10 or (2) D2–D5 diff peak latency difference was greater than 0.5ms,11 or (3) motor diff onset latency difference was greater than 1.7ms.10 Midpalm was our first choice for the electrodiagnostic method of diagnosing CTS. If midpalm was unobtainable, then D2–D5 diff and/or motor diff were used to diagnose CTS. All electrodiagnostic testing was performed with a Nicolet Viking IVD.a For sensory responses, we used a bandpass of 20Hz to 2kHz and sensitivity between 5 and 50V/div appropriate for the size of the response. Palm temperatures were assessed immediately before electrodiagnostic testing. The extremity was warmed to greater than 31.5°C whenever a temperature less than this was recorded. Clinical Standard In consideration of the controversy surrounding the electrodiagnostic diagnosis of CTS, we also assessed patients with clinical criteria. Brief hand symptom diagram instructions were given by 1 examiner (MPK). The hand diagrams were scored blindly by the same examiner according to criteria described by Katz et al.3 We used their method of grouping results into possible or unlikely versus classic or probable groups. For our purposes, there were no unlikely category patients because our inclusion criteria required median inclusive symptoms. We considered that subjects with a classic or probable diagram rating qualified for the diagnosis of CTS. A classic designation is given when a diagram shows tingling, numbness, or decreased sensation with or without pain in at least 2 of digits 1, 2, or 3; symptoms in the palm and dorsum of the hand excluded; and fifth finger symptoms, wrist pain, or radiation proximal to the wrist allowed. A probable rating is defined as classic symptoms plus symptoms in the median aspect of the palm (table 2). Analysis The data were analyzed with 2 ⫻ 2 tables for both the electrodiagnostic and clinical (hand diagram) criteria to determine prevalence, sensitivity, specificity, positive predictive value, and negative predictive value (table 3).
Table 3: 2 ⴛ 2 Tables of the LPT vs Electrodiagnostic and vs Clinical Electrodiagnosis and Hand Diagram Standard Table 2: Internal Spectrum Categorized by Electrodiagnosis and Hand Symptom Diagram
Electrodiagnosis
Results of LPT Symptom Spectrum
Level (n ⫽ 80) Mild Moderate Severe Category (n ⫽ 96) Classic Probable Possible
n
%
23 38 19
29 47 24
19 26 51
20 27 53
Arch Phys Med Rehabil Vol 82, July 2001
LPT (⫹) LPT (⫺) Total
CTS(⫹)
CTS(⫺)
Total
19 32 51
13 32 45
32 64 96
Clinical Hand Diagram
LPT (⫹) LPT (⫺) Total
CTS(⫹)
CTS(⫺)
Total
20 27 47
14 35 49
34 62 96
THE LUMBRICAL PROVOCATION TEST, Karl
RESULTS A total of 96 consecutive patients with median inclusive paresthesia were assessed. The mean age ⫾ standard deviation was 52.6 ⫾ 13.8 years. Other baseline demographic and clinical features of the patients in our study are summarized in table 1. Electrodiagnostic Results In 18 cases, palm diff could not be used for CTS diagnosis. In these subjects, 8 cases were diagnosed with CTS by D2–D5 diff and 10 cases were diagnosed with motor diff. In all cases in which CTS was excluded, the midpalm assessment was used. Fifty-one patients (53%) were diagnosed by electrodiagnostic criteria as having CTS in at least 1 upper extremity. The sensitivity (proportion of patients with CTS who had a positive LPT) was .37. The specificity (proportion of those without CTS who had a negative LPT) was .71. The positive predictive value (proportion of those with a positive LPT who had CTS) was .59. The negative predictive value (proportion of those with a negative LPT who did not have CTS) was .50. These values were calculated from the 2 ⫻ 2 tables listed in table 3. Clinical Hand Diagram Results The sensitivity of the LPT compared with the clinical hand diagram was .43, the specificity was .71, the positive predictive value was .59, and the negative predictive value was .56. These values were calculated from the 2 ⫻ 2 tables listed in table 3. DISCUSSION The prevalence of CTS in our sample population was .53. For a diagnostic test to be valuable, the predictive value of a test must be substantially greater than the sample prevalence. The positive predictive value of the LPT was .59, compared with both the electrodiagnostic and hand diagram results. Thus, the positive predictive value of the LPT is essentially no different than the prevalence of CTS in our sample, and, therefore, the test is of no diagnostic value in distinguishing outcomes among subjects with a positive test. For a negative test to be of value, the negative predictive value should be substantially greater than the prevalence of subjects without the CTS. Our negative predictive value was .50 with electromyography gold standard, and .56 with the hand symptom diagram. This added no value in predicting outcomes among subjects with negative test results. Our findings differ substantially from those reported in other evaluations of the LPT.8,9 We think that this is in large measure caused by differences in methodology. In contrast to previous research on the LPT, our study compared outcomes among patients with a broad range of symptom severity and a broad spectrum of clinical presentation. Our approach limited the sensitivity bias that occurs when a diagnostic test is performed on patients with severe or classic disease, and then the test is assumed to be equally effective among patients with less classic disease. Sensitivity may be substantially lower in the more equivocal population. To control for sensitivity bias, the disease-positive group must have a broad spectrum of clinical presentations and a broad range of severity. We documented both of these for our disease-positive group. Previous investigators studying the LPT acknowledge their study’s limitation in this regard.9 We think this accounts for the substantially lower sensitivity of the LPT found in our study (.37 with electromyography, .43 with the hand diagram).
