- Email: [email protected]
Mid-thoracic tenderness: a comparison of pressure pain threshold between spinal regions, in asymptomatic subjects
Manual Therapy (2001) 6(1), 34–39 # 2001 Harcourt Publishers Ltd doi:10.1054/math.2000.0377, available online at http://www.idealibrary.com on
Original article
Mid-thoracic tenderness: a comparison of pressure pain threshold between spinal regions, in asymptomatic subjects L. Keating, C. Lubke, V. Powell, T. Young, T. Souvlis, G. Jull Department of Physiotherapy, The University of Queensland, Brisbane, Australia
SUMMARY. Palpation for tenderness forms an important part of the manual therapy assessment for musculoskeletal dysfunction. In conjunction with other testing procedures it assists in establishing the clinical diagnosis. Tenderness in the thoracic spine has been reported in the literature as a clinical feature in musculoskeletal conditions where pain and dysfunction are located primarily in the upper quadrant. This study aimed to establish whether pressure pain thresholds (PPTs) of the mid-thoracic region of asymptomatic subjects were naturally lower than those of the cervical and lumbar areas. A within-subject study design was used to examine PPT at four spinal levels C6, T4, T6, and L4 in 50 asymptomatic volunteers. Results showed significant (P50.001) regional differences. PPT values increased in a caudal direction. The cervical region had the lowest PPT scores, that is was the most tender. Values increased in the thoracic region and were highest in the lumbar region. This study contributes to the normative data on spinal PPT values and demonstrates that mid-thoracic tenderness relative to the cervical spine is not a normal finding in asymptomatic subjects. # 2001 Harcourt Publishers Ltd.
such as a thermal, mechanical or chemical stimulus and a subjective response. Cohen (1995) considers that, clinically, the term hyperalgesia, which is defined as an increased response to a stimulus that is normally noxious (Merskey & Bogduk 1994), also includes allodynia, the response of pain to a stimulus that is not normally noxious. Tenderness in musculoskeletal disorders may be a result of local tissue damage or may indicate an alteration in afferent processing of mechanical stimuli by second order neurones within the dorsal horn (Sheather-Reid & Cohen, 1998). Thus, hyperalgesia may be described as primary or secondary depending on whether it originates in tissue which has been damaged (primary) or in normal tissue which may be neuroanatomically related to the area of damage (secondary). Sensitization of both the peripheral and central nervous systems have been demonstrated to cause the decrease in threshold or increased response to mechanical stimuli that result in hyperalgesia (Meyer et al. 1994). Palpation to identify areas of tenderness may provide us with important information in the assessment of varied musculoskeletal dysfunctions and pain syndromes as it provides relevant information about both the underlying tissue and the nociceptive system.
INTRODUCTION An important element of manual therapy assessment is the use of passive manual examination to detect the presence of musculoskeletal dysfunction. Palpation of the spine for tenderness is part of this manual examination (Maitland 1986). In the thoracic spine a challenge arises in interpreting the diagnostic relevance of tenderness. Controversy exists regarding whether pain and tenderness within the mid-thoracic spine is an expected finding in patients who present with symptoms in the upper quadrant. Is the midthoracic spine normally tender or does increased sensitivity in this area indicate local dysfunction? Cohen (1995) describes tenderness or more accurately, hyperalgesia, as a psychophysical concept that establishes a relationship between a physical stimulus Received 6 March 2000 Revised 4 August 2000 Accepted 11 November 2000 L. Keating, BPhty, MPhty St. Catherine Lubke BPhty, M.Phty St, Vicki Powell BPhty (Hons), MPhty, St, Tanya Young BPhty MPhty St, Tina Souvlis BPhty (Hons), Gwendolen Jull MPhty, Grad Dip Adv Manip Ther, FACP, Associate Professor, Department of Physiotherapy, University of Queensland, Brisbane, Qld 4072, Australia. Correspondence to GJ. Tel: +61 7 3365 2275; Fax: +61 7 335 2775. E-mail: [email protected] 34
Mid-thoracic tenderness: a comparison of pressure pain threshold 35
Pressure algometry has been demonstrated to be a reliable and repeatable tool to quantify local pain and tenderness (Fischer 1987, Magora et al. 1992, Nussbaum & Downes 1998). Pressure pain threshold is defined as the least stimulus intensity at which a subject perceives discomfort (Fischer 1986a). PPT has been used as a measure to quantify pain and tenderness in asymptomatic subjects thereby establishing normative PPT values (Fischer 1986a, Fischer 1986b, Reeves et al. 1986, Fischer 1987, Ohrbach & Gale 1989, Hogeweg et al. 1992). Subsequently, PPT has been measured in symptomatic patients presenting with a variety of musculoskeletal conditions (Bovim 1992, Vatine et al. 1998). It has also been suggested that comparisons of PPT values can be used as an aid in diagnosis (Fischer 1987), and additionally, they have been used as an outcome measure for therapeutic interventions (Vicenzino et al. 1998, Dhondt et al. 1999). Significant regional differences in PPT of the paravertebral tissues of the spine have been observed in symptomatic subjects (Vanderweee¨n et al. 1996) and in asymptomatic subjects (Hogeweg et al. 1992, Kosek et al. 1993) with values increasing in a caudal direction. These studies have more frequently addressed muscle and soft tissue tenderness, with reference to trigger point pathology. However, manual examination techniques, such as posteroanterior glides involve pressure over bony prominences. Kosek et al. (1993) compared PPT values over bone versus muscle and found no difference, whereas Fischer (1986a) reported a difference but did not analyse for statistical significance. With respect to the thoracic spine, Minucci (1987), in a small study of asymptomatic subjects, demonstrated that testing the thoracic spine with posteroanterior glides produced a subjective report of discomfort or pain in 37% of this group. This was found to be significantly more frequent at T3, T4 and T5. In an asymptomatic population, several theories can be offered to explain the presence of mid-thoracic tenderness. Maigne et al. (1991) performed dissections of T1–T5 and postulated that local tenderness over the spinous processes is due to the location of the medial branch of the dorsal rami over the apex of these prominences. Kosek et al. (1999) found lower PPT values for muscle sites where the nerve became superficial (‘muscle/nerve sites’) than in ‘pure’ muscle and bony sites. In the light of these findings, Minucci’s results might be explained by compression of several dorsal cutaneous rami under the tendinous insertion of splenius cervicis on the spinous processes of T3, T4 and T5. Thoracic pain and tenderness has been reported as a feature of a variety of musculoskeletal conditions involving both the thoracic spine and structures capable of pain referral to the thoracic spine. Injection studies of cervical zygapophysial joints # 2001 Harcourt Publishers Ltd
