12 SA JOURNAL OF PHYSIOTHERAPY 2006 VOL 62 NO 2 DO CLINICAL FEATURES AND MRI SUGGEST THE SAME NERVE ROOT IN ACUTE CERVICAL RADICULOPATHY? R E S E A R C H A R T I C L E INTRODUCTION The optimal, universally acceptable diagnostic criteria for cervical radiculo- pathy (CR) are debatable (Wainner & Gill 2000). Magnetic resonance imaging (MRI) and needle electromyography (EMG) are considered to be the most accurate means of diagnosis available. These tests are, however, very expen- sive, not universally available and the diagnostic accuracy thereof is not infal- lible (Wainner & Gill 2000). It has been demonstrated that a high percentage of subjects with structural lesions on MRI were asymptomatic (Schellhas et al 1996). According to many authors the diagnosis of CR remains largely a clinical diagnosis (Wainner & Gill 2000). Different diagnostic tests have been purported to be useful for the clinical diagnosis of CR (Maitland 1993; Butler 2000). Clinicians rely on the distribu- tion patterns of neurological signs and radicular symptoms like pain and paraesthesia, to determine the level of NR involvement, since these tend to have a certain predictive value (Howe et al 1977; Butler 2000). Although the interview, forms an important part of clinical decision making, very weak diagnostic value has been demonstrated. Wainner et al (2003) demonstrated that only two historical questions, one per- taining to the most bothersome area being the scapula (Sp = 0.84, Sn = 0.38, LR+ = 2.3) and the other to the influence of neck movement on the symptoms (Sp = 0.71, Sn = 0.65, LR+ = 2.23), had acceptable diagnostic value (LR > 2). The neurological examination demon- strated moderate interexaminer reliability (Kappa between 0.16 and 0.73) and few isolated tests demonstrated acceptable diagnostic accuracy (LR+ = 2). Butler (2000) proposed that a combination of neurological signs provide a moderately accurate diagnosis, as opposed to one sign which may not offer substantial diagnostic value by itself, supporting the concept of combining tests. Wainner et al (2003) also found however that a combination of four tests was more useful in establishing a diagnosis than any single test. The suggested areas of radicular pain representing each NR level (dermatomes) vary in the literature. Keegan & Garrett (1948) proposed that radicular symp- toms such as pain are distributed in neat anatomically-fixed, non-overlapping dermatomes. However, Lundsford et al (1980) reported that only 40-60% of 295 subjects had single-level neurological signs, whereas 20-30% had multiple- level dermatomal sensory and myotomal deficits. Supporting this, Slipman et al (1998) demonstrated dermatomal over- ABSTRACT: Different proposed pathophysiological mechanisms can result in variable clinical presentations of cervical radicu- lopathy (CR), often making it difficult to detect minor nerve root (NR) conditions. This descriptive study determined (1) the level(s) of NR involvement suggested by the distribution patterns of clinical features and detected by magnetic resonance imaging (MRI) and (2) the most common associations between the different variables in patients diagnosed with acute CR by a neurosurgeon. A physiotherapist blinded to the level(s) of NR involvement performed a standardized interview on 21 subjects to determine the distribution patterns of pain and paraesthesia, and a neurological examination. The Fisher exact test was used to determine associations between the different variables. Only seven subjects presented clinically and radiologically with the same single-level NR involvement. Multiple- level presentations occurred which might be due to dermatomal overlapping, central sensitization or the possible involvement of two adjacent NR levels. Distribution patterns of motor weakness, pain and paraesthesia, and to a lesser extent sensory and reflex changes, have value in identifying the compressed NR level. For this sample the distri- bution patterns of radicular features identified C6 and C8 with more certainty than C7. KEY WORDS: CERVICAL RADICULOPATHY; CLINICAL ASSESSMENT; DIAGNOSIS; NERVE ROOT LEVEL. Conradie M*, MScPhysio (Stell)1, Bester MM, MScPhysio (UWC)1, Crous LC, MScPhysio (Stell)1, Kidd M, PhD (Stell)2 1 Department of Physiotherapy, Faculty of Health Sciences, Stellenbosch