937
Just as important as the spectrum of disease in the diseasepositive group is the selection of subjects for the diseasenegative group. Specificity bias has been shown in many different studies in which asymptomatic comparison subjects were used to determine the specificity of diagnostic or clinical tests. This false elevation of specificity can be substantial. For example, the specificity of the Phalen’s test in diagnosing CTS was found to be .97 when the study was performed with asymptomatic controls.2 However, when the same study was performed with symptomatic non-CTS controls, the specificity dropped to .61. We controlled for specificity bias by including a symptomatic non-CTS control group. Because it is among patients suspected of having CTS that the LPT would be used, it is critical to determine whether the test performs differently among different control populations. One prior investigator of the LPT considered this. However, the symptomatic non-CTS control sample was too small (12 subjects) to draw valuable conclusions.8 We think our results are generalizable to the predominately male veteran population based on a number of lines of reasoning. Our hospital does not have a built-in disincentive to ordering electrodiagnostic testing. Surgeons in our hospital generally require preoperative electrodiagnostic confirmation of CTS in all surgical candidates, even in patients with classic symptoms. CONCLUSION The LPT does not have use in predicting CTS in veterans with symptoms suspicious for CTS compared with electrodiagnosis and hand diagrams, 2 commonly accepted methods of diagnosing CTS. References 1. Dawson DM. Current concepts: entrapment neuropathies of the upper extremities. N Engl J Med 1993;27:2013-8. 2. Gerr F, Letz R. The sensitivity and specificity of tests for carpal tunnel syndrome vary with the comparison subjects. J Hand Surg [Br] 1998;23:151-5. 3. Katz JN, Larson MG, Sabra A, Karup C, Stirrat CR, Sethi R, et al. The carpal tunnel syndrome: diagnostic utility of the history and physical examination findings. Ann Intern Med 1990;112:321-7. 4. de Krom MC, Knipschild PG, Kester AD, Spaans F. Efficacy of provocative tests for diagnosis of carpal tunnel syndrome. Lancet 1990;335:393-5. 5. Buch-Jaeger N, Foucher G. Correlation of clinical signs with nerve conduction test in the diagnosis of carpal tunnel syndrome. J Hand Surg [Br] 1994;19:720-4. 6. Golding DN, Rose DM, Selvarajah K. Clinical tests for carpal tunnel syndrome: an evaluation. Br J Rheumatol 1986;25:388-90. 7. Heller L, Ring H, Costeff H, Solzi P. Evaluation of Tinel’s and Phalen’s signs in diagnosis of the carpal tunnel syndrome. Eur Neurol 1986;25:40-2. 8. De Smet L, Steenwerckx A, Van den Bogaert G, Cnudde P, Fabry G. Value of clinical provocative tests in carpal tunnel syndrome. Acta Orthop Belg 1995;61:177-82. 9. Yii NW, Elliot D. A study of the dynamic relationship of the lumbrical muscles and the carpal tunnel. J Hand Surg [Br] 1994; 19:439-43. 10. Stevens JC. AAEM Minimonograph #26. The electrodiagnosis of carpal tunnel syndrome. Muscle Nerve 1997;20:1477-86. 11. Stevens JC. AAEM Minimonograph #26. The electrodiagnosis of carpal tunnel syndrome. Muscle Nerve 1987;10:99-113. Supplier a. Nicolet Biomedical Inc, 5225 Verona Rd, Bldg 2, Madison, WI 53711-4495.
Arch Phys Med Rehabil Vol 82, July 2001