(Dwyer et al. 1990) and cervical discs (Cloward 1959) have both demonstrated pain referral to the thoracic region. Cervical spondylosis may produce pain and movement dysfunction in the upper thoracic spine (Edmonston & Singer 1997). Carrothers (1994) observed a similar relationship between whiplash associated cervical injury and the thoracic spine. However, Giglia-Smith et al. (1997), in a group of chronic cervical pain patients without thoracic pain, demonstrated that there was no difference in pressure pain threshold (PPT) values in the thoracic spine when compared to an asymptomatic control group. Focusing specifically on the mid-thoracic region, the T4 syndrome is a clinical entity that is purported to have tenderness on palpation as one of several diagnostic signs (Matthews 1986, DeFranca & Levine 1995, Evans 1997). Despite its name, T4 to T7 levels have been documented as being implicated in the T4 syndrome. Scientific validation of this clinical syndrome has as yet, not been forthcoming. The primary aim of this study was to investigate whether the mid-thoracic area exhibits greater tenderness than other regions of the spine in an asymptomatic population. Comparison of PPTs over the cervical, thoracic and lumbar spinous processes would assist in determining the clinical relevance of thoracic tenderness when found in patients with upper quadrant pain syndromes. Pressure algometry was used to measure PPTs to gain a quantitative estimation of tenderness. Two segments of the midthoracic spine (T4 and T6) were chosen to encompass the region in the thoracic spine that is often cited as tender in asymptomatic subjects as well as in patients with upper quadrant clinical syndromes. Studies have shown the spinal cord to be relatively fixed at C6, T6 and L4 (Louis 1981). For T6, as a tension point (Butler 1991), there is a direct link to a proposed mechanism of thoracic tenderness whereas T4 is not considered a tension point. Yet both are associated with a specific pain syndrome. Both segments share similar biomechanical properties and anatomical relationships to the sympathetic nervous system (Grieve 1988). The levels of C6 and L4 were chosen to counterbalance the possible general influence of tension points on PPT values. The hypothesis tested in this study was that reduced pressure pain threshold is present in one or more segments in the mid-thoracic spine (T4 or T6) in asymptomatic subjects, when compared with cervical and lumbar segments in the same subject.
METHODOLOGY Subjects Fifty volunteers (23 males and 27 females) aged 19 to 56 (mean=25.7+7.1years) were recruited from The Manual Therapy (2001) 6(1), 34–39
36 Manual Therapy
University of Queensland and local community. Subjects were required to have no history of headaches, or pain in the cervical, thoracic or lumbo-sacral regions. Subjects were excluded on the basis of the criteria outlined in Table 1. Prior to commencement of the study, the subjects received a full explanation of the experimental procedure and signed a consent form. The Human Ethics Committee of The University of Queensland, in accordance with the National Health and Medical Research Council’s guidelines, approved this study. Apparatus and measurement PPT was measured using a hand held electronic pressure algometer (Somedic, Sweden). The circular probe had a surface area of 1cm2. The algometer was calibrated at the start of each session and the same investigator took all PPT measurements. Pressure was applied to the spinous processes at a rate of 40 kPa/s and the subject was instructed to push a hand held button immediately the sensation changed from pressure to pain. This instantaneously froze the pressure reading, which was then recorded by a second investigator. Three measurements were taken at each level, with an interval of 20 sec between readings. The PPT was calculated as the mean of these 3 trials. A standard treatment plinth was used to position the subject. A third examiner identified spinal levels each time, using a standardized protocol counting in cranial and caudal directions and checking with lateral bony landmarks. The relevant interspinous spaces were then marked with a pen. Procedure Following an explanation of the experiment, the subject signed an informed consent statement. To familiarize the patient with pressure algometry, a demonstration was performed, firstly on the examiner’s wrist, and then on the subject’s ulnar styloid process. The subject was instructed to push the button when the pressure became a discomfort.
The subject undressed to expose the spine and was positioned prone on the plinth with 1 pillow each underneath the hips and feet. The spinal levels, C6, T4, T6 and L4, were then marked. The order of levels to be tested was systematically randomized prior to testing by drawing lots. Confounding variables were controlled by standardizing equipment set up, noise and light levels, subject instructions, surface markings and patient position. The pressure algometer was positioned perpendicular to the body surface at the specified spinous process. The algometer pressure tip was stabilized between the examiner’s thumb and first finger. Pressure readings were recorded once the subject had pressed the indicator button. On completion of the measurement of all spinal levels, the subject’s skin was cleaned to remove the surface markings. The subjects attended one session only. Repeatability Repeatability of PPT measurement by the examiner was verified by conducting a second experimental session on 10 subjects (5 female and 5 male), 30 min after the completion of the first session. An Intraclass Correlation Coefficient (ICC) was calculated using a 2-way random effect model (for consistency agreement). The results of the analysis are presented in Table 2. The ICCs showed that the reproducibility of PPTs in the cervical and thoracic levels was excellent (ICC40.9), and good at the lumbar level (ICC40.75). Inspection of the coefficient of variance (CV) indicated that PPTs varied substantially between subjects (CV ¼ 28–45%), however this was similar to other published PPT studies (Hogeweg et al. 1992, Fischer 1987). Data management For the main study, the analysis was undertaken on the mean of the three PPT readings taken at each spinal level. The main analysis compared the mean PPT readings taken from different levels of the spine. The comparison was performed using a mixed-model analysis of variance, where level (C6, T4, T6, and L4) was a within-groups factor and gender (male, female)
Table 1. Exclusion criteria for subjects History of spinal pain and/or radiculopathy Spinal congenital abnormality Deformity or history of spinal fracture History of chronic respiratory disease e.g. asthma Diabetes Peripheral vascular disease Osteoporosis Rheumatoid arthritis Ankylosing spondylitis Pregnancy Medication: NSAIDs, analgesics, anti-coagulants and steroids
Manual Therapy (2001) 6(1), 34–39
Table 2. Intraclass correlation coefficient and coefficient of variance for the spinal levels tested in repeatability study Spinal level
C6 T4 T6 L4
Mean PPT (kPa/cm2) (+standard deviation)
Coefficient of variance
Test 1
Test 2
Test 1
Test 2
335+150 349+141 373+147 504+182
323+117 390+129 373+103 513+164
45% 40% 39% 36%
36% 33% 28% 32%
ICC
0.93 0.93 0.90 0.84
# 2001 Harcourt Publishers Ltd
Mid-thoracic tenderness: a comparison of pressure pain threshold 37
was a between-groups factor. A series of pairwise comparisons were undertaken to investigate differences between levels, and Tukey’s Least Significant Difference (LSD) test was used to adjust for multiple comparisons. All analyses were undertaken using SPSS 8.0 for Windows. P50.05 was considered to be statistically significant.