University, PO Box 19063, Tygerberg 7505, South Africa 2 Centre for Statistical Consultation, Stellenbosch University, South Africa. CORRESPONDENCE TO: Marlen Conradie Department of Physiotherapy, Faculty of Health Sciences, Stellenbosch University, PO Box 19063, Tygerberg 7505, South Africa Tel: 27-21 938-9300 Fax: 27-21 931-7810 mc6@sun.ac.za (Marlen Conradie) * This study was conducted in partial fulfillment of the requirements for a master’s degree at the Stellenbosch University, South Africa JRLJUN 2006 Short 31/5/06 10:27 am Page 12 SA JOURNAL OF PHYSIOTHERAPY 2006 VOL 62 NO 2 13 lapping by indicating that pain was frequently experienced in widespread, non-dermatomal areas. According to Hall & Elvey (1999) two forms of peripheral neuropathic pain have been identified following nerve injury: dysaesthetic pain and nerve trunk pain. Dysaesthetic pain results from damaged nociceptive afferent axons and is accompanied by axonal damage. In contrast nerve trunk pain results from heightened activity in chemically or mechanically sensitized nociceptive sensory fibres that innervate connective tissue of peripheral nerve trunks. This results in increased mecha- nosensitivity without altered conduction. It is thus possible that patients with nerve injuries can experience dysaesthetic pain or nerve trunk pain in isolation or more commonly, in combination, and this could explain the variable symptom presentation. The purpose of this study was to determine (1) the level(s) of NR involve- ment suggested by the distribution patterns of the clinical features and detected by MRI and (2) the most com- mon associations between MRI and clinical findings in patients with acute cervical radiculopathy (CR). Findings of this study can assist clinicians in their clinical reasoning by suggesting the value for the various radicular features in identifying the involved NR level(s). METHODS Subjects and Study design A prospective study was performed on twenty-one consecutive patients referred by a neurosurgeon after the diagnosis of acute CR was confirmed by MRI. A sample of convenience, referred from private medical practices, was utilized. Exclusion criteria: Chronic CR (longer than 3 months), previous cervical spinal surgery, cervical myelopathy without any radicular involvement, known malignant disease, diabetes mellitus, diagnosed inflammatory joint disease, work related compensation claims and/or pending litigation. Ethical clearance was obtained from the relevant Research Ethics Committee (number 2003/027/N) and all subjects provided informed consent. Procedure A physiotherapist blinded to the suspected level(s) of NR involvement detected by MRI, performed a standardized inter- view and neurological examination as described by Butler (2000) to determine the level(s) of NR involvement suggested by the distribution patterns of clinical features. Findings of the neurological examination tests were graded as either normal or altered in comparison to the opposite extremity. Determining level(s) of NR involvement Criteria for analyses of clinical find- ings. Since dermatomal maps presented in several textbooks differ slightly from each other (Maitland 1993; Patten 1995; Butler 2000), the most commonly docu- mented areas representing each NR level were summarized. From these, the researcher compiled specific criteria (Table 1) for determining the involved NR by selecting the “signature zones” of each spinal nerve (Nitta et al 1993). To avoid dermatomal overlapping, the index and ring fingers were excluded from the criteria. The clinical findings were analysed to determine the involved NR level(s) for each radicular symptom and sign separately. MRI findings. The radiologists’ reports were accessed after completion of the clinical examination. Compression of NR’s was considered if the radio- logists report indicated that MRI detected definite or slight NR compression. Ashkan et al (2002) demonstrated that the sensitivity of MRI for diagnosing CR was 93%, with a positive predictive value of 91% and a negative predictive value of 25% in 48 patients who under- went MRI and neurophysiological studies preoperatively. Instrumentation Data were recorded on a self-designed data capturing sheet that consisted of three sections: section A - demographic data; section B - descriptors