significant difference (P50.05) in the average PPT increasing in a caudal direction from cervical (C6: 255 kPa/cm2), to thoracic, to lumbar spine (L4: 445 kPa/cm2). There was no significant difference (P=0.184) for PPT values within the thoracic spine (T4: 324 kPa/cm2 and T6: 302 kPa/cm2).
DISCUSSION RESULTS Distributions of average PPT readings were examined using boxplots. These determined that at each spinal level, the scores were normally distributed with the exception of four outliers. The average PPT scores were transformed by a log scale, which failed to remove the outliers. These were subsequently removed from the initial analysis. The analysis found a significant gender effect when the outliers were included but not when they were removed. Due to the instability of the gender effect, it was decided to interpret the results conservatively. Inspection of Box’s M indicated that the data conformed with the homogeneity of variancecovariance assumption (M ¼ 11.22, F(10, 7888)=1.01, P=0.434), whilst the Levene’s tests indicated that the variance of scores in each gender was similar at each of the four levels. Regional comparison The mean PPTs for each spinal region are summarized in Figure 1. These means were compared using a mixed model analysis of variance. The results of the analysis revealed a significant main effect for region (F(3,23)=20.13, P50.001), but no main effect for gender (F(1,25)=2.187, P=0.146) and no interaction between gender and region (F(3,23)=0.96, P=0.791). The effect of region accounted for around 60.6% of the variance in average PPT. Post hoc pairwise comparisons, using Tukey’s test, determined which levels accounted for the significant difference between the regions. These showed a
Fig. 1—Mean PPT values for tested spinal level. # 2001 Harcourt Publishers Ltd
Significant regional differences in spinal PPT values were noted in the study, with values increasing from the cervical to the lumbar spine. It was shown that mid-thoracic segments were less tender than the cervical segment but more tender than the lumbar segment. Wright et al. (1995) also demonstrated significantly lower PPTs from C6 (P50.01) than from sites on the upper limb. Lower PPTs evident in the cervical spine may indicate that protection from noxious stimulation is important in this area due to close proximity of vascular and neural bundles. There was no significant difference between the PPT values within the mid-thoracic region at T4 and T6. Based on these results, palpation that elicits greater tenderness or hyperalgesia in the mid-thoracic spine than in the cervical spine, is not a normal finding in the asymptomatic individual. The average PPT values compliment the results of previous studies of paraspinal tissue in healthy (Hogeweg et al. 1992, Kosek et al. 1993) and symptomatic subjects (Vanderweee¨n et al. 1996). Based on the results of this study and in the presence of relevant subjective and physical findings, mid-thoracic tenderness may be considered relevant. Thoracic PPT values that differ significantly from the values, or regional pattern, reported in this study warrant further clinical consideration. As would be expected in a normal distribution, some asymptomatic subjects deviated from the regional pattern, therefore thoracic tenderness, in isolation, should not constitute a diagnosis of dysfunction. The findings of the current study suggest that the use of intra-subject regional patterns of tenderness should be further explored. With the normative values and regional pattern provided in this study, pressure algometry might be considered for use in the clinical setting to measure PPT and to quantify tenderness. Whilst this study measured PPT at two thoracic spinal levels, it is not unreasonable to extrapolate this data to the rest of the thoracic spine as Hogeweg et al. (1992) reported an unremarkable steady increase in PPT values measured at T1, T3, T6 and T10. The averaged values were similar to this study despite the expected differences between PPT values over soft tissue and bone. As Hogeweg et al. (1992) reported very similar readings at three lumbar paraspinal levels (L1, L3, L5), it would also be reasonable to extrapolate the single L4 value to the rest of the Manual Therapy (2001) 6(1), 34–39
38 Manual Therapy
lumbar region. As investigation of the cervical region has been limited to the C6 spinous process (Hogeweg et al. 1992) and the C7 spinous process (Kosek et al. 1993), care must be taken inferring PPT values to the whole area. The results of this quantitative study cannot be compared to those of Minucci (1987) where PAIVM tests were used to elicit tenderness. PAIVM testing is not replicated with pressure algometry as the former technique also uses accessory joint motion for pain provocation. There may be a relationship between the two assessment techniques but further study is required. Whilst this study has demonstrated that relative mid-thoracic tenderness is not a normal finding in asymptomatic subjects, future studies must confirm the presence of mid-thoracic tenderness in varied symptomatic populations. Lowered PPT values or a change in the regional pattern would represent this. The presence of thoracic tenderness should be quantified in relevant musculoskeletal dysfunctions such as cervical spondylosis, cervicogenic headache, whiplash-associated disorders and cervical disc pathology. A preliminary study of patients with chronic cervical pain without thoracic symptoms has found thoracic pain threshold was similar to a control group (Giglia-Smith et al. 1997). This data should be interpreted with caution because of the small sample size investigated (n=29). Future studies of symptomatic subjects should not only assess thoracic tenderness but additionally consider the pathogenesis of clinical pain presentations. Monitoring changes in cervical and thoracic intervertebral motion, sympathetic nervous system outflow and neuromeningeal compliance in addition to PPT could begin to decipher the pathogenesis of various clinical syndromes that have mid-thoracic pain/tenderness as a common feature. Tenderness of the thoracic spine may also indicate the presence of secondary hyperalgesia or sensitization of the central nervous system causing a heightened response or lowered threshold to mechanical stimulation in normal tissue. This scenario should also be considered as it may offer an alternative explanation for thoracic tenderness in upper quadrant musculoskeletal dysfunction.