of pain and paraesthesia and section C - the neuro- logical findings. The distribution pat- terns of pain, paraesthesia, and altered sensation were mapped out on separate body charts. Statistical analysis Descriptive and inferential statistics were used for the analysis of data. Associations between the suggested level(s) of NR involvement for different variables were assessed with the Fisher exact test. A statistical significance level of 5% was selected and all significant associations (p<0.05) are indicated by * in the presented histograms. RESULTS The demographic information, the level(s) of NR involvement suggested by MRI, as well as clinically detected radicular features are outlined in Table 2, indicating a variable clinical presentation. Level(s) of NR involvement MRI MRI detected compression of a single NR in 13 subjects and multiple NR’s in 8 subjects. The majority of subjects had compression of C7, followed by C6, C8 and C5. Clinical signs and symptoms Neurological sign(s) were detected in 19 subjects. Not all three neurological signs occurred in all the subjects. Motor weakness was detected in 17 subjects of whom 12 subjects presented with a single-level and 5 subjects with multi- ple-level myotomal deficits. Reduced reflexes were detected in 14 subjects. Sensory changes occurred in 14 subjects of whom 11 had a single-level and 3 multiple-level sensory deficits. NR Myotome Reflex Only Pain Pain, Paraesthesia, Sensory changes C5 Shoulder abduction Biceps Deltoid, lateral upper, radial forearm, but not into hand C6 Elbow Flexion Biceps Anterolateral upper arm (biceps) Radial forearm, thumb C7 Elbow Extension Triceps Posterolateral upper arm (triceps) Posterior forearm, middle finger C8 Thumb Extension Triceps Medial upper arm Ulnar forearm, little finger Table 1: Criteria for determining suggested level(s) of NR involvement regarding myotomes, reflexes, pattern(s) of sensory disturbances and pain pattern(s) (Compiled from: Keegan & Garrett 1948; Maitland 1993; Nitta et al 1993; Butler 2000). JRLJUN 2006 Short 31/5/06 10:27 am Page 13 14 SA JOURNAL OF PHYSIOTHERAPY 2006 VOL 62 NO 2 Demographic information Level(s) of NR involvement suggested by: No Age Gender Extre- Referred MRI Pain Paraesthesia Motor Reflex Sensory mity Rx C5 C6 C7 C8 UA FA Finger UA FA Finger LT SP 1 49 Female Left Conser- No No Yes No C7 C7 N/A N/A N/A C7 C7 Triceps N/A N/A vative 2 43 Male Right C5/6 & No Yes Yes No C5 N/A N/A N/A C6&C7 C6&C7 C7 Triceps C6&C7 C6&C7 C6/7 3 38 Female Right C5/6 No Yes No No C6 C6 C6 C6 C6 C6 C6 Biceps C6 C6 4 39 Female Left C6/7 No No Yes No C8 C8 C8 N/A N/A N/A C7 Triceps N/A N/A 5 59 Female Right Conser- No Yes Poss Poss N/A C8 N/A N/A C8 C7&C8 N/A N/A C8 C8 vative 6 55 Male Right Conser- No No Yes No N/A C7 N/A N/A C7 N/A N/A N/A N/A N/A vative 7 55 Male Right Conser- No No Yes Poss C8&C7 N/A N/A N/A C8 C8 C7&C8 N/A N/A C8 vative 8 33 Male Left Conser- No No Slight No C8 C8 N/A N/A N/A N/A C7 N/A C8 N/A vative 9 42 Male Left C6/7 No No Yes No N/A C6&C7 C6&C7 N/A N/A C6 C7 Triceps C6 C6 10 39 Male Left C6/7 No Yes Yes No C5&C6 C5&C6 N/A N/A N/A C6&7 C7 Triceps C6&C7 C6&C7 11 63 Female Right C6/7 Yes No Slight No C5&C7 N/A N/A N/A N/A N/A C7 Triceps N/A N/A 12 57 Female Right C7/T1 No No No Slight N/A C7&C8 C7&C8 N/A N/A C8 C8 Triceps C8 C8 13 54 Female Left C6/7 No No Yes No N/A C7 C7 N/A N/A C7 C7 Triceps N/A C7 14 40 Male Left C6/7 No No Yes Poss N/A C6&C7 C6&C7 N/A N/A C6&C7 C6&C7 Triceps N/A N/A 15 50 Female Left C5/6 No Slight Poss No C6 C6 C6 C6 C6 C6 C6&C7 Biceps C6 C6 16 51 Male Left C5/6 & No Yes Slight No C5&C6 C6&C7 N/A N/A N/A C6 C6&C7 N/A C6&C7 C7 C6/7 C7 17 41 Male Left C6/7 No No Yes No C7 C7 N/A N/A C6&C7 C6&C7 C7 Triceps N/A N/A 18 54 Male Right Conser- No No Slight No N/A C7 N/A N/A N/A N/A N/A N/A N/A N/A vative 19 44 Female Left Conser- No No Slight No C7 C7 C7 N/A C7 C7 N/A N/A N/A C7 vative 20 59 Female Right C5/6 No Yes No No C6 C6 C6 N/A C6 C6 C6&C7 Biceps C6 C6 21 40 Female Right C5/6 No Yes No No C6 C6 C6 C6 C6 C6 C6 Biceps C6 C6 Key for table 2: UA = Upper arm; FA = Forearm; Sing = Single; Multi = Multiple Single = Single-level NR involvement; Multiple = Multiple-level NR involevement N/A = not applicable due to the absence of the radicular sign or symptom; Compression of NR was divided into definite compression (Yes), slight compression (Slight), possible dynamic compression with no detectable static compression (Poss) and no compression (No) according to the MRI LT = Light touch sensory testing; SP = Superficial pain sensory testing Table 2: Demographic data for the 21 subjects and suggested level(s) of NR involvement for MRI, radicular symptoms and signs. All 21 subjects experienced pain of whom 14 subjects reported a single- level and 7 multiple-level dermatomal pain patterns. In only 12 subjects the pain was distributed in the correspond- ing dermatome of compressed NR level identified by MRI. Pain was commonly experienced over the upper trapezius, posterior neck and medial scapular border, and was more frequently experienced over the upper- and forearm, than the distal fingers. Over- lapping between C6 and C7 NR’s regarding the index finger and C7 and C8 NR’s regarding the ring fingers occurred. Paraesthesia was reported by 17 subjects, of whom 13 had a single-level and 4 multiple-level distribution patterns. JRLJUN 2006 Short 31/5/06 10:27 am Page 14 SA JOURNAL OF PHYSIOTHERAPY 2006 VOL 62 NO 2 15 Paraesthesia was always reported in the most distal parts of the dermatomes, reaching the fingers. The area/s most commonly reported representing each NR level were: thumb and index finger (C6), index, middle and ring fingers (C7), little and ring fingers (C8). Over- lapping between NR levels regarding the index and ring fingers were reported. Comparison between MRI and clinical features Only 7 out of 21 subjects presented clini- cally and radiologically with the same single-level NR involvement. ASSOCIATIONS BETWEEN MRI AND CLINICAL FEATURES Only the results of C6, C7 and C8 NR levels are presented since the clinical features were only reported in these distribution patterns. In the histograms, the MRI ‘yes’ group (right) refers to the subjects in whom MRI was able to detect compression of the mentioned NR level and the MRI ‘no’ group (left) represents the subjects without evident compression. Motor testing (Figure 1a-c) Statistically significant associations were found between MRI compression and motor weakness in the corresponding myotome for C6 (p<0.01) and C8 (p=0.03), but not for C7 (p=0.09). More than half of the sample with detectable compression of C6, C7 and C8 presented with corresponding myotomal weakness. Motor weakness occurred in one subject without compression of C6 and C7. Sensory testing The C6 NR level demonstrated signifi- cant associations between MRI compres- sion and sensory changes with both light touch (LT) (p<0.01) and superficial pain (SP) (p<0.01) sensory testing. The C8 NR level showed a significant asso- ciation for SP (p<0.01) but not for LT testing (p=0.08). No significant asso- ciation was found for the C7 NR level with both LT (p=0.51) and SP (p=0.30) sensory testing. Three-quarters or more of the sub- jects with C6 (6 subjects) and C8 (3 subjects) compression presented with sensory changes in the corresponding dermatome. Twelve of the 17 subjects with C7 compression and all 4 subjects without C7 compression had normal SP sensation in the C7 dermatome, indicat- ing a low incidence of altered sensation in the C7 distribution pattern. Altered SP sensation was also noted in two subjects without C6 NR compression. Reflex testing Only half of the sample with NR com- pression had hypoactivity of the corre- sponding reflex, indicating a low incidence of altered reflexes. The association between the level of NR compression and hypoactivity of the corresponding reflex was significant for C6 (p<0.01), but not for C8 (p=0.67) or C7 (p=0.33). Radicular pain Figure 2(a-c) The association between the level of MRI compression and pain in the corre- sponding dermatomal distribution pattern was significant for C6 (p=0.04) and C8 (p=0.03), but not for C7 (p=0.41). Between 71% and 75% of the subjects with detectable MRI compression expe- rienced pain in the corresponding dermatome. Pain was experienced in dermatomes without detectable compres- sion of the corresponding NR level for C6 (3 subjects), C7 (2 subjects) and C8 (2 subjects). Paraesthesia (Figure 3a-c) Significant associations were found for C6 (p<0.01) and C8 (p<0.01) between MRI compression and paraesthesia in the corresponding dermatome, but not for C7 (p=0.41). Three-quarters or more of the subjects with either C6 (6 subjects) or C8 (3 subjects) compression expe- rienced paraesthesia in the corresponding dermatome. A low incidence of paraes- thesia (47%) in the C7 distribution pat- tern occurred amongst the subjects who had C7 compression. Paraesthesia was reported in the C6 (3 subjects) and C7 (1 subject) dermatomes without detect- able compression of the corresponding NR level. DISCUSSION In this study only 7 subjects out of 21 with acute CR presented clinically and radiologically with the same single-level NR involvement. Current results indicate that radicular pain can occur in isolation, supporting the proposal of Greening & Lynn (1998) that neuropathic pain may be due to nerve sheath inflammation when rela- tively minor or no axonal damage is present. This is in contrast to previous beliefs that the diagnosis of CR could only be made when the radicular pain was accompanied by evident neurolo- gical signs and that all radicular pain is due to NR compression. Furthermore all three neurological signs were not always present simultaneously, confirming pre- vious findings (Lundsford et al 1980; Radhakrishnan et al 1994). These observations can be justified if neurophysiological evidence is considered (Gifford 2001). In a study by Howe et al (1977) mechanical compression of the lumbar NR elicited paraesthesia and numbness, and caused neurological signs due to altered conduction, but not pain. It is therefore possible that any one or more of the three neurological signs can be present, depending on which part of the NR complex is affected. On the other hand, radicular pain will only be expe- rienced once the NR is subjected to intraneural inflammation (Saal 1995), due to peripheral sensitization (Greening & Lynn 1998) or central sensitization (Butler 2000; Scholtz & Woolf 2002). This could explain why only 12 of the 21 subjects reported pain in the corre- sponding dermatome of the compressed NR. A combination of mechanical and chemical mechanisms can cause any combination of radicular features. It is evident that radicular pain and signs should be recognized as two separate pathophysiological entities which can occur in isolation or combination, depending whether the mechanism of mechanical deformation and/or chemical irritation is responsible for the clinical presentation. In this study, NR compression was also detected by MRI without corre- sponding clinical signs and symptoms and vice versa. Although MRI is the imaging method of choice in identifying structural lesions, it has been demon- strated that MRI detected compression of NR’s in 20-30% of asymptomatic subjects. Schellhas et al (1996) therefore suggested that MRI used in isolation, is inadequate for the reliable identification of the source of pain. The utilization of MRI as the reference criterion, which is unable to detect intraneural inflam- mation, might be one of the reasons why radicular symptoms were reported in the dermatomal distribution(s) of the NR’s of which compression was not detected. Current results demonstrated that the distribution patterns of clinical features suggested multiple-level NR involve- ment in approximately one-third of the sample. This is in agreement with the findings of Lundsford et al (1980) and Radhakrishnan et al (1994). The occur- rence of multiple-level clinical presenta- JRLJUN 2006 Short 31/5/06 10:27 am Page 15 16 SA JOURNAL OF PHYSIOTHERAPY 2006 VOL 62 NO 2 93% 7% 29% 71% 14 12 10 8 6 4 2 0 N um be r of s ub je ct s normal reduced normal reduced 0% 18 16 14 12 10 8 6 4 2 0 100% 50% 50% N um be r of s ub je ct s normal reduced normal reduced 75% 25% 24% 76% 14 12 10 8 6 4 2 0 N um be r of s ub je ct s normal reduced normal reduced Figure 1: Comparison between MRI findings and clinically demonstrated myotomal motor weakness for (a) C6, (b) C8 and (c) C7 NR (Fisher exact, one-tailed test and categorized histograms). Figure 3: Comparison between MRI findings and dermatomal area(s) of paraesthesia as determined during the interview for (a) C6, (b) C8 and (c) C7 NR involvement (Fisher exact, one-tailed test and categorized histograms). Figure 2: Comparison between MRI findings and dermatomal area(s) of radicular pain as determined during the interview for (a) C6, (b) C8 and (c) C7 NR involvement (Fisher