CONCLUSION The results of this study of asymptomatic subjects, showed significant regional differences in spinal PPT, with values increasing in a caudal direction. The midthoracic segments, T4 and T6, were less tender than cervical, C6 segment but more tender than the lumbar, L4 segment. Anecdotally, physiotherapists report tenderness on palpation of the mid-thoracic spine in patients with a variety of musculoskeletal Manual Therapy (2001) 6(1), 34–39
disorders. This study clarifies that mid-thoracic tenderness is not a normal finding in asymptomatic subjects. Future studies should confirm the presence of thoracic tenderness in patients presenting with upper quadrant dysfunction.
Acknowledgments The authors wish to thank the subjects for their contribution to this study and acknowledge the assistance of Mr Craig Shaw with statistical analyses.
References Bovim G 1992 Cervicogenic headache, migraine and tension-type headache. Pressure pain threshold measurements. Pain 51: 169– 173 Butler D 1991 Mobilization of the nervous system. Churchill Livingstone, Melbourne Carrothers C 1994 The effect of whiplash injury on the thoracic spine. Physiotherapy Canada 46 (Suppl 1): S143 Cloward R 1959 Cervical discography. Annals of Surgery 150: 1052–1057 Cohen ML, 1995 The clinical challenge of secondary hyperalgesia. In: Proceedings of Moving in on Pain Conference: 21–26. Butterworth-Heinemann, Adelaide DeFranca G, Levine L 1995 The T4 syndrome. Journal of Manipulation and Physiological Therapeutics 18(1): 34–37 Dhondt W, Willaeys T, Verbruggen LA, Oostendorp RA, Duquet W 1999 Pain threshold in patients with rheumatoid arthritis and effect of manual oscillation. Scandinavian Journal of Rheumatology 28(2): 88–93 Dwyer A, Aprill C, Bogduk N 1990 Cervical zygapophyseal joint pain patterns 1: a study in normal volunteers. Spine 15(6): 453–457 Edmonston SJ, Singer KP 1997 Thoracic spine: anatomical and biomechanical considerations for manual therapy. Manual Therapy 2(3): 132–143 Evans P 1997 The T4 syndrome: some basic science aspects. Physiotherapy 83(4): 186–189 Fischer AA 1986a Pressure tolerance over muscles and bones in normal subjects. Archives of Physical Medicine Rehabilitation 67(June): 406–409 Fischer AA 1986b Pressure threshold meter: its use for quantification of tender spots. Archives of Physical Medicine Rehabilitation 67(November): 836–838 Fischer AA 1987 Pressure algometry over normal muscles. Standard values, validity and reproducibility of pressure threshold. Pain 30: 115–126 Giglia-Smith L, Coll F, McKenzie A, Wessel J, Allison G 1997 Is thoracic tenderness significant in chronic cervical pain? 10th Biennial MPAA Conference proceedings, Melbourne Grieve GP 1988 Common vertebral joint problems, 2nd edn. Churchill Livingstone, Edinburgh Hogeweg JA, Langereis MJ, Bernards ATM, Faber JAJ, Helders PJM 1992 Algometry-measuring pain threshold, method and characteristics in healthy subjects. Scandinavian Journal of Rehabilitation Medicine 24: 99–103 Kosek E, Ekholm J, Nordemar R 1993 A comparison of pressure pain thresholds in different tissues and body regions. Scandinavian Journal of Rehabilitative Medicine 25: 117–124 Kosek E, Ekholm J, Hannson P 1999 Pressure pain thresholds in different tissues in one body region: the influence of skin sensitivity in pressure algometry. Scandinavian Journal of Rehabilitative Medicine 31(2): 89–93 Louis R 1981 Vertebroradicular and vertebromedullar dynamics. Anatomica Clinica 3: 1–11 Maitland G 1986 Vertebral manipulation, 5th edn. Butterworths, London Magora A, Vatine JJ, Magora F 1992 Quantification of musculoskeletal pain by pressure algometry. The Pain clinic 5(2): 101–104 # 2001 Harcourt Publishers Ltd
Mid-thoracic tenderness: a comparison of pressure pain threshold 39 Maigne J-Y, Maigne R, Guerin-Surville H 1991 Upper thoracic dorsal rami: anatomic study of their medial cutaneous branches. Surgical and Radiologic Anatomy 13(2): 109–112 Matthews M 1986 T4 syndrome; clinical notes. Australian Journal of Physiotherapy 32(2): 124–125 Merskey H, Bogduk N (eds) 1994 Classification of chronic pain: descriptions of chronic pain syndromes and definitions of pain terms, 2nd edn. IASP Press, Seattle Meyer RA, Campbell JN, Raja SN 1994 Peripheral neural mechanisms of nociception. In: Wall PD, Melzack R (eds) Textbook of pain, 3rd edn. Churchill Livingstone, New York ch1 p 13–44 Minucci A, 1987 Palpation of the thoracic spine (T1 to T8). In: Proceedings of Fifth Biennial Conference of Manipulative Therapists Association of Australia, Melbourne; 367–376 Nussbaum EL, Downes L 1998 Reliability of clinical pressure-pain measurements obtained on consecutive days. Physical Therapy 78(2): 160–169 Ohrbach R, Gale EN 1989 Pressure pain thresholds in normal muscles: reliability measurement effects, and topographic differences. Pain 37: 257–263
# 2001 Harcourt Publishers Ltd
Reeves JL, Jaeger B, Graff-Radford SB 1986 Reliability of the pressure algometer as a measure of myofascial trigger point sensitivity. Pain 24: 313–321 Sheather-Reid RB, Cohen ML 1998 Psychophysical evidence for a neuropathic component of chronic neck pain. Pain 75: 341–347 Vanderweee¨n L, Oostendorp RAB, Vaes P, Duquet W 1996 Pressure algometry in manual therapy. Manual Therapy 1(5): 258–265 Vatine J-J, Tsenter J, Nirel R 1998 Experimental pressure pain in patients with complex regional pain syndrome, type I (reflex sympathetic dystrophy). American Journal of Physical Medicine and Rehabilitation 77(5): 382–387 Vicenzino B, Collins D, Benson H, Wright A 1998 An investigation of the interrelationship between manipulative therapy induced hypalgesia and sympathoexcitation. Journal of Manipulative and Physiological Therapeutics 21(7): 448–453 Wright A, Robins L, McGuire B 1995 Pressure and thermal pain thresholds in the forequarter: normative data. Journal of Musculoskeletal Pain 3(1): 91–103
Manual Therapy (2001) 6(1), 34–39
Original article
Mid-thoracic tenderness: a comparison of pressure pain threshold between spinal regions, in asymptomatic subjects L. Keating, C. Lubke, V. Powell, T. Young, T. Souvlis, G. Jull Department of Physiotherapy, The University of Queensland, Brisbane, Australia
SUMMARY. Palpation for tenderness forms an important part of the manual therapy assessment for musculoskeletal dysfunction. In conjunction with other testing procedures it assists in establishing the clinical diagnosis. Tenderness in the thoracic spine has been reported in the literature as a clinical feature in musculoskeletal conditions where pain and dysfunction are located primarily in the upper quadrant. This study aimed to establish whether pressure pain thresholds (PPTs) of the mid-thoracic region of asymptomatic subjects were naturally lower than those of the cervical and lumbar areas. A within-subject study design was used to examine PPT at four spinal levels C6, T4, T6, and L4 in 50 asymptomatic volunteers. Results showed significant (P50.001) regional differences. PPT values increased in a caudal direction. The cervical region had the lowest PPT scores, that is was the most tender. Values increased in the thoracic region and were highest in the lumbar region. This study contributes to the normative data on spinal PPT values and demonstrates that mid-thoracic tenderness relative to the cervical spine is not a normal finding in asymptomatic subjects. # 2001 Harcourt Publishers Ltd.
such as a thermal, mechanical or chemical stimulus and a subjective response. Cohen (1995) considers that, clinically, the term hyperalgesia, which is defined as an increased response to a stimulus that is normally noxious (Merskey & Bogduk 1994), also includes allodynia, the response of pain to a stimulus that is not normally noxious. Tenderness in musculoskeletal disorders may be a result of local tissue damage or may indicate an alteration in afferent processing of mechanical stimuli by second order neurones within the dorsal horn (Sheather-Reid & Cohen, 1998). Thus, hyperalgesia may be described as primary or secondary depending on whether it originates in tissue which has been damaged (primary) or in normal tissue which may be neuroanatomically related to the area of damage (secondary). Sensitization of both the peripheral and central nervous systems have been demonstrated to cause the decrease in threshold or increased response to mechanical stimuli that result in hyperalgesia (Meyer et al. 1994). Palpation to identify areas of tenderness may provide us with important information in the assessment of varied musculoskeletal dysfunctions and pain syndromes as it provides relevant information about both the underlying tissue and the nociceptive system.
INTRODUCTION An important element of manual therapy assessment is the use of passive manual examination to detect the presence of musculoskeletal dysfunction. Palpation of the spine for tenderness is part of this manual examination (Maitland 1986). In the thoracic spine a challenge arises in interpreting the diagnostic relevance of tenderness. Controversy exists regarding whether pain and tenderness within the mid-thoracic spine is an expected finding in patients who present with symptoms in the upper quadrant. Is the midthoracic spine normally tender or does increased sensitivity in this area indicate local dysfunction? Cohen (1995) describes tenderness or more accurately, hyperalgesia, as a psychophysical concept that establishes a relationship between a physical stimulus Received 6 March 2000 Revised 4 August 2000 Accepted 11 November 2000 L. Keating, BPhty, MPhty St. Catherine Lubke BPhty, M.Phty St, Vicki Powell BPhty (Hons), MPhty, St, Tanya Young BPhty MPhty St, Tina Souvlis BPhty (Hons), Gwendolen Jull MPhty, Grad Dip Adv Manip Ther, FACP, Associate Professor, Department of Physiotherapy, University of Queensland, Brisbane, Qld 4072, Australia. Correspondence to GJ. Tel: +61 7 3365 2275; Fax: +61 7 335 2775. E-mail: [email protected] 34
Mid-thoracic tenderness: a comparison of pressure pain threshold 35
Pressure algometry has been demonstrated to be a reliable and repeatable tool to quantify local pain and tenderness (Fischer 1987, Magora et al. 1992, Nussbaum & Downes 1998). Pressure pain threshold is defined as the least stimulus intensity at which a subject perceives discomfort (Fischer 1986a). PPT has been used as a measure to quantify pain and tenderness in asymptomatic subjects thereby establishing normative PPT values (Fischer 1986a, Fischer 1986b, Reeves et al. 1986, Fischer 1987, Ohrbach & Gale 1989, Hogeweg et al. 1992). Subsequently, PPT has been measured in symptomatic patients presenting with a variety of musculoskeletal conditions (Bovim 1992, Vatine et al. 1998). It has also been suggested that comparisons of PPT values can be used as an aid in diagnosis (Fischer 1987), and additionally, they have been used as an outcome measure for therapeutic interventions (Vicenzino et al. 1998, Dhondt et al. 1999). Significant regional differences in PPT of the paravertebral tissues of the spine have been observed in symptomatic subjects (Vanderweee¨n et al. 1996) and in asymptomatic subjects (Hogeweg et al. 1992, Kosek et al. 1993) with values increasing in a caudal direction. These studies have more frequently addressed muscle and soft tissue tenderness, with reference to trigger point pathology. However, manual examination techniques, such as posteroanterior glides involve pressure over bony prominences. Kosek et al. (1993) compared PPT values over bone versus muscle and found no difference, whereas Fischer (1986a) reported a difference but did not analyse for statistical significance. With respect to the thoracic spine, Minucci (1987), in a small study of asymptomatic subjects, demonstrated that testing the thoracic spine with posteroanterior glides produced a subjective report of discomfort or pain in 37% of this group. This was found to be significantly more frequent at T3, T4 and T5. In an asymptomatic population, several theories can be offered to explain the presence of mid-thoracic tenderness. Maigne et al. (1991) performed dissections of T1–T5 and postulated that local tenderness over the spinous processes is due to the location of the medial branch of the dorsal rami over the apex of these prominences. Kosek et al. (1999) found lower PPT values for muscle sites where the nerve became superficial (‘muscle/nerve sites’) than in ‘pure’ muscle and bony sites. In the light of these findings, Minucci’s results might be explained by compression of several dorsal cutaneous rami under the tendinous insertion of splenius cervicis on the spinous processes of T3, T4 and T5. Thoracic pain and tenderness has been reported as a feature of a variety of musculoskeletal conditions involving both the thoracic spine and structures capable of pain referral to the thoracic spine. Injection studies of cervical zygapophysial joints # 2001 Harcourt Publishers Ltd
(Dwyer et al. 1990) and cervical discs (Cloward 1959) have both demonstrated pain referral to the thoracic region. Cervical spondylosis may produce pain and movement dysfunction in the upper thoracic spine (Edmonston & Singer 1997). Carrothers (1994) observed a similar relationship between whiplash associated cervical injury and the thoracic spine. However, Giglia-Smith et al. (1997), in a group of chronic cervical pain patients without thoracic pain, demonstrated that there was no difference in pressure pain threshold (PPT) values in the thoracic spine when compared to an asymptomatic control group. Focusing specifically on the mid-thoracic region, the T4 syndrome is a clinical entity that is purported to have tenderness on palpation as one of several diagnostic signs (Matthews 1986, DeFranca & Levine 1995, Evans 1997). Despite its name, T4 to T7 levels have been documented as being implicated in the T4 syndrome. Scientific validation of this clinical syndrome has as yet, not been forthcoming. The primary aim of this study was to investigate whether the mid-thoracic area exhibits greater tenderness