exact, one-tailed test and categorized histograms). MRI: yes Figure 1(a) MRI vs motor weakness C6 p<0.01* MRI: no MRI: yes Figure 1(b) MRI vs motor weakness C8 p=0.03* MRI: no MRI: yes Figure 1(c) MRI vs motor weakness C7 p=0.09 MRI: no 79% 21% 29% 71% 12 10 8 6 4 2 0 N um be r of s ub je ct s no yes no yes 16 14 12 10 8 6 4 2 0 25% 75% no yes no yes N um be r of s ub je ct s 88% 12% 14 12 10 8 6 4 2 0 50% 50% no yes no yes N um be r of s ub je ct s 71% 29% MRI: yes Figure 2(a) MRI vs Pain C6 p=0.04* MRI: no MRI: yes Figure 2(b) MRI vs Pain C8 p=0.03* MRI: no MRI: yes Figure 2(c) MRI vs Pain C7 p=0.41 MRI: no 79% 21% 14% 86% 12 10 8 6 4 2 0 N um be r of s ub je ct s no yes no yes 0% 18 16 14 12 10 8 6 4 2 0 25% 75% 100% N um be r of s ub je ct s no yes no yes 10 9 8 7 6 5 4 3 2 1 0 75% 25% 53% N um be r of s ub je ct s no yes no yes 47% MRI: yes Figure 3(a) MRI vs Paraesthesia C6 p<0.01* MRI: no MRI: yes Figure 3(b) MRI vs Paraesthesia C8 p<0.01* MRI: no MRI: yes Figure 3(c) MRI vs Paraesthesia C7 p=0.41 MRI: no JRLJUN 2006 Short 31/5/06 10:27 am Page 16 SA JOURNAL OF PHYSIOTHERAPY 2006 VOL 62 NO 2 17 tions could be due to several factors: Slipman et al (1998) have suggested the concept of dermatomal overlapping. These investigators demonstrated that direct stimulation of NR’s in patients with radicular pain caused symptoms in non-dermatomal, widespread distribu- tions and overlapping of dermatomes, especially regarding the fingers. More recently, Butler (2000) proposed that the distribution patterns for peripheral neu- rogenic pain may not be clear if central sensitivity coexists. Central sensitization refers to a lowered activation threshold of the central nervous system when inputs are magnified (Winkelstein 2004). Another possible reason for the mul- tiple-level presentations might be the simultaneous involvement of two NR levels. The researcher propose that inflammation, surrounding the original disc herniation that compress a NR level, might spread, leading to sensiti- zation of the adjacent level above or below. Thus, neurological signs could occur as a result of the compressed NR, whereas pain may be due to mechano- sensitivity of the adjacent level, even when MRI was unable to detect com- pression of this level. Current results suggest that certain radicular features when used in isolation have more value than others in identify- ing the involved cervical NR level detected by MRI. It is proposed from the results that motor weakness has the most value in determining the compressed NR level(s), followed by pain, paraes- thesia, sensory changes and lastly altered reflexes. The observed trend that motor weakness had better predictive value than sensory changes correlates to previous recommendations. Motor find- ings in the upper extremity are more reliable than sensory findings, since each movement of the arm is controlled almost exclusively by a single NR. In this sample both pain and paraesthesia had value in predicting the compressed NR level(s). From the results it seems that, when present, paraesthesia provides a better indication of the compressed NR than pain which did not often extend into the fingers. Significant associations (p<0.05) were found between all the variables for C6, most of the variables for C8, and none for C7 NR, indicating that for this sample the distribution pat- terns of radicular features identified C6, and to a lesser extent C8 NR levels with greater certainty than C7. Consequently, current results indicate that the diagnosis of CR and the identi- fication of the involved NR level(s) can- not be based on the distribution pattern of one isolated radicular feature. The variable clinical presentations, might explain the inability to formulate optimal, universally accepted diagnostic criteria and why very few isolated tests have demonstrated acceptable diagnostic accuracy. The researcher proposes that it is essential to combine findings from different diagnostic tests and the symp- tom presentation, emphasizing Maitland’s recommendation of “making all the fea- tures fit” (Maitland 1993, p. 55). Findings of this study can not be generalized due to the small sample of convenience. This small, descriptive study indicates areas for further work on NR conditions performed on larger and more representative samples. Before