than other regions of the spine in an asymptomatic population. Comparison of PPTs over the cervical, thoracic and lumbar spinous processes would assist in determining the clinical relevance of thoracic tenderness when found in patients with upper quadrant pain syndromes. Pressure algometry was used to measure PPTs to gain a quantitative estimation of tenderness. Two segments of the midthoracic spine (T4 and T6) were chosen to encompass the region in the thoracic spine that is often cited as tender in asymptomatic subjects as well as in patients with upper quadrant clinical syndromes. Studies have shown the spinal cord to be relatively fixed at C6, T6 and L4 (Louis 1981). For T6, as a tension point (Butler 1991), there is a direct link to a proposed mechanism of thoracic tenderness whereas T4 is not considered a tension point. Yet both are associated with a specific pain syndrome. Both segments share similar biomechanical properties and anatomical relationships to the sympathetic nervous system (Grieve 1988). The levels of C6 and L4 were chosen to counterbalance the possible general influence of tension points on PPT values. The hypothesis tested in this study was that reduced pressure pain threshold is present in one or more segments in the mid-thoracic spine (T4 or T6) in asymptomatic subjects, when compared with cervical and lumbar segments in the same subject.
METHODOLOGY Subjects Fifty volunteers (23 males and 27 females) aged 19 to 56 (mean=25.7+7.1years) were recruited from The Manual Therapy (2001) 6(1), 34–39
36 Manual Therapy
University of Queensland and local community. Subjects were required to have no history of headaches, or pain in the cervical, thoracic or lumbo-sacral regions. Subjects were excluded on the basis of the criteria outlined in Table 1. Prior to commencement of the study, the subjects received a full explanation of the experimental procedure and signed a consent form. The Human Ethics Committee of The University of Queensland, in accordance with the National Health and Medical Research Council’s guidelines, approved this study. Apparatus and measurement PPT was measured using a hand held electronic pressure algometer (Somedic, Sweden). The circular probe had a surface area of 1cm2. The algometer was calibrated at the start of each session and the same investigator took all PPT measurements. Pressure was applied to the spinous processes at a rate of 40 kPa/s and the subject was instructed to push a hand held button immediately the sensation changed from pressure to pain. This instantaneously froze the pressure reading, which was then recorded by a second investigator. Three measurements were taken at each level, with an interval of 20 sec between readings. The PPT was calculated as the mean of these 3 trials. A standard treatment plinth was used to position the subject. A third examiner identified spinal levels each time, using a standardized protocol counting in cranial and caudal directions and checking with lateral bony landmarks. The relevant interspinous spaces were then marked with a pen. Procedure Following an explanation of the experiment, the subject signed an informed consent statement. To familiarize the patient with pressure algometry, a demonstration was performed, firstly on the examiner’s wrist, and then on the subject’s ulnar styloid process. The subject was instructed to push the button when the pressure became a discomfort.
The subject undressed to expose the spine and was positioned prone on the plinth with 1 pillow each underneath the hips and feet. The spinal levels, C6, T4, T6 and L4, were then marked. The order of levels to be tested was systematically randomized prior to testing by drawing lots. Confounding variables were controlled by standardizing equipment set up, noise and light levels, subject instructions, surface markings and patient position. The pressure algometer was positioned perpendicular to the body surface at the specified spinous process. The algometer pressure tip was stabilized between the examiner’s thumb and first finger. Pressure readings were recorded once the subject had pressed the indicator button. On completion of the measurement of all spinal levels, the subject’s skin was cleaned to remove the surface markings. The subjects attended one session only. Repeatability Repeatability of PPT measurement by the examiner was verified by conducting a second experimental session on 10 subjects (5 female and 5 male), 30 min after the completion of the first session. An Intraclass Correlation Coefficient (ICC) was calculated using a 2-way random effect model (for consistency agreement). The results of the analysis are presented in Table 2. The ICCs showed that the reproducibility of PPTs in the cervical and thoracic levels was excellent (ICC40.9), and good at the lumbar level (ICC40.75). Inspection of the coefficient of variance (CV) indicated that PPTs varied substantially between subjects (CV ¼ 28–45%), however this was similar to other published PPT studies (Hogeweg et al. 1992, Fischer 1987). Data management For the main study, the analysis was undertaken on the mean of the three PPT readings taken at each spinal level. The main analysis compared the mean PPT readings taken from different levels of the spine. The comparison was performed using a mixed-model analysis of variance, where level (C6, T4, T6, and L4) was a within-groups factor and gender (male, female)
Table 1. Exclusion criteria for subjects History of spinal pain and/or radiculopathy Spinal congenital abnormality Deformity or history of spinal fracture History of chronic respiratory disease e.g. asthma Diabetes Peripheral vascular disease Osteoporosis Rheumatoid arthritis Ankylosing spondylitis Pregnancy Medication: NSAIDs, analgesics, anti-coagulants and steroids
Manual Therapy (2001) 6(1), 34–39
Table 2. Intraclass correlation coefficient and coefficient of variance for the spinal levels tested in repeatability study Spinal level
C6 T4 T6 L4
Mean PPT (kPa/cm2) (+standard deviation)
Coefficient of variance
Test 1
Test 2
Test 1
Test 2
335+150 349+141 373+147 504+182
323+117 390+129 373+103 513+164
45% 40% 39% 36%
36% 33% 28% 32%
ICC
0.93 0.93 0.90 0.84
# 2001 Harcourt Publishers Ltd
Mid-thoracic tenderness: a comparison of pressure pain threshold 37
was a between-groups factor. A series of pairwise comparisons were undertaken to investigate differences between levels, and Tukey’s Least Significant Difference (LSD) test was used to adjust for multiple comparisons. All analyses were undertaken using SPSS 8.0 for Windows. P50.05 was considered to be statistically significant.
significant difference (P50.05) in the average PPT increasing in a caudal direction from cervical (C6: 255 kPa/cm2), to thoracic, to lumbar spine (L4: 445 kPa/cm2). There was no significant difference (P=0.184) for PPT values within the thoracic spine (T4: 324 kPa/cm2 and T6: 302 kPa/cm2).
DISCUSSION RESULTS Distributions of average PPT readings were examined using boxplots. These determined that at each spinal level, the scores were normally distributed with the exception of four outliers. The average PPT scores were transformed by a log scale, which failed to remove the outliers. These were subsequently removed from the initial analysis. The analysis found a significant gender effect when the outliers were included but not when they were removed. Due to the instability of the gender effect, it was decided to interpret the results conservatively. Inspection of Box’s M indicated that the data conformed with the homogeneity of variancecovariance assumption (M ¼ 11.22, F(10, 7888)=1.01, P=0.434), whilst the Levene’s tests indicated that the variance of scores in each gender was similar at each of the four levels. Regional comparison The mean PPTs for each spinal region are summarized in Figure 1. These means were compared using a mixed model analysis of variance. The results of the analysis revealed a significant main effect for region (F(3,23)=20.13, P50.001), but no main effect for gender (F(1,25)=2.187, P=0.146) and no interaction between gender and region (F(3,23)=0.96, P=0.791). The effect of region accounted for around 60.6% of the variance in average PPT. Post hoc pairwise comparisons, using Tukey’s test, determined which levels accounted for the significant difference between the regions. These showed a
Fig. 1—Mean PPT values for tested spinal level. # 2001 Harcourt Publishers Ltd
Significant regional differences in spinal PPT values were noted in the study, with values increasing from the cervical to the lumbar spine. It was shown that mid-thoracic segments were less tender than the cervical segment but more tender than the lumbar segment. Wright et al. (1995) also demonstrated significantly lower PPTs from C6 (P50.01) than from sites on the upper limb. Lower PPTs evident in the cervical spine may indicate that protection from noxious stimulation is important in this area due to close proximity of vascular and neural bundles. There was no significant difference between the PPT values within the mid-thoracic region at T4 and T6. Based on these results, palpation that elicits greater tenderness or hyperalgesia in the mid-thoracic spine than in the cervical spine, is not a normal finding in the asymptomatic individual. The average PPT values compliment the results of previous studies of paraspinal tissue in healthy (Hogeweg et al. 1992, Kosek et al. 1993) and symptomatic subjects (Vanderweee¨n et al. 1996). Based on the results of this study and in the presence of relevant subjective and physical findings, mid-thoracic tenderness may be considered relevant. Thoracic PPT values that differ significantly from the values, or regional pattern, reported in this study warrant further clinical consideration. As would be expected in a normal distribution, some asymptomatic subjects deviated from the regional pattern, therefore thoracic tenderness, in isolation, should not constitute a diagnosis of dysfunction. The findings of the current study suggest that the use of intra-subject regional patterns of tenderness should be further explored. With the normative values and regional pattern provided in this study, pressure algometry might be considered for use in the clinical setting to measure PPT and to quantify tenderness. Whilst this study measured PPT at two thoracic spinal levels, it is not unreasonable to extrapolate this data to the rest of the thoracic spine as Hogeweg et al. (1992) reported an unremarkable steady increase in PPT values measured at T1, T3, T6 and T10. The averaged values were similar to this study despite the expected differences between PPT values over soft tissue and bone. As Hogeweg et al. (1992) reported very similar readings at three lumbar paraspinal levels (L1, L3, L5), it would also be reasonable to extrapolate the single L4 value to the rest of the Manual Therapy (2001) 6(1), 34–39
38 Manual Therapy
lumbar region. As investigation of the cervical region has been limited to the C6 spinous process (Hogeweg et al. 1992) and the C7 spinous process (Kosek et al. 1993), care must be taken inferring PPT values to the whole area. The results of this quantitative study cannot be compared to those of Minucci (1987) where PAIVM tests were used to elicit tenderness. PAIVM testing is not replicated with pressure algometry as the former technique also uses accessory joint motion for pain provocation. There may be a relationship between the two assessment techniques but further study is required. Whilst this study has demonstrated that relative mid-thoracic tenderness is not a normal finding in asymptomatic subjects, future studies must confirm the presence of mid-thoracic tenderness in varied symptomatic populations. Lowered PPT values or a change in the regional pattern would represent this. The presence of thoracic tenderness should be quantified in relevant musculoskeletal dysfunctions such as cervical spondylosis, cervicogenic headache, whiplash-associated disorders and cervical disc pathology. A preliminary study of patients with chronic cervical pain without thoracic symptoms has found thoracic pain threshold was similar to a control group (Giglia-Smith et al. 1997). This data should be interpreted with caution because of the small sample size investigated (n=29). Future studies of symptomatic subjects should not only assess thoracic tenderness but additionally consider the pathogenesis of clinical pain presentations. Monitoring changes in cervical and thoracic intervertebral motion, sympathetic nervous system outflow and neuromeningeal compliance in addition to PPT could begin to decipher the pathogenesis of various clinical syndromes that have mid-thoracic pain/tenderness as a common feature. Tenderness of the thoracic spine may also indicate the presence of secondary hyperalgesia or sensitization of the central nervous system causing a heightened response or lowered threshold to mechanical stimulation in normal tissue. This scenario should also be considered as it may offer an alternative explanation for thoracic tenderness in upper quadrant musculoskeletal dysfunction.