very specific diagnostic criteria for CR can be formulated, it will be necessary to determine the diagnostic accuracy of different combinations of clinical, radiological, imaging and surgical find- ings. Future research could explore the possibility of radicular pain and neuro- logical signs being recognized as two pathophysiological entities, even distin- guishing between nerve trunk pain and dysaesthetic pain. Non-invasive ultra- sound imaging could be utilized for the recognition of intra-neural inflammation when assessing neuropathic pain. CONCLUSION Current results indicate that the clinical presentation of acute CR is variable, since multiple-level distribution patterns of clinical features occurred. The vari- able, multiple-level presentations may be due to dermatomal overlapping, central sensitization or the possible involvement of two adjacent NR levels. The distribution patterns of motor changes, pain and paraesthesia has value in determining the compressed NR level(s) and to a lesser extent sensory and reflex changes. Radicular features may identify C6 and C8 NR levels with more certainty than C7. ACKNOWLEDGEMENTS The researcher would like to thank all neurosurgeons i.e Dr CAJ Botha, NP Fisher-Jeffs & MP Small-Smith, for the referral of patients and all subjects who participated in this study. REFERENCES Ashkan K, Johnston P, Moore AJ 2002 A comparison of magnetic resonance imaging and neurophysiological studies in the assessment of cervical radiculopathy. British Journal of Neurosurgery 16(2): 146-148 Butler DS 2000 Manual assessment of nerve conduction In: The sensitive nervous system, ch 9 p 211-255. Noigroup Publications, Adelaide Gifford L 2001 Acute low cervical nerve root conditions: Symptom presentations and pathobiological reasoning. Manual Therapy 6(2): 106-115 Greening J, Lynn B 1998 Minor peripheral nerve injuries: An underestimated pain? Manual Therapy 3(4):187-194 Hall TM, Elvey RL 1999 Nerve trunk pain: Physical diagnosis and treatment. Manual Therapy 4(2): 63-73 Howe JF, Loeser JD, Calvin WH 1977 Mechanosensitivity of dorsal root ganglia and chronically injured axons: A physiological basis for the radicular pain of nerve root compression. Pain 3: 25-41 Keegan JJ, Garrett FD 1948 The segmental distribution of the cutaneous nerves in the limbs of man. Anatomical Record 102: 409-437 Lundsford LD, Bissonette DJ, Jannetta PJ, Sheptak PE, Zorub DS 1980 Anterior surgery for cervical disc disease. Part 1: Treatment of lateral cervical disc herniation in 253 cases. Journal of Neurosurgery 53: 1-11 Maitland GD 1993 Vertebral Manipulation, 5th edn. Butterworth-Heinemann, London Nitta H, Tajima T, Sugiyama H 1993 Study on dermatomes by means of selected lumbar spinal nerve block. Spine 10: 1782-1786 Radhakrishnan K, Litchy WJ, O’Fallon WM, Kurland T 1994 Epidemiology of cervical radi- culopathy: A population-based study from Rochester, Minnesoto, 1976 through 1990. Brain 117: 325-335 Saal JS 1995 The role of inflammation in lumbar pain. Spine 20(16): 1821-1827 Schellhas KP, Smith MD, Gundry CR, Pollei SR 1996 Cervical discogenic pain: Prospective correlation of magnetic resonance imaging and discography in asymptomatic subjects and pain sufferers. Spine 21(3): 300-312 Scholz J, Woolf CJ 2002 Can we conquer pain? Nature Neuroscience Supplement 5:1062-1067 Slipman CW, Plastaras CT, Palmitier RA, Huston CW, Sterenfeld EB 1998 Symptom provocation of fluoroscopically guided cervical nerve root stimulation: Are dynatomal maps identical to dermatomal maps? Spine 23(20): 2235-2242 Takahashi N, Yabuki S, Aoki Y, Kikuchi S 2003 Pathomechanisms of nerve root injury caused by disc herniation. Spine 28(5): 435-441 Wainner RS, Gill H 2000 Diagnosis and non- operative management of cervical radiculopathy. Journal of Orthopaedic & Sports Physical Therapy 30(12): 728-744 Wainner RS, Fritz JM, Irrgang JJ, Boninger ML, Delitto A, Allison S 2003 Reliability and diag- nostic accuracy of the clinical examination and patient self-report measures for cervical radi- culopathy. Spine 28(1): 52-62 Winkelstein BA 2004 Mechanisms of central sensitization, neuroimmunology & injury biome- chanics in persistent pain: implications for musculoskeletal disorders. Journal of Electro- myography and Kinesiology 14(1):87-93 JRLJUN 2006 Short 31/5/06 10:27 am Page 17