CONCLUSION The results of this study of asymptomatic subjects, showed significant regional differences in spinal PPT, with values increasing in a caudal direction. The midthoracic segments, T4 and T6, were less tender than cervical, C6 segment but more tender than the lumbar, L4 segment. Anecdotally, physiotherapists report tenderness on palpation of the mid-thoracic spine in patients with a variety of musculoskeletal Manual Therapy (2001) 6(1), 34–39
disorders. This study clarifies that mid-thoracic tenderness is not a normal finding in asymptomatic subjects. Future studies should confirm the presence of thoracic tenderness in patients presenting with upper quadrant dysfunction.
Acknowledgments The authors wish to thank the subjects for their contribution to this study and acknowledge the assistance of Mr Craig Shaw with statistical analyses.
References Bovim G 1992 Cervicogenic headache, migraine and tension-type headache. Pressure pain threshold measurements. Pain 51: 169– 173 Butler D 1991 Mobilization of the nervous system. Churchill Livingstone, Melbourne Carrothers C 1994 The effect of whiplash injury on the thoracic spine. Physiotherapy Canada 46 (Suppl 1): S143 Cloward R 1959 Cervical discography. Annals of Surgery 150: 1052–1057 Cohen ML, 1995 The clinical challenge of secondary hyperalgesia. In: Proceedings of Moving in on Pain Conference: 21–26. Butterworth-Heinemann, Adelaide DeFranca G, Levine L 1995 The T4 syndrome. Journal of Manipulation and Physiological Therapeutics 18(1): 34–37 Dhondt W, Willaeys T, Verbruggen LA, Oostendorp RA, Duquet W 1999 Pain threshold in patients with rheumatoid arthritis and effect of manual oscillation. Scandinavian Journal of Rheumatology 28(2): 88–93 Dwyer A, Aprill C, Bogduk N 1990 Cervical zygapophyseal joint pain patterns 1: a study in normal volunteers. Spine 15(6): 453–457 Edmonston SJ, Singer KP 1997 Thoracic spine: anatomical and biomechanical considerations for manual therapy. Manual Therapy 2(3): 132–143 Evans P 1997 The T4 syndrome: some basic science aspects. Physiotherapy 83(4): 186–189 Fischer AA 1986a Pressure tolerance over muscles and bones in normal subjects. Archives of Physical Medicine Rehabilitation 67(June): 406–409 Fischer AA 1986b Pressure threshold meter: its use for quantification of tender spots. Archives of Physical Medicine Rehabilitation 67(November): 836–838 Fischer AA 1987 Pressure algometry over normal muscles. Standard values, validity and reproducibility of pressure threshold. Pain 30: 115–126 Giglia-Smith L, Coll F, McKenzie A, Wessel J, Allison G 1997 Is thoracic tenderness significant in chronic cervical pain? 10th Biennial MPAA Conference proceedings, Melbourne Grieve GP 1988 Common vertebral joint problems, 2nd edn. Churchill Livingstone, Edinburgh Hogeweg JA, Langereis MJ, Bernards ATM, Faber JAJ, Helders PJM 1992 Algometry-measuring pain threshold, method and characteristics in healthy subjects. Scandinavian Journal of Rehabilitation Medicine 24: 99–103 Kosek E, Ekholm J, Nordemar R 1993 A comparison of pressure pain thresholds in different tissues and body regions. Scandinavian Journal of Rehabilitative Medicine 25: 117–124 Kosek E, Ekholm J, Hannson P 1999 Pressure pain thresholds in different tissues in one body region: the influence of skin sensitivity in pressure algometry. Scandinavian Journal of Rehabilitative Medicine 31(2): 89–93 Louis R 1981 Vertebroradicular and vertebromedullar dynamics. Anatomica Clinica 3: 1–11 Maitland G 1986 Vertebral manipulation, 5th edn. Butterworths, London Magora A, Vatine JJ, Magora F 1992 Quantification of musculoskeletal pain by pressure algometry. The Pain clinic 5(2): 101–104 # 2001 Harcourt Publishers Ltd
Mid-thoracic tenderness: a comparison of pressure pain threshold 39 Maigne J-Y, Maigne R, Guerin-Surville H 1991 Upper thoracic dorsal rami: anatomic study of their medial cutaneous branches. Surgical and Radiologic Anatomy 13(2): 109–112 Matthews M 1986 T4 syndrome; clinical notes. Australian Journal of Physiotherapy 32(2): 124–125 Merskey H, Bogduk N (eds) 1994 Classification of chronic pain: descriptions of chronic pain syndromes and definitions of pain terms, 2nd edn. IASP Press, Seattle Meyer RA, Campbell JN, Raja SN 1994 Peripheral neural mechanisms of nociception. In: Wall PD, Melzack R (eds) Textbook of pain, 3rd edn. Churchill Livingstone, New York ch1 p 13–44 Minucci A, 1987 Palpation of the thoracic spine (T1 to T8). In: Proceedings of Fifth Biennial Conference of Manipulative Therapists Association of Australia, Melbourne; 367–376 Nussbaum EL, Downes L 1998 Reliability of clinical pressure-pain measurements obtained on consecutive days. Physical Therapy 78(2): 160–169 Ohrbach R, Gale EN 1989 Pressure pain thresholds in normal muscles: reliability measurement effects, and topographic differences. Pain 37: 257–263
# 2001 Harcourt Publishers Ltd
Reeves JL, Jaeger B, Graff-Radford SB 1986 Reliability of the pressure algometer as a measure of myofascial trigger point sensitivity. Pain 24: 313–321 Sheather-Reid RB, Cohen ML 1998 Psychophysical evidence for a neuropathic component of chronic neck pain. Pain 75: 341–347 Vanderweee¨n L, Oostendorp RAB, Vaes P, Duquet W 1996 Pressure algometry in manual therapy. Manual Therapy 1(5): 258–265 Vatine J-J, Tsenter J, Nirel R 1998 Experimental pressure pain in patients with complex regional pain syndrome, type I (reflex sympathetic dystrophy). American Journal of Physical Medicine and Rehabilitation 77(5): 382–387 Vicenzino B, Collins D, Benson H, Wright A 1998 An investigation of the interrelationship between manipulative therapy induced hypalgesia and sympathoexcitation. Journal of Manipulative and Physiological Therapeutics 21(7): 448–453 Wright A, Robins L, McGuire B 1995 Pressure and thermal pain thresholds in the forequarter: normative data. Journal of Musculoskeletal Pain 3(1): 91–103
Manual Therapy (2001) 6